Adding untrusted crash dump / stack trace tests.

src/untrusted/crash_dump/decode_dump.py

+#!/usr/bin/python
+# Copyright (c) 2012 The Native Client Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+"""Utility to decode a crash dump generated by untrusted_crash_dump.[ch]
+
+Currently this produces a simple stack trace.
+"""
+
+import json
+import optparse
+import os
+import posixpath
+import subprocess
+import sys
+
+
+def SelectModulePath(options, filename):
+  """Select which path to get a module from.
+
+  Args:
+    options: option object.
+    filename: filename of a module (as appears in phdrs).
+  Returns:
+    Full local path to the file.
+    Derived by consulting the manifest.
+  """
+  # For some names try the main nexe.
+  # NaClMain is the argv[0] setup in sel_main.c
+  # (null) shows up in chrome.
+  if options.main_nexe and filename in ['NaClMain', '(null)']:

[Mark Seaborn, 2012/02/16 18:42:50]

"if options.main_nexe is not None ..."

[bradn, 2012/02/16 20:13:10]

Done.

+    return options.main_nexe
+  filepart = posixpath.basename(filename)
+  nmf_entry = options.nmf_data.get('files', {}).get(filepart, {})
+  # TODO(bradnelson): support x86-64 + arm.
+  nmf_url = nmf_entry.get('x86-32', {}).get('url')
+  # Try filename directly if not in manifest.
+  if nmf_url is None:
+    return filename
+  # Look for the module relative to the manifest (if any), then toolchain.
+  paths = []
+  if options.nmf is not None:
+    paths.append(os.path.dirname(options.nmf))
+  if options.toolchain_libs is not None:
+    paths.append(options.toolchain_libs)
+  for path in paths:
+    pfilename = os.path.join(path, nmf_url)
+    if os.path.exists(pfilename):
+      return pfilename
+  # If nothing else, try the path directly.
+  return filename
+
+
+def DecodeAddressSegment(segments, address):
+  """Convert an address to a segment relative one, plus filename.
+
+  Args:
+    segments: a list of phdr segments.
+    address: a process wide code address.
+  Returns:
+    A tuple of filename and segment relative address.
+  """
+  for segment in segments:
+    for phdr in segment['dlpi_phdr']:
+      start = segment['dlpi_addr'] + phdr['p_vaddr']
+      end = start + phdr['p_memsz']
+      if address >= start and address < end:
+        return (segment['dlpi_name'], address - segment['dlpi_addr'])
+  return ('(null)', address)
+
+
+def Addr2Line(options, segments, address):
+  """Use addr2line to decode a code address.
+
+  Args:
+    options: option object.
+    segments: A list of phdr segments.
+    address: a code address.
+  Returns:
+    A list of dicts containing: function, filename, lineno.
+  """
+  filename, address = DecodeAddressSegment(segments, address)
+  filename = SelectModulePath(options, filename)
+  if not os.path.exists(filename):
+    return [{
+        'function': 'Unknown_function',
+        'filename': 'unknown_file',
+        'lineno': -1,
+    }]
+  cmd = [
+      options.addr2line, '-f', '--inlines', '-e', filename, '0x%08x' % address,
+  ]
+  process = subprocess.Popen(cmd, stdout=subprocess.PIPE)
+  process_stdout, _ = process.communicate()
+  assert process.returncode == 0
+  lines = process_stdout.splitlines()
+  assert len(lines) % 2 == 0
+  results = []
+  for index in range(len(lines) / 2):

[Mark Seaborn, 2012/02/16 18:42:50]

Nit: xrange() is preferred over range() in general (although it makes little
difference here)

[bradn, 2012/02/16 20:13:10]

Done.

+    func = lines[index * 2]
+    afilename, lineno = lines[index * 2 + 1].split(':')

[Mark Seaborn, 2012/02/16 18:42:50]

Nit: split(':', 1) is preferable if you're expecting 2 results (although it
makes little difference here)

[bradn, 2012/02/16 20:13:10]

Done.

+    results.append({
+        'function': func,
+        'filename': afilename,
+        'lineno': int(lineno),
+    })
+  return results
+
+
+def LoadAndDecode(options, core_path):
+  """Given a core.json file, load and embellish with decoded addresses.
+
+  Args:
+    options: options object.
+    core_path: source file containing a dump.
+  Returns:
+
+  """
+  if options.nmf:
+    options.nmf_data = json.load(open(options.nmf))
+  else:
+    options.nmf_data = {}
+  core = json.load(open(core_path))
+  for frame in core['frames']:
+    frame['scopes'] = Addr2Line(
+        options, core['segments'], frame['ip'])
+  return core
+
+
+def StackTrace(info):
+  """Convert a decoded core.json dump to a simple stack trace.
+
+  Args:
+    info: core.json info with decoded code addresses.
+  Returns:
+    A list of dicts with filename, lineno, function (deepest first).
+  """
+  trace = []
+  for frame in info['frames']:
+    for scope in frame['scopes']:
+      trace.append(scope)
+  return trace
+
+
+def PrintTrace(trace, out):
+  """Print a trace to a file like object.
+
+  Args:
+    trace: A list of [filename, lineno, function] (deepest first).
+    out: file like object to output the trace to.
+  """
+  for scope in trace:
+    out.write('%s at %s:%d\n' % (
+        scope['function'],
+        scope['filename'],
+        scope['lineno']))
+
+
+def Main(args):
+  parser = optparse.OptionParser(
+      usage='USAGE: %prog [options] <core.json>')
+  parser.add_option('-m', '--main-nexe', dest='main_nexe',
+                    help='nexe to resolve NaClMain references from')
+  parser.add_option('-n', '--nmf', dest='nmf',

[Mark Seaborn, 2012/02/16 18:42:50]

dest='nmf_filename' would be a little more descriptive given that you also have
options.nmf_data.

[bradn, 2012/02/16 20:13:10]

Done.

+                    help='nmf to resolve references from')
+  parser.add_option('-a', '--addr2line', dest='addr2line',
+                    help='path to appropriate addr2line')
+  parser.add_option('-l', '--toolchain-libs', dest='toolchain_libs',
+                    help='path to the toolchain libraries')
+  options, args = parser.parse_args(args)
+  if len(args) != 1:
+    parser.print_help()
+    sys.exit(1)
+  info = LoadAndDecode(options, args[0])
+  trace = StackTrace(info)
+  PrintTrace(trace, sys.stdout)
+
+
+if __name__ == '__main__':
+  Main(sys.argv[1:])