afdo_redaction: add a remove_indirect_calls script

This CL adds a script that removes all indirect call info from a textual
AFDO profile. This is useful for reasons mentioned in the module's
docstring.

BUG=chromium:1005023
TEST=Ran on a profile. All indirect call info was removed, but the
profiles were otherwise identical.

Change-Id: I36caf2040d011e5cabd6f2058db784c71c42c25e
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/third_party/toolchain-utils/+/1856827
Commit-Queue: George Burgess <gbiv@chromium.org>
Legacy-Commit-Queue: Commit Bot <commit-bot@chromium.org>
Reviewed-by: Tiancong Wang <tcwang@google.com>
Tested-by: George Burgess <gbiv@chromium.org>

2 files changed, 148 insertions, 0 deletions

diff --git a/afdo_redaction/remove_indirect_calls.py b/afdo_redaction/remove_indirect_calls.py
new file mode 100755
index 00000000..b879b2f0
--- /dev/null
+++ b/afdo_redaction/remove_indirect_calls.py
@@ -0,0 +1,81 @@
+#!/usr/bin/env python2
+# -*- coding: utf-8 -*-
+# Copyright 2019 The Chromium OS Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+"""Script to remove all indirect call targets from textual AFDO profiles.
+
+Indirect call samples can cause code to appear 'live' when it otherwise
+wouldn't be. This resurrection can happen either by the way of profile-based
+speculative devirtualization, or because of imprecision in LLVM's liveness
+calculations when performing LTO.
+
+This generally isn't a problem when an AFDO profile is applied to the binary it
+was collected on. However, because we e.g., build NaCl from the same set of
+objects as Chrome, this can become problematic, and lead to NaCl doubling in
+size (or worse). See crbug.com/1005023 and crbug.com/916130.
+"""
+
+from __future__ import division, print_function
+
+import argparse
+import re
+import sys
+
+
+def _remove_indirect_call_targets(lines):
+  # Lines with indirect call targets look like:
+  #   1.1: 1234 foo:111 bar:122
+  #
+  # Where 1.1 is the line info/discriminator, 1234 is the total number of
+  # samples seen for that line/discriminator, foo:111 is "111 of the calls here
+  # went to foo," and bar:122 is "122 of the calls here went to bar."
+  call_target_re = re.compile(
+      r"""
+      ^\s+                    # Top-level lines are function records.
+      \d+(?:\.\d+)?:          # Line info/discriminator
+      \s+
+      \d+                     # Total sample count
+      \s+
+      ((?:[^\s:]+:\d+\s*)+)   # Indirect call target(s)
+      $
+  """, re.VERBOSE)
+  for line in lines:
+    line = line.rstrip()
+
+    match = call_target_re.match(line)
+    if not match:
+      yield line + '\n'
+      continue
+
+    group_start, group_end = match.span(1)
+    assert group_end == len(line)
+    yield line[:group_start].rstrip() + '\n'
+
+
+def run(input_stream, output_stream):
+  for line in _remove_indirect_call_targets(input_stream):
+    output_stream.write(line)
+
+
+def main():
+  parser = argparse.ArgumentParser(
+      description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
+  parser.add_argument(
+      '--input',
+      default='/dev/stdin',
+      help='File to read from. Defaults to stdin.')
+  parser.add_argument(
+      '--output',
+      default='/dev/stdout',
+      help='File to write to. Defaults to stdout.')
+  args = parser.parse_args()
+
+  with open(args.input) as stdin:
+    with open(args.output, 'w') as stdout:
+      run(stdin, stdout)
+
+
+if __name__ == '__main__':
+  main()
diff --git a/afdo_redaction/remove_indirect_calls_test.py b/afdo_redaction/remove_indirect_calls_test.py
new file mode 100755
index 00000000..1499af25
--- /dev/null
+++ b/afdo_redaction/remove_indirect_calls_test.py
@@ -0,0 +1,67 @@
+#!/usr/bin/env python2
+# -*- coding: utf-8 -*-
+# Copyright 2019 The Chromium OS Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+"""Tests for remove_indirect_calls"""
+
+from __future__ import print_function
+
+import io
+import unittest
+
+import remove_indirect_calls
+
+
+def _run_test(input_lines):
+  input_buf = io.BytesIO('\n'.join(input_lines))
+  output_buf = io.BytesIO()
+  remove_indirect_calls.run(input_buf, output_buf)
+  return output_buf.getvalue().splitlines()
+
+
+class Test(unittest.TestCase):
+  """Tests"""
+
+  def test_empty_profile(self):
+    self.assertEqual(_run_test([]), [])
+
+  def test_removal_on_real_world_code(self):
+    # These are copied from an actual textual AFDO profile, but the names made
+    # lints unhappy due to their length, so I had to be creative.
+    profile_lines = """_ZLongSymbolName:52862:1766
+ 14: 2483
+ 8.1: _SomeInlinedSym:45413
+  11: _AndAnother:35481
+   2: 2483
+   2.1: _YetAnother:25549
+    3: 2483
+    3.1: 351
+    3.3: 2526 IndirectTarg1:675 Targ2:397 Targ3:77
+  13.2: Whee:9932
+   1.1: Whoo:9932
+    0: BleepBloop:9932
+     0: 2483
+  """.strip().splitlines()
+
+    expected_lines = """_ZLongSymbolName:52862:1766
+ 14: 2483
+ 8.1: _SomeInlinedSym:45413
+  11: _AndAnother:35481
+   2: 2483
+   2.1: _YetAnother:25549
+    3: 2483
+    3.1: 351
+    3.3: 2526
+  13.2: Whee:9932
+   1.1: Whoo:9932
+    0: BleepBloop:9932
+     0: 2483
+  """.strip().splitlines()
+
+    self.assertEqual(_run_test(profile_lines), expected_lines)
+
+
+if __name__ == '__main__':
+  unittest.main()