Delete old x86 validator.

src/trusted/validator_x86/testscripts/run_rdfa_validator_tests.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.
-
-import glob
-import optparse
-import os
-import re
-import struct
-import subprocess
-import sys
-import tempfile
-
-import test_format
-
-
-BUNDLE_SIZE = 32
-
-
-def AssertEquals(actual, expected):
-  if actual != expected:
-    raise AssertionError('\nEXPECTED:\n"""\n%s"""\n\nACTUAL:\n"""\n%s"""'
-                         % (expected, actual))
-
-
-def ParseHex(hex_content):
-  """Parse content of @hex section and return binary data
-
-  Args:
-    hex_content: Content of @hex section as a string.
-
-  Yields:
-    Chunks of binary data corresponding to lines of given @hex section (as
-    strings). If line ends with r'\\', chunk is continued on the following line.
-  """
-
-  bytes = []
-  for line in hex_content.split('\n'):
-    line, sep, comment = line.partition('#')
-    line = line.strip()
-    if line == '':
-      continue
-
-    if line.endswith(r'\\'):
-      line = line[:-2]
-      continuation = True
-    else:
-      continuation = False
-
-    for byte in line.split():
-      assert len(byte) == 2
-      bytes.append(chr(int(byte, 16)))
-
-    if not continuation:
-      assert len(bytes) > 0
-      yield ''.join(bytes)
-      bytes = []
-
-  assert bytes == [], r'r"\\" should not appear on the last line'
-
-
-def CreateElfContent(bits, text_segment):
-  e_ident = {
-      32: '\177ELF\1',
-      64: '\177ELF\2'}[bits]
-  e_machine = {
-      32: 3,
-      64: 62}[bits]
-
-  e_phoff = 256
-  e_phnum = 1
-  e_phentsize = 0
-
-  elf_header_fmt = {
-      32: '<16sHHIIIIIHHHHHH',
-      64: '<16sHHIQQQIHHHHHH'}[bits]
-
-  elf_header = struct.pack(
-      elf_header_fmt,
-      e_ident, 0, e_machine, 0, 0, e_phoff, 0, 0, 0,
-      e_phentsize, e_phnum, 0, 0, 0)
-
-  p_type = 1  # PT_LOAD
-  p_flags = 5  # r-x
-  p_filesz = len(text_segment)
-  p_memsz = p_filesz
-  p_vaddr = 0
-  p_offset = 512
-  p_align = 0
-  p_paddr = 0
-
-  pheader_fmt = {
-      32: '<IIIIIIII',
-      64: '<IIQQQQQQ'}[bits]
-
-  pheader_fields = {
-      32: (p_type, p_offset, p_vaddr, p_paddr,
-           p_filesz, p_memsz, p_flags, p_align),
-      64: (p_type, p_flags, p_offset, p_vaddr,
-           p_paddr, p_filesz, p_memsz, p_align)}[bits]
-
-  pheader = struct.pack(pheader_fmt, *pheader_fields)
-
-  result = elf_header
-  assert len(result) <= e_phoff
-  result += '\0' * (e_phoff - len(result))
-  result += pheader
-  assert len(result) <= p_offset
-  result += '\0' * (p_offset - len(result))
-  result += text_segment
-
-  return result
-
-
-def RunRdfaValidator(options, data):
-  # Add nops to make it bundle-sized.
-  data += (-len(data) % BUNDLE_SIZE) * '\x90'
-  assert len(data) % BUNDLE_SIZE == 0
-
-  tmp = tempfile.NamedTemporaryFile(
-      prefix='tmp_legacy_validator_', mode='wb', delete=False)
-  try:
-    tmp.write(CreateElfContent(options.bits, data))
-    tmp.close()
-
-    proc = subprocess.Popen([options.rdfaval, tmp.name],
-                            stdout=subprocess.PIPE,
-                            stderr=subprocess.PIPE)
-    stdout, stderr = proc.communicate()
-    assert stderr == '', stderr
-    return_code = proc.wait()
-  finally:
-    tmp.close()
-    os.remove(tmp.name)
-
-  # Remove the carriage return characters that we get on Windows.
-  stdout = stdout.replace('\r', '')
-  return return_code, stdout
-
-
-def ParseRdfaMessages(stdout):
-  """Get (offset, message) pairs from rdfa validator output.
-
-  Args:
-    stdout: Output of rdfa validator as string.
-
-  Yields:
-    Pairs (offset, message).
-  """
-  for line in stdout.split('\n'):
-    line = line.strip()
-    if line == '':
-      continue
-    if re.match(r"(Valid|Invalid)\.$", line):
-      continue
-
-    m = re.match(r'([0-9a-f]+): (.*)$', line, re.IGNORECASE)
-    assert m is not None, "can't parse line '%s'" % line
-    offset = int(m.group(1), 16)
-    message = m.group(2)
-
-    if not message.startswith('warning - '):
-      yield offset, message
-
-
-def RunRdfaWithNopPatching(options, data_chunks):
-  r"""Run RDFA validator with NOP patching for better error reporting.
-
-  If the RDFA validator encounters an invalid instruction, it resumes validation
-  from the beginning of the next bundle, while the original, non-DFA-based
-  validators skip maybe one or two bytes and recover. And there are plenty of
-  tests where there are more than one error in a single bundle. To mitigate such
-  spurious disagreements, the following procedure is used: when RDFA complaints
-  that particular piece can't be decoded, the problematic line in @hex section
-  (which usually corresponds to one instruction) is replaced with NOPs and the
-  validator is rerun from the beginning. This process may take several
-  iterations (it seems it always converges in practice). All errors reported on
-  all such runs (sans duplicate ones) are taken as validation result. So, in a
-  sense, this trick is to emulate line-level recovery as opposed to bundle-
-  level. In practice it turns out ok, and lots of spurious errors are
-  eliminated. To each error message we add the stage at which it was produced,
-  so we can destinguish 'primary' errors from additional ones.
-
-  Example. Suppose DE AD and BE EF machine codes correspond to invalid
-  instructions. Lets take a look at what happens when we invoke
-  RunRdfaWithNopPatching(options, ['\de\ad', '\be\ef']). First the RDFA
-  validator is run on the code '\de\ad\be\ef\90\90\90...'. It encounters an
-  undecipherable instruction, produces an error message at offset zero and
-  stops. Now we replace what is at offset zero ('\de\ad') with corresponding
-  amount of nops, and run the RDFA validator again on
-  '\90\90\be\ef\90\90\90...'. This time it decodes first two NOPs sucessfully
-  and reports problem at offset 2. In the next iteration of NOP patching BE EF
-  is replaced with 90 90 as well, no decoding errors are reported on the next
-  run so the whole process stops. Finally the combined output looks like
-  following:
-
-    0: [0] unrecognized instruction  <- produced at stage 0
-    2: [1] unrecognized instruction  <- produced at stage 1
-    return code: 1                   <- return code at stage 0
-
-  Args:
-    options: Options as produced by optparse.
-        Relevant fields are .bits and .update.
-    data_chunks: List of strings containing binary data. For the described
-        heuristic to work better it is desirable (although not absolutelty
-        required) that strings correspond to singular instructions, as it
-        usually happens in @hex section.
-
-  Returns:
-    String representing combined output from all stages. Error messages are
-    of the form
-      <offset in hex>: [<stage>] <message>
-  """
-
-  data_chunks = list(data_chunks)
-
-  offset_to_chunk = {}
-  offset = 0
-  for i, chunk in enumerate(data_chunks):
-    offset_to_chunk[offset] = i
-    offset += len(chunk)
-
-  first_return_code = None
-  messages = []  # list of triples (offset, stage, message)
-  messages_set = set()  # set of pairs (offset, message)
-  stage = 0
-
-  while True:
-    return_code, stdout = RunRdfaValidator(options, ''.join(data_chunks))
-    if first_return_code is None:
-      first_return_code = return_code
-
-    nop_patched = False
-
-    for offset, message in ParseRdfaMessages(stdout):
-      if (offset, message) in messages_set:
-        continue
-      messages.append((offset, stage, message))
-      messages_set.add((offset, message))
-
-      if offset in offset_to_chunk and message == 'unrecognized instruction':
-        chunk_no = offset_to_chunk[offset]
-        nops_chunk = '\x90' * len(data_chunks[chunk_no])
-        if nops_chunk != data_chunks[chunk_no]:
-          data_chunks[chunk_no] = nops_chunk
-          nop_patched = True
-
-    if not nop_patched:
-      break
-    stage += 1
-
-  messages.sort(key=lambda (offset, stage, _): (offset, stage))
-
-  result = ''.join('%x: [%d] %s\n' % (offset, stage, message)
-                   for offset, stage, message in messages)
-  result += 'return code: %d\n' % first_return_code
-  return result
-
-
-def CheckValidJumpTargets(options, data_chunks):
-  """
-  Check that the validator infers valid jump targets correctly.
-
-  This test checks that the validator identifies instruction boundaries and
-  superinstructions correctly. In order to do that, it attempts to append a jump
-  to each byte at the end of the given code. Jump should be valid if and only if
-  it goes to the boundary between data chunks.
-
-  Note that the same chunks as in RunRdfaWithNopPatching are used, but here they
-  play a different role. In RunRdfaWithNopPatching the partitioning into chunks
-  is only relevant when the whole snippet is invalid. Here, on the other hand,
-  we only care about valid snippets, and we use chunks to mark valid jump
-  targets.
-
-  Args:
-    options: Options as produced by optparse.
-    data_chunks: List of strings containing binary data. Each such chunk is
-        expected to correspond to indivisible instruction or superinstruction.
-
-  Returns:
-    None.
-  """
-  data = ''.join(data_chunks)
-  # Add nops to make it bundle-sized.
-  data += (-len(data) % BUNDLE_SIZE) * '\x90'
-  assert len(data) % BUNDLE_SIZE == 0
-
-  # Since we check validity of jump target by adding jump and validating
-  # resulting piece, we rely on validity of original snippet.
-  return_code, _ = RunRdfaValidator(options, data)
-  assert return_code == 0, 'Can only validate jump targets on valid snippet'
-
-  valid_jump_targets = set()
-  pos = 0
-  for data_chunk in data_chunks:
-    valid_jump_targets.add(pos)
-    pos += len(data_chunk)
-  valid_jump_targets.add(pos)
-
-  for i in range(pos + 1):
-    # Encode JMP with 32-bit relative target.
-    jump = '\xe9' + struct.pack('<i', i - (len(data) + 5))
-    return_code, _ = RunRdfaValidator(options, data + jump)
-    if return_code == 0:
-      assert i in valid_jump_targets, (
-          'Offset 0x%x was reported valid jump target' % i)
-    else:
-      assert i not in valid_jump_targets, (
-          'Offset 0x%x was reported invalid jump target' % i)
-
-
-def Test(options, items_list):
-  info = dict(items_list)
-
-  if 'rdfa_output' in info:
-    data_chunks = list(ParseHex(info['hex']))
-    stdout = RunRdfaWithNopPatching(options, data_chunks)
-    print '  Checking rdfa_output field...'
-    if options.update:
-      if stdout != info['rdfa_output']:
-        print '  Updating rdfa_output field...'
-        info['rdfa_output'] = stdout
-    else:
-      AssertEquals(stdout, info['rdfa_output'])
-
-    last_line = re.search('return code: (-?\d+)\n$', info['rdfa_output'])
-    expected_return_code = int(last_line.group(1))
-
-    # This test only works for valid snippets, see CheckValidJumpTargets
-    # for details.
-    if expected_return_code == 0:
-      print '  Checking jump targets...'
-      CheckValidJumpTargets(options, data_chunks)
-
-  # Update field values, but preserve their order.
-  items_list = [(field, info[field]) for field, _ in items_list]
-
-  return items_list
-
-
-def main(args):
-  parser = optparse.OptionParser()
-  parser.add_option('--rdfaval', default='validator_test',
-                    help='Path to the ncval validator executable')
-  parser.add_option('--bits',
-                    type=int,
-                    help='The subarchitecture to run tests against: 32 or 64')
-  parser.add_option('--update',
-                    default=False,
-                    action='store_true',
-                    help='Regenerate golden fields instead of testing')
-
-  options, args = parser.parse_args(args)
-
-  if options.bits not in [32, 64]:
-    parser.error('specify --bits 32 or --bits 64')
-
-  if len(args) == 0:
-    parser.error('No test files specified')
-  processed = 0
-  for glob_expr in args:
-    test_files = sorted(glob.glob(glob_expr))
-    if len(test_files) == 0:
-      raise AssertionError(
-          '%r matched no files, which was probably not intended' % glob_expr)
-    for test_file in test_files:
-      print 'Testing %s...' % test_file
-      tests = test_format.LoadTestFile(test_file)
-      tests = [Test(options, test) for test in tests]
-      if options.update:
-        test_format.SaveTestFile(tests, test_file)
-      processed += 1
-  print '%s test files were processed.' % processed
-
-
-if __name__ == '__main__':
-  main(sys.argv[1:])