1
2
3
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
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
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
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
|
#!/usr/bin/env vpython
# Copyright 2017 The Chromium Authors. All rights reserved.
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.
"""Compares two orderfiles, from filenames or a commit.
This shows some statistics about two orderfiles, possibly extracted from an
updating commit made by the orderfile bot.
"""
from __future__ import print_function
import argparse
import collections
import logging
import os
import subprocess
import sys
def ParseOrderfile(filename):
"""Parses an orderfile into a list of symbols.
Args:
filename: (str) Path to the orderfile.
Returns:
[str] List of symbols.
"""
symbols = []
lines = []
already_seen = set()
with open(filename, 'r') as f:
lines = [line.strip() for line in f]
# The (new) orderfiles that are oriented at the LLD linker contain only symbol
# names (i.e. not prefixed with '.text'). The (old) orderfiles aimed at the
# Gold linker were patched by duplicating symbols prefixed with '.text.hot.',
# '.text.unlikely.' and '.text.', hence the appearance of '.text' on the first
# symbol indicates such a legacy orderfile.
if not lines[0].startswith('.text.'):
for entry in lines:
symbol_name = entry.rstrip('\n')
assert symbol_name != '*' and symbol_name != '.text'
already_seen.add(symbol_name)
symbols.append(symbol_name)
else:
for entry in lines:
# Keep only (input) section names, not symbol names (only rare special
# symbols contain '.'). We could only keep symbols, but then some even
# older orderfiles would not be parsed.
if '.' not in entry:
continue
# Example: .text.startup.BLA
symbol_name = entry[entry.rindex('.'):]
if symbol_name in already_seen or symbol_name == '*' or entry == '.text':
continue
already_seen.add(symbol_name)
symbols.append(symbol_name)
return symbols
def CommonSymbolsToOrder(symbols, common_symbols):
"""Returns s -> index for all s in common_symbols."""
result = {}
index = 0
for s in symbols:
if s not in common_symbols:
continue
result[s] = index
index += 1
return result
CompareResult = collections.namedtuple(
'CompareResult', ('first_count', 'second_count',
'new_count', 'removed_count',
'average_fractional_distance'))
def Compare(first_filename, second_filename):
"""Outputs a comparison of two orderfiles to stdout.
Args:
first_filename: (str) First orderfile.
second_filename: (str) Second orderfile.
Returns:
An instance of CompareResult.
"""
first_symbols = ParseOrderfile(first_filename)
second_symbols = ParseOrderfile(second_filename)
print('Symbols count:\n\tfirst:\t%d\n\tsecond:\t%d' % (len(first_symbols),
len(second_symbols)))
first_symbols = set(first_symbols)
second_symbols = set(second_symbols)
new_symbols = second_symbols - first_symbols
removed_symbols = first_symbols - second_symbols
common_symbols = first_symbols & second_symbols
# Distance between orderfiles.
first_to_ordering = CommonSymbolsToOrder(first_symbols, common_symbols)
second_to_ordering = CommonSymbolsToOrder(second_symbols, common_symbols)
total_distance = sum(abs(first_to_ordering[s] - second_to_ordering[s])\
for s in first_to_ordering)
# Each distance is in [0, len(common_symbols)] and there are
# len(common_symbols) entries, hence the normalization.
average_fractional_distance = float(total_distance) / (len(common_symbols)**2)
print('New symbols = %d' % len(new_symbols))
print('Removed symbols = %d' % len(removed_symbols))
print('Average fractional distance = %.2f%%' %
(100. * average_fractional_distance))
return CompareResult(len(first_symbols), len(second_symbols),
len(new_symbols), len(removed_symbols),
average_fractional_distance)
def CheckOrderfileCommit(commit_hash, clank_path):
"""Asserts that a commit is an orderfile update from the bot.
Args:
commit_hash: (str) Git hash of the orderfile roll commit.
clank_path: (str) Path to the clank repository.
"""
output = subprocess.check_output(['git', 'show', r'--format=%s', commit_hash],
cwd=clank_path)
first_line = output.split('\n')[0]
# Capitalization changed at some point. Not checking the bot name because it
# changed too.
assert first_line.upper().endswith(
'Update Orderfile.'.upper()), ('Not an orderfile commit')
def GetBeforeAfterOrderfileHashes(commit_hash, clank_path):
"""Downloads the orderfiles before and afer an orderfile roll.
Args:
commit_hash: (str) Git hash of the orderfile roll commit.
clank_path: (str) Path to the clank repository.
Returns:
(str, str) Path to the before and after commit orderfiles.
"""
orderfile_hash_relative_path = 'orderfiles/orderfile.arm.out.sha1'
before_output = subprocess.check_output(
['git', 'show', '%s^:%s' % (commit_hash, orderfile_hash_relative_path)],
cwd=clank_path)
before_hash = before_output.split('\n')[0]
after_output = subprocess.check_output(
['git', 'show', '%s:%s' % (commit_hash, orderfile_hash_relative_path)],
cwd=clank_path)
after_hash = after_output.split('\n')[0]
assert before_hash != after_hash
return (before_hash, after_hash)
def DownloadOrderfile(orderfile_hash, output_filename):
"""Downloads an orderfile with a given hash to a given destination."""
cloud_storage_path = (
'gs://clank-archive/orderfile-clankium/%s' % orderfile_hash)
subprocess.check_call(
['gsutil.py', 'cp', cloud_storage_path, output_filename])
def GetOrderfilesFromCommit(commit_hash):
"""Returns paths to the before and after orderfiles for a commit."""
clank_path = os.path.join(os.path.dirname(__file__), os.pardir, os.pardir,
'clank')
logging.info('Checking that the commit is an orderfile')
CheckOrderfileCommit(commit_hash, clank_path)
(before_hash, after_hash) = GetBeforeAfterOrderfileHashes(
commit_hash, clank_path)
logging.info('Before / after hashes: %s %s', before_hash, after_hash)
before_filename = os.path.join('/tmp/', before_hash)
after_filename = os.path.join('/tmp/', after_hash)
logging.info('Downloading files')
DownloadOrderfile(before_hash, before_filename)
DownloadOrderfile(after_hash, after_filename)
return (before_filename, after_filename)
def CreateArgumentParser():
"""Returns the argumeng parser."""
parser = argparse.ArgumentParser()
parser.add_argument('--first', help='First orderfile')
parser.add_argument('--second', help='Second orderfile')
parser.add_argument('--keep', default=False, action='store_true',
help='Keep the downloaded orderfiles')
parser.add_argument('--from-commit', help='Analyze the difference in the '
'orderfile from an orderfile bot commit.')
parser.add_argument('--csv-output', help='Appends the result to a CSV file.')
return parser
def main():
logging.basicConfig(level=logging.INFO)
parser = CreateArgumentParser()
args = parser.parse_args()
if args.first or args.second:
assert args.first and args.second, 'Need both files.'
Compare(args.first, args.second)
elif args.from_commit:
first, second = GetOrderfilesFromCommit(args.from_commit)
try:
logging.info('Comparing the orderfiles')
result = Compare(first, second)
if args.csv_output:
with open(args.csv_output, 'a') as f:
f.write('%s,%d,%d,%d,%d,%f\n' % tuple(
[args.from_commit] + list(result)))
finally:
if not args.keep:
os.remove(first)
os.remove(second)
else:
return False
return True
if __name__ == '__main__':
sys.exit(0 if main() else 1)
|
