emsymbolizer.py

#!/usr/bin/env python3

# This is a utility for looking up the symbol names and/or file+line numbers
# of code addresses. There are several possible sources of this information,
# with varying granularity (listed here in approximate preference order).

# If the wasm has DWARF info, llvm-symbolizer can show the symbol, file, and
# line/column number, potentially including inlining.
# If the wasm has separate DWARF info, do the above with the side file
# If there is a source map, we can parse it to get file and line number.
# If there is an emscripten symbol map, we can parse that to get the symbol name
# If there is a name section or symbol table, llvm-nm can show the symbol name.

import argparse
from collections import namedtuple
import json
import os
import re
import subprocess
import sys
from tools import shared
from tools import webassembly

LLVM_SYMBOLIZER = os.path.expanduser(
    shared.build_llvm_tool_path(shared.exe_suffix('llvm-symbolizer')))


class Error(BaseException):
  pass


# Class to treat location info in a uniform way across information sources.
class LocationInfo(object):
  def __init__(self, source=None, line=0, column=0, func=None):
    self.source = source
    self.line = line
    self.column = column
    self.func = func

  def print(self):
    source = self.source if self.source else '??'
    func = self.func if self.func else '??'
    print(f'{func}\n{source}:{self.line}:{self.column}')


def get_codesec_offset(module):
  sec = module.get_section(webassembly.SecType.CODE)
  if not sec:
    raise Error(f'No code section found in {module.filename}')
  return sec.offset


def has_debug_line_section(module):
  for sec in module.sections():
    if sec.name == ".debug_line":
      return True
  return False


def symbolize_address_dwarf(module, address):
  vma_adjust = get_codesec_offset(module)
  cmd = [LLVM_SYMBOLIZER, '-e', module.filename, f'--adjust-vma={vma_adjust}',
         str(address)]
  out = shared.run_process(cmd, stdout=subprocess.PIPE).stdout.strip()
  out_lines = out.splitlines()
  # Source location regex, e.g., /abc/def.c:3:5
  SOURCE_LOC_RE = re.compile(r'(.+):(\d+):(\d+)$')
  # llvm-dwarfdump prints two lines per location. The first line contains a
  # function name, and the second contains a source location like
  # '/abc/def.c:3:5'. If the function or source info is not available, it will
  # be printed as '??', in which case we store None. If the line and column info
  # is not available, they will be printed as 0, which we store as is.
  for i in range(0, len(out_lines), 2):
    func, loc_str = out_lines[i], out_lines[i + 1]
    m = SOURCE_LOC_RE.match(loc_str)
    source, line, column = m.group(1), m.group(2), m.group(3)
    if func == '??':
      func = None
    if source == '??':
      source = None
    LocationInfo(source, line, column, func).print()


def get_sourceMappingURL_section(module):
  for sec in module.sections():
    if sec.name == "sourceMappingURL":
      return sec
  return None


class WasmSourceMap(object):
  # This implementation is derived from emscripten's sourcemap-support.js
  Location = namedtuple('Location', ['source', 'line', 'column'])

  def __init__(self):
    self.version = None
    self.sources = []
    self.mappings = {}
    self.offsets = []

  def parse(self, filename):
    with open(filename) as f:
      source_map_json = json.loads(f.read())
      if shared.DEBUG:
        print(source_map_json)

    self.version = source_map_json['version']
    self.sources = source_map_json['sources']

    vlq_map = {}
    chars = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/='
    for i, c in enumerate(chars):
      vlq_map[c] = i

    def decodeVLQ(string):
      result = []
      shift = 0
      value = 0
      for c in string:
        try:
          integer = vlq_map[c]
        except ValueError:
          raise Error(f'Invalid character ({c}) in VLQ')
        value += (integer & 31) << shift
        if integer & 32:
          shift += 5
        else:
          negate = value & 1
          value >>= 1
          result.append(-value if negate else value)
          value = shift = 0
      return result

    offset = 0
    src = 0
    line = 1
    col = 1
    for segment in source_map_json['mappings'].split(','):
      data = decodeVLQ(segment)
      info = []

      offset += data[0]
      if len(data) >= 2:
        src += data[1]
        info.append(src)
      if len(data) >= 3:
        line += data[2]
        info.append(line)
      if len(data) >= 4:
        col += data[3]
        info.append(col)
      # TODO: see if we need the name, which is the next field (data[4])

      self.mappings[offset] = WasmSourceMap.Location(*info)
      self.offsets.append(offset)
    self.offsets.sort()

  def find_offset(self, offset):
    # Find the largest mapped offset <= the search offset
    lo = 0
    hi = len(self.offsets)

    while lo < hi:
      mid = (lo + hi) // 2
      if self.offsets[mid] > offset:
        hi = mid
      else:
        lo = mid + 1
    return self.offsets[lo - 1]

  def lookup(self, offset):
    nearest = self.find_offset(offset)
    assert nearest in self.mappings, 'Sourcemap has an offset with no mapping'
    info = self.mappings[nearest]
    return LocationInfo(
        self.sources[info.source] if info.source is not None else None,
        info.line,
        info.column
      )


def symbolize_address_sourcemap(module, address, force_file):
  URL = force_file
  if not URL:
    # If a sourcemap file is not forced, read it from the wasm module
    section = get_sourceMappingURL_section(module)
    assert section
    module.seek(section.offset)
    assert module.read_string() == 'sourceMappingURL'
    # TODO: support stripping/replacing a prefix from the URL
    URL = module.read_string()

  if shared.DEBUG:
    print(f'Source Mapping URL: {URL}')
  sm = WasmSourceMap()
  sm.parse(URL)
  if shared.DEBUG:
    csoff = get_codesec_offset(module)
    print(sm.mappings)
    # Print with section offsets to easily compare against dwarf
    for k, v in sm.mappings.items():
      print(f'{k-csoff:x}: {v}')
  sm.lookup(address).print()


def main(args):
  with webassembly.Module(args.wasm_file) as module:
    base = 16 if args.address.lower().startswith('0x') else 10
    address = int(args.address, base)
    symbolized = 0

    if args.addrtype == 'code':
      address += get_codesec_offset(module)

    if ((has_debug_line_section(module) and not args.source) or
       'dwarf' in args.source):
      symbolize_address_dwarf(module, address)
      symbolized += 1

    if ((get_sourceMappingURL_section(module) and not args.source) or
       'sourcemap' in args.source):
      symbolize_address_sourcemap(module, address, args.file)
      symbolized += 1

    if not symbolized:
      raise Error('No .debug_line or sourceMappingURL section found in '
                  f'{module.filename}.'
                  " I don't know how to symbolize this file yet")


def get_args():
  parser = argparse.ArgumentParser()
  parser.add_argument('-s', '--source', choices=['dwarf', 'sourcemap'],
                      help='Force debug info source type', default=())
  parser.add_argument('-f', '--file', action='store',
                      help='Force debug info source file')
  parser.add_argument('-t', '--addrtype', choices=['code', 'file'],
                      default='file',
                      help='Address type (code section or file offset)')
  parser.add_argument('-v', '--verbose', action='store_true',
                      help='Print verbose info for debugging this script')
  parser.add_argument('wasm_file', help='Wasm file')
  parser.add_argument('address', help='Address to lookup')
  args = parser.parse_args()
  if args.verbose:
    shared.PRINT_STAGES = 1
    shared.DEBUG = True
  return args


if __name__ == '__main__':
  try:
    rv = main(get_args())
  except (Error, webassembly.InvalidWasmError, OSError) as e:
    print(f'{sys.argv[0]}: {str(e)}', file=sys.stderr)
    rv = 1
  sys.exit(rv)