grokdump.py 111 KB
Newer Older
1 2
#!/usr/bin/env python
#
3
# Copyright 2012 the V8 project authors. All rights reserved.
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
#     * Redistributions of source code must retain the above copyright
#       notice, this list of conditions and the following disclaimer.
#     * Redistributions in binary form must reproduce the above
#       copyright notice, this list of conditions and the following
#       disclaimer in the documentation and/or other materials provided
#       with the distribution.
#     * Neither the name of Google Inc. nor the names of its
#       contributors may be used to endorse or promote products derived
#       from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

30
import BaseHTTPServer
31
import bisect
32
import cgi
33
import cmd
34
import codecs
35
import ctypes
36
import datetime
37
import disasm
38 39 40 41
import mmap
import optparse
import os
import re
42 43
import sys
import types
44 45
import urllib
import urlparse
46
import v8heapconst
47 48 49 50
import webbrowser

PORT_NUMBER = 8081

51

52
USAGE="""usage: %prog [OPTIONS] [DUMP-FILE]
53 54 55 56 57 58

Minidump analyzer.

Shows the processor state at the point of exception including the
stack of the active thread and the referenced objects in the V8
heap. Code objects are disassembled and the addresses linked from the
59
stack (e.g. pushed return addresses) are marked with "=>".
60 61

Examples:
62
  $ %prog 12345678-1234-1234-1234-123456789abcd-full.dmp"""
63

64

65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117
DEBUG=False


def DebugPrint(s):
  if not DEBUG: return
  print s


class Descriptor(object):
  """Descriptor of a structure in a memory."""

  def __init__(self, fields):
    self.fields = fields
    self.is_flexible = False
    for _, type_or_func in fields:
      if isinstance(type_or_func, types.FunctionType):
        self.is_flexible = True
        break
    if not self.is_flexible:
      self.ctype = Descriptor._GetCtype(fields)
      self.size = ctypes.sizeof(self.ctype)

  def Read(self, memory, offset):
    if self.is_flexible:
      fields_copy = self.fields[:]
      last = 0
      for name, type_or_func in fields_copy:
        if isinstance(type_or_func, types.FunctionType):
          partial_ctype = Descriptor._GetCtype(fields_copy[:last])
          partial_object = partial_ctype.from_buffer(memory, offset)
          type = type_or_func(partial_object)
          if type is not None:
            fields_copy[last] = (name, type)
            last += 1
        else:
          last += 1
      complete_ctype = Descriptor._GetCtype(fields_copy[:last])
    else:
      complete_ctype = self.ctype
    return complete_ctype.from_buffer(memory, offset)

  @staticmethod
  def _GetCtype(fields):
    class Raw(ctypes.Structure):
      _fields_ = fields
      _pack_ = 1

      def __str__(self):
        return "{" + ", ".join("%s: %s" % (field, self.__getattribute__(field))
                               for field, _ in Raw._fields_) + "}"
    return Raw


118
def FullDump(reader, heap):
119 120
  """Dump all available memory regions."""
  def dump_region(reader, start, size, location):
121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137
    print
    while start & 3 != 0:
      start += 1
      size -= 1
      location += 1
    is_executable = reader.IsProbableExecutableRegion(location, size)
    is_ascii = reader.IsProbableASCIIRegion(location, size)

    if is_executable is not False:
      lines = reader.GetDisasmLines(start, size)
      for line in lines:
        print FormatDisasmLine(start, heap, line)
      print

    if is_ascii is not False:
      # Output in the same format as the Unix hd command
      addr = start
138 139
      for i in xrange(0, size, 16):
        slot = i + location
140 141
        hex_line = ""
        asc_line = ""
142
        for i in xrange(16):
143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161
          if slot + i < location + size:
            byte = ctypes.c_uint8.from_buffer(reader.minidump, slot + i).value
            if byte >= 0x20 and byte < 0x7f:
              asc_line += chr(byte)
            else:
              asc_line += "."
            hex_line += " %02x" % (byte)
          else:
            hex_line += "   "
          if i == 7:
            hex_line += " "
        print "%s  %s |%s|" % (reader.FormatIntPtr(addr),
                               hex_line,
                               asc_line)
        addr += 16

    if is_executable is not True and is_ascii is not True:
      print "%s - %s" % (reader.FormatIntPtr(start),
                         reader.FormatIntPtr(start + size))
162 163 164
      print start + size + 1;
      for i in xrange(0, size, reader.PointerSize()):
        slot = start + i
165 166 167 168 169 170 171
        maybe_address = reader.ReadUIntPtr(slot)
        heap_object = heap.FindObject(maybe_address)
        print "%s: %s" % (reader.FormatIntPtr(slot),
                          reader.FormatIntPtr(maybe_address))
        if heap_object:
          heap_object.Print(Printer())
          print
172 173 174

  reader.ForEachMemoryRegion(dump_region)

175 176 177 178 179
# Heap constants generated by 'make grokdump' in v8heapconst module.
INSTANCE_TYPES = v8heapconst.INSTANCE_TYPES
KNOWN_MAPS = v8heapconst.KNOWN_MAPS
KNOWN_OBJECTS = v8heapconst.KNOWN_OBJECTS

180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197
# Set of structures and constants that describe the layout of minidump
# files. Based on MSDN and Google Breakpad.

MINIDUMP_HEADER = Descriptor([
  ("signature", ctypes.c_uint32),
  ("version", ctypes.c_uint32),
  ("stream_count", ctypes.c_uint32),
  ("stream_directories_rva", ctypes.c_uint32),
  ("checksum", ctypes.c_uint32),
  ("time_date_stampt", ctypes.c_uint32),
  ("flags", ctypes.c_uint64)
])

MINIDUMP_LOCATION_DESCRIPTOR = Descriptor([
  ("data_size", ctypes.c_uint32),
  ("rva", ctypes.c_uint32)
])

198 199 200 201 202
MINIDUMP_STRING = Descriptor([
  ("length", ctypes.c_uint32),
  ("buffer", lambda t: ctypes.c_uint8 * (t.length + 2))
])

203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
MINIDUMP_DIRECTORY = Descriptor([
  ("stream_type", ctypes.c_uint32),
  ("location", MINIDUMP_LOCATION_DESCRIPTOR.ctype)
])

MD_EXCEPTION_MAXIMUM_PARAMETERS = 15

MINIDUMP_EXCEPTION = Descriptor([
  ("code", ctypes.c_uint32),
  ("flags", ctypes.c_uint32),
  ("record", ctypes.c_uint64),
  ("address", ctypes.c_uint64),
  ("parameter_count", ctypes.c_uint32),
  ("unused_alignment", ctypes.c_uint32),
  ("information", ctypes.c_uint64 * MD_EXCEPTION_MAXIMUM_PARAMETERS)
])

MINIDUMP_EXCEPTION_STREAM = Descriptor([
  ("thread_id", ctypes.c_uint32),
  ("unused_alignment", ctypes.c_uint32),
  ("exception", MINIDUMP_EXCEPTION.ctype),
  ("thread_context", MINIDUMP_LOCATION_DESCRIPTOR.ctype)
])

# Stream types.
MD_UNUSED_STREAM = 0
MD_RESERVED_STREAM_0 = 1
MD_RESERVED_STREAM_1 = 2
MD_THREAD_LIST_STREAM = 3
MD_MODULE_LIST_STREAM = 4
MD_MEMORY_LIST_STREAM = 5
MD_EXCEPTION_STREAM = 6
MD_SYSTEM_INFO_STREAM = 7
MD_THREAD_EX_LIST_STREAM = 8
MD_MEMORY_64_LIST_STREAM = 9
MD_COMMENT_STREAM_A = 10
MD_COMMENT_STREAM_W = 11
MD_HANDLE_DATA_STREAM = 12
MD_FUNCTION_TABLE_STREAM = 13
MD_UNLOADED_MODULE_LIST_STREAM = 14
MD_MISC_INFO_STREAM = 15
MD_MEMORY_INFO_LIST_STREAM = 16
MD_THREAD_INFO_LIST_STREAM = 17
MD_HANDLE_OPERATION_LIST_STREAM = 18

MD_FLOATINGSAVEAREA_X86_REGISTERAREA_SIZE = 80

MINIDUMP_FLOATING_SAVE_AREA_X86 = Descriptor([
  ("control_word", ctypes.c_uint32),
  ("status_word", ctypes.c_uint32),
  ("tag_word", ctypes.c_uint32),
  ("error_offset", ctypes.c_uint32),
  ("error_selector", ctypes.c_uint32),
  ("data_offset", ctypes.c_uint32),
  ("data_selector", ctypes.c_uint32),
  ("register_area", ctypes.c_uint8 * MD_FLOATINGSAVEAREA_X86_REGISTERAREA_SIZE),
  ("cr0_npx_state", ctypes.c_uint32)
])

MD_CONTEXT_X86_EXTENDED_REGISTERS_SIZE = 512

# Context flags.
MD_CONTEXT_X86 = 0x00010000
MD_CONTEXT_X86_CONTROL = (MD_CONTEXT_X86 | 0x00000001)
MD_CONTEXT_X86_INTEGER = (MD_CONTEXT_X86 | 0x00000002)
MD_CONTEXT_X86_SEGMENTS = (MD_CONTEXT_X86 | 0x00000004)
MD_CONTEXT_X86_FLOATING_POINT = (MD_CONTEXT_X86 | 0x00000008)
MD_CONTEXT_X86_DEBUG_REGISTERS = (MD_CONTEXT_X86 | 0x00000010)
MD_CONTEXT_X86_EXTENDED_REGISTERS = (MD_CONTEXT_X86 | 0x00000020)

def EnableOnFlag(type, flag):
  return lambda o: [None, type][int((o.context_flags & flag) != 0)]

MINIDUMP_CONTEXT_X86 = Descriptor([
  ("context_flags", ctypes.c_uint32),
  # MD_CONTEXT_X86_DEBUG_REGISTERS.
  ("dr0", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_DEBUG_REGISTERS)),
  ("dr1", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_DEBUG_REGISTERS)),
  ("dr2", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_DEBUG_REGISTERS)),
  ("dr3", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_DEBUG_REGISTERS)),
  ("dr6", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_DEBUG_REGISTERS)),
  ("dr7", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_DEBUG_REGISTERS)),
  # MD_CONTEXT_X86_FLOATING_POINT.
  ("float_save", EnableOnFlag(MINIDUMP_FLOATING_SAVE_AREA_X86.ctype,
                              MD_CONTEXT_X86_FLOATING_POINT)),
  # MD_CONTEXT_X86_SEGMENTS.
  ("gs", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_SEGMENTS)),
  ("fs", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_SEGMENTS)),
  ("es", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_SEGMENTS)),
  ("ds", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_SEGMENTS)),
  # MD_CONTEXT_X86_INTEGER.
  ("edi", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_INTEGER)),
  ("esi", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_INTEGER)),
  ("ebx", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_INTEGER)),
  ("edx", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_INTEGER)),
  ("ecx", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_INTEGER)),
  ("eax", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_INTEGER)),
  # MD_CONTEXT_X86_CONTROL.
  ("ebp", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_CONTROL)),
  ("eip", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_CONTROL)),
  ("cs", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_CONTROL)),
  ("eflags", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_CONTROL)),
  ("esp", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_CONTROL)),
  ("ss", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_X86_CONTROL)),
  # MD_CONTEXT_X86_EXTENDED_REGISTERS.
  ("extended_registers",
   EnableOnFlag(ctypes.c_uint8 * MD_CONTEXT_X86_EXTENDED_REGISTERS_SIZE,
                MD_CONTEXT_X86_EXTENDED_REGISTERS))
])

313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348
MD_CONTEXT_ARM = 0x40000000
MD_CONTEXT_ARM_INTEGER = (MD_CONTEXT_ARM | 0x00000002)
MD_CONTEXT_ARM_FLOATING_POINT = (MD_CONTEXT_ARM | 0x00000004)
MD_FLOATINGSAVEAREA_ARM_FPR_COUNT = 32
MD_FLOATINGSAVEAREA_ARM_FPEXTRA_COUNT = 8

MINIDUMP_FLOATING_SAVE_AREA_ARM = Descriptor([
  ("fpscr", ctypes.c_uint64),
  ("regs", ctypes.c_uint64 * MD_FLOATINGSAVEAREA_ARM_FPR_COUNT),
  ("extra", ctypes.c_uint64 * MD_FLOATINGSAVEAREA_ARM_FPEXTRA_COUNT)
])

MINIDUMP_CONTEXT_ARM = Descriptor([
  ("context_flags", ctypes.c_uint32),
  # MD_CONTEXT_ARM_INTEGER.
  ("r0", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r1", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r2", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r3", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r4", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r5", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r6", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r7", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r8", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r9", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r10", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r11", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("r12", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("sp", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("lr", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("pc", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_ARM_INTEGER)),
  ("cpsr", ctypes.c_uint32),
  ("float_save", EnableOnFlag(MINIDUMP_FLOATING_SAVE_AREA_ARM.ctype,
                              MD_CONTEXT_ARM_FLOATING_POINT))
])

349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401

MD_CONTEXT_ARM64 =  0x80000000
MD_CONTEXT_ARM64_INTEGER = (MD_CONTEXT_ARM64 | 0x00000002)
MD_CONTEXT_ARM64_FLOATING_POINT = (MD_CONTEXT_ARM64 | 0x00000004)
MD_FLOATINGSAVEAREA_ARM64_FPR_COUNT = 64

MINIDUMP_FLOATING_SAVE_AREA_ARM = Descriptor([
  ("fpscr", ctypes.c_uint64),
  ("regs", ctypes.c_uint64 * MD_FLOATINGSAVEAREA_ARM64_FPR_COUNT),
])

MINIDUMP_CONTEXT_ARM64 = Descriptor([
  ("context_flags", ctypes.c_uint64),
  # MD_CONTEXT_ARM64_INTEGER.
  ("r0", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r1", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r2", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r3", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r4", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r5", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r6", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r7", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r8", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r9", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r10", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r11", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r12", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r13", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r14", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r15", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r16", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r17", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r18", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r19", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r20", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r21", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r22", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r23", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r24", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r25", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r26", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r27", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("r28", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("fp", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("lr", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("sp", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("pc", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_ARM64_INTEGER)),
  ("cpsr", ctypes.c_uint32),
  ("float_save", EnableOnFlag(MINIDUMP_FLOATING_SAVE_AREA_ARM.ctype,
                              MD_CONTEXT_ARM64_FLOATING_POINT))
])


402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
MD_CONTEXT_AMD64 = 0x00100000
MD_CONTEXT_AMD64_CONTROL = (MD_CONTEXT_AMD64 | 0x00000001)
MD_CONTEXT_AMD64_INTEGER = (MD_CONTEXT_AMD64 | 0x00000002)
MD_CONTEXT_AMD64_SEGMENTS = (MD_CONTEXT_AMD64 | 0x00000004)
MD_CONTEXT_AMD64_FLOATING_POINT = (MD_CONTEXT_AMD64 | 0x00000008)
MD_CONTEXT_AMD64_DEBUG_REGISTERS = (MD_CONTEXT_AMD64 | 0x00000010)

MINIDUMP_CONTEXT_AMD64 = Descriptor([
  ("p1_home", ctypes.c_uint64),
  ("p2_home", ctypes.c_uint64),
  ("p3_home", ctypes.c_uint64),
  ("p4_home", ctypes.c_uint64),
  ("p5_home", ctypes.c_uint64),
  ("p6_home", ctypes.c_uint64),
  ("context_flags", ctypes.c_uint32),
  ("mx_csr", ctypes.c_uint32),
  # MD_CONTEXT_AMD64_CONTROL.
  ("cs", EnableOnFlag(ctypes.c_uint16, MD_CONTEXT_AMD64_CONTROL)),
  # MD_CONTEXT_AMD64_SEGMENTS
  ("ds", EnableOnFlag(ctypes.c_uint16, MD_CONTEXT_AMD64_SEGMENTS)),
  ("es", EnableOnFlag(ctypes.c_uint16, MD_CONTEXT_AMD64_SEGMENTS)),
  ("fs", EnableOnFlag(ctypes.c_uint16, MD_CONTEXT_AMD64_SEGMENTS)),
  ("gs", EnableOnFlag(ctypes.c_uint16, MD_CONTEXT_AMD64_SEGMENTS)),
  # MD_CONTEXT_AMD64_CONTROL.
  ("ss", EnableOnFlag(ctypes.c_uint16, MD_CONTEXT_AMD64_CONTROL)),
  ("eflags", EnableOnFlag(ctypes.c_uint32, MD_CONTEXT_AMD64_CONTROL)),
  # MD_CONTEXT_AMD64_DEBUG_REGISTERS.
  ("dr0", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_DEBUG_REGISTERS)),
  ("dr1", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_DEBUG_REGISTERS)),
  ("dr2", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_DEBUG_REGISTERS)),
  ("dr3", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_DEBUG_REGISTERS)),
  ("dr6", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_DEBUG_REGISTERS)),
  ("dr7", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_DEBUG_REGISTERS)),
  # MD_CONTEXT_AMD64_INTEGER.
  ("rax", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("rcx", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("rdx", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("rbx", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  # MD_CONTEXT_AMD64_CONTROL.
  ("rsp", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_CONTROL)),
  # MD_CONTEXT_AMD64_INTEGER.
  ("rbp", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("rsi", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("rdi", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("r8", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("r9", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("r10", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("r11", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("r12", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("r13", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("r14", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  ("r15", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_INTEGER)),
  # MD_CONTEXT_AMD64_CONTROL.
  ("rip", EnableOnFlag(ctypes.c_uint64, MD_CONTEXT_AMD64_CONTROL)),
  # MD_CONTEXT_AMD64_FLOATING_POINT
  ("sse_registers", EnableOnFlag(ctypes.c_uint8 * (16 * 26),
                                 MD_CONTEXT_AMD64_FLOATING_POINT)),
  ("vector_registers", EnableOnFlag(ctypes.c_uint8 * (16 * 26),
                                    MD_CONTEXT_AMD64_FLOATING_POINT)),
  ("vector_control", EnableOnFlag(ctypes.c_uint64,
                                  MD_CONTEXT_AMD64_FLOATING_POINT)),
  # MD_CONTEXT_AMD64_DEBUG_REGISTERS.
  ("debug_control", EnableOnFlag(ctypes.c_uint64,
                                 MD_CONTEXT_AMD64_DEBUG_REGISTERS)),
  ("last_branch_to_rip", EnableOnFlag(ctypes.c_uint64,
                                      MD_CONTEXT_AMD64_DEBUG_REGISTERS)),
  ("last_branch_from_rip", EnableOnFlag(ctypes.c_uint64,
                                        MD_CONTEXT_AMD64_DEBUG_REGISTERS)),
  ("last_exception_to_rip", EnableOnFlag(ctypes.c_uint64,
                                         MD_CONTEXT_AMD64_DEBUG_REGISTERS)),
  ("last_exception_from_rip", EnableOnFlag(ctypes.c_uint64,
                                           MD_CONTEXT_AMD64_DEBUG_REGISTERS))
])

476 477 478 479 480 481 482 483 484 485 486 487 488 489 490
MINIDUMP_MEMORY_DESCRIPTOR = Descriptor([
  ("start", ctypes.c_uint64),
  ("memory", MINIDUMP_LOCATION_DESCRIPTOR.ctype)
])

MINIDUMP_MEMORY_DESCRIPTOR64 = Descriptor([
  ("start", ctypes.c_uint64),
  ("size", ctypes.c_uint64)
])

MINIDUMP_MEMORY_LIST = Descriptor([
  ("range_count", ctypes.c_uint32),
  ("ranges", lambda m: MINIDUMP_MEMORY_DESCRIPTOR.ctype * m.range_count)
])

491 492 493 494 495 496
MINIDUMP_MEMORY_LIST_Mac = Descriptor([
  ("range_count", ctypes.c_uint32),
  ("junk", ctypes.c_uint32),
  ("ranges", lambda m: MINIDUMP_MEMORY_DESCRIPTOR.ctype * m.range_count)
])

497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517
MINIDUMP_MEMORY_LIST64 = Descriptor([
  ("range_count", ctypes.c_uint64),
  ("base_rva", ctypes.c_uint64),
  ("ranges", lambda m: MINIDUMP_MEMORY_DESCRIPTOR64.ctype * m.range_count)
])

MINIDUMP_THREAD = Descriptor([
  ("id", ctypes.c_uint32),
  ("suspend_count", ctypes.c_uint32),
  ("priority_class", ctypes.c_uint32),
  ("priority", ctypes.c_uint32),
  ("ted", ctypes.c_uint64),
  ("stack", MINIDUMP_MEMORY_DESCRIPTOR.ctype),
  ("context", MINIDUMP_LOCATION_DESCRIPTOR.ctype)
])

MINIDUMP_THREAD_LIST = Descriptor([
  ("thread_count", ctypes.c_uint32),
  ("threads", lambda t: MINIDUMP_THREAD.ctype * t.thread_count)
])

518 519 520 521 522 523
MINIDUMP_THREAD_LIST_Mac = Descriptor([
  ("thread_count", ctypes.c_uint32),
  ("junk", ctypes.c_uint32),
  ("threads", lambda t: MINIDUMP_THREAD.ctype * t.thread_count)
])

524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539
MINIDUMP_VS_FIXEDFILEINFO = Descriptor([
  ("dwSignature", ctypes.c_uint32),
  ("dwStrucVersion", ctypes.c_uint32),
  ("dwFileVersionMS", ctypes.c_uint32),
  ("dwFileVersionLS", ctypes.c_uint32),
  ("dwProductVersionMS", ctypes.c_uint32),
  ("dwProductVersionLS", ctypes.c_uint32),
  ("dwFileFlagsMask", ctypes.c_uint32),
  ("dwFileFlags", ctypes.c_uint32),
  ("dwFileOS", ctypes.c_uint32),
  ("dwFileType", ctypes.c_uint32),
  ("dwFileSubtype", ctypes.c_uint32),
  ("dwFileDateMS", ctypes.c_uint32),
  ("dwFileDateLS", ctypes.c_uint32)
])

540 541 542 543 544 545
MINIDUMP_RAW_MODULE = Descriptor([
  ("base_of_image", ctypes.c_uint64),
  ("size_of_image", ctypes.c_uint32),
  ("checksum", ctypes.c_uint32),
  ("time_date_stamp", ctypes.c_uint32),
  ("module_name_rva", ctypes.c_uint32),
546
  ("version_info", MINIDUMP_VS_FIXEDFILEINFO.ctype),
547 548 549 550 551 552 553 554 555 556 557
  ("cv_record", MINIDUMP_LOCATION_DESCRIPTOR.ctype),
  ("misc_record", MINIDUMP_LOCATION_DESCRIPTOR.ctype),
  ("reserved0", ctypes.c_uint32 * 2),
  ("reserved1", ctypes.c_uint32 * 2)
])

MINIDUMP_MODULE_LIST = Descriptor([
  ("number_of_modules", ctypes.c_uint32),
  ("modules", lambda t: MINIDUMP_RAW_MODULE.ctype * t.number_of_modules)
])

558 559 560 561 562 563
MINIDUMP_MODULE_LIST_Mac = Descriptor([
  ("number_of_modules", ctypes.c_uint32),
  ("junk", ctypes.c_uint32),
  ("modules", lambda t: MINIDUMP_RAW_MODULE.ctype * t.number_of_modules)
])

564 565 566 567 568
MINIDUMP_RAW_SYSTEM_INFO = Descriptor([
  ("processor_architecture", ctypes.c_uint16)
])

MD_CPU_ARCHITECTURE_X86 = 0
569
MD_CPU_ARCHITECTURE_ARM = 5
570
MD_CPU_ARCHITECTURE_ARM64 = 0x8003
571
MD_CPU_ARCHITECTURE_AMD64 = 9
572

573 574 575 576 577 578 579 580 581 582 583 584 585 586
class FuncSymbol:
  def __init__(self, start, size, name):
    self.start = start
    self.end = self.start + size
    self.name = name

  def __cmp__(self, other):
    if isinstance(other, FuncSymbol):
      return self.start - other.start
    return self.start - other

  def Covers(self, addr):
    return (self.start <= addr) and (addr < self.end)

587 588 589 590 591 592 593 594 595 596 597
class MinidumpReader(object):
  """Minidump (.dmp) reader."""

  _HEADER_MAGIC = 0x504d444d

  def __init__(self, options, minidump_name):
    self.minidump_name = minidump_name
    self.minidump_file = open(minidump_name, "r")
    self.minidump = mmap.mmap(self.minidump_file.fileno(), 0, mmap.MAP_PRIVATE)
    self.header = MINIDUMP_HEADER.Read(self.minidump, 0)
    if self.header.signature != MinidumpReader._HEADER_MAGIC:
598
      print >>sys.stderr, "Warning: Unsupported minidump header magic!"
599 600 601 602 603 604
    DebugPrint(self.header)
    directories = []
    offset = self.header.stream_directories_rva
    for _ in xrange(self.header.stream_count):
      directories.append(MINIDUMP_DIRECTORY.Read(self.minidump, offset))
      offset += MINIDUMP_DIRECTORY.size
605
    self.arch = None
606 607 608
    self.exception = None
    self.exception_context = None
    self.memory_list = None
609
    self.memory_list64 = None
610
    self.module_list = None
611
    self.thread_map = {}
612

613 614 615 616
    self.symdir = options.symdir
    self.modules_with_symbols = []
    self.symbols = []

617 618 619 620 621 622
    # Find MDRawSystemInfo stream and determine arch.
    for d in directories:
      if d.stream_type == MD_SYSTEM_INFO_STREAM:
        system_info = MINIDUMP_RAW_SYSTEM_INFO.Read(
            self.minidump, d.location.rva)
        self.arch = system_info.processor_architecture
623 624
        assert self.arch in [MD_CPU_ARCHITECTURE_AMD64,
                             MD_CPU_ARCHITECTURE_ARM,
625
                             MD_CPU_ARCHITECTURE_ARM64,
626
                             MD_CPU_ARCHITECTURE_X86]
627 628
    assert not self.arch is None

629 630 631 632 633 634
    for d in directories:
      DebugPrint(d)
      if d.stream_type == MD_EXCEPTION_STREAM:
        self.exception = MINIDUMP_EXCEPTION_STREAM.Read(
          self.minidump, d.location.rva)
        DebugPrint(self.exception)
635 636 637 638 639 640
        if self.arch == MD_CPU_ARCHITECTURE_X86:
          self.exception_context = MINIDUMP_CONTEXT_X86.Read(
              self.minidump, self.exception.thread_context.rva)
        elif self.arch == MD_CPU_ARCHITECTURE_AMD64:
          self.exception_context = MINIDUMP_CONTEXT_AMD64.Read(
              self.minidump, self.exception.thread_context.rva)
641 642 643
        elif self.arch == MD_CPU_ARCHITECTURE_ARM:
          self.exception_context = MINIDUMP_CONTEXT_ARM.Read(
              self.minidump, self.exception.thread_context.rva)
644 645 646
        elif self.arch == MD_CPU_ARCHITECTURE_ARM64:
          self.exception_context = MINIDUMP_CONTEXT_ARM64.Read(
              self.minidump, self.exception.thread_context.rva)
647 648 649
        DebugPrint(self.exception_context)
      elif d.stream_type == MD_THREAD_LIST_STREAM:
        thread_list = MINIDUMP_THREAD_LIST.Read(self.minidump, d.location.rva)
650 651 652
        if ctypes.sizeof(thread_list) + 4 == d.location.data_size:
          thread_list = MINIDUMP_THREAD_LIST_Mac.Read(
              self.minidump, d.location.rva)
653 654 655 656 657
        assert ctypes.sizeof(thread_list) == d.location.data_size
        DebugPrint(thread_list)
        for thread in thread_list.threads:
          DebugPrint(thread)
          self.thread_map[thread.id] = thread
658 659 660 661
      elif d.stream_type == MD_MODULE_LIST_STREAM:
        assert self.module_list is None
        self.module_list = MINIDUMP_MODULE_LIST.Read(
          self.minidump, d.location.rva)
662 663 664
        if ctypes.sizeof(self.module_list) + 4 == d.location.data_size:
          self.module_list = MINIDUMP_MODULE_LIST_Mac.Read(
              self.minidump, d.location.rva)
665
        assert ctypes.sizeof(self.module_list) == d.location.data_size
666
        DebugPrint(self.module_list)
667
      elif d.stream_type == MD_MEMORY_LIST_STREAM:
668
        print >>sys.stderr, "Warning: This is not a full minidump!"
669
        assert self.memory_list is None
670
        self.memory_list = MINIDUMP_MEMORY_LIST.Read(
671
          self.minidump, d.location.rva)
672 673 674
        if ctypes.sizeof(self.memory_list) + 4 == d.location.data_size:
          self.memory_list = MINIDUMP_MEMORY_LIST_Mac.Read(
              self.minidump, d.location.rva)
675 676
        assert ctypes.sizeof(self.memory_list) == d.location.data_size
        DebugPrint(self.memory_list)
677 678 679 680 681 682
      elif d.stream_type == MD_MEMORY_64_LIST_STREAM:
        assert self.memory_list64 is None
        self.memory_list64 = MINIDUMP_MEMORY_LIST64.Read(
          self.minidump, d.location.rva)
        assert ctypes.sizeof(self.memory_list64) == d.location.data_size
        DebugPrint(self.memory_list64)
683 684 685 686 687 688 689 690 691 692 693 694

  def IsValidAddress(self, address):
    return self.FindLocation(address) is not None

  def ReadU8(self, address):
    location = self.FindLocation(address)
    return ctypes.c_uint8.from_buffer(self.minidump, location).value

  def ReadU32(self, address):
    location = self.FindLocation(address)
    return ctypes.c_uint32.from_buffer(self.minidump, location).value

695 696 697 698 699 700 701
  def ReadU64(self, address):
    location = self.FindLocation(address)
    return ctypes.c_uint64.from_buffer(self.minidump, location).value

  def ReadUIntPtr(self, address):
    if self.arch == MD_CPU_ARCHITECTURE_AMD64:
      return self.ReadU64(address)
702 703
    elif self.arch == MD_CPU_ARCHITECTURE_ARM:
      return self.ReadU32(address)
704 705
    elif self.arch == MD_CPU_ARCHITECTURE_ARM64:
      return self.ReadU64(address)
706 707 708
    elif self.arch == MD_CPU_ARCHITECTURE_X86:
      return self.ReadU32(address)

709 710 711 712
  def ReadBytes(self, address, size):
    location = self.FindLocation(address)
    return self.minidump[location:location + size]

713 714 715
  def _ReadWord(self, location):
    if self.arch == MD_CPU_ARCHITECTURE_AMD64:
      return ctypes.c_uint64.from_buffer(self.minidump, location).value
716 717
    elif self.arch == MD_CPU_ARCHITECTURE_ARM:
      return ctypes.c_uint32.from_buffer(self.minidump, location).value
718 719
    elif self.arch == MD_CPU_ARCHITECTURE_ARM64:
      return ctypes.c_uint64.from_buffer(self.minidump, location).value
720 721 722
    elif self.arch == MD_CPU_ARCHITECTURE_X86:
      return ctypes.c_uint32.from_buffer(self.minidump, location).value

723 724 725
  def IsProbableASCIIRegion(self, location, length):
    ascii_bytes = 0
    non_ascii_bytes = 0
726 727
    for i in xrange(length):
      loc = location + i
728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747
      byte = ctypes.c_uint8.from_buffer(self.minidump, loc).value
      if byte >= 0x7f:
        non_ascii_bytes += 1
      if byte < 0x20 and byte != 0:
        non_ascii_bytes += 1
      if byte < 0x7f and byte >= 0x20:
        ascii_bytes += 1
      if byte == 0xa:  # newline
        ascii_bytes += 1
    if ascii_bytes * 10 <= length:
      return False
    if length > 0 and ascii_bytes > non_ascii_bytes * 7:
      return True
    if ascii_bytes > non_ascii_bytes * 3:
      return None  # Maybe
    return False

  def IsProbableExecutableRegion(self, location, length):
    opcode_bytes = 0
    sixty_four = self.arch == MD_CPU_ARCHITECTURE_AMD64
748 749
    for i in xrange(length):
      loc = location + i
750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782
      byte = ctypes.c_uint8.from_buffer(self.minidump, loc).value
      if (byte == 0x8b or           # mov
          byte == 0x89 or           # mov reg-reg
          (byte & 0xf0) == 0x50 or  # push/pop
          (sixty_four and (byte & 0xf0) == 0x40) or  # rex prefix
          byte == 0xc3 or           # return
          byte == 0x74 or           # jeq
          byte == 0x84 or           # jeq far
          byte == 0x75 or           # jne
          byte == 0x85 or           # jne far
          byte == 0xe8 or           # call
          byte == 0xe9 or           # jmp far
          byte == 0xeb):            # jmp near
        opcode_bytes += 1
    opcode_percent = (opcode_bytes * 100) / length
    threshold = 20
    if opcode_percent > threshold + 2:
      return True
    if opcode_percent > threshold - 2:
      return None  # Maybe
    return False

  def FindRegion(self, addr):
    answer = [-1, -1]
    def is_in(reader, start, size, location):
      if addr >= start and addr < start + size:
        answer[0] = start
        answer[1] = size
    self.ForEachMemoryRegion(is_in)
    if answer[0] == -1:
      return None
    return answer

783 784 785 786 787 788 789 790 791 792 793
  def ForEachMemoryRegion(self, cb):
    if self.memory_list64 is not None:
      for r in self.memory_list64.ranges:
        location = self.memory_list64.base_rva + offset
        cb(self, r.start, r.size, location)
        offset += r.size

    if self.memory_list is not None:
      for r in self.memory_list.ranges:
        cb(self, r.start, r.memory.data_size, r.memory.rva)

794
  def FindWord(self, word, alignment=0):
795
    def search_inside_region(reader, start, size, location):
796
      location = (location + alignment) & ~alignment
797 798
      for i in xrange(size - self.PointerSize()):
        loc = location + i
799 800 801 802 803 804
        if reader._ReadWord(loc) == word:
          slot = start + (loc - location)
          print "%s: %s" % (reader.FormatIntPtr(slot),
                            reader.FormatIntPtr(word))
    self.ForEachMemoryRegion(search_inside_region)

805 806 807 808
  def FindWordList(self, word):
    aligned_res = []
    unaligned_res = []
    def search_inside_region(reader, start, size, location):
809 810
      for i in xrange(size - self.PointerSize()):
        loc = location + i
811 812 813 814 815 816 817 818 819
        if reader._ReadWord(loc) == word:
          slot = start + (loc - location)
          if slot % self.PointerSize() == 0:
            aligned_res.append(slot)
          else:
            unaligned_res.append(slot)
    self.ForEachMemoryRegion(search_inside_region)
    return (aligned_res, unaligned_res)

820 821
  def FindLocation(self, address):
    offset = 0
822 823 824 825
    if self.memory_list64 is not None:
      for r in self.memory_list64.ranges:
        if r.start <= address < r.start + r.size:
          return self.memory_list64.base_rva + offset + address - r.start
826
        offset += r.size
827 828 829 830
    if self.memory_list is not None:
      for r in self.memory_list.ranges:
        if r.start <= address < r.start + r.memory.data_size:
          return r.memory.rva + address - r.start
831 832 833
    return None

  def GetDisasmLines(self, address, size):
834 835 836 837
    def CountUndefinedInstructions(lines):
      pattern = "<UNDEFINED>"
      return sum([line.count(pattern) for (ignore, line) in lines])

838 839
    location = self.FindLocation(address)
    if location is None: return []
840
    arch = None
841
    possible_objdump_flags = [""]
842 843
    if self.arch == MD_CPU_ARCHITECTURE_X86:
      arch = "ia32"
844 845 846
    elif self.arch == MD_CPU_ARCHITECTURE_ARM:
      arch = "arm"
      possible_objdump_flags = ["", "--disassembler-options=force-thumb"]
847 848 849
    elif self.arch == MD_CPU_ARCHITECTURE_ARM64:
      arch = "arm64"
      possible_objdump_flags = ["", "--disassembler-options=force-thumb"]
850 851
    elif self.arch == MD_CPU_ARCHITECTURE_AMD64:
      arch = "x64"
852 853 854 855 856 857 858 859
    results = [ disasm.GetDisasmLines(self.minidump_name,
                                     location,
                                     size,
                                     arch,
                                     False,
                                     objdump_flags)
                for objdump_flags in possible_objdump_flags ]
    return min(results, key=CountUndefinedInstructions)
860 861 862 863 864 865


  def Dispose(self):
    self.minidump.close()
    self.minidump_file.close()

866 867 868
  def ExceptionIP(self):
    if self.arch == MD_CPU_ARCHITECTURE_AMD64:
      return self.exception_context.rip
869 870
    elif self.arch == MD_CPU_ARCHITECTURE_ARM:
      return self.exception_context.pc
871 872
    elif self.arch == MD_CPU_ARCHITECTURE_ARM64:
      return self.exception_context.pc
873 874 875 876 877 878
    elif self.arch == MD_CPU_ARCHITECTURE_X86:
      return self.exception_context.eip

  def ExceptionSP(self):
    if self.arch == MD_CPU_ARCHITECTURE_AMD64:
      return self.exception_context.rsp
879 880
    elif self.arch == MD_CPU_ARCHITECTURE_ARM:
      return self.exception_context.sp
881 882
    elif self.arch == MD_CPU_ARCHITECTURE_ARM64:
      return self.exception_context.sp
883
    elif self.arch == MD_CPU_ARCHITECTURE_X86:
884
      return self.exception_context.esp
885

886 887 888 889 890
  def ExceptionFP(self):
    if self.arch == MD_CPU_ARCHITECTURE_AMD64:
      return self.exception_context.rbp
    elif self.arch == MD_CPU_ARCHITECTURE_ARM:
      return None
891 892
    elif self.arch == MD_CPU_ARCHITECTURE_ARM64:
      return self.exception_context.fp
893 894 895
    elif self.arch == MD_CPU_ARCHITECTURE_X86:
      return self.exception_context.ebp

896 897 898
  def FormatIntPtr(self, value):
    if self.arch == MD_CPU_ARCHITECTURE_AMD64:
      return "%016x" % value
899 900
    elif self.arch == MD_CPU_ARCHITECTURE_ARM:
      return "%08x" % value
901 902
    elif self.arch == MD_CPU_ARCHITECTURE_ARM64:
      return "%016x" % value
903 904 905 906 907 908
    elif self.arch == MD_CPU_ARCHITECTURE_X86:
      return "%08x" % value

  def PointerSize(self):
    if self.arch == MD_CPU_ARCHITECTURE_AMD64:
      return 8
909 910
    elif self.arch == MD_CPU_ARCHITECTURE_ARM:
      return 4
911 912
    elif self.arch == MD_CPU_ARCHITECTURE_ARM64:
      return 8
913 914 915 916 917 918
    elif self.arch == MD_CPU_ARCHITECTURE_X86:
      return 4

  def Register(self, name):
    return self.exception_context.__getattribute__(name)

919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946
  def ReadMinidumpString(self, rva):
    string = bytearray(MINIDUMP_STRING.Read(self.minidump, rva).buffer)
    string = string.decode("utf16")
    return string[0:len(string) - 1]

  # Load FUNC records from a BreakPad symbol file
  #
  #    http://code.google.com/p/google-breakpad/wiki/SymbolFiles
  #
  def _LoadSymbolsFrom(self, symfile, baseaddr):
    print "Loading symbols from %s" % (symfile)
    funcs = []
    with open(symfile) as f:
      for line in f:
        result = re.match(
            r"^FUNC ([a-f0-9]+) ([a-f0-9]+) ([a-f0-9]+) (.*)$", line)
        if result is not None:
          start = int(result.group(1), 16)
          size = int(result.group(2), 16)
          name = result.group(4).rstrip()
          bisect.insort_left(self.symbols,
                             FuncSymbol(baseaddr + start, size, name))
    print " ... done"

  def TryLoadSymbolsFor(self, modulename, module):
    try:
      symfile = os.path.join(self.symdir,
                             modulename.replace('.', '_') + ".pdb.sym")
947 948 949
      if os.path.isfile(symfile):
        self._LoadSymbolsFrom(symfile, module.base_of_image)
        self.modules_with_symbols.append(module)
950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979
    except Exception as e:
      print "  ... failure (%s)" % (e)

  # Returns true if address is covered by some module that has loaded symbols.
  def _IsInModuleWithSymbols(self, addr):
    for module in self.modules_with_symbols:
      start = module.base_of_image
      end = start + module.size_of_image
      if (start <= addr) and (addr < end):
        return True
    return False

  # Find symbol covering the given address and return its name in format
  #     <symbol name>+<offset from the start>
  def FindSymbol(self, addr):
    if not self._IsInModuleWithSymbols(addr):
      return None

    i = bisect.bisect_left(self.symbols, addr)
    symbol = None
    if (0 < i) and self.symbols[i - 1].Covers(addr):
      symbol = self.symbols[i - 1]
    elif (i < len(self.symbols)) and self.symbols[i].Covers(addr):
      symbol = self.symbols[i]
    else:
      return None
    diff = addr - symbol.start
    return "%s+0x%x" % (symbol.name, diff)


980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012
class Printer(object):
  """Printer with indentation support."""

  def __init__(self):
    self.indent = 0

  def Indent(self):
    self.indent += 2

  def Dedent(self):
    self.indent -= 2

  def Print(self, string):
    print "%s%s" % (self._IndentString(), string)

  def PrintLines(self, lines):
    indent = self._IndentString()
    print "\n".join("%s%s" % (indent, line) for line in lines)

  def _IndentString(self):
    return self.indent * " "


ADDRESS_RE = re.compile(r"0x[0-9a-fA-F]+")


def FormatDisasmLine(start, heap, line):
  line_address = start + line[0]
  stack_slot = heap.stack_map.get(line_address)
  marker = "  "
  if stack_slot:
    marker = "=>"
  code = AnnotateAddresses(heap, line[1])
1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025

  # Compute the actual call target which the disassembler is too stupid
  # to figure out (it adds the call offset to the disassembly offset rather
  # than the absolute instruction address).
  if heap.reader.arch == MD_CPU_ARCHITECTURE_X86:
    if code.startswith("e8"):
      words = code.split()
      if len(words) > 6 and words[5] == "call":
        offset = int(words[4] + words[3] + words[2] + words[1], 16)
        target = (line_address + offset + 5) & 0xFFFFFFFF
        code = code.replace(words[6], "0x%08x" % target)
  # TODO(jkummerow): port this hack to ARM and x64.

1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052
  return "%s%08x %08x: %s" % (marker, line_address, line[0], code)


def AnnotateAddresses(heap, line):
  extra = []
  for m in ADDRESS_RE.finditer(line):
    maybe_address = int(m.group(0), 16)
    object = heap.FindObject(maybe_address)
    if not object: continue
    extra.append(str(object))
  if len(extra) == 0: return line
  return "%s  ;; %s" % (line, ", ".join(extra))


class HeapObject(object):
  def __init__(self, heap, map, address):
    self.heap = heap
    self.map = map
    self.address = address

  def Is(self, cls):
    return isinstance(self, cls)

  def Print(self, p):
    p.Print(str(self))

  def __str__(self):
1053 1054 1055
    instance_type = "???"
    if self.map is not None:
      instance_type = INSTANCE_TYPES[self.map.instance_type]
1056
    return "HeapObject(%s, %s)" % (self.heap.reader.FormatIntPtr(self.address),
1057
                                   instance_type)
1058 1059

  def ObjectField(self, offset):
1060
    field_value = self.heap.reader.ReadUIntPtr(self.address + offset)
1061 1062 1063
    return self.heap.FindObjectOrSmi(field_value)

  def SmiField(self, offset):
1064
    field_value = self.heap.reader.ReadUIntPtr(self.address + offset)
1065 1066 1067
    if (field_value & 1) == 0:
      return field_value / 2
    return None
1068 1069 1070


class Map(HeapObject):
1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093
  def Decode(self, offset, size, value):
    return (value >> offset) & ((1 << size) - 1)

  # Instance Sizes
  def InstanceSizesOffset(self):
    return self.heap.PointerSize()

  def InstanceSizeOffset(self):
    return self.InstanceSizesOffset()

  def InObjectProperties(self):
    return self.InstanceSizeOffset() + 1

  def PreAllocatedPropertyFields(self):
    return self.InObjectProperties() + 1

  def VisitorId(self):
    return self.PreAllocatedPropertyFields() + 1

  # Instance Attributes
  def InstanceAttributesOffset(self):
    return self.InstanceSizesOffset() + self.heap.IntSize()

1094
  def InstanceTypeOffset(self):
1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148
    return self.InstanceAttributesOffset()

  def UnusedPropertyFieldsOffset(self):
    return self.InstanceTypeOffset() + 1

  def BitFieldOffset(self):
    return self.UnusedPropertyFieldsOffset() + 1

  def BitField2Offset(self):
    return self.BitFieldOffset() + 1

  # Other fields
  def PrototypeOffset(self):
    return self.InstanceAttributesOffset() + self.heap.IntSize()

  def ConstructorOffset(self):
    return self.PrototypeOffset() + self.heap.PointerSize()

  def TransitionsOrBackPointerOffset(self):
    return self.ConstructorOffset() + self.heap.PointerSize()

  def DescriptorsOffset(self):
    return self.TransitionsOrBackPointerOffset() + self.heap.PointerSize()

  def CodeCacheOffset(self):
    return self.DescriptorsOffset() + self.heap.PointerSize()

  def DependentCodeOffset(self):
    return self.CodeCacheOffset() + self.heap.PointerSize()

  def BitField3Offset(self):
    return self.DependentCodeOffset() + self.heap.PointerSize()

  def ReadByte(self, offset):
    return self.heap.reader.ReadU8(self.address + offset)

  def Print(self, p):
    p.Print("Map(%08x)" % (self.address))
    p.Print("- size: %d, inobject: %d, preallocated: %d, visitor: %d" % (
        self.ReadByte(self.InstanceSizeOffset()),
        self.ReadByte(self.InObjectProperties()),
        self.ReadByte(self.PreAllocatedPropertyFields()),
        self.VisitorId()))

    bitfield = self.ReadByte(self.BitFieldOffset())
    bitfield2 = self.ReadByte(self.BitField2Offset())
    p.Print("- %s, unused: %d, bf: %d, bf2: %d" % (
        INSTANCE_TYPES[self.ReadByte(self.InstanceTypeOffset())],
        self.ReadByte(self.UnusedPropertyFieldsOffset()),
        bitfield, bitfield2))

    p.Print("- kind: %s" % (self.Decode(3, 5, bitfield2)))

    bitfield3 = self.ObjectField(self.BitField3Offset())
1149 1150 1151 1152 1153
    p.Print(
        "- EnumLength: %d NumberOfOwnDescriptors: %d OwnsDescriptors: %s" % (
            self.Decode(0, 11, bitfield3),
            self.Decode(11, 11, bitfield3),
            self.Decode(25, 1, bitfield3)))
1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168
    p.Print("- IsShared: %s" % (self.Decode(22, 1, bitfield3)))
    p.Print("- FunctionWithPrototype: %s" % (self.Decode(23, 1, bitfield3)))
    p.Print("- DictionaryMap: %s" % (self.Decode(24, 1, bitfield3)))

    descriptors = self.ObjectField(self.DescriptorsOffset())
    if descriptors.__class__ == FixedArray:
      DescriptorArray(descriptors).Print(p)
    else:
      p.Print("Descriptors: %s" % (descriptors))

    transitions = self.ObjectField(self.TransitionsOrBackPointerOffset())
    if transitions.__class__ == FixedArray:
      TransitionArray(transitions).Print(p)
    else:
      p.Print("TransitionsOrBackPointer: %s" % (transitions))
1169 1170 1171 1172

  def __init__(self, heap, map, address):
    HeapObject.__init__(self, heap, map, address)
    self.instance_type = \
1173
        heap.reader.ReadU8(self.address + self.InstanceTypeOffset())
1174 1175 1176


class String(HeapObject):
1177
  def LengthOffset(self):
1178 1179
    # First word after the map is the hash, the second is the length.
    return self.heap.PointerSize() * 2
1180 1181 1182

  def __init__(self, heap, map, address):
    HeapObject.__init__(self, heap, map, address)
1183
    self.length = self.SmiField(self.LengthOffset())
1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195

  def GetChars(self):
    return "?string?"

  def Print(self, p):
    p.Print(str(self))

  def __str__(self):
    return "\"%s\"" % self.GetChars()


class SeqString(String):
1196 1197
  def CharsOffset(self):
    return self.heap.PointerSize() * 3
1198 1199 1200

  def __init__(self, heap, map, address):
    String.__init__(self, heap, map, address)
1201
    self.chars = heap.reader.ReadBytes(self.address + self.CharsOffset(),
1202 1203 1204 1205 1206 1207 1208
                                       self.length)

  def GetChars(self):
    return self.chars


class ExternalString(String):
1209
  # TODO(vegorov) fix ExternalString for X64 architecture
1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238
  RESOURCE_OFFSET = 12

  WEBKIT_RESOUCE_STRING_IMPL_OFFSET = 4
  WEBKIT_STRING_IMPL_CHARS_OFFSET = 8

  def __init__(self, heap, map, address):
    String.__init__(self, heap, map, address)
    reader = heap.reader
    self.resource = \
        reader.ReadU32(self.address + ExternalString.RESOURCE_OFFSET)
    self.chars = "?external string?"
    if not reader.IsValidAddress(self.resource): return
    string_impl_address = self.resource + \
        ExternalString.WEBKIT_RESOUCE_STRING_IMPL_OFFSET
    if not reader.IsValidAddress(string_impl_address): return
    string_impl = reader.ReadU32(string_impl_address)
    chars_ptr_address = string_impl + \
        ExternalString.WEBKIT_STRING_IMPL_CHARS_OFFSET
    if not reader.IsValidAddress(chars_ptr_address): return
    chars_ptr = reader.ReadU32(chars_ptr_address)
    if not reader.IsValidAddress(chars_ptr): return
    raw_chars = reader.ReadBytes(chars_ptr, 2 * self.length)
    self.chars = codecs.getdecoder("utf16")(raw_chars)[0]

  def GetChars(self):
    return self.chars


class ConsString(String):
1239 1240 1241 1242 1243
  def LeftOffset(self):
    return self.heap.PointerSize() * 3

  def RightOffset(self):
    return self.heap.PointerSize() * 4
1244 1245 1246

  def __init__(self, heap, map, address):
    String.__init__(self, heap, map, address)
1247 1248
    self.left = self.ObjectField(self.LeftOffset())
    self.right = self.ObjectField(self.RightOffset())
1249 1250

  def GetChars(self):
1251 1252 1253 1254
    try:
      return self.left.GetChars() + self.right.GetChars()
    except:
      return "***CAUGHT EXCEPTION IN GROKDUMP***"
1255 1256 1257


class Oddball(HeapObject):
1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268
  # Should match declarations in objects.h
  KINDS = [
    "False",
    "True",
    "TheHole",
    "Null",
    "ArgumentMarker",
    "Undefined",
    "Other"
  ]

1269 1270
  def ToStringOffset(self):
    return self.heap.PointerSize()
1271

1272 1273 1274 1275 1276 1277
  def ToNumberOffset(self):
    return self.ToStringOffset() + self.heap.PointerSize()

  def KindOffset(self):
    return self.ToNumberOffset() + self.heap.PointerSize()

1278 1279
  def __init__(self, heap, map, address):
    HeapObject.__init__(self, heap, map, address)
1280
    self.to_string = self.ObjectField(self.ToStringOffset())
1281
    self.kind = self.SmiField(self.KindOffset())
1282 1283 1284 1285 1286

  def Print(self, p):
    p.Print(str(self))

  def __str__(self):
1287
    if self.to_string:
1288
      return "Oddball(%08x, <%s>)" % (self.address, str(self.to_string))
1289
    else:
1290 1291 1292 1293
      kind = "???"
      if 0 <= self.kind < len(Oddball.KINDS):
        kind = Oddball.KINDS[self.kind]
      return "Oddball(%08x, kind=%s)" % (self.address, kind)
1294 1295 1296


class FixedArray(HeapObject):
1297 1298 1299 1300 1301
  def LengthOffset(self):
    return self.heap.PointerSize()

  def ElementsOffset(self):
    return self.heap.PointerSize() * 2
1302

1303 1304 1305 1306 1307 1308
  def MemberOffset(self, i):
    return self.ElementsOffset() + self.heap.PointerSize() * i

  def Get(self, i):
    return self.ObjectField(self.MemberOffset(i))

1309 1310
  def __init__(self, heap, map, address):
    HeapObject.__init__(self, heap, map, address)
1311
    self.length = self.SmiField(self.LengthOffset())
1312 1313

  def Print(self, p):
1314
    p.Print("FixedArray(%s) {" % self.heap.reader.FormatIntPtr(self.address))
1315 1316
    p.Indent()
    p.Print("length: %d" % self.length)
1317
    base_offset = self.ElementsOffset()
1318
    for i in xrange(self.length):
1319
      offset = base_offset + 4 * i
1320 1321 1322 1323 1324 1325 1326
      try:
        p.Print("[%08d] = %s" % (i, self.ObjectField(offset)))
      except TypeError:
        p.Dedent()
        p.Print("...")
        p.Print("}")
        return
1327 1328 1329 1330 1331 1332 1333
    p.Dedent()
    p.Print("}")

  def __str__(self):
    return "FixedArray(%08x, length=%d)" % (self.address, self.length)


1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364
class DescriptorArray(object):
  def __init__(self, array):
    self.array = array

  def Length(self):
    return self.array.Get(0)

  def Decode(self, offset, size, value):
    return (value >> offset) & ((1 << size) - 1)

  TYPES = [
      "normal",
      "field",
      "function",
      "callbacks"
  ]

  def Type(self, value):
    return DescriptorArray.TYPES[self.Decode(0, 3, value)]

  def Attributes(self, value):
    attributes = self.Decode(3, 3, value)
    result = []
    if (attributes & 0): result += ["ReadOnly"]
    if (attributes & 1): result += ["DontEnum"]
    if (attributes & 2): result += ["DontDelete"]
    return "[" + (",".join(result)) + "]"

  def Deleted(self, value):
    return self.Decode(6, 1, value) == 1

1365 1366
  def FieldIndex(self, value):
    return self.Decode(20, 11, value)
1367 1368

  def Pointer(self, value):
1369
    return self.Decode(6, 11, value)
1370 1371 1372 1373 1374 1375

  def Details(self, di, value):
    return (
        di,
        self.Type(value),
        self.Attributes(value),
1376
        self.FieldIndex(value),
1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391
        self.Pointer(value)
    )


  def Print(self, p):
    length = self.Length()
    array = self.array

    p.Print("Descriptors(%08x, length=%d)" % (array.address, length))
    p.Print("[et] %s" % (array.Get(1)))

    for di in xrange(length):
      i = 2 + di * 3
      p.Print("0x%x" % (array.address + array.MemberOffset(i)))
      p.Print("[%i] name:    %s" % (di, array.Get(i + 0)))
1392
      p.Print("[%i] details: %s %s field-index %i pointer %i" % \
1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438
              self.Details(di, array.Get(i + 1)))
      p.Print("[%i] value:   %s" % (di, array.Get(i + 2)))

    end = self.array.length // 3
    if length != end:
      p.Print("[%i-%i] slack descriptors" % (length, end))


class TransitionArray(object):
  def __init__(self, array):
    self.array = array

  def IsSimpleTransition(self):
    return self.array.length <= 2

  def Length(self):
    # SimpleTransition cases
    if self.IsSimpleTransition():
      return self.array.length - 1
    return (self.array.length - 3) // 2

  def Print(self, p):
    length = self.Length()
    array = self.array

    p.Print("Transitions(%08x, length=%d)" % (array.address, length))
    p.Print("[backpointer] %s" % (array.Get(0)))
    if self.IsSimpleTransition():
      if length == 1:
        p.Print("[simple target] %s" % (array.Get(1)))
      return

    elements = array.Get(1)
    if elements is not None:
      p.Print("[elements   ] %s" % (elements))

    prototype = array.Get(2)
    if prototype is not None:
      p.Print("[prototype  ] %s" % (prototype))

    for di in xrange(length):
      i = 3 + di * 2
      p.Print("[%i] symbol: %s" % (di, array.Get(i + 0)))
      p.Print("[%i] target: %s" % (di, array.Get(i + 1)))


1439
class JSFunction(HeapObject):
1440 1441 1442 1443 1444
  def CodeEntryOffset(self):
    return 3 * self.heap.PointerSize()

  def SharedOffset(self):
    return 5 * self.heap.PointerSize()
1445 1446 1447 1448

  def __init__(self, heap, map, address):
    HeapObject.__init__(self, heap, map, address)
    code_entry = \
1449 1450 1451
        heap.reader.ReadU32(self.address + self.CodeEntryOffset())
    self.code = heap.FindObject(code_entry - Code.HeaderSize(heap) + 1)
    self.shared = self.ObjectField(self.SharedOffset())
1452 1453 1454

  def Print(self, p):
    source = "\n".join("  %s" % line for line in self._GetSource().split("\n"))
1455
    p.Print("JSFunction(%s) {" % self.heap.reader.FormatIntPtr(self.address))
1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471
    p.Indent()
    p.Print("inferred name: %s" % self.shared.inferred_name)
    if self.shared.script.Is(Script) and self.shared.script.name.Is(String):
      p.Print("script name: %s" % self.shared.script.name)
    p.Print("source:")
    p.PrintLines(self._GetSource().split("\n"))
    p.Print("code:")
    self.code.Print(p)
    if self.code != self.shared.code:
      p.Print("unoptimized code:")
      self.shared.code.Print(p)
    p.Dedent()
    p.Print("}")

  def __str__(self):
    inferred_name = ""
1472
    if self.shared is not None and self.shared.Is(SharedFunctionInfo):
1473
      inferred_name = self.shared.inferred_name
1474
    return "JSFunction(%s, %s) " % \
1475
          (self.heap.reader.FormatIntPtr(self.address), inferred_name)
1476 1477 1478 1479 1480 1481 1482 1483

  def _GetSource(self):
    source = "?source?"
    start = self.shared.start_position
    end = self.shared.end_position
    if not self.shared.script.Is(Script): return source
    script_source = self.shared.script.source
    if not script_source.Is(String): return source
1484 1485 1486
    if start and end:
      source = script_source.GetChars()[start:end]
    return source
1487 1488 1489


class SharedFunctionInfo(HeapObject):
1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503
  def CodeOffset(self):
    return 2 * self.heap.PointerSize()

  def ScriptOffset(self):
    return 7 * self.heap.PointerSize()

  def InferredNameOffset(self):
    return 9 * self.heap.PointerSize()

  def EndPositionOffset(self):
    return 12 * self.heap.PointerSize() + 4 * self.heap.IntSize()

  def StartPositionAndTypeOffset(self):
    return 12 * self.heap.PointerSize() + 5 * self.heap.IntSize()
1504 1505 1506

  def __init__(self, heap, map, address):
    HeapObject.__init__(self, heap, map, address)
1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519
    self.code = self.ObjectField(self.CodeOffset())
    self.script = self.ObjectField(self.ScriptOffset())
    self.inferred_name = self.ObjectField(self.InferredNameOffset())
    if heap.PointerSize() == 8:
      start_position_and_type = \
          heap.reader.ReadU32(self.StartPositionAndTypeOffset())
      self.start_position = start_position_and_type >> 2
      pseudo_smi_end_position = \
          heap.reader.ReadU32(self.EndPositionOffset())
      self.end_position = pseudo_smi_end_position >> 2
    else:
      start_position_and_type = \
          self.SmiField(self.StartPositionAndTypeOffset())
1520 1521 1522 1523
      if start_position_and_type:
        self.start_position = start_position_and_type >> 2
      else:
        self.start_position = None
1524 1525
      self.end_position = \
          self.SmiField(self.EndPositionOffset())
1526 1527 1528


class Script(HeapObject):
1529 1530 1531 1532 1533
  def SourceOffset(self):
    return self.heap.PointerSize()

  def NameOffset(self):
    return self.SourceOffset() + self.heap.PointerSize()
1534 1535 1536

  def __init__(self, heap, map, address):
    HeapObject.__init__(self, heap, map, address)
1537 1538
    self.source = self.ObjectField(self.SourceOffset())
    self.name = self.ObjectField(self.NameOffset())
1539 1540


1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561
class CodeCache(HeapObject):
  def DefaultCacheOffset(self):
    return self.heap.PointerSize()

  def NormalTypeCacheOffset(self):
    return self.DefaultCacheOffset() + self.heap.PointerSize()

  def __init__(self, heap, map, address):
    HeapObject.__init__(self, heap, map, address)
    self.default_cache = self.ObjectField(self.DefaultCacheOffset())
    self.normal_type_cache = self.ObjectField(self.NormalTypeCacheOffset())

  def Print(self, p):
    p.Print("CodeCache(%s) {" % self.heap.reader.FormatIntPtr(self.address))
    p.Indent()
    p.Print("default cache: %s" % self.default_cache)
    p.Print("normal type cache: %s" % self.normal_type_cache)
    p.Dedent()
    p.Print("}")


1562
class Code(HeapObject):
1563 1564 1565 1566 1567 1568 1569 1570 1571
  CODE_ALIGNMENT_MASK = (1 << 5) - 1

  def InstructionSizeOffset(self):
    return self.heap.PointerSize()

  @staticmethod
  def HeaderSize(heap):
    return (heap.PointerSize() + heap.IntSize() + \
        4 * heap.PointerSize() + 3 * heap.IntSize() + \
1572
        Code.CODE_ALIGNMENT_MASK) & ~Code.CODE_ALIGNMENT_MASK
1573 1574 1575

  def __init__(self, heap, map, address):
    HeapObject.__init__(self, heap, map, address)
1576
    self.entry = self.address + Code.HeaderSize(heap)
1577
    self.instruction_size = \
1578
        heap.reader.ReadU32(self.address + self.InstructionSizeOffset())
1579 1580 1581

  def Print(self, p):
    lines = self.heap.reader.GetDisasmLines(self.entry, self.instruction_size)
1582
    p.Print("Code(%s) {" % self.heap.reader.FormatIntPtr(self.address))
1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595
    p.Indent()
    p.Print("instruction_size: %d" % self.instruction_size)
    p.PrintLines(self._FormatLine(line) for line in lines)
    p.Dedent()
    p.Print("}")

  def _FormatLine(self, line):
    return FormatDisasmLine(self.entry, self.heap, line)


class V8Heap(object):
  CLASS_MAP = {
    "SYMBOL_TYPE": SeqString,
1596
    "ONE_BYTE_SYMBOL_TYPE": SeqString,
1597
    "CONS_SYMBOL_TYPE": ConsString,
1598
    "CONS_ONE_BYTE_SYMBOL_TYPE": ConsString,
1599
    "EXTERNAL_SYMBOL_TYPE": ExternalString,
1600 1601
    "EXTERNAL_SYMBOL_WITH_ONE_BYTE_DATA_TYPE": ExternalString,
    "EXTERNAL_ONE_BYTE_SYMBOL_TYPE": ExternalString,
1602
    "SHORT_EXTERNAL_SYMBOL_TYPE": ExternalString,
1603 1604
    "SHORT_EXTERNAL_SYMBOL_WITH_ONE_BYTE_DATA_TYPE": ExternalString,
    "SHORT_EXTERNAL_ONE_BYTE_SYMBOL_TYPE": ExternalString,
1605
    "STRING_TYPE": SeqString,
1606
    "ONE_BYTE_STRING_TYPE": SeqString,
1607
    "CONS_STRING_TYPE": ConsString,
1608
    "CONS_ONE_BYTE_STRING_TYPE": ConsString,
1609
    "EXTERNAL_STRING_TYPE": ExternalString,
1610 1611
    "EXTERNAL_STRING_WITH_ONE_BYTE_DATA_TYPE": ExternalString,
    "EXTERNAL_ONE_BYTE_STRING_TYPE": ExternalString,
1612 1613 1614 1615 1616 1617
    "MAP_TYPE": Map,
    "ODDBALL_TYPE": Oddball,
    "FIXED_ARRAY_TYPE": FixedArray,
    "JS_FUNCTION_TYPE": JSFunction,
    "SHARED_FUNCTION_INFO_TYPE": SharedFunctionInfo,
    "SCRIPT_TYPE": Script,
1618 1619
    "CODE_CACHE_TYPE": CodeCache,
    "CODE_TYPE": Code,
1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633
  }

  def __init__(self, reader, stack_map):
    self.reader = reader
    self.stack_map = stack_map
    self.objects = {}

  def FindObjectOrSmi(self, tagged_address):
    if (tagged_address & 1) == 0: return tagged_address / 2
    return self.FindObject(tagged_address)

  def FindObject(self, tagged_address):
    if tagged_address in self.objects:
      return self.objects[tagged_address]
1634
    if (tagged_address & self.ObjectAlignmentMask()) != 1: return None
1635 1636
    address = tagged_address - 1
    if not self.reader.IsValidAddress(address): return None
1637
    map_tagged_address = self.reader.ReadUIntPtr(address)
1638 1639 1640 1641 1642 1643 1644 1645
    if tagged_address == map_tagged_address:
      # Meta map?
      meta_map = Map(self, None, address)
      instance_type_name = INSTANCE_TYPES.get(meta_map.instance_type)
      if instance_type_name != "MAP_TYPE": return None
      meta_map.map = meta_map
      object = meta_map
    else:
1646
      map = self.FindMap(map_tagged_address)
1647 1648 1649 1650 1651 1652 1653 1654
      if map is None: return None
      instance_type_name = INSTANCE_TYPES.get(map.instance_type)
      if instance_type_name is None: return None
      cls = V8Heap.CLASS_MAP.get(instance_type_name, HeapObject)
      object = cls(self, map, address)
    self.objects[tagged_address] = object
    return object

1655 1656 1657 1658 1659 1660 1661 1662 1663 1664
  def FindMap(self, tagged_address):
    if (tagged_address & self.MapAlignmentMask()) != 1: return None
    address = tagged_address - 1
    if not self.reader.IsValidAddress(address): return None
    object = Map(self, None, address)
    return object

  def IntSize(self):
    return 4

1665 1666 1667
  def PointerSize(self):
    return self.reader.PointerSize()

1668 1669 1670 1671 1672 1673
  def ObjectAlignmentMask(self):
    return self.PointerSize() - 1

  def MapAlignmentMask(self):
    if self.reader.arch == MD_CPU_ARCHITECTURE_AMD64:
      return (1 << 4) - 1
1674 1675
    elif self.reader.arch == MD_CPU_ARCHITECTURE_ARM:
      return (1 << 4) - 1
1676 1677
    elif self.reader.arch == MD_CPU_ARCHITECTURE_ARM64:
      return (1 << 4) - 1
1678 1679
    elif self.reader.arch == MD_CPU_ARCHITECTURE_X86:
      return (1 << 5) - 1
1680

1681 1682
  def PageAlignmentMask(self):
    return (1 << 20) - 1
1683 1684


1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703
class KnownObject(HeapObject):
  def __init__(self, heap, known_name):
    HeapObject.__init__(self, heap, None, None)
    self.known_name = known_name

  def __str__(self):
    return "<%s>" % self.known_name


class KnownMap(HeapObject):
  def __init__(self, heap, known_name, instance_type):
    HeapObject.__init__(self, heap, None, None)
    self.instance_type = instance_type
    self.known_name = known_name

  def __str__(self):
    return "<%s>" % self.known_name


1704 1705
COMMENT_RE = re.compile(r"^C (0x[0-9a-fA-F]+) (.*)$")
PAGEADDRESS_RE = re.compile(
1706
    r"^P (mappage|oldpage) (0x[0-9a-fA-F]+)$")
1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776


class InspectionInfo(object):
  def __init__(self, minidump_name, reader):
    self.comment_file = minidump_name + ".comments"
    self.address_comments = {}
    self.page_address = {}
    if os.path.exists(self.comment_file):
      with open(self.comment_file, "r") as f:
        lines = f.readlines()
        f.close()

        for l in lines:
          m = COMMENT_RE.match(l)
          if m:
            self.address_comments[int(m.group(1), 0)] = m.group(2)
          m = PAGEADDRESS_RE.match(l)
          if m:
            self.page_address[m.group(1)] = int(m.group(2), 0)
    self.reader = reader
    self.styles = {}
    self.color_addresses()
    return

  def get_page_address(self, page_kind):
    return self.page_address.get(page_kind, 0)

  def save_page_address(self, page_kind, address):
    with open(self.comment_file, "a") as f:
      f.write("P %s 0x%x\n" % (page_kind, address))
      f.close()

  def color_addresses(self):
    # Color all stack addresses.
    exception_thread = self.reader.thread_map[self.reader.exception.thread_id]
    stack_top = self.reader.ExceptionSP()
    stack_bottom = exception_thread.stack.start + \
        exception_thread.stack.memory.data_size
    frame_pointer = self.reader.ExceptionFP()
    self.styles[frame_pointer] = "frame"
    for slot in xrange(stack_top, stack_bottom, self.reader.PointerSize()):
      self.styles[slot] = "stackaddress"
    for slot in xrange(stack_top, stack_bottom, self.reader.PointerSize()):
      maybe_address = self.reader.ReadUIntPtr(slot)
      self.styles[maybe_address] = "stackval"
      if slot == frame_pointer:
        self.styles[slot] = "frame"
        frame_pointer = maybe_address
    self.styles[self.reader.ExceptionIP()] = "pc"

  def get_style_class(self, address):
    return self.styles.get(address, None)

  def get_style_class_string(self, address):
    style = self.get_style_class(address)
    if style != None:
      return " class=\"%s\" " % style
    else:
      return ""

  def set_comment(self, address, comment):
    self.address_comments[address] = comment
    with open(self.comment_file, "a") as f:
      f.write("C 0x%x %s\n" % (address, comment))
      f.close()

  def get_comment(self, address):
    return self.address_comments.get(address, "")


1777 1778 1779 1780 1781 1782
class InspectionPadawan(object):
  """The padawan can improve annotations by sensing well-known objects."""
  def __init__(self, reader, heap):
    self.reader = reader
    self.heap = heap
    self.known_first_map_page = 0
1783
    self.known_first_old_page = 0
1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798

  def __getattr__(self, name):
    """An InspectionPadawan can be used instead of V8Heap, even though
       it does not inherit from V8Heap (aka. mixin)."""
    return getattr(self.heap, name)

  def GetPageOffset(self, tagged_address):
    return tagged_address & self.heap.PageAlignmentMask()

  def IsInKnownMapSpace(self, tagged_address):
    page_address = tagged_address & ~self.heap.PageAlignmentMask()
    return page_address == self.known_first_map_page

  def IsInKnownOldSpace(self, tagged_address):
    page_address = tagged_address & ~self.heap.PageAlignmentMask()
1799
    return page_address == self.known_first_old_page
1800 1801 1802

  def ContainingKnownOldSpaceName(self, tagged_address):
    page_address = tagged_address & ~self.heap.PageAlignmentMask()
1803
    if page_address == self.known_first_old_page: return "OLD_SPACE"
1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817
    return None

  def SenseObject(self, tagged_address):
    if self.IsInKnownOldSpace(tagged_address):
      offset = self.GetPageOffset(tagged_address)
      lookup_key = (self.ContainingKnownOldSpaceName(tagged_address), offset)
      known_obj_name = KNOWN_OBJECTS.get(lookup_key)
      if known_obj_name:
        return KnownObject(self, known_obj_name)
    if self.IsInKnownMapSpace(tagged_address):
      known_map = self.SenseMap(tagged_address)
      if known_map:
        return known_map
    found_obj = self.heap.FindObject(tagged_address)
1818
    if found_obj: return found_obj
1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859
    address = tagged_address - 1
    if self.reader.IsValidAddress(address):
      map_tagged_address = self.reader.ReadUIntPtr(address)
      map = self.SenseMap(map_tagged_address)
      if map is None: return None
      instance_type_name = INSTANCE_TYPES.get(map.instance_type)
      if instance_type_name is None: return None
      cls = V8Heap.CLASS_MAP.get(instance_type_name, HeapObject)
      return cls(self, map, address)
    return None

  def SenseMap(self, tagged_address):
    if self.IsInKnownMapSpace(tagged_address):
      offset = self.GetPageOffset(tagged_address)
      known_map_info = KNOWN_MAPS.get(offset)
      if known_map_info:
        known_map_type, known_map_name = known_map_info
        return KnownMap(self, known_map_name, known_map_type)
    found_map = self.heap.FindMap(tagged_address)
    if found_map: return found_map
    return None

  def FindObjectOrSmi(self, tagged_address):
    """When used as a mixin in place of V8Heap."""
    found_obj = self.SenseObject(tagged_address)
    if found_obj: return found_obj
    if (tagged_address & 1) == 0:
      return "Smi(%d)" % (tagged_address / 2)
    else:
      return "Unknown(%s)" % self.reader.FormatIntPtr(tagged_address)

  def FindObject(self, tagged_address):
    """When used as a mixin in place of V8Heap."""
    raise NotImplementedError

  def FindMap(self, tagged_address):
    """When used as a mixin in place of V8Heap."""
    raise NotImplementedError

  def PrintKnowledge(self):
    print "  known_first_map_page = %s\n"\
1860
          "  known_first_old_page = %s" % (
1861
          self.reader.FormatIntPtr(self.known_first_map_page),
1862
          self.reader.FormatIntPtr(self.known_first_old_page))
1863

1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176
WEB_HEADER = """
<!DOCTYPE html>
<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="content-type">
<style media="screen" type="text/css">

.code {
  font-family: monospace;
}

.dmptable {
  border-collapse : collapse;
  border-spacing : 0px;
}

.codedump {
  border-collapse : collapse;
  border-spacing : 0px;
}

.addrcomments {
  border : 0px;
}

.register {
  padding-right : 1em;
}

.header {
  clear : both;
}

.header .navigation {
  float : left;
}

.header .dumpname {
  float : right;
}

tr.highlight-line {
  background-color : yellow;
}

.highlight {
  background-color : magenta;
}

tr.inexact-highlight-line {
  background-color : pink;
}

input {
  background-color: inherit;
  border: 1px solid LightGray;
}

.dumpcomments {
  border : 1px solid LightGray;
  width : 32em;
}

.regions td {
  padding:0 15px 0 15px;
}

.stackframe td {
  background-color : cyan;
}

.stackaddress {
  background-color : LightGray;
}

.stackval {
  background-color : LightCyan;
}

.frame {
  background-color : cyan;
}

.commentinput {
  width : 20em;
}

a.nodump:visited {
  color : black;
  text-decoration : none;
}

a.nodump:link {
  color : black;
  text-decoration : none;
}

a:visited {
  color : blueviolet;
}

a:link {
  color : blue;
}

.disasmcomment {
  color : DarkGreen;
}

</style>

<script type="application/javascript">

var address_str = "address-";
var address_len = address_str.length;

function comment() {
  var s = event.srcElement.id;
  var index = s.indexOf(address_str);
  if (index >= 0) {
    send_comment(s.substring(index + address_len), event.srcElement.value);
  }
}

function send_comment(address, comment) {
  xmlhttp = new XMLHttpRequest();
  address = encodeURIComponent(address)
  comment = encodeURIComponent(comment)
  xmlhttp.open("GET",
      "setcomment?%(query_dump)s&address=" + address +
      "&comment=" + comment, true);
  xmlhttp.send();
}

var dump_str = "dump-";
var dump_len = dump_str.length;

function dump_comment() {
  var s = event.srcElement.id;
  var index = s.indexOf(dump_str);
  if (index >= 0) {
    send_dump_desc(s.substring(index + dump_len), event.srcElement.value);
  }
}

function send_dump_desc(name, desc) {
  xmlhttp = new XMLHttpRequest();
  name = encodeURIComponent(name)
  desc = encodeURIComponent(desc)
  xmlhttp.open("GET",
      "setdumpdesc?dump=" + name +
      "&description=" + desc, true);
  xmlhttp.send();
}

function onpage(kind, address) {
  xmlhttp = new XMLHttpRequest();
  kind = encodeURIComponent(kind)
  address = encodeURIComponent(address)
  xmlhttp.onreadystatechange = function() {
    if (xmlhttp.readyState==4 && xmlhttp.status==200) {
      location.reload(true)
    }
  };
  xmlhttp.open("GET",
      "setpageaddress?%(query_dump)s&kind=" + kind +
      "&address=" + address);
  xmlhttp.send();
}

</script>

<title>Dump %(dump_name)s</title>
</head>

<body>
  <div class="header">
    <form class="navigation" action="search.html">
      <a href="summary.html?%(query_dump)s">Context info</a>&nbsp;&nbsp;&nbsp;
      <a href="info.html?%(query_dump)s">Dump info</a>&nbsp;&nbsp;&nbsp;
      <a href="modules.html?%(query_dump)s">Modules</a>&nbsp;&nbsp;&nbsp;
      &nbsp;
      <input type="search" name="val">
      <input type="submit" name="search" value="Search">
      <input type="hidden" name="dump" value="%(dump_name)s">
    </form>
    <form class="navigation" action="disasm.html#highlight">
      &nbsp;
      &nbsp;
      &nbsp;
      <input type="search" name="val">
      <input type="submit" name="disasm" value="Disasm">
      &nbsp;
      &nbsp;
      &nbsp;
      <a href="dumps.html">Dumps...</a>
    </form>
  </div>
  <br>
  <hr>
"""


WEB_FOOTER = """
</body>
</html>
"""


class WebParameterError(Exception):
  def __init__(self, message):
    Exception.__init__(self, message)


class InspectionWebHandler(BaseHTTPServer.BaseHTTPRequestHandler):
  def formatter(self, query_components):
    name = query_components.get("dump", [None])[0]
    return self.server.get_dump_formatter(name)

  def send_success_html_headers(self):
    self.send_response(200)
    self.send_header("Cache-Control", "no-cache, no-store, must-revalidate")
    self.send_header("Pragma", "no-cache")
    self.send_header("Expires", "0")
    self.send_header('Content-type','text/html')
    self.end_headers()
    return

  def do_GET(self):
    try:
      parsedurl = urlparse.urlparse(self.path)
      query_components = urlparse.parse_qs(parsedurl.query)
      if parsedurl.path == "/dumps.html":
        self.send_success_html_headers()
        self.server.output_dumps(self.wfile)
      elif parsedurl.path == "/summary.html":
        self.send_success_html_headers()
        self.formatter(query_components).output_summary(self.wfile)
      elif parsedurl.path == "/info.html":
        self.send_success_html_headers()
        self.formatter(query_components).output_info(self.wfile)
      elif parsedurl.path == "/modules.html":
        self.send_success_html_headers()
        self.formatter(query_components).output_modules(self.wfile)
      elif parsedurl.path == "/search.html":
        address = query_components.get("val", [])
        if len(address) != 1:
          self.send_error(404, "Invalid params")
          return
        self.send_success_html_headers()
        self.formatter(query_components).output_search_res(
            self.wfile, address[0])
      elif parsedurl.path == "/disasm.html":
        address = query_components.get("val", [])
        exact = query_components.get("exact", ["on"])
        if len(address) != 1:
          self.send_error(404, "Invalid params")
          return
        self.send_success_html_headers()
        self.formatter(query_components).output_disasm(
            self.wfile, address[0], exact[0])
      elif parsedurl.path == "/data.html":
        address = query_components.get("val", [])
        datakind = query_components.get("type", ["address"])
        if len(address) == 1 and len(datakind) == 1:
          self.send_success_html_headers()
          self.formatter(query_components).output_data(
              self.wfile, address[0], datakind[0])
        else:
          self.send_error(404,'Invalid params')
      elif parsedurl.path == "/setdumpdesc":
        name = query_components.get("dump", [""])
        description = query_components.get("description", [""])
        if len(name) == 1 and len(description) == 1:
          name = name[0]
          description = description[0]
          if self.server.set_dump_desc(name, description):
            self.send_success_html_headers()
            self.wfile.write("OK")
            return
        self.send_error(404,'Invalid params')
      elif parsedurl.path == "/setcomment":
        address = query_components.get("address", [])
        comment = query_components.get("comment", [""])
        if len(address) == 1 and len(comment) == 1:
          address = address[0]
          comment = comment[0]
          self.formatter(query_components).set_comment(address, comment)
          self.send_success_html_headers()
          self.wfile.write("OK")
        else:
          self.send_error(404,'Invalid params')
      elif parsedurl.path == "/setpageaddress":
        kind = query_components.get("kind", [])
        address = query_components.get("address", [""])
        if len(kind) == 1 and len(address) == 1:
          kind = kind[0]
          address = address[0]
          self.formatter(query_components).set_page_address(kind, address)
          self.send_success_html_headers()
          self.wfile.write("OK")
        else:
          self.send_error(404,'Invalid params')
      else:
        self.send_error(404,'File Not Found: %s' % self.path)

    except IOError:
      self.send_error(404,'File Not Found: %s' % self.path)

    except WebParameterError as e:
      self.send_error(404, 'Web parameter error: %s' % e.message)


2177
HTML_REG_FORMAT = "<span class=\"register\"><b>%s</b>:&nbsp;%s</span><br/>\n"
2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203


class InspectionWebFormatter(object):
  CONTEXT_FULL = 0
  CONTEXT_SHORT = 1

  def __init__(self, switches, minidump_name, http_server):
    self.dumpfilename = os.path.split(minidump_name)[1]
    self.encfilename = urllib.urlencode({ 'dump' : self.dumpfilename })
    self.reader = MinidumpReader(switches, minidump_name)
    self.server = http_server

    # Set up the heap
    exception_thread = self.reader.thread_map[self.reader.exception.thread_id]
    stack_top = self.reader.ExceptionSP()
    stack_bottom = exception_thread.stack.start + \
        exception_thread.stack.memory.data_size
    stack_map = {self.reader.ExceptionIP(): -1}
    for slot in xrange(stack_top, stack_bottom, self.reader.PointerSize()):
      maybe_address = self.reader.ReadUIntPtr(slot)
      if not maybe_address in stack_map:
        stack_map[maybe_address] = slot
    self.heap = V8Heap(self.reader, stack_map)

    self.padawan = InspectionPadawan(self.reader, self.heap)
    self.comments = InspectionInfo(minidump_name, self.reader)
2204 2205
    self.padawan.known_first_old_page = (
        self.comments.get_page_address("oldpage"))
2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218
    self.padawan.known_first_map_page = (
        self.comments.get_page_address("mappage"))

  def set_comment(self, straddress, comment):
    try:
      address = int(straddress, 0)
      self.comments.set_comment(address, comment)
    except ValueError:
      print "Invalid address"

  def set_page_address(self, kind, straddress):
    try:
      address = int(straddress, 0)
2219 2220
      if kind == "oldpage":
        self.padawan.known_first_old_page = address
2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341
      elif kind == "mappage":
        self.padawan.known_first_map_page = address
      self.comments.save_page_address(kind, address)
    except ValueError:
      print "Invalid address"

  def td_from_address(self, f, address):
    f.write("<td %s>" % self.comments.get_style_class_string(address))

  def format_address(self, maybeaddress, straddress = None):
    if maybeaddress is None:
      return "not in dump"
    else:
      if straddress is None:
        straddress = "0x" + self.reader.FormatIntPtr(maybeaddress)
      style_class = ""
      if not self.reader.IsValidAddress(maybeaddress):
        style_class = " class=\"nodump\""
      return ("<a %s href=\"search.html?%s&amp;val=%s\">%s</a>" %
              (style_class, self.encfilename, straddress, straddress))

  def output_header(self, f):
    f.write(WEB_HEADER %
        { "query_dump" : self.encfilename,
          "dump_name"  : cgi.escape(self.dumpfilename) })

  def output_footer(self, f):
    f.write(WEB_FOOTER)

  MAX_CONTEXT_STACK = 4096

  def output_summary(self, f):
    self.output_header(f)
    f.write('<div class="code">')
    self.output_context(f, InspectionWebFormatter.CONTEXT_SHORT)
    self.output_disasm_pc(f)

    # Output stack
    exception_thread = self.reader.thread_map[self.reader.exception.thread_id]
    stack_bottom = exception_thread.stack.start + \
        min(exception_thread.stack.memory.data_size, self.MAX_CONTEXT_STACK)
    stack_top = self.reader.ExceptionSP()
    self.output_words(f, stack_top - 16, stack_bottom, stack_top, "Stack")

    f.write('</div>')
    self.output_footer(f)
    return

  def output_info(self, f):
    self.output_header(f)
    f.write("<h3>Dump info</h3>\n")
    f.write("Description: ")
    self.server.output_dump_desc_field(f, self.dumpfilename)
    f.write("<br>\n")
    f.write("Filename: ")
    f.write("<span class=\"code\">%s</span><br>\n" % (self.dumpfilename))
    dt = datetime.datetime.fromtimestamp(self.reader.header.time_date_stampt)
    f.write("Timestamp: %s<br>\n" % dt.strftime('%Y-%m-%d %H:%M:%S'))
    self.output_context(f, InspectionWebFormatter.CONTEXT_FULL)
    self.output_address_ranges(f)
    self.output_footer(f)
    return

  def output_address_ranges(self, f):
    regions = {}
    def print_region(_reader, start, size, _location):
      regions[start] = size
    self.reader.ForEachMemoryRegion(print_region)
    f.write("<h3>Available memory regions</h3>\n")
    f.write('<div class="code">')
    f.write("<table class=\"regions\">\n")
    f.write("<thead><tr>")
    f.write("<th>Start address</th>")
    f.write("<th>End address</th>")
    f.write("<th>Number of bytes</th>")
    f.write("</tr></thead>\n")
    for start in sorted(regions):
      size = regions[start]
      f.write("<tr>")
      f.write("<td>%s</td>" % self.format_address(start))
      f.write("<td>&nbsp;%s</td>" % self.format_address(start + size))
      f.write("<td>&nbsp;%d</td>" % size)
      f.write("</tr>\n")
    f.write("</table>\n")
    f.write('</div>')
    return

  def output_module_details(self, f, module):
    f.write("<b>%s</b>" % GetModuleName(self.reader, module))
    file_version = GetVersionString(module.version_info.dwFileVersionMS,
                                    module.version_info.dwFileVersionLS)
    product_version = GetVersionString(module.version_info.dwProductVersionMS,
                                       module.version_info.dwProductVersionLS)
    f.write("<br>&nbsp;&nbsp;\n")
    f.write("base: %s" % self.reader.FormatIntPtr(module.base_of_image))
    f.write("<br>&nbsp;&nbsp;\n")
    f.write("  end: %s" % self.reader.FormatIntPtr(module.base_of_image +
                                            module.size_of_image))
    f.write("<br>&nbsp;&nbsp;\n")
    f.write("  file version: %s" % file_version)
    f.write("<br>&nbsp;&nbsp;\n")
    f.write("  product version: %s" % product_version)
    f.write("<br>&nbsp;&nbsp;\n")
    time_date_stamp = datetime.datetime.fromtimestamp(module.time_date_stamp)
    f.write("  timestamp: %s" % time_date_stamp)
    f.write("<br>\n");

  def output_modules(self, f):
    self.output_header(f)
    f.write('<div class="code">')
    for module in self.reader.module_list.modules:
      self.output_module_details(f, module)
    f.write("</div>")
    self.output_footer(f)
    return

  def output_context(self, f, details):
    exception_thread = self.reader.thread_map[self.reader.exception.thread_id]
    f.write("<h3>Exception context</h3>")
    f.write('<div class="code">\n')
    f.write("Thread id: %d" % exception_thread.id)
2342
    f.write("&nbsp;&nbsp; Exception code: %08X<br/>\n" %
2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354
            self.reader.exception.exception.code)
    if details == InspectionWebFormatter.CONTEXT_FULL:
      if self.reader.exception.exception.parameter_count > 0:
        f.write("&nbsp;&nbsp; Exception parameters: \n")
        for i in xrange(0, self.reader.exception.exception.parameter_count):
          f.write("%08x" % self.reader.exception.exception.information[i])
        f.write("<br><br>\n")

    for r in CONTEXT_FOR_ARCH[self.reader.arch]:
      f.write(HTML_REG_FORMAT %
              (r, self.format_address(self.reader.Register(r))))
    # TODO(vitalyr): decode eflags.
2355
    if self.reader.arch in [MD_CPU_ARCHITECTURE_ARM, MD_CPU_ARCHITECTURE_ARM64]:
2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422
      f.write("<b>cpsr</b>: %s" % bin(self.reader.exception_context.cpsr)[2:])
    else:
      f.write("<b>eflags</b>: %s" %
              bin(self.reader.exception_context.eflags)[2:])
    f.write('</div>\n')
    return

  def align_down(self, a, size):
    alignment_correction = a % size
    return a - alignment_correction

  def align_up(self, a, size):
    alignment_correction = (size - 1) - ((a + size - 1) % size)
    return a + alignment_correction

  def format_object(self, address):
    heap_object = self.padawan.SenseObject(address)
    return cgi.escape(str(heap_object or ""))

  def output_data(self, f, straddress, datakind):
    try:
      self.output_header(f)
      address = int(straddress, 0)
      if not self.reader.IsValidAddress(address):
        f.write("<h3>Address 0x%x not found in the dump.</h3>" % address)
        return
      region = self.reader.FindRegion(address)
      if datakind == "address":
        self.output_words(f, region[0], region[0] + region[1], address, "Dump")
      elif datakind == "ascii":
        self.output_ascii(f, region[0], region[0] + region[1], address)
      self.output_footer(f)

    except ValueError:
      f.write("<h3>Unrecognized address format \"%s\".</h3>" % straddress)
    return

  def output_words(self, f, start_address, end_address,
                   highlight_address, desc):
    region = self.reader.FindRegion(highlight_address)
    if region is None:
      f.write("<h3>Address 0x%x not found in the dump.</h3>\n" %
              (highlight_address))
      return
    size = self.heap.PointerSize()
    start_address = self.align_down(start_address, size)
    low = self.align_down(region[0], size)
    high = self.align_up(region[0] + region[1], size)
    if start_address < low:
      start_address = low
    end_address = self.align_up(end_address, size)
    if end_address > high:
      end_address = high

    expand = ""
    if start_address != low or end_address != high:
      expand = ("(<a href=\"data.html?%s&amp;val=0x%x#highlight\">"
                " more..."
                " </a>)" %
                (self.encfilename, highlight_address))

    f.write("<h3>%s 0x%x - 0x%x, "
            "highlighting <a href=\"#highlight\">0x%x</a> %s</h3>\n" %
            (desc, start_address, end_address, highlight_address, expand))
    f.write('<div class="code">')
    f.write("<table class=\"codedump\">\n")

2423 2424
    for j in xrange(0, end_address - start_address, size):
      slot = start_address + j
2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460
      heap_object = ""
      maybe_address = None
      end_region = region[0] + region[1]
      if slot < region[0] or slot + size > end_region:
        straddress = "0x"
        for i in xrange(end_region, slot + size):
          straddress += "??"
        for i in reversed(
            xrange(max(slot, region[0]), min(slot + size, end_region))):
          straddress += "%02x" % self.reader.ReadU8(i)
        for i in xrange(slot, region[0]):
          straddress += "??"
      else:
        maybe_address = self.reader.ReadUIntPtr(slot)
        straddress = self.format_address(maybe_address)
        if maybe_address:
          heap_object = self.format_object(maybe_address)

      address_fmt = "%s&nbsp;</td>\n"
      if slot == highlight_address:
        f.write("<tr class=\"highlight-line\">\n")
        address_fmt = "<a id=\"highlight\"></a>%s&nbsp;</td>\n"
      elif slot < highlight_address and highlight_address < slot + size:
        f.write("<tr class=\"inexact-highlight-line\">\n")
        address_fmt = "<a id=\"highlight\"></a>%s&nbsp;</td>\n"
      else:
        f.write("<tr>\n")

      f.write("  <td>")
      self.output_comment_box(f, "da-", slot)
      f.write("</td>\n")
      f.write("  ")
      self.td_from_address(f, slot)
      f.write(address_fmt % self.format_address(slot))
      f.write("  ")
      self.td_from_address(f, maybe_address)
2461
      f.write(":&nbsp;%s&nbsp;</td>\n" % straddress)
2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498
      f.write("  <td>")
      if maybe_address != None:
        self.output_comment_box(
            f, "sv-" + self.reader.FormatIntPtr(slot), maybe_address)
      f.write("  </td>\n")
      f.write("  <td>%s</td>\n" % (heap_object or ''))
      f.write("</tr>\n")
    f.write("</table>\n")
    f.write("</div>")
    return

  def output_ascii(self, f, start_address, end_address, highlight_address):
    region = self.reader.FindRegion(highlight_address)
    if region is None:
      f.write("<h3>Address %x not found in the dump.</h3>" %
          highlight_address)
      return
    if start_address < region[0]:
      start_address = region[0]
    if end_address > region[0] + region[1]:
      end_address = region[0] + region[1]

    expand = ""
    if start_address != region[0] or end_address != region[0] + region[1]:
      link = ("data.html?%s&amp;val=0x%x&amp;type=ascii#highlight" %
              (self.encfilename, highlight_address))
      expand = "(<a href=\"%s\">more...</a>)" % link

    f.write("<h3>ASCII dump 0x%x - 0x%x, highlighting 0x%x %s</h3>" %
            (start_address, end_address, highlight_address, expand))

    line_width = 64

    f.write('<div class="code">')

    start = self.align_down(start_address, line_width)

2499 2500
    for i in xrange(end_address - start):
      address = start + i
2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568
      if address % 64 == 0:
        if address != start:
          f.write("<br>")
        f.write("0x%08x:&nbsp;" % address)
      if address < start_address:
        f.write("&nbsp;")
      else:
        if address == highlight_address:
          f.write("<span class=\"highlight\">")
        code = self.reader.ReadU8(address)
        if code < 127 and code >= 32:
          f.write("&#")
          f.write(str(code))
          f.write(";")
        else:
          f.write("&middot;")
        if address == highlight_address:
          f.write("</span>")
    f.write("</div>")
    return

  def output_disasm(self, f, straddress, strexact):
    try:
      self.output_header(f)
      address = int(straddress, 0)
      if not self.reader.IsValidAddress(address):
        f.write("<h3>Address 0x%x not found in the dump.</h3>" % address)
        return
      region = self.reader.FindRegion(address)
      self.output_disasm_range(
          f, region[0], region[0] + region[1], address, strexact == "on")
      self.output_footer(f)
    except ValueError:
      f.write("<h3>Unrecognized address format \"%s\".</h3>" % straddress)
    return

  def output_disasm_range(
      self, f, start_address, end_address, highlight_address, exact):
    region = self.reader.FindRegion(highlight_address)
    if start_address < region[0]:
      start_address = region[0]
    if end_address > region[0] + region[1]:
      end_address = region[0] + region[1]
    count = end_address - start_address
    lines = self.reader.GetDisasmLines(start_address, count)
    found = False
    if exact:
      for line in lines:
        if line[0] + start_address == highlight_address:
          found = True
          break
      if not found:
        start_address = highlight_address
        count = end_address - start_address
        lines = self.reader.GetDisasmLines(highlight_address, count)
    expand = ""
    if start_address != region[0] or end_address != region[0] + region[1]:
      exactness = ""
      if exact and not found and end_address == region[0] + region[1]:
        exactness = "&amp;exact=off"
      expand = ("(<a href=\"disasm.html?%s%s"
                "&amp;val=0x%x#highlight\">more...</a>)" %
                (self.encfilename, exactness, highlight_address))

    f.write("<h3>Disassembling 0x%x - 0x%x, highlighting 0x%x %s</h3>" %
            (start_address, end_address, highlight_address, expand))
    f.write('<div class="code">')
    f.write("<table class=\"codedump\">\n");
2569
    for i in xrange(len(lines)):
2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694
      line = lines[i]
      next_address = count
      if i + 1 < len(lines):
        next_line = lines[i + 1]
        next_address = next_line[0]
      self.format_disasm_line(
          f, start_address, line, next_address, highlight_address)
    f.write("</table>\n")
    f.write("</div>")
    return

  def annotate_disasm_addresses(self, line):
    extra = []
    for m in ADDRESS_RE.finditer(line):
      maybe_address = int(m.group(0), 16)
      formatted_address = self.format_address(maybe_address, m.group(0))
      line = line.replace(m.group(0), formatted_address)
      object_info = self.padawan.SenseObject(maybe_address)
      if not object_info:
        continue
      extra.append(cgi.escape(str(object_info)))
    if len(extra) == 0:
      return line
    return ("%s <span class=\"disasmcomment\">;; %s</span>" %
            (line, ", ".join(extra)))

  def format_disasm_line(
      self, f, start, line, next_address, highlight_address):
    line_address = start + line[0]
    address_fmt = "  <td>%s</td>\n"
    if line_address == highlight_address:
      f.write("<tr class=\"highlight-line\">\n")
      address_fmt = "  <td><a id=\"highlight\">%s</a></td>\n"
    elif (line_address < highlight_address and
          highlight_address < next_address + start):
      f.write("<tr class=\"inexact-highlight-line\">\n")
      address_fmt = "  <td><a id=\"highlight\">%s</a></td>\n"
    else:
      f.write("<tr>\n")
    num_bytes = next_address - line[0]
    stack_slot = self.heap.stack_map.get(line_address)
    marker = ""
    if stack_slot:
      marker = "=>"
    op_offset = 3 * num_bytes - 1

    code = line[1]
    # Compute the actual call target which the disassembler is too stupid
    # to figure out (it adds the call offset to the disassembly offset rather
    # than the absolute instruction address).
    if self.heap.reader.arch == MD_CPU_ARCHITECTURE_X86:
      if code.startswith("e8"):
        words = code.split()
        if len(words) > 6 and words[5] == "call":
          offset = int(words[4] + words[3] + words[2] + words[1], 16)
          target = (line_address + offset + 5) & 0xFFFFFFFF
          code = code.replace(words[6], "0x%08x" % target)
    # TODO(jkummerow): port this hack to ARM and x64.

    opcodes = code[:op_offset]
    code = self.annotate_disasm_addresses(code[op_offset:])
    f.write("  <td>")
    self.output_comment_box(f, "codel-", line_address)
    f.write("</td>\n")
    f.write(address_fmt % marker)
    f.write("  ")
    self.td_from_address(f, line_address)
    f.write("%s (+0x%x)</td>\n" %
            (self.format_address(line_address), line[0]))
    f.write("  <td>:&nbsp;%s&nbsp;</td>\n" % opcodes)
    f.write("  <td>%s</td>\n" % code)
    f.write("</tr>\n")

  def output_comment_box(self, f, prefix, address):
    f.write("<input type=\"text\" class=\"commentinput\" "
            "id=\"%s-address-0x%s\" onchange=\"comment()\" value=\"%s\">" %
            (prefix,
             self.reader.FormatIntPtr(address),
             cgi.escape(self.comments.get_comment(address)) or ""))

  MAX_FOUND_RESULTS = 100

  def output_find_results(self, f, results):
    f.write("Addresses")
    toomany = len(results) > self.MAX_FOUND_RESULTS
    if toomany:
      f.write("(found %i results, displaying only first %i)" %
              (len(results), self.MAX_FOUND_RESULTS))
    f.write(": \n")
    results = sorted(results)
    results = results[:min(len(results), self.MAX_FOUND_RESULTS)]
    for address in results:
      f.write("<span %s>%s</span>\n" %
              (self.comments.get_style_class_string(address),
               self.format_address(address)))
    if toomany:
      f.write("...\n")


  def output_page_info(self, f, page_kind, page_address, my_page_address):
    if my_page_address == page_address and page_address != 0:
      f.write("Marked first %s page.\n" % page_kind)
    else:
      f.write("<span id=\"%spage\" style=\"display:none\">" % page_kind)
      f.write("Marked first %s page." % page_kind)
      f.write("</span>\n")
      f.write("<button onclick=\"onpage('%spage', '0x%x')\">" %
              (page_kind, my_page_address))
      f.write("Mark as first %s page</button>\n" % page_kind)
    return

  def output_search_res(self, f, straddress):
    try:
      self.output_header(f)
      f.write("<h3>Search results for %s</h3>" % straddress)

      address = int(straddress, 0)

      f.write("Comment: ")
      self.output_comment_box(f, "search-", address)
      f.write("<br>\n")

      page_address = address & ~self.heap.PageAlignmentMask()

      f.write("Page info: \n")
2695
      self.output_page_info(f, "old", self.padawan.known_first_old_page, \
2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885
                            page_address)
      self.output_page_info(f, "map", self.padawan.known_first_map_page, \
                            page_address)

      if not self.reader.IsValidAddress(address):
        f.write("<h3>The contents at address %s not found in the dump.</h3>" % \
                straddress)
      else:
        # Print as words
        self.output_words(f, address - 8, address + 32, address, "Dump")

        # Print as ASCII
        f.write("<hr>\n")
        self.output_ascii(f, address, address + 256, address)

        # Print as code
        f.write("<hr>\n")
        self.output_disasm_range(f, address - 16, address + 16, address, True)

      aligned_res, unaligned_res = self.reader.FindWordList(address)

      if len(aligned_res) > 0:
        f.write("<h3>Occurrences of 0x%x at aligned addresses</h3>\n" %
                address)
        self.output_find_results(f, aligned_res)

      if len(unaligned_res) > 0:
        f.write("<h3>Occurrences of 0x%x at unaligned addresses</h3>\n" % \
                address)
        self.output_find_results(f, unaligned_res)

      if len(aligned_res) + len(unaligned_res) == 0:
        f.write("<h3>No occurences of 0x%x found in the dump</h3>\n" % address)

      self.output_footer(f)

    except ValueError:
      f.write("<h3>Unrecognized address format \"%s\".</h3>" % straddress)
    return

  def output_disasm_pc(self, f):
    address = self.reader.ExceptionIP()
    if not self.reader.IsValidAddress(address):
      return
    self.output_disasm_range(f, address - 16, address + 16, address, True)


WEB_DUMPS_HEADER = """
<!DOCTYPE html>
<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="content-type">
<style media="screen" type="text/css">

.dumplist {
  border-collapse : collapse;
  border-spacing : 0px;
  font-family: monospace;
}

.dumpcomments {
  border : 1px solid LightGray;
  width : 32em;
}

</style>

<script type="application/javascript">

var dump_str = "dump-";
var dump_len = dump_str.length;

function dump_comment() {
  var s = event.srcElement.id;
  var index = s.indexOf(dump_str);
  if (index >= 0) {
    send_dump_desc(s.substring(index + dump_len), event.srcElement.value);
  }
}

function send_dump_desc(name, desc) {
  xmlhttp = new XMLHttpRequest();
  name = encodeURIComponent(name)
  desc = encodeURIComponent(desc)
  xmlhttp.open("GET",
      "setdumpdesc?dump=" + name +
      "&description=" + desc, true);
  xmlhttp.send();
}

</script>

<title>Dump list</title>
</head>

<body>
"""

WEB_DUMPS_FOOTER = """
</body>
</html>
"""

DUMP_FILE_RE = re.compile(r"[-_0-9a-zA-Z][-\._0-9a-zA-Z]*\.dmp$")


class InspectionWebServer(BaseHTTPServer.HTTPServer):
  def __init__(self, port_number, switches, minidump_name):
    BaseHTTPServer.HTTPServer.__init__(
        self, ('', port_number), InspectionWebHandler)
    splitpath = os.path.split(minidump_name)
    self.dumppath = splitpath[0]
    self.dumpfilename = splitpath[1]
    self.default_formatter = InspectionWebFormatter(
        switches, minidump_name, self)
    self.formatters = { self.dumpfilename : self.default_formatter }
    self.switches = switches

  def output_dump_desc_field(self, f, name):
    try:
      descfile = open(os.path.join(self.dumppath, name + ".desc"), "r")
      desc = descfile.readline()
      descfile.close()
    except IOError:
      desc = ""
    f.write("<input type=\"text\" class=\"dumpcomments\" "
            "id=\"dump-%s\" onchange=\"dump_comment()\" value=\"%s\">\n" %
            (cgi.escape(name), desc))

  def set_dump_desc(self, name, description):
    if not DUMP_FILE_RE.match(name):
      return False
    fname = os.path.join(self.dumppath, name)
    if not os.path.isfile(fname):
      return False
    fname = fname + ".desc"
    descfile = open(fname, "w")
    descfile.write(description)
    descfile.close()
    return True

  def get_dump_formatter(self, name):
    if name is None:
      return self.default_formatter
    else:
      if not DUMP_FILE_RE.match(name):
        raise WebParameterError("Invalid name '%s'" % name)
      formatter = self.formatters.get(name, None)
      if formatter is None:
        try:
          formatter = InspectionWebFormatter(
              self.switches, os.path.join(self.dumppath, name), self)
          self.formatters[name] = formatter
        except IOError:
          raise WebParameterError("Could not open dump '%s'" % name)
      return formatter

  def output_dumps(self, f):
    f.write(WEB_DUMPS_HEADER)
    f.write("<h3>List of available dumps</h3>")
    f.write("<table class=\"dumplist\">\n")
    f.write("<thead><tr>")
    f.write("<th>Name</th>")
    f.write("<th>File time</th>")
    f.write("<th>Comment</th>")
    f.write("</tr></thead>")
    dumps_by_time = {}
    for fname in os.listdir(self.dumppath):
      if DUMP_FILE_RE.match(fname):
        mtime = os.stat(os.path.join(self.dumppath, fname)).st_mtime
        fnames = dumps_by_time.get(mtime, [])
        fnames.append(fname)
        dumps_by_time[mtime] = fnames

    for mtime in sorted(dumps_by_time, reverse=True):
      fnames = dumps_by_time[mtime]
      for fname in fnames:
        f.write("<tr>\n")
        f.write("<td><a href=\"summary.html?%s\">%s</a></td>\n" % (
            (urllib.urlencode({ 'dump' : fname }), fname)))
        f.write("<td>&nbsp;&nbsp;&nbsp;")
        f.write(datetime.datetime.fromtimestamp(mtime))
        f.write("</td>")
        f.write("<td>&nbsp;&nbsp;&nbsp;")
        self.output_dump_desc_field(f, fname)
        f.write("</td>")
        f.write("</tr>\n")
    f.write("</table>\n")
    f.write(WEB_DUMPS_FOOTER)
    return
2886

2887 2888 2889 2890 2891
class InspectionShell(cmd.Cmd):
  def __init__(self, reader, heap):
    cmd.Cmd.__init__(self)
    self.reader = reader
    self.heap = heap
2892
    self.padawan = InspectionPadawan(reader, heap)
2893 2894
    self.prompt = "(grok) "

2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912
  def do_da(self, address):
    """
     Print ASCII string starting at specified address.
    """
    address = int(address, 16)
    string = ""
    while self.reader.IsValidAddress(address):
      code = self.reader.ReadU8(address)
      if code < 128:
        string += chr(code)
      else:
        break
      address += 1
    if string == "":
      print "Not an ASCII string at %s" % self.reader.FormatIntPtr(address)
    else:
      print "%s\n" % string

2913
  def do_dd(self, args):
2914
    """
2915 2916 2917 2918
     Interpret memory in the given region [address, address + num * word_size)
     (if available) as a sequence of words. Automatic alignment is not performed.
     If the num is not specified, a default value of 16 words is used.
     Synopsis: dd 0x<address> 0x<num>
2919
    """
2920 2921 2922
    args = args.split(' ')
    start = int(args[0], 16)
    num = int(args[1], 16) if len(args) > 1 else 0x10
2923 2924
    if (start & self.heap.ObjectAlignmentMask()) != 0:
      print "Warning: Dumping un-aligned memory, is this what you had in mind?"
2925 2926 2927 2928
    for i in xrange(0,
                    self.reader.PointerSize() * num,
                    self.reader.PointerSize()):
      slot = start + i
2929 2930 2931
      if not self.reader.IsValidAddress(slot):
        print "Address is not contained within the minidump!"
        return
2932
      maybe_address = self.reader.ReadUIntPtr(slot)
2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955
      heap_object = self.padawan.SenseObject(maybe_address)
      print "%s: %s %s" % (self.reader.FormatIntPtr(slot),
                           self.reader.FormatIntPtr(maybe_address),
                           heap_object or '')

  def do_do(self, address):
    """
     Interpret memory at the given address as a V8 object. Automatic
     alignment makes sure that you can pass tagged as well as un-tagged
     addresses.
    """
    address = int(address, 16)
    if (address & self.heap.ObjectAlignmentMask()) == 0:
      address = address + 1
    elif (address & self.heap.ObjectAlignmentMask()) != 1:
      print "Address doesn't look like a valid pointer!"
      return
    heap_object = self.padawan.SenseObject(address)
    if heap_object:
      heap_object.Print(Printer())
    else:
      print "Address cannot be interpreted as object!"

2956 2957 2958 2959 2960
  def do_do_desc(self, address):
    """
      Print a descriptor array in a readable format.
    """
    start = int(address, 16)
2961
    if ((start & 1) == 1): start = start - 1
2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979
    DescriptorArray(FixedArray(self.heap, None, start)).Print(Printer())

  def do_do_map(self, address):
    """
      Print a descriptor array in a readable format.
    """
    start = int(address, 16)
    if ((start & 1) == 1): start = start - 1
    Map(self.heap, None, start).Print(Printer())

  def do_do_trans(self, address):
    """
      Print a transition array in a readable format.
    """
    start = int(address, 16)
    if ((start & 1) == 1): start = start - 1
    TransitionArray(FixedArray(self.heap, None, start)).Print(Printer())

2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000
  def do_dp(self, address):
    """
     Interpret memory at the given address as being on a V8 heap page
     and print information about the page header (if available).
    """
    address = int(address, 16)
    page_address = address & ~self.heap.PageAlignmentMask()
    if self.reader.IsValidAddress(page_address):
      raise NotImplementedError
    else:
      print "Page header is not available!"

  def do_k(self, arguments):
    """
     Teach V8 heap layout information to the inspector. This increases
     the amount of annotations the inspector can produce while dumping
     data. The first page of each heap space is of particular interest
     because it contains known objects that do not move.
    """
    self.padawan.PrintKnowledge()

3001
  def do_ko(self, address):
3002 3003
    """
     Teach V8 heap layout information to the inspector. Set the first
3004
     old space page by passing any pointer into that page.
3005 3006 3007
    """
    address = int(address, 16)
    page_address = address & ~self.heap.PageAlignmentMask()
3008
    self.padawan.known_first_old_page = page_address
3009

3010
  def do_km(self, address):
3011 3012
    """
     Teach V8 heap layout information to the inspector. Set the first
3013
     map-space page by passing any pointer into that page.
3014 3015 3016
    """
    address = int(address, 16)
    page_address = address & ~self.heap.PageAlignmentMask()
3017
    self.padawan.known_first_map_page = page_address
3018

3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029
  def do_list(self, smth):
    """
     List all available memory regions.
    """
    def print_region(reader, start, size, location):
      print "  %s - %s (%d bytes)" % (reader.FormatIntPtr(start),
                                      reader.FormatIntPtr(start + size),
                                      size)
    print "Available memory regions:"
    self.reader.ForEachMemoryRegion(print_region)

3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044
  def do_lm(self, arg):
    """
     List details for all loaded modules in the minidump. An argument can
     be passed to limit the output to only those modules that contain the
     argument as a substring (case insensitive match).
    """
    for module in self.reader.module_list.modules:
      if arg:
        name = GetModuleName(self.reader, module).lower()
        if name.find(arg.lower()) >= 0:
          PrintModuleDetails(self.reader, module)
      else:
        PrintModuleDetails(self.reader, module)
    print

3045
  def do_s(self, word):
3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057
    """
     Search for a given word in available memory regions. The given word
     is expanded to full pointer size and searched at aligned as well as
     un-aligned memory locations. Use 'sa' to search aligned locations
     only.
    """
    try:
      word = int(word, 0)
    except ValueError:
      print "Malformed word, prefix with '0x' to use hexadecimal format."
      return
    print "Searching for word %d/0x%s:" % (word, self.reader.FormatIntPtr(word))
3058 3059
    self.reader.FindWord(word)

3060 3061 3062 3063 3064 3065 3066 3067
  def do_sh(self, none):
    """
     Search for the V8 Heap object in all available memory regions. You
     might get lucky and find this rare treasure full of invaluable
     information.
    """
    raise NotImplementedError

3068
  def do_u(self, args):
3069
    """
3070 3071 3072
     Unassemble memory in the region [address, address + size). If the
     size is not specified, a default value of 32 bytes is used.
     Synopsis: u 0x<address> 0x<size>
3073
    """
3074 3075
    args = args.split(' ')
    start = int(args[0], 16)
3076 3077 3078 3079
    size = int(args[1], 16) if len(args) > 1 else 0x20
    if not self.reader.IsValidAddress(start):
      print "Address is not contained within the minidump!"
      return
3080 3081 3082 3083
    lines = self.reader.GetDisasmLines(start, size)
    for line in lines:
      print FormatDisasmLine(start, self.heap, line)
    print
3084

3085 3086 3087
  def do_EOF(self, none):
    raise KeyboardInterrupt

3088 3089 3090 3091
EIP_PROXIMITY = 64

CONTEXT_FOR_ARCH = {
    MD_CPU_ARCHITECTURE_AMD64:
3092 3093
      ['rax', 'rbx', 'rcx', 'rdx', 'rdi', 'rsi', 'rbp', 'rsp', 'rip',
       'r8', 'r9', 'r10', 'r11', 'r12', 'r13', 'r14', 'r15'],
3094 3095 3096
    MD_CPU_ARCHITECTURE_ARM:
      ['r0', 'r1', 'r2', 'r3', 'r4', 'r5', 'r6', 'r7', 'r8', 'r9',
       'r10', 'r11', 'r12', 'sp', 'lr', 'pc'],
3097 3098 3099 3100 3101
    MD_CPU_ARCHITECTURE_ARM64:
      ['r0', 'r1', 'r2', 'r3', 'r4', 'r5', 'r6', 'r7', 'r8', 'r9',
       'r10', 'r11', 'r12', 'r13', 'r14', 'r15', 'r16', 'r17', 'r18', 'r19',
       'r20', 'r21', 'r22', 'r23', 'r24', 'r25', 'r26', 'r27', 'r28',
       'fp', 'lr', 'sp', 'pc'],
3102 3103 3104 3105
    MD_CPU_ARCHITECTURE_X86:
      ['eax', 'ebx', 'ecx', 'edx', 'edi', 'esi', 'ebp', 'esp', 'eip']
}

3106 3107
KNOWN_MODULES = {'chrome.exe', 'chrome.dll'}

3108 3109 3110 3111
def GetVersionString(ms, ls):
  return "%d.%d.%d.%d" % (ms >> 16, ms & 0xffff, ls >> 16, ls & 0xffff)


3112 3113
def GetModuleName(reader, module):
  name = reader.ReadMinidumpString(module.module_name_rva)
3114 3115
  # simplify for path manipulation
  name = name.encode('utf-8')
3116
  return str(os.path.basename(str(name).replace("\\", "/")))
3117

3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133

def PrintModuleDetails(reader, module):
  print "%s" % GetModuleName(reader, module)
  file_version = GetVersionString(module.version_info.dwFileVersionMS,
                                  module.version_info.dwFileVersionLS)
  product_version = GetVersionString(module.version_info.dwProductVersionMS,
                                     module.version_info.dwProductVersionLS)
  print "  base: %s" % reader.FormatIntPtr(module.base_of_image)
  print "  end: %s" % reader.FormatIntPtr(module.base_of_image +
                                          module.size_of_image)
  print "  file version: %s" % file_version
  print "  product version: %s" % product_version
  time_date_stamp = datetime.datetime.fromtimestamp(module.time_date_stamp)
  print "  timestamp: %s" % time_date_stamp


3134 3135
def AnalyzeMinidump(options, minidump_name):
  reader = MinidumpReader(options, minidump_name)
3136
  heap = None
3137 3138 3139
  DebugPrint("========================================")
  if reader.exception is None:
    print "Minidump has no exception info"
3140 3141 3142 3143 3144 3145 3146 3147 3148
  else:
    print "Exception info:"
    exception_thread = reader.thread_map[reader.exception.thread_id]
    print "  thread id: %d" % exception_thread.id
    print "  code: %08X" % reader.exception.exception.code
    print "  context:"
    for r in CONTEXT_FOR_ARCH[reader.arch]:
      print "    %s: %s" % (r, reader.FormatIntPtr(reader.Register(r)))
    # TODO(vitalyr): decode eflags.
3149
    if reader.arch in [MD_CPU_ARCHITECTURE_ARM, MD_CPU_ARCHITECTURE_ARM64]:
3150 3151 3152 3153
      print "    cpsr: %s" % bin(reader.exception_context.cpsr)[2:]
    else:
      print "    eflags: %s" % bin(reader.exception_context.eflags)[2:]

3154 3155 3156 3157 3158 3159 3160
    print
    print "  modules:"
    for module in reader.module_list.modules:
      name = GetModuleName(reader, module)
      if name in KNOWN_MODULES:
        print "    %s at %08X" % (name, module.base_of_image)
        reader.TryLoadSymbolsFor(name, module)
3161
    print
3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173

    stack_top = reader.ExceptionSP()
    stack_bottom = exception_thread.stack.start + \
        exception_thread.stack.memory.data_size
    stack_map = {reader.ExceptionIP(): -1}
    for slot in xrange(stack_top, stack_bottom, reader.PointerSize()):
      maybe_address = reader.ReadUIntPtr(slot)
      if not maybe_address in stack_map:
        stack_map[maybe_address] = slot
    heap = V8Heap(reader, stack_map)

    print "Disassembly around exception.eip:"
3174 3175 3176
    eip_symbol = reader.FindSymbol(reader.ExceptionIP())
    if eip_symbol is not None:
      print eip_symbol
3177 3178 3179 3180 3181 3182 3183 3184 3185 3186
    disasm_start = reader.ExceptionIP() - EIP_PROXIMITY
    disasm_bytes = 2 * EIP_PROXIMITY
    if (options.full):
      full_range = reader.FindRegion(reader.ExceptionIP())
      if full_range is not None:
        disasm_start = full_range[0]
        disasm_bytes = full_range[1]

    lines = reader.GetDisasmLines(disasm_start, disasm_bytes)

3187 3188 3189 3190
    if not lines:
      print "Could not disassemble using %s." % OBJDUMP_BIN
      print "Pass path to architecture specific objdump via --objdump?"

3191 3192 3193 3194 3195 3196
    for line in lines:
      print FormatDisasmLine(disasm_start, heap, line)
    print

  if heap is None:
    heap = V8Heap(reader, None)
3197

3198
  if options.full:
3199
    FullDump(reader, heap)
3200

3201 3202 3203
  if options.command:
    InspectionShell(reader, heap).onecmd(options.command)

3204
  if options.shell:
3205 3206 3207 3208
    try:
      InspectionShell(reader, heap).cmdloop("type help to get help")
    except KeyboardInterrupt:
      print "Kthxbye."
3209
  elif not options.command:
3210
    if reader.exception is not None:
3211
      frame_pointer = reader.ExceptionFP()
3212
      in_oom_dump_area = False
3213 3214
      print "Annotated stack (from exception.esp to bottom):"
      for slot in xrange(stack_top, stack_bottom, reader.PointerSize()):
3215 3216
        ascii_content = [c if c >= '\x20' and c <  '\x7f' else '.'
                         for c in reader.ReadBytes(slot, reader.PointerSize())]
3217
        maybe_address = reader.ReadUIntPtr(slot)
3218 3219 3220 3221 3222
        maybe_address_contents = None
        if maybe_address >= stack_top and maybe_address <= stack_bottom:
          maybe_address_contents = reader.ReadUIntPtr(maybe_address)
          if maybe_address_contents == 0xdecade00:
            in_oom_dump_area = True
3223
        heap_object = heap.FindObject(maybe_address)
3224
        maybe_symbol = reader.FindSymbol(maybe_address)
3225 3226 3227 3228 3229 3230
        oom_comment = ""
        if in_oom_dump_area:
          if maybe_address_contents == 0xdecade00:
            oom_comment = " <----- HeapStats start marker"
          elif maybe_address_contents == 0xdecade01:
            oom_comment = " <----- HeapStats end marker"
3231
          elif maybe_address_contents is not None:
3232 3233
            oom_comment = " %d (%d Mbytes)" % (maybe_address_contents,
                                            maybe_address_contents >> 20)
3234 3235 3236
        if slot == frame_pointer:
          maybe_symbol = "<---- frame pointer"
          frame_pointer = maybe_address
3237 3238 3239 3240 3241 3242 3243
        print "%s: %s %s %s%s" % (reader.FormatIntPtr(slot),
                                  reader.FormatIntPtr(maybe_address),
                                   "".join(ascii_content),
                                   maybe_symbol or "",
                                   oom_comment)
        if maybe_address_contents == 0xdecade01:
          in_oom_dump_area = False
3244 3245 3246
        if heap_object:
          heap_object.Print(Printer())
          print
3247 3248 3249 3250 3251 3252

  reader.Dispose()


if __name__ == "__main__":
  parser = optparse.OptionParser(USAGE)
3253 3254
  parser.add_option("-s", "--shell", dest="shell", action="store_true",
                    help="start an interactive inspector shell")
3255 3256
  parser.add_option("-w", "--web", dest="web", action="store_true",
                    help="start a web server on localhost:%i" % PORT_NUMBER)
3257 3258
  parser.add_option("-c", "--command", dest="command", default="",
                    help="run an interactive inspector shell command and exit")
3259 3260
  parser.add_option("-f", "--full", dest="full", action="store_true",
                    help="dump all information contained in the minidump")
3261 3262
  parser.add_option("--symdir", dest="symdir", default=".",
                    help="directory containing *.pdb.sym file with symbols")
3263 3264 3265
  parser.add_option("--objdump",
                    default="/usr/bin/objdump",
                    help="objdump tool to use [default: %default]")
3266
  options, args = parser.parse_args()
3267 3268 3269 3270 3271
  if os.path.exists(options.objdump):
    disasm.OBJDUMP_BIN = options.objdump
    OBJDUMP_BIN = options.objdump
  else:
    print "Cannot find %s, falling back to default objdump" % options.objdump
3272 3273 3274
  if len(args) != 1:
    parser.print_help()
    sys.exit(1)
3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285
  if options.web:
    try:
      server = InspectionWebServer(PORT_NUMBER, options, args[0])
      print 'Started httpserver on port ' , PORT_NUMBER
      webbrowser.open('http://localhost:%i/summary.html' % PORT_NUMBER)
      server.serve_forever()
    except KeyboardInterrupt:
      print '^C received, shutting down the web server'
      server.socket.close()
  else:
    AnalyzeMinidump(options, args[0])