Merge pull request #658 from littlefs-project/no-recursion

Restructure littlefs to not use recursion, measure stack usage
This commit is contained in:
Christopher Haster
2022-04-10 23:10:49 -05:00
14 changed files with 2584 additions and 752 deletions

View File

@@ -15,7 +15,7 @@ import csv
import collections as co
OBJ_PATHS = ['*.o', 'bd/*.o']
OBJ_PATHS = ['*.o']
def collect(paths, **args):
results = co.defaultdict(lambda: 0)
@@ -31,7 +31,8 @@ def collect(paths, **args):
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True)
universal_newlines=True,
errors='replace')
for line in proc.stdout:
m = pattern.match(line)
if m:
@@ -48,16 +49,30 @@ def collect(paths, **args):
# map to source files
if args.get('build_dir'):
file = re.sub('%s/*' % re.escape(args['build_dir']), '', file)
# replace .o with .c, different scripts report .o/.c, we need to
# choose one if we want to deduplicate csv files
file = re.sub('\.o$', '.c', file)
# discard internal functions
if func.startswith('__'):
continue
if not args.get('everything'):
if func.startswith('__'):
continue
# discard .8449 suffixes created by optimizer
func = re.sub('\.[0-9]+', '', func)
flat_results.append((file, func, size))
return flat_results
def main(**args):
def openio(path, mode='r'):
if path == '-':
if 'r' in mode:
return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
else:
return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
else:
return open(path, mode)
# find sizes
if not args.get('use', None):
# find .o files
@@ -75,13 +90,14 @@ def main(**args):
results = collect(paths, **args)
else:
with open(args['use']) as f:
with openio(args['use']) as f:
r = csv.DictReader(f)
results = [
( result['file'],
result['function'],
int(result['size']))
for result in r]
result['name'],
int(result['code_size']))
for result in r
if result.get('code_size') not in {None, ''}]
total = 0
for _, _, size in results:
@@ -89,13 +105,17 @@ def main(**args):
# find previous results?
if args.get('diff'):
with open(args['diff']) as f:
r = csv.DictReader(f)
prev_results = [
( result['file'],
result['function'],
int(result['size']))
for result in r]
try:
with openio(args['diff']) as f:
r = csv.DictReader(f)
prev_results = [
( result['file'],
result['name'],
int(result['code_size']))
for result in r
if result.get('code_size') not in {None, ''}]
except FileNotFoundError:
prev_results = []
prev_total = 0
for _, _, size in prev_results:
@@ -103,14 +123,34 @@ def main(**args):
# write results to CSV
if args.get('output'):
with open(args['output'], 'w') as f:
w = csv.writer(f)
w.writerow(['file', 'function', 'size'])
for file, func, size in sorted(results):
w.writerow((file, func, size))
merged_results = co.defaultdict(lambda: {})
other_fields = []
# merge?
if args.get('merge'):
try:
with openio(args['merge']) as f:
r = csv.DictReader(f)
for result in r:
file = result.pop('file', '')
func = result.pop('name', '')
result.pop('code_size', None)
merged_results[(file, func)] = result
other_fields = result.keys()
except FileNotFoundError:
pass
for file, func, size in results:
merged_results[(file, func)]['code_size'] = size
with openio(args['output'], 'w') as f:
w = csv.DictWriter(f, ['file', 'name', *other_fields, 'code_size'])
w.writeheader()
for (file, func), result in sorted(merged_results.items()):
w.writerow({'file': file, 'name': func, **result})
# print results
def dedup_entries(results, by='function'):
def dedup_entries(results, by='name'):
entries = co.defaultdict(lambda: 0)
for file, func, size in results:
entry = (file if by == 'file' else func)
@@ -126,45 +166,67 @@ def main(**args):
diff[name] = (old, new, new-old, (new-old)/old if old else 1.0)
return diff
def sorted_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1], x))
else:
return sorted(entries)
def sorted_diff_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1][1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1][1], x))
else:
return sorted(entries, key=lambda x: (-x[1][3], x))
def print_header(by=''):
if not args.get('diff'):
print('%-36s %7s' % (by, 'size'))
else:
print('%-36s %7s %7s %7s' % (by, 'old', 'new', 'diff'))
def print_entries(by='function'):
def print_entry(name, size):
print("%-36s %7d" % (name, size))
def print_diff_entry(name, old, new, diff, ratio):
print("%-36s %7s %7s %+7d%s" % (name,
old or "-",
new or "-",
diff,
' (%+.1f%%)' % (100*ratio) if ratio else ''))
def print_entries(by='name'):
entries = dedup_entries(results, by=by)
if not args.get('diff'):
print_header(by=by)
for name, size in sorted(entries.items()):
print("%-36s %7d" % (name, size))
for name, size in sorted_entries(entries.items()):
print_entry(name, size)
else:
prev_entries = dedup_entries(prev_results, by=by)
diff = diff_entries(prev_entries, entries)
print_header(by='%s (%d added, %d removed)' % (by,
sum(1 for old, _, _, _ in diff.values() if not old),
sum(1 for _, new, _, _ in diff.values() if not new)))
for name, (old, new, diff, ratio) in sorted(diff.items(),
key=lambda x: (-x[1][3], x)):
for name, (old, new, diff, ratio) in sorted_diff_entries(
diff.items()):
if ratio or args.get('all'):
print("%-36s %7s %7s %+7d%s" % (name,
old or "-",
new or "-",
diff,
' (%+.1f%%)' % (100*ratio) if ratio else ''))
print_diff_entry(name, old, new, diff, ratio)
def print_totals():
if not args.get('diff'):
print("%-36s %7d" % ('TOTAL', total))
print_entry('TOTAL', total)
else:
ratio = (total-prev_total)/prev_total if prev_total else 1.0
print("%-36s %7s %7s %+7d%s" % (
'TOTAL',
prev_total if prev_total else '-',
total if total else '-',
ratio = (0.0 if not prev_total and not total
else 1.0 if not prev_total
else (total-prev_total)/prev_total)
print_diff_entry('TOTAL',
prev_total, total,
total-prev_total,
' (%+.1f%%)' % (100*ratio) if ratio else ''))
ratio)
if args.get('quiet'):
pass
@@ -175,7 +237,7 @@ def main(**args):
print_entries(by='file')
print_totals()
else:
print_entries(by='function')
print_entries(by='name')
print_totals()
if __name__ == "__main__":
@@ -188,22 +250,30 @@ if __name__ == "__main__":
or a list of paths. Defaults to %r." % OBJ_PATHS)
parser.add_argument('-v', '--verbose', action='store_true',
help="Output commands that run behind the scenes.")
parser.add_argument('-q', '--quiet', action='store_true',
help="Don't show anything, useful with -o.")
parser.add_argument('-o', '--output',
help="Specify CSV file to store results.")
parser.add_argument('-u', '--use',
help="Don't compile and find code sizes, instead use this CSV file.")
parser.add_argument('-d', '--diff',
help="Specify CSV file to diff code size against.")
parser.add_argument('-m', '--merge',
help="Merge with an existing CSV file when writing to output.")
parser.add_argument('-a', '--all', action='store_true',
help="Show all functions, not just the ones that changed.")
parser.add_argument('--files', action='store_true',
parser.add_argument('-A', '--everything', action='store_true',
help="Include builtin and libc specific symbols.")
parser.add_argument('-s', '--size-sort', action='store_true',
help="Sort by size.")
parser.add_argument('-S', '--reverse-size-sort', action='store_true',
help="Sort by size, but backwards.")
parser.add_argument('-F', '--files', action='store_true',
help="Show file-level code sizes. Note this does not include padding! "
"So sizes may differ from other tools.")
parser.add_argument('-s', '--summary', action='store_true',
parser.add_argument('-Y', '--summary', action='store_true',
help="Only show the total code size.")
parser.add_argument('-q', '--quiet', action='store_true',
help="Don't show anything, useful with -o.")
parser.add_argument('--type', default='tTrRdDbB',
parser.add_argument('--type', default='tTrRdD',
help="Type of symbols to report, this uses the same single-character "
"type-names emitted by nm. Defaults to %(default)r.")
parser.add_argument('--nm-tool', default=['nm'], type=lambda x: x.split(),

View File

@@ -55,8 +55,9 @@ def collect(paths, **args):
for (file, func), (hits, count) in reduced_funcs.items():
# discard internal/testing functions (test_* injected with
# internal testing)
if func.startswith('__') or func.startswith('test_'):
continue
if not args.get('everything'):
if func.startswith('__') or func.startswith('test_'):
continue
# discard .8449 suffixes created by optimizer
func = re.sub('\.[0-9]+', '', func)
results.append((file, func, hits, count))
@@ -65,6 +66,15 @@ def collect(paths, **args):
def main(**args):
def openio(path, mode='r'):
if path == '-':
if 'r' in mode:
return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
else:
return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
else:
return open(path, mode)
# find coverage
if not args.get('use'):
# find *.info files
@@ -82,14 +92,16 @@ def main(**args):
results = collect(paths, **args)
else:
with open(args['use']) as f:
with openio(args['use']) as f:
r = csv.DictReader(f)
results = [
( result['file'],
result['function'],
int(result['hits']),
int(result['count']))
for result in r]
result['name'],
int(result['coverage_hits']),
int(result['coverage_count']))
for result in r
if result.get('coverage_hits') not in {None, ''}
if result.get('coverage_count') not in {None, ''}]
total_hits, total_count = 0, 0
for _, _, hits, count in results:
@@ -98,14 +110,19 @@ def main(**args):
# find previous results?
if args.get('diff'):
with open(args['diff']) as f:
r = csv.DictReader(f)
prev_results = [
( result['file'],
result['function'],
int(result['hits']),
int(result['count']))
for result in r]
try:
with openio(args['diff']) as f:
r = csv.DictReader(f)
prev_results = [
( result['file'],
result['name'],
int(result['coverage_hits']),
int(result['coverage_count']))
for result in r
if result.get('coverage_hits') not in {None, ''}
if result.get('coverage_count') not in {None, ''}]
except FileNotFoundError:
prev_results = []
prev_total_hits, prev_total_count = 0, 0
for _, _, hits, count in prev_results:
@@ -114,14 +131,36 @@ def main(**args):
# write results to CSV
if args.get('output'):
with open(args['output'], 'w') as f:
w = csv.writer(f)
w.writerow(['file', 'function', 'hits', 'count'])
for file, func, hits, count in sorted(results):
w.writerow((file, func, hits, count))
merged_results = co.defaultdict(lambda: {})
other_fields = []
# merge?
if args.get('merge'):
try:
with openio(args['merge']) as f:
r = csv.DictReader(f)
for result in r:
file = result.pop('file', '')
func = result.pop('name', '')
result.pop('coverage_hits', None)
result.pop('coverage_count', None)
merged_results[(file, func)] = result
other_fields = result.keys()
except FileNotFoundError:
pass
for file, func, hits, count in results:
merged_results[(file, func)]['coverage_hits'] = hits
merged_results[(file, func)]['coverage_count'] = count
with openio(args['output'], 'w') as f:
w = csv.DictWriter(f, ['file', 'name', *other_fields, 'coverage_hits', 'coverage_count'])
w.writeheader()
for (file, func), result in sorted(merged_results.items()):
w.writerow({'file': file, 'name': func, **result})
# print results
def dedup_entries(results, by='function'):
def dedup_entries(results, by='name'):
entries = co.defaultdict(lambda: (0, 0))
for file, func, hits, count in results:
entry = (file if by == 'file' else func)
@@ -147,23 +186,59 @@ def main(**args):
- (old_hits/old_count if old_count else 1.0)))
return diff
def sorted_entries(entries):
if args.get('coverage_sort'):
return sorted(entries, key=lambda x: (-(x[1][0]/x[1][1] if x[1][1] else -1), x))
elif args.get('reverse_coverage_sort'):
return sorted(entries, key=lambda x: (+(x[1][0]/x[1][1] if x[1][1] else -1), x))
else:
return sorted(entries)
def sorted_diff_entries(entries):
if args.get('coverage_sort'):
return sorted(entries, key=lambda x: (-(x[1][2]/x[1][3] if x[1][3] else -1), x))
elif args.get('reverse_coverage_sort'):
return sorted(entries, key=lambda x: (+(x[1][2]/x[1][3] if x[1][3] else -1), x))
else:
return sorted(entries, key=lambda x: (-x[1][6], x))
def print_header(by=''):
if not args.get('diff'):
print('%-36s %19s' % (by, 'hits/line'))
else:
print('%-36s %19s %19s %11s' % (by, 'old', 'new', 'diff'))
def print_entries(by='function'):
def print_entry(name, hits, count):
print("%-36s %11s %7s" % (name,
'%d/%d' % (hits, count)
if count else '-',
'%.1f%%' % (100*hits/count)
if count else '-'))
def print_diff_entry(name,
old_hits, old_count,
new_hits, new_count,
diff_hits, diff_count,
ratio):
print("%-36s %11s %7s %11s %7s %11s%s" % (name,
'%d/%d' % (old_hits, old_count)
if old_count else '-',
'%.1f%%' % (100*old_hits/old_count)
if old_count else '-',
'%d/%d' % (new_hits, new_count)
if new_count else '-',
'%.1f%%' % (100*new_hits/new_count)
if new_count else '-',
'%+d/%+d' % (diff_hits, diff_count),
' (%+.1f%%)' % (100*ratio) if ratio else ''))
def print_entries(by='name'):
entries = dedup_entries(results, by=by)
if not args.get('diff'):
print_header(by=by)
for name, (hits, count) in sorted(entries.items()):
print("%-36s %11s %7s" % (name,
'%d/%d' % (hits, count)
if count else '-',
'%.1f%%' % (100*hits/count)
if count else '-'))
for name, (hits, count) in sorted_entries(entries.items()):
print_entry(name, hits, count)
else:
prev_entries = dedup_entries(prev_results, by=by)
diff = diff_entries(prev_entries, entries)
@@ -173,45 +248,28 @@ def main(**args):
for name, (
old_hits, old_count,
new_hits, new_count,
diff_hits, diff_count, ratio) in sorted(diff.items(),
key=lambda x: (-x[1][6], x)):
diff_hits, diff_count, ratio) in sorted_diff_entries(
diff.items()):
if ratio or args.get('all'):
print("%-36s %11s %7s %11s %7s %11s%s" % (name,
'%d/%d' % (old_hits, old_count)
if old_count else '-',
'%.1f%%' % (100*old_hits/old_count)
if old_count else '-',
'%d/%d' % (new_hits, new_count)
if new_count else '-',
'%.1f%%' % (100*new_hits/new_count)
if new_count else '-',
'%+d/%+d' % (diff_hits, diff_count),
' (%+.1f%%)' % (100*ratio) if ratio else ''))
print_diff_entry(name,
old_hits, old_count,
new_hits, new_count,
diff_hits, diff_count,
ratio)
def print_totals():
if not args.get('diff'):
print("%-36s %11s %7s" % ('TOTAL',
'%d/%d' % (total_hits, total_count)
if total_count else '-',
'%.1f%%' % (100*total_hits/total_count)
if total_count else '-'))
print_entry('TOTAL', total_hits, total_count)
else:
ratio = ((total_hits/total_count
if total_count else 1.0)
- (prev_total_hits/prev_total_count
if prev_total_count else 1.0))
print("%-36s %11s %7s %11s %7s %11s%s" % ('TOTAL',
'%d/%d' % (prev_total_hits, prev_total_count)
if prev_total_count else '-',
'%.1f%%' % (100*prev_total_hits/prev_total_count)
if prev_total_count else '-',
'%d/%d' % (total_hits, total_count)
if total_count else '-',
'%.1f%%' % (100*total_hits/total_count)
if total_count else '-',
'%+d/%+d' % (total_hits-prev_total_hits,
total_count-prev_total_count),
' (%+.1f%%)' % (100*ratio) if ratio else ''))
print_diff_entry('TOTAL',
prev_total_hits, prev_total_count,
total_hits, total_count,
total_hits-prev_total_hits, total_count-prev_total_count,
ratio)
if args.get('quiet'):
pass
@@ -222,7 +280,7 @@ def main(**args):
print_entries(by='file')
print_totals()
else:
print_entries(by='function')
print_entries(by='name')
print_totals()
if __name__ == "__main__":
@@ -243,12 +301,23 @@ if __name__ == "__main__":
help="Don't do any work, instead use this CSV file.")
parser.add_argument('-d', '--diff',
help="Specify CSV file to diff code size against.")
parser.add_argument('-m', '--merge',
help="Merge with an existing CSV file when writing to output.")
parser.add_argument('-a', '--all', action='store_true',
help="Show all functions, not just the ones that changed.")
parser.add_argument('--files', action='store_true',
parser.add_argument('-A', '--everything', action='store_true',
help="Include builtin and libc specific symbols.")
parser.add_argument('-s', '--coverage-sort', action='store_true',
help="Sort by coverage.")
parser.add_argument('-S', '--reverse-coverage-sort', action='store_true',
help="Sort by coverage, but backwards.")
parser.add_argument('-F', '--files', action='store_true',
help="Show file-level coverage.")
parser.add_argument('-s', '--summary', action='store_true',
parser.add_argument('-Y', '--summary', action='store_true',
help="Only show the total coverage.")
parser.add_argument('-q', '--quiet', action='store_true',
help="Don't show anything, useful with -o.")
parser.add_argument('--build-dir',
help="Specify the relative build directory. Used to map object files \
to the correct source files.")
sys.exit(main(**vars(parser.parse_args())))

283
scripts/data.py Executable file
View File

@@ -0,0 +1,283 @@
#!/usr/bin/env python3
#
# Script to find data size at the function level. Basically just a bit wrapper
# around nm with some extra conveniences for comparing builds. Heavily inspired
# by Linux's Bloat-O-Meter.
#
import os
import glob
import itertools as it
import subprocess as sp
import shlex
import re
import csv
import collections as co
OBJ_PATHS = ['*.o']
def collect(paths, **args):
results = co.defaultdict(lambda: 0)
pattern = re.compile(
'^(?P<size>[0-9a-fA-F]+)' +
' (?P<type>[%s])' % re.escape(args['type']) +
' (?P<func>.+?)$')
for path in paths:
# note nm-tool may contain extra args
cmd = args['nm_tool'] + ['--size-sort', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace')
for line in proc.stdout:
m = pattern.match(line)
if m:
results[(path, m.group('func'))] += int(m.group('size'), 16)
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
for line in proc.stderr:
sys.stdout.write(line)
sys.exit(-1)
flat_results = []
for (file, func), size in results.items():
# map to source files
if args.get('build_dir'):
file = re.sub('%s/*' % re.escape(args['build_dir']), '', file)
# replace .o with .c, different scripts report .o/.c, we need to
# choose one if we want to deduplicate csv files
file = re.sub('\.o$', '.c', file)
# discard internal functions
if not args.get('everything'):
if func.startswith('__'):
continue
# discard .8449 suffixes created by optimizer
func = re.sub('\.[0-9]+', '', func)
flat_results.append((file, func, size))
return flat_results
def main(**args):
def openio(path, mode='r'):
if path == '-':
if 'r' in mode:
return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
else:
return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
else:
return open(path, mode)
# find sizes
if not args.get('use', None):
# find .o files
paths = []
for path in args['obj_paths']:
if os.path.isdir(path):
path = path + '/*.o'
for path in glob.glob(path):
paths.append(path)
if not paths:
print('no .obj files found in %r?' % args['obj_paths'])
sys.exit(-1)
results = collect(paths, **args)
else:
with openio(args['use']) as f:
r = csv.DictReader(f)
results = [
( result['file'],
result['name'],
int(result['data_size']))
for result in r
if result.get('data_size') not in {None, ''}]
total = 0
for _, _, size in results:
total += size
# find previous results?
if args.get('diff'):
try:
with openio(args['diff']) as f:
r = csv.DictReader(f)
prev_results = [
( result['file'],
result['name'],
int(result['data_size']))
for result in r
if result.get('data_size') not in {None, ''}]
except FileNotFoundError:
prev_results = []
prev_total = 0
for _, _, size in prev_results:
prev_total += size
# write results to CSV
if args.get('output'):
merged_results = co.defaultdict(lambda: {})
other_fields = []
# merge?
if args.get('merge'):
try:
with openio(args['merge']) as f:
r = csv.DictReader(f)
for result in r:
file = result.pop('file', '')
func = result.pop('name', '')
result.pop('data_size', None)
merged_results[(file, func)] = result
other_fields = result.keys()
except FileNotFoundError:
pass
for file, func, size in results:
merged_results[(file, func)]['data_size'] = size
with openio(args['output'], 'w') as f:
w = csv.DictWriter(f, ['file', 'name', *other_fields, 'data_size'])
w.writeheader()
for (file, func), result in sorted(merged_results.items()):
w.writerow({'file': file, 'name': func, **result})
# print results
def dedup_entries(results, by='name'):
entries = co.defaultdict(lambda: 0)
for file, func, size in results:
entry = (file if by == 'file' else func)
entries[entry] += size
return entries
def diff_entries(olds, news):
diff = co.defaultdict(lambda: (0, 0, 0, 0))
for name, new in news.items():
diff[name] = (0, new, new, 1.0)
for name, old in olds.items():
_, new, _, _ = diff[name]
diff[name] = (old, new, new-old, (new-old)/old if old else 1.0)
return diff
def sorted_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1], x))
else:
return sorted(entries)
def sorted_diff_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1][1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1][1], x))
else:
return sorted(entries, key=lambda x: (-x[1][3], x))
def print_header(by=''):
if not args.get('diff'):
print('%-36s %7s' % (by, 'size'))
else:
print('%-36s %7s %7s %7s' % (by, 'old', 'new', 'diff'))
def print_entry(name, size):
print("%-36s %7d" % (name, size))
def print_diff_entry(name, old, new, diff, ratio):
print("%-36s %7s %7s %+7d%s" % (name,
old or "-",
new or "-",
diff,
' (%+.1f%%)' % (100*ratio) if ratio else ''))
def print_entries(by='name'):
entries = dedup_entries(results, by=by)
if not args.get('diff'):
print_header(by=by)
for name, size in sorted_entries(entries.items()):
print_entry(name, size)
else:
prev_entries = dedup_entries(prev_results, by=by)
diff = diff_entries(prev_entries, entries)
print_header(by='%s (%d added, %d removed)' % (by,
sum(1 for old, _, _, _ in diff.values() if not old),
sum(1 for _, new, _, _ in diff.values() if not new)))
for name, (old, new, diff, ratio) in sorted_diff_entries(
diff.items()):
if ratio or args.get('all'):
print_diff_entry(name, old, new, diff, ratio)
def print_totals():
if not args.get('diff'):
print_entry('TOTAL', total)
else:
ratio = (0.0 if not prev_total and not total
else 1.0 if not prev_total
else (total-prev_total)/prev_total)
print_diff_entry('TOTAL',
prev_total, total,
total-prev_total,
ratio)
if args.get('quiet'):
pass
elif args.get('summary'):
print_header()
print_totals()
elif args.get('files'):
print_entries(by='file')
print_totals()
else:
print_entries(by='name')
print_totals()
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Find data size at the function level.")
parser.add_argument('obj_paths', nargs='*', default=OBJ_PATHS,
help="Description of where to find *.o files. May be a directory \
or a list of paths. Defaults to %r." % OBJ_PATHS)
parser.add_argument('-v', '--verbose', action='store_true',
help="Output commands that run behind the scenes.")
parser.add_argument('-q', '--quiet', action='store_true',
help="Don't show anything, useful with -o.")
parser.add_argument('-o', '--output',
help="Specify CSV file to store results.")
parser.add_argument('-u', '--use',
help="Don't compile and find data sizes, instead use this CSV file.")
parser.add_argument('-d', '--diff',
help="Specify CSV file to diff data size against.")
parser.add_argument('-m', '--merge',
help="Merge with an existing CSV file when writing to output.")
parser.add_argument('-a', '--all', action='store_true',
help="Show all functions, not just the ones that changed.")
parser.add_argument('-A', '--everything', action='store_true',
help="Include builtin and libc specific symbols.")
parser.add_argument('-s', '--size-sort', action='store_true',
help="Sort by size.")
parser.add_argument('-S', '--reverse-size-sort', action='store_true',
help="Sort by size, but backwards.")
parser.add_argument('-F', '--files', action='store_true',
help="Show file-level data sizes. Note this does not include padding! "
"So sizes may differ from other tools.")
parser.add_argument('-Y', '--summary', action='store_true',
help="Only show the total data size.")
parser.add_argument('--type', default='dDbB',
help="Type of symbols to report, this uses the same single-character "
"type-names emitted by nm. Defaults to %(default)r.")
parser.add_argument('--nm-tool', default=['nm'], type=lambda x: x.split(),
help="Path to the nm tool to use.")
parser.add_argument('--build-dir',
help="Specify the relative build directory. Used to map object files \
to the correct source files.")
sys.exit(main(**vars(parser.parse_args())))

430
scripts/stack.py Executable file
View File

@@ -0,0 +1,430 @@
#!/usr/bin/env python3
#
# Script to find stack usage at the function level. Will detect recursion and
# report as infinite stack usage.
#
import os
import glob
import itertools as it
import re
import csv
import collections as co
import math as m
CI_PATHS = ['*.ci']
def collect(paths, **args):
# parse the vcg format
k_pattern = re.compile('([a-z]+)\s*:', re.DOTALL)
v_pattern = re.compile('(?:"(.*?)"|([a-z]+))', re.DOTALL)
def parse_vcg(rest):
def parse_vcg(rest):
node = []
while True:
rest = rest.lstrip()
m = k_pattern.match(rest)
if not m:
return (node, rest)
k, rest = m.group(1), rest[m.end(0):]
rest = rest.lstrip()
if rest.startswith('{'):
v, rest = parse_vcg(rest[1:])
assert rest[0] == '}', "unexpected %r" % rest[0:1]
rest = rest[1:]
node.append((k, v))
else:
m = v_pattern.match(rest)
assert m, "unexpected %r" % rest[0:1]
v, rest = m.group(1) or m.group(2), rest[m.end(0):]
node.append((k, v))
node, rest = parse_vcg(rest)
assert rest == '', "unexpected %r" % rest[0:1]
return node
# collect into functions
results = co.defaultdict(lambda: (None, None, 0, set()))
f_pattern = re.compile(
r'([^\\]*)\\n([^:]*)[^\\]*\\n([0-9]+) bytes \((.*)\)')
for path in paths:
with open(path) as f:
vcg = parse_vcg(f.read())
for k, graph in vcg:
if k != 'graph':
continue
for k, info in graph:
if k == 'node':
info = dict(info)
m = f_pattern.match(info['label'])
if m:
function, file, size, type = m.groups()
if not args.get('quiet') and type != 'static':
print('warning: found non-static stack for %s (%s)'
% (function, type))
_, _, _, targets = results[info['title']]
results[info['title']] = (
file, function, int(size), targets)
elif k == 'edge':
info = dict(info)
_, _, _, targets = results[info['sourcename']]
targets.add(info['targetname'])
else:
continue
if not args.get('everything'):
for source, (s_file, s_function, _, _) in list(results.items()):
# discard internal functions
if s_file.startswith('<') or s_file.startswith('/usr/include'):
del results[source]
# find maximum stack size recursively, this requires also detecting cycles
# (in case of recursion)
def find_limit(source, seen=None):
seen = seen or set()
if source not in results:
return 0
_, _, frame, targets = results[source]
limit = 0
for target in targets:
if target in seen:
# found a cycle
return float('inf')
limit_ = find_limit(target, seen | {target})
limit = max(limit, limit_)
return frame + limit
def find_deps(targets):
deps = set()
for target in targets:
if target in results:
t_file, t_function, _, _ = results[target]
deps.add((t_file, t_function))
return deps
# flatten into a list
flat_results = []
for source, (s_file, s_function, frame, targets) in results.items():
limit = find_limit(source)
deps = find_deps(targets)
flat_results.append((s_file, s_function, frame, limit, deps))
return flat_results
def main(**args):
def openio(path, mode='r'):
if path == '-':
if 'r' in mode:
return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
else:
return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
else:
return open(path, mode)
# find sizes
if not args.get('use', None):
# find .ci files
paths = []
for path in args['ci_paths']:
if os.path.isdir(path):
path = path + '/*.ci'
for path in glob.glob(path):
paths.append(path)
if not paths:
print('no .ci files found in %r?' % args['ci_paths'])
sys.exit(-1)
results = collect(paths, **args)
else:
with openio(args['use']) as f:
r = csv.DictReader(f)
results = [
( result['file'],
result['name'],
int(result['stack_frame']),
float(result['stack_limit']), # note limit can be inf
set())
for result in r
if result.get('stack_frame') not in {None, ''}
if result.get('stack_limit') not in {None, ''}]
total_frame = 0
total_limit = 0
for _, _, frame, limit, _ in results:
total_frame += frame
total_limit = max(total_limit, limit)
# find previous results?
if args.get('diff'):
try:
with openio(args['diff']) as f:
r = csv.DictReader(f)
prev_results = [
( result['file'],
result['name'],
int(result['stack_frame']),
float(result['stack_limit']),
set())
for result in r
if result.get('stack_frame') not in {None, ''}
if result.get('stack_limit') not in {None, ''}]
except FileNotFoundError:
prev_results = []
prev_total_frame = 0
prev_total_limit = 0
for _, _, frame, limit, _ in prev_results:
prev_total_frame += frame
prev_total_limit = max(prev_total_limit, limit)
# write results to CSV
if args.get('output'):
merged_results = co.defaultdict(lambda: {})
other_fields = []
# merge?
if args.get('merge'):
try:
with openio(args['merge']) as f:
r = csv.DictReader(f)
for result in r:
file = result.pop('file', '')
func = result.pop('name', '')
result.pop('stack_frame', None)
result.pop('stack_limit', None)
merged_results[(file, func)] = result
other_fields = result.keys()
except FileNotFoundError:
pass
for file, func, frame, limit, _ in results:
merged_results[(file, func)]['stack_frame'] = frame
merged_results[(file, func)]['stack_limit'] = limit
with openio(args['output'], 'w') as f:
w = csv.DictWriter(f, ['file', 'name', *other_fields, 'stack_frame', 'stack_limit'])
w.writeheader()
for (file, func), result in sorted(merged_results.items()):
w.writerow({'file': file, 'name': func, **result})
# print results
def dedup_entries(results, by='name'):
entries = co.defaultdict(lambda: (0, 0, set()))
for file, func, frame, limit, deps in results:
entry = (file if by == 'file' else func)
entry_frame, entry_limit, entry_deps = entries[entry]
entries[entry] = (
entry_frame + frame,
max(entry_limit, limit),
entry_deps | {file if by == 'file' else func
for file, func in deps})
return entries
def diff_entries(olds, news):
diff = co.defaultdict(lambda: (None, None, None, None, 0, 0, 0, set()))
for name, (new_frame, new_limit, deps) in news.items():
diff[name] = (
None, None,
new_frame, new_limit,
new_frame, new_limit,
1.0,
deps)
for name, (old_frame, old_limit, _) in olds.items():
_, _, new_frame, new_limit, _, _, _, deps = diff[name]
diff[name] = (
old_frame, old_limit,
new_frame, new_limit,
(new_frame or 0) - (old_frame or 0),
0 if m.isinf(new_limit or 0) and m.isinf(old_limit or 0)
else (new_limit or 0) - (old_limit or 0),
0.0 if m.isinf(new_limit or 0) and m.isinf(old_limit or 0)
else +float('inf') if m.isinf(new_limit or 0)
else -float('inf') if m.isinf(old_limit or 0)
else +0.0 if not old_limit and not new_limit
else +1.0 if not old_limit
else ((new_limit or 0) - (old_limit or 0))/(old_limit or 0),
deps)
return diff
def sorted_entries(entries):
if args.get('limit_sort'):
return sorted(entries, key=lambda x: (-x[1][1], x))
elif args.get('reverse_limit_sort'):
return sorted(entries, key=lambda x: (+x[1][1], x))
elif args.get('frame_sort'):
return sorted(entries, key=lambda x: (-x[1][0], x))
elif args.get('reverse_frame_sort'):
return sorted(entries, key=lambda x: (+x[1][0], x))
else:
return sorted(entries)
def sorted_diff_entries(entries):
if args.get('limit_sort'):
return sorted(entries, key=lambda x: (-(x[1][3] or 0), x))
elif args.get('reverse_limit_sort'):
return sorted(entries, key=lambda x: (+(x[1][3] or 0), x))
elif args.get('frame_sort'):
return sorted(entries, key=lambda x: (-(x[1][2] or 0), x))
elif args.get('reverse_frame_sort'):
return sorted(entries, key=lambda x: (+(x[1][2] or 0), x))
else:
return sorted(entries, key=lambda x: (-x[1][6], x))
def print_header(by=''):
if not args.get('diff'):
print('%-36s %7s %7s' % (by, 'frame', 'limit'))
else:
print('%-36s %15s %15s %15s' % (by, 'old', 'new', 'diff'))
def print_entry(name, frame, limit):
print("%-36s %7d %7s" % (name,
frame, '' if m.isinf(limit) else int(limit)))
def print_diff_entry(name,
old_frame, old_limit,
new_frame, new_limit,
diff_frame, diff_limit,
ratio):
print('%-36s %7s %7s %7s %7s %+7d %7s%s' % (name,
old_frame if old_frame is not None else "-",
('' if m.isinf(old_limit) else int(old_limit))
if old_limit is not None else "-",
new_frame if new_frame is not None else "-",
('' if m.isinf(new_limit) else int(new_limit))
if new_limit is not None else "-",
diff_frame,
('+∞' if diff_limit > 0 and m.isinf(diff_limit)
else '-∞' if diff_limit < 0 and m.isinf(diff_limit)
else '%+d' % diff_limit),
'' if not ratio
else ' (+∞%)' if ratio > 0 and m.isinf(ratio)
else ' (-∞%)' if ratio < 0 and m.isinf(ratio)
else ' (%+.1f%%)' % (100*ratio)))
def print_entries(by='name'):
# build optional tree of dependencies
def print_deps(entries, depth, print,
filter=lambda _: True,
prefixes=('', '', '', '')):
entries = entries if isinstance(entries, list) else list(entries)
filtered_entries = [(name, entry)
for name, entry in entries
if filter(name)]
for i, (name, entry) in enumerate(filtered_entries):
last = (i == len(filtered_entries)-1)
print(prefixes[0+last] + name, entry)
if depth > 0:
deps = entry[-1]
print_deps(entries, depth-1, print,
lambda name: name in deps,
( prefixes[2+last] + "|-> ",
prefixes[2+last] + "'-> ",
prefixes[2+last] + "| ",
prefixes[2+last] + " "))
entries = dedup_entries(results, by=by)
if not args.get('diff'):
print_header(by=by)
print_deps(
sorted_entries(entries.items()),
args.get('depth') or 0,
lambda name, entry: print_entry(name, *entry[:-1]))
else:
prev_entries = dedup_entries(prev_results, by=by)
diff = diff_entries(prev_entries, entries)
print_header(by='%s (%d added, %d removed)' % (by,
sum(1 for _, old, _, _, _, _, _, _ in diff.values() if old is None),
sum(1 for _, _, _, new, _, _, _, _ in diff.values() if new is None)))
print_deps(
filter(
lambda x: x[1][6] or args.get('all'),
sorted_diff_entries(diff.items())),
args.get('depth') or 0,
lambda name, entry: print_diff_entry(name, *entry[:-1]))
def print_totals():
if not args.get('diff'):
print_entry('TOTAL', total_frame, total_limit)
else:
diff_frame = total_frame - prev_total_frame
diff_limit = (
0 if m.isinf(total_limit or 0) and m.isinf(prev_total_limit or 0)
else (total_limit or 0) - (prev_total_limit or 0))
ratio = (
0.0 if m.isinf(total_limit or 0) and m.isinf(prev_total_limit or 0)
else +float('inf') if m.isinf(total_limit or 0)
else -float('inf') if m.isinf(prev_total_limit or 0)
else 0.0 if not prev_total_limit and not total_limit
else 1.0 if not prev_total_limit
else ((total_limit or 0) - (prev_total_limit or 0))/(prev_total_limit or 0))
print_diff_entry('TOTAL',
prev_total_frame, prev_total_limit,
total_frame, total_limit,
diff_frame, diff_limit,
ratio)
if args.get('quiet'):
pass
elif args.get('summary'):
print_header()
print_totals()
elif args.get('files'):
print_entries(by='file')
print_totals()
else:
print_entries(by='name')
print_totals()
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Find stack usage at the function level.")
parser.add_argument('ci_paths', nargs='*', default=CI_PATHS,
help="Description of where to find *.ci files. May be a directory \
or a list of paths. Defaults to %r." % CI_PATHS)
parser.add_argument('-v', '--verbose', action='store_true',
help="Output commands that run behind the scenes.")
parser.add_argument('-q', '--quiet', action='store_true',
help="Don't show anything, useful with -o.")
parser.add_argument('-o', '--output',
help="Specify CSV file to store results.")
parser.add_argument('-u', '--use',
help="Don't parse callgraph files, instead use this CSV file.")
parser.add_argument('-d', '--diff',
help="Specify CSV file to diff against.")
parser.add_argument('-m', '--merge',
help="Merge with an existing CSV file when writing to output.")
parser.add_argument('-a', '--all', action='store_true',
help="Show all functions, not just the ones that changed.")
parser.add_argument('-A', '--everything', action='store_true',
help="Include builtin and libc specific symbols.")
parser.add_argument('-s', '--limit-sort', action='store_true',
help="Sort by stack limit.")
parser.add_argument('-S', '--reverse-limit-sort', action='store_true',
help="Sort by stack limit, but backwards.")
parser.add_argument('--frame-sort', action='store_true',
help="Sort by stack frame size.")
parser.add_argument('--reverse-frame-sort', action='store_true',
help="Sort by stack frame size, but backwards.")
parser.add_argument('-L', '--depth', default=0, type=lambda x: int(x, 0),
nargs='?', const=float('inf'),
help="Depth of dependencies to show.")
parser.add_argument('-F', '--files', action='store_true',
help="Show file-level calls.")
parser.add_argument('-Y', '--summary', action='store_true',
help="Only show the total stack size.")
parser.add_argument('--build-dir',
help="Specify the relative build directory. Used to map object files \
to the correct source files.")
sys.exit(main(**vars(parser.parse_args())))

331
scripts/structs.py Executable file
View File

@@ -0,0 +1,331 @@
#!/usr/bin/env python3
#
# Script to find struct sizes.
#
import os
import glob
import itertools as it
import subprocess as sp
import shlex
import re
import csv
import collections as co
OBJ_PATHS = ['*.o']
def collect(paths, **args):
decl_pattern = re.compile(
'^\s+(?P<no>[0-9]+)'
'\s+(?P<dir>[0-9]+)'
'\s+.*'
'\s+(?P<file>[^\s]+)$')
struct_pattern = re.compile(
'^(?:.*DW_TAG_(?P<tag>[a-z_]+).*'
'|^.*DW_AT_name.*:\s*(?P<name>[^:\s]+)\s*'
'|^.*DW_AT_decl_file.*:\s*(?P<decl>[0-9]+)\s*'
'|^.*DW_AT_byte_size.*:\s*(?P<size>[0-9]+)\s*)$')
results = co.defaultdict(lambda: 0)
for path in paths:
# find decl, we want to filter by structs in .h files
decls = {}
# note objdump-tool may contain extra args
cmd = args['objdump_tool'] + ['--dwarf=rawline', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace')
for line in proc.stdout:
# find file numbers
m = decl_pattern.match(line)
if m:
decls[int(m.group('no'))] = m.group('file')
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
for line in proc.stderr:
sys.stdout.write(line)
sys.exit(-1)
# collect structs as we parse dwarf info
found = False
name = None
decl = None
size = None
# note objdump-tool may contain extra args
cmd = args['objdump_tool'] + ['--dwarf=info', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace')
for line in proc.stdout:
# state machine here to find structs
m = struct_pattern.match(line)
if m:
if m.group('tag'):
if (name is not None
and decl is not None
and size is not None):
decl = decls.get(decl, '?')
results[(decl, name)] = size
found = (m.group('tag') == 'structure_type')
name = None
decl = None
size = None
elif found and m.group('name'):
name = m.group('name')
elif found and name and m.group('decl'):
decl = int(m.group('decl'))
elif found and name and m.group('size'):
size = int(m.group('size'))
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
for line in proc.stderr:
sys.stdout.write(line)
sys.exit(-1)
flat_results = []
for (file, struct), size in results.items():
# map to source files
if args.get('build_dir'):
file = re.sub('%s/*' % re.escape(args['build_dir']), '', file)
# only include structs declared in header files in the current
# directory, ignore internal-only # structs (these are represented
# in other measurements)
if not args.get('everything'):
if not file.endswith('.h'):
continue
# replace .o with .c, different scripts report .o/.c, we need to
# choose one if we want to deduplicate csv files
file = re.sub('\.o$', '.c', file)
flat_results.append((file, struct, size))
return flat_results
def main(**args):
def openio(path, mode='r'):
if path == '-':
if 'r' in mode:
return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
else:
return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
else:
return open(path, mode)
# find sizes
if not args.get('use', None):
# find .o files
paths = []
for path in args['obj_paths']:
if os.path.isdir(path):
path = path + '/*.o'
for path in glob.glob(path):
paths.append(path)
if not paths:
print('no .obj files found in %r?' % args['obj_paths'])
sys.exit(-1)
results = collect(paths, **args)
else:
with openio(args['use']) as f:
r = csv.DictReader(f)
results = [
( result['file'],
result['name'],
int(result['struct_size']))
for result in r
if result.get('struct_size') not in {None, ''}]
total = 0
for _, _, size in results:
total += size
# find previous results?
if args.get('diff'):
try:
with openio(args['diff']) as f:
r = csv.DictReader(f)
prev_results = [
( result['file'],
result['name'],
int(result['struct_size']))
for result in r
if result.get('struct_size') not in {None, ''}]
except FileNotFoundError:
prev_results = []
prev_total = 0
for _, _, size in prev_results:
prev_total += size
# write results to CSV
if args.get('output'):
merged_results = co.defaultdict(lambda: {})
other_fields = []
# merge?
if args.get('merge'):
try:
with openio(args['merge']) as f:
r = csv.DictReader(f)
for result in r:
file = result.pop('file', '')
struct = result.pop('name', '')
result.pop('struct_size', None)
merged_results[(file, struct)] = result
other_fields = result.keys()
except FileNotFoundError:
pass
for file, struct, size in results:
merged_results[(file, struct)]['struct_size'] = size
with openio(args['output'], 'w') as f:
w = csv.DictWriter(f, ['file', 'name', *other_fields, 'struct_size'])
w.writeheader()
for (file, struct), result in sorted(merged_results.items()):
w.writerow({'file': file, 'name': struct, **result})
# print results
def dedup_entries(results, by='name'):
entries = co.defaultdict(lambda: 0)
for file, struct, size in results:
entry = (file if by == 'file' else struct)
entries[entry] += size
return entries
def diff_entries(olds, news):
diff = co.defaultdict(lambda: (0, 0, 0, 0))
for name, new in news.items():
diff[name] = (0, new, new, 1.0)
for name, old in olds.items():
_, new, _, _ = diff[name]
diff[name] = (old, new, new-old, (new-old)/old if old else 1.0)
return diff
def sorted_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1], x))
else:
return sorted(entries)
def sorted_diff_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1][1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1][1], x))
else:
return sorted(entries, key=lambda x: (-x[1][3], x))
def print_header(by=''):
if not args.get('diff'):
print('%-36s %7s' % (by, 'size'))
else:
print('%-36s %7s %7s %7s' % (by, 'old', 'new', 'diff'))
def print_entry(name, size):
print("%-36s %7d" % (name, size))
def print_diff_entry(name, old, new, diff, ratio):
print("%-36s %7s %7s %+7d%s" % (name,
old or "-",
new or "-",
diff,
' (%+.1f%%)' % (100*ratio) if ratio else ''))
def print_entries(by='name'):
entries = dedup_entries(results, by=by)
if not args.get('diff'):
print_header(by=by)
for name, size in sorted_entries(entries.items()):
print_entry(name, size)
else:
prev_entries = dedup_entries(prev_results, by=by)
diff = diff_entries(prev_entries, entries)
print_header(by='%s (%d added, %d removed)' % (by,
sum(1 for old, _, _, _ in diff.values() if not old),
sum(1 for _, new, _, _ in diff.values() if not new)))
for name, (old, new, diff, ratio) in sorted_diff_entries(
diff.items()):
if ratio or args.get('all'):
print_diff_entry(name, old, new, diff, ratio)
def print_totals():
if not args.get('diff'):
print_entry('TOTAL', total)
else:
ratio = (0.0 if not prev_total and not total
else 1.0 if not prev_total
else (total-prev_total)/prev_total)
print_diff_entry('TOTAL',
prev_total, total,
total-prev_total,
ratio)
if args.get('quiet'):
pass
elif args.get('summary'):
print_header()
print_totals()
elif args.get('files'):
print_entries(by='file')
print_totals()
else:
print_entries(by='name')
print_totals()
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Find struct sizes.")
parser.add_argument('obj_paths', nargs='*', default=OBJ_PATHS,
help="Description of where to find *.o files. May be a directory \
or a list of paths. Defaults to %r." % OBJ_PATHS)
parser.add_argument('-v', '--verbose', action='store_true',
help="Output commands that run behind the scenes.")
parser.add_argument('-q', '--quiet', action='store_true',
help="Don't show anything, useful with -o.")
parser.add_argument('-o', '--output',
help="Specify CSV file to store results.")
parser.add_argument('-u', '--use',
help="Don't compile and find struct sizes, instead use this CSV file.")
parser.add_argument('-d', '--diff',
help="Specify CSV file to diff struct size against.")
parser.add_argument('-m', '--merge',
help="Merge with an existing CSV file when writing to output.")
parser.add_argument('-a', '--all', action='store_true',
help="Show all functions, not just the ones that changed.")
parser.add_argument('-A', '--everything', action='store_true',
help="Include builtin and libc specific symbols.")
parser.add_argument('-s', '--size-sort', action='store_true',
help="Sort by size.")
parser.add_argument('-S', '--reverse-size-sort', action='store_true',
help="Sort by size, but backwards.")
parser.add_argument('-F', '--files', action='store_true',
help="Show file-level struct sizes.")
parser.add_argument('-Y', '--summary', action='store_true',
help="Only show the total struct size.")
parser.add_argument('--objdump-tool', default=['objdump'], type=lambda x: x.split(),
help="Path to the objdump tool to use.")
parser.add_argument('--build-dir',
help="Specify the relative build directory. Used to map object files \
to the correct source files.")
sys.exit(main(**vars(parser.parse_args())))

279
scripts/summary.py Executable file
View File

@@ -0,0 +1,279 @@
#!/usr/bin/env python3
#
# Script to summarize the outputs of other scripts. Operates on CSV files.
#
import functools as ft
import collections as co
import os
import csv
import re
import math as m
# displayable fields
Field = co.namedtuple('Field', 'name,parse,acc,key,fmt,repr,null,ratio')
FIELDS = [
# name, parse, accumulate, fmt, print, null
Field('code',
lambda r: int(r['code_size']),
sum,
lambda r: r,
'%7s',
lambda r: r,
'-',
lambda old, new: (new-old)/old),
Field('data',
lambda r: int(r['data_size']),
sum,
lambda r: r,
'%7s',
lambda r: r,
'-',
lambda old, new: (new-old)/old),
Field('stack',
lambda r: float(r['stack_limit']),
max,
lambda r: r,
'%7s',
lambda r: '' if m.isinf(r) else int(r),
'-',
lambda old, new: (new-old)/old),
Field('structs',
lambda r: int(r['struct_size']),
sum,
lambda r: r,
'%8s',
lambda r: r,
'-',
lambda old, new: (new-old)/old),
Field('coverage',
lambda r: (int(r['coverage_hits']), int(r['coverage_count'])),
lambda rs: ft.reduce(lambda a, b: (a[0]+b[0], a[1]+b[1]), rs),
lambda r: r[0]/r[1],
'%19s',
lambda r: '%11s %7s' % ('%d/%d' % (r[0], r[1]), '%.1f%%' % (100*r[0]/r[1])),
'%11s %7s' % ('-', '-'),
lambda old, new: ((new[0]/new[1]) - (old[0]/old[1])))
]
def main(**args):
def openio(path, mode='r'):
if path == '-':
if 'r' in mode:
return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
else:
return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
else:
return open(path, mode)
# find results
results = co.defaultdict(lambda: {})
for path in args.get('csv_paths', '-'):
try:
with openio(path) as f:
r = csv.DictReader(f)
for result in r:
file = result.pop('file', '')
name = result.pop('name', '')
prev = results[(file, name)]
for field in FIELDS:
try:
r = field.parse(result)
if field.name in prev:
results[(file, name)][field.name] = field.acc(
[prev[field.name], r])
else:
results[(file, name)][field.name] = r
except (KeyError, ValueError):
pass
except FileNotFoundError:
pass
# find fields
if args.get('all_fields'):
fields = FIELDS
elif args.get('fields') is not None:
fields_dict = {field.name: field for field in FIELDS}
fields = [fields_dict[f] for f in args['fields']]
else:
fields = []
for field in FIELDS:
if any(field.name in result for result in results.values()):
fields.append(field)
# find total for every field
total = {}
for result in results.values():
for field in fields:
if field.name in result and field.name in total:
total[field.name] = field.acc(
[total[field.name], result[field.name]])
elif field.name in result:
total[field.name] = result[field.name]
# find previous results?
if args.get('diff'):
prev_results = co.defaultdict(lambda: {})
try:
with openio(args['diff']) as f:
r = csv.DictReader(f)
for result in r:
file = result.pop('file', '')
name = result.pop('name', '')
prev = prev_results[(file, name)]
for field in FIELDS:
try:
r = field.parse(result)
if field.name in prev:
prev_results[(file, name)][field.name] = field.acc(
[prev[field.name], r])
else:
prev_results[(file, name)][field.name] = r
except (KeyError, ValueError):
pass
except FileNotFoundError:
pass
prev_total = {}
for result in prev_results.values():
for field in fields:
if field.name in result and field.name in prev_total:
prev_total[field.name] = field.acc(
[prev_total[field.name], result[field.name]])
elif field.name in result:
prev_total[field.name] = result[field.name]
# print results
def dedup_entries(results, by='name'):
entries = co.defaultdict(lambda: {})
for (file, func), result in results.items():
entry = (file if by == 'file' else func)
prev = entries[entry]
for field in fields:
if field.name in result and field.name in prev:
entries[entry][field.name] = field.acc(
[prev[field.name], result[field.name]])
elif field.name in result:
entries[entry][field.name] = result[field.name]
return entries
def sorted_entries(entries):
if args.get('sort') is not None:
field = {field.name: field for field in FIELDS}[args['sort']]
return sorted(entries, key=lambda x: (
-(field.key(x[1][field.name])) if field.name in x[1] else -1, x))
elif args.get('reverse_sort') is not None:
field = {field.name: field for field in FIELDS}[args['reverse_sort']]
return sorted(entries, key=lambda x: (
+(field.key(x[1][field.name])) if field.name in x[1] else -1, x))
else:
return sorted(entries)
def print_header(by=''):
if not args.get('diff'):
print('%-36s' % by, end='')
for field in fields:
print((' '+field.fmt) % field.name, end='')
print()
else:
print('%-36s' % by, end='')
for field in fields:
print((' '+field.fmt) % field.name, end='')
print(' %-9s' % '', end='')
print()
def print_entry(name, result):
print('%-36s' % name, end='')
for field in fields:
r = result.get(field.name)
if r is not None:
print((' '+field.fmt) % field.repr(r), end='')
else:
print((' '+field.fmt) % '-', end='')
print()
def print_diff_entry(name, old, new):
print('%-36s' % name, end='')
for field in fields:
n = new.get(field.name)
if n is not None:
print((' '+field.fmt) % field.repr(n), end='')
else:
print((' '+field.fmt) % '-', end='')
o = old.get(field.name)
ratio = (
0.0 if m.isinf(o or 0) and m.isinf(n or 0)
else +float('inf') if m.isinf(n or 0)
else -float('inf') if m.isinf(o or 0)
else 0.0 if not o and not n
else +1.0 if not o
else -1.0 if not n
else field.ratio(o, n))
print(' %-9s' % (
'' if not ratio
else '(+∞%)' if ratio > 0 and m.isinf(ratio)
else '(-∞%)' if ratio < 0 and m.isinf(ratio)
else '(%+.1f%%)' % (100*ratio)), end='')
print()
def print_entries(by='name'):
entries = dedup_entries(results, by=by)
if not args.get('diff'):
print_header(by=by)
for name, result in sorted_entries(entries.items()):
print_entry(name, result)
else:
prev_entries = dedup_entries(prev_results, by=by)
print_header(by='%s (%d added, %d removed)' % (by,
sum(1 for name in entries if name not in prev_entries),
sum(1 for name in prev_entries if name not in entries)))
for name, result in sorted_entries(entries.items()):
if args.get('all') or result != prev_entries.get(name, {}):
print_diff_entry(name, prev_entries.get(name, {}), result)
def print_totals():
if not args.get('diff'):
print_entry('TOTAL', total)
else:
print_diff_entry('TOTAL', prev_total, total)
if args.get('summary'):
print_header()
print_totals()
elif args.get('files'):
print_entries(by='file')
print_totals()
else:
print_entries(by='name')
print_totals()
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Summarize measurements")
parser.add_argument('csv_paths', nargs='*', default='-',
help="Description of where to find *.csv files. May be a directory \
or list of paths. *.csv files will be merged to show the total \
coverage.")
parser.add_argument('-d', '--diff',
help="Specify CSV file to diff against.")
parser.add_argument('-a', '--all', action='store_true',
help="Show all objects, not just the ones that changed.")
parser.add_argument('-e', '--all-fields', action='store_true',
help="Show all fields, even those with no results.")
parser.add_argument('-f', '--fields', type=lambda x: re.split('\s*,\s*', x),
help="Comma separated list of fields to print, by default all fields \
that are found in the CSV files are printed.")
parser.add_argument('-s', '--sort',
help="Sort by this field.")
parser.add_argument('-S', '--reverse-sort',
help="Sort by this field, but backwards.")
parser.add_argument('-F', '--files', action='store_true',
help="Show file-level calls.")
parser.add_argument('-Y', '--summary', action='store_true',
help="Only show the totals.")
sys.exit(main(**vars(parser.parse_args())))

View File

@@ -784,10 +784,13 @@ def main(**args):
stdout=sp.PIPE if not args.get('verbose') else None,
stderr=sp.STDOUT if not args.get('verbose') else None,
universal_newlines=True)
stdout = []
for line in proc.stdout:
stdout.append(line)
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
for line in proc.stdout:
for line in stdout:
sys.stdout.write(line)
sys.exit(-1)
@@ -803,9 +806,9 @@ def main(**args):
failure.case.test(failure=failure, **args)
sys.exit(0)
print('tests passed %d/%d (%.2f%%)' % (passed, total,
print('tests passed %d/%d (%.1f%%)' % (passed, total,
100*(passed/total if total else 1.0)))
print('tests failed %d/%d (%.2f%%)' % (failed, total,
print('tests failed %d/%d (%.1f%%)' % (failed, total,
100*(failed/total if total else 1.0)))
return 1 if failed > 0 else 0