Merge changes I1f5cebbc,I43a031a5 am: 99c24dc5b6 am: 46d947a2be am: df3ab60e7c am: b029d0772a

Original change: https://android-review.googlesource.com/c/platform/development/+/1376076

Change-Id: I891944c1c2ded85ae0f8c5377695dc9625809118
This commit is contained in:
Treehugger Robot
2020-07-31 04:42:31 +00:00
committed by Automerger Merge Worker
3 changed files with 106 additions and 116 deletions

View File

@@ -41,26 +41,26 @@ readonly CHECK_DIFF_RESULT
. build/envsetup.sh . build/envsetup.sh
lunch aosp_arm64 lunch aosp_arm64
build/soong/soong_ui.bash --make-mode compare_images build/soong/soong_ui.bash --make-mode compare_images
COMMON_WHITELIST=development/vndk/tools/image-diff-tool/whitelist.txt COMMON_ALLOWLIST=development/vndk/tools/image-diff-tool/allowlist.txt
out/host/linux-x86/bin/compare_images $1 -u -w "${COMMON_WHITELIST}" out/host/linux-x86/bin/compare_images $1 -u -w "${COMMON_ALLOWLIST}"
if [ -v DIST_DIR ]; then if [ -v DIST_DIR ]; then
cp common.csv diff.csv whitelisted_diff.csv "${DIST_DIR}" cp common.csv diff.csv allowlisted_diff.csv "${DIST_DIR}"
fi fi
echo >&2 echo >&2
echo >&2 " - Different parts (that are whitelisted)" echo >&2 " - Different parts (that are allowlisted)"
cat >&2 whitelisted_diff.csv cat >&2 allowlisted_diff.csv
echo >&2 echo >&2
echo >&2 " - Different parts (that are not whitelisted)" echo >&2 " - Different parts (that are not allowlisted)"
cat >&2 diff.csv cat >&2 diff.csv
if [ "$(wc -l diff.csv | cut -d' ' -f1)" -gt "1" ]; then if [ "$(wc -l diff.csv | cut -d' ' -f1)" -gt "1" ]; then
cat >&2 <<EOF cat >&2 <<EOF
[ERROR] Unexpected diffing files. [ERROR] Unexpected diffing files.
There are diffing files that are not ignored by whitelist. There are diffing files that are not ignored by allowlist.
The whitelist may need to be updated. The allowlist may need to be updated.
EOF EOF
if [ -n "${CHECK_DIFF_RESULT}" ]; then if [ -n "${CHECK_DIFF_RESULT}" ]; then
exit 1 exit 1

View File

@@ -12,41 +12,38 @@
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and # See the License for the specific language governing permissions and
# limitations under the License. # limitations under the License.
import os
import subprocess
import sys
from collections import defaultdict
from pathlib import Path
import hashlib
import argparse import argparse
import zipfile import collections
import fnmatch import fnmatch
import hashlib
import os
import pathlib
import subprocess
import tempfile import tempfile
import zipfile
def silent_call(cmd): def silent_call(cmd):
return subprocess.call(cmd, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL) == 0 return subprocess.call(
cmd, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL) == 0
def sha1sum(f): def sha1sum(f):
with open(f, 'rb') as fin: with open(f, "rb") as fin:
return hashlib.sha1(fin.read()).hexdigest() return hashlib.sha1(fin.read()).hexdigest()
def sha1sum_without_signing_key(filepath): def sha1sum_without_signing_key(filepath):
apk = zipfile.ZipFile(filepath) apk = zipfile.ZipFile(filepath)
l = [] l = []
for f in sorted(apk.namelist()): for f in sorted(apk.namelist()):
if f.startswith('META-INF/'): if f.startswith("META-INF/"):
continue continue
l.append(hashlib.sha1(apk.read(f)).hexdigest()) l.append(hashlib.sha1(apk.read(f)).hexdigest())
l.append(f) l.append(f)
return hashlib.sha1(",".join(l).encode()).hexdigest() return hashlib.sha1(",".join(l).encode()).hexdigest()
def strip_and_sha1sum(filepath): def strip_and_sha1sum(filepath):
# TODO: save striped file in tmp directory to support readonly directory. """Strip informations of elf file and calculate sha1 hash."""
tmp_filepath = filepath + '.tmp.no-build-id' tmp_filepath = filepath + ".tmp.no-build-id"
strip_all_and_remove_build_id = lambda: silent_call( llvm_strip = [
["llvm-strip", "--strip-all", "--keep-section=.ARM.attributes", "llvm-strip", "--strip-all", "--keep-section=.ARM.attributes",
"--remove-section=.note.gnu.build-id", filepath, "-o", tmp_filepath]) "--remove-section=.note.gnu.build-id", filepath, "-o", tmp_filepath
]
strip_all_and_remove_build_id = lambda: silent_call(llvm_strip)
try: try:
if strip_all_and_remove_build_id(): if strip_all_and_remove_build_id():
return sha1sum(tmp_filepath) return sha1sum(tmp_filepath)
@@ -55,65 +52,51 @@ def strip_and_sha1sum(filepath):
finally: finally:
if os.path.exists(tmp_filepath): if os.path.exists(tmp_filepath):
os.remove(tmp_filepath) os.remove(tmp_filepath)
return sha1sum(filepath) return sha1sum(filepath)
def make_filter_from_allowlists(allowlists, all_targets):
"""Creates a callable filter from a list of allowlist files.
def make_filter_from_whitelists(whitelists, all_targets): Allowlist can contain pathname patterns or skipped lines. Pathnames are case
"""Creates a callable filter from a list of whitelist files.
Whitelist can contain pathname patterns or ignored lines. Pathnames are case
insensitive. insensitive.
Allowlist can contain single-line comments. Comment lines begin with #
Whitelist can contain single-line comments. Comment lines begin with #
For example, this ignores the file "system/build.prop": For example, this ignores the file "system/build.prop":
SYSTEM/build.prop SYSTEM/build.prop
This ignores txt files: This ignores txt files:
*.txt *.txt
This ignores files in directory "system/dontcare/" This ignores files in directory "system/dontcare/"
SYSTEM/dontcare/* SYSTEM/dontcare/*
This ignores lines prefixed with pat1 or pat2 in file "system/build.prop": This ignores lines prefixed with pat1 or pat2 in file "system/build.prop":
SYSTEM/build.prop=pat1 pat2 SYSTEM/build.prop=pat1 pat2
Args: Args:
whitelists: A list of whitelist filenames. allowlists: A list of allowlist filenames.
all_targets: A list of targets to compare. all_targets: A list of targets to compare.
Returns: Returns:
A callable object that accepts a file pathname and returns True if the file A callable object that accepts a file pathname and returns True if the file
is ignored by the whitelists and False when it is not. is skipped by the allowlists and False when it is not.
""" """
ignored_patterns = set() skipped_patterns = set()
ignored_lines = defaultdict(list) skipped_lines = collections.defaultdict(list)
for whitelist in whitelists: for allowlist in allowlists:
if not os.path.isfile(whitelist): if not os.path.isfile(allowlist):
continue continue
with open(whitelist, 'rb') as f: with open(allowlist, "rb") as f:
for line in f: for line in f:
pat = line.strip().decode() pat = line.strip().decode()
if pat.startswith('#'): if pat.startswith("#"):
continue continue
if pat and pat[-1] == '\\': if pat and pat[-1] == "\\":
pat = pat.rstrip('\\') pat = pat.rstrip("\\")
if '=' in pat: if "=" in pat:
filename, prefixes = pat.split('=', 1) filename, prefixes = pat.split("=", 1)
prefixes = prefixes.split() prefixes = prefixes.split()
if prefixes: if prefixes:
ignored_lines[filename.lower()].extend(prefixes) skipped_lines[filename.lower()].extend(prefixes)
elif pat: elif pat:
ignored_patterns.add(pat.lower()) skipped_patterns.add(pat.lower())
def diff_with_skipped_lines(filename, prefixes):
def diff_with_ignored_lines(filename, prefixes):
"""Compares sha1 digest of file while ignoring lines. """Compares sha1 digest of file while ignoring lines.
Args: Args:
filename: File to compare among each target. filename: File to compare among each target.
prefixes: A list of prefixes. Lines that start with prefix are ignored. prefixes: A list of prefixes. Lines that start with prefix are skipped.
Returns: Returns:
True if file is identical among each target. True if file is identical among each target.
""" """
@@ -123,7 +106,7 @@ def make_filter_from_whitelists(whitelists, all_targets):
if not os.path.isfile(pathname): if not os.path.isfile(pathname):
return False return False
sha1 = hashlib.sha1() sha1 = hashlib.sha1()
with open(pathname, 'rb') as f: with open(pathname, "rb") as f:
for line in f: for line in f:
line_text = line.decode() line_text = line.decode()
if not any(line_text.startswith(prefix) for prefix in prefixes): if not any(line_text.startswith(prefix) for prefix in prefixes):
@@ -131,109 +114,111 @@ def make_filter_from_whitelists(whitelists, all_targets):
file_digest_respect_ignore.append(sha1.hexdigest()) file_digest_respect_ignore.append(sha1.hexdigest())
return (len(file_digest_respect_ignore) == len(all_targets) and return (len(file_digest_respect_ignore) == len(all_targets) and
len(set(file_digest_respect_ignore)) == 1) len(set(file_digest_respect_ignore)) == 1)
def allowlist_filter(filename):
def whitelist_filter(filename):
norm_filename = filename.lower() norm_filename = filename.lower()
for pattern in ignored_patterns: for pattern in skipped_patterns:
if fnmatch.fnmatch(norm_filename, pattern): if fnmatch.fnmatch(norm_filename, pattern):
return True return True
if norm_filename in ignored_lines: if norm_filename in skipped_lines:
ignored_prefixes = ignored_lines[norm_filename] skipped_prefixes = skipped_lines[norm_filename]
return diff_with_ignored_lines(filename, ignored_prefixes) return diff_with_skipped_lines(filename, skipped_prefixes)
return False return False
return allowlist_filter
return whitelist_filter def main(all_targets,
search_paths,
allowlists,
def main(all_targets, search_paths, whitelists, ignore_signing_key=False, list_only=False): ignore_signing_key=False,
list_only=False):
def run(path): def run(path):
is_native_component = silent_call(["llvm-objdump", "-a", path]) is_executable_component = silent_call(["llvm-objdump", "-a", path])
is_apk = path.endswith('.apk') is_apk = path.endswith(".apk")
if is_native_component: if is_executable_component:
return strip_and_sha1sum(path) return strip_and_sha1sum(path)
elif is_apk and ignore_signing_key: elif is_apk and ignore_signing_key:
return sha1sum_without_signing_key(path) return sha1sum_without_signing_key(path)
else: else:
return sha1sum(path) return sha1sum(path)
# artifact_sha1_target_map[filename][sha1] = list of targets # artifact_sha1_target_map[filename][sha1] = list of targets
artifact_sha1_target_map = defaultdict(lambda: defaultdict(list)) artifact_sha1_target_map = collections.defaultdict(
lambda: collections.defaultdict(list))
for target in all_targets: for target in all_targets:
paths = [] paths = []
for search_path in search_paths: for search_path in search_paths:
for path in Path(target, search_path).glob('**/*'): for path in pathlib.Path(target, search_path).glob("**/*"):
if path.exists() and not path.is_dir(): if path.exists() and not path.is_dir():
paths.append((str(path), str(path.relative_to(target)))) paths.append((str(path), str(path.relative_to(target))))
target_basename = os.path.basename(os.path.normpath(target)) target_basename = os.path.basename(os.path.normpath(target))
for path, filename in paths: for path, filename in paths:
sha1 = 0 sha1 = 0
if not list_only: if not list_only:
sha1 = run(path) sha1 = run(path)
artifact_sha1_target_map[filename][sha1].append(target_basename) artifact_sha1_target_map[filename][sha1].append(target_basename)
def pretty_print(sha1, filename, targets, exclude_sha1): def pretty_print(sha1, filename, targets, exclude_sha1):
if exclude_sha1: if exclude_sha1:
return '{}, {}\n'.format(filename, ';'.join(targets)) return "{}, {}\n".format(filename, ";".join(targets))
return "{}, {}, {}\n".format(filename, sha1[:10], ";".join(targets))
return '{}, {}, {}\n'.format(filename, sha1[:10], ';'.join(targets))
def is_common(sha1_target_map): def is_common(sha1_target_map):
for sha1, targets in sha1_target_map.items(): for _, targets in sha1_target_map.items():
return len(sha1_target_map) == 1 and len(targets) == len(all_targets) return len(sha1_target_map) == 1 and len(targets) == len(all_targets)
return False return False
allowlist_filter = make_filter_from_allowlists(allowlists, all_targets)
whitelist_filter = make_filter_from_whitelists(whitelists, all_targets)
common = [] common = []
diff = [] diff = []
whitelisted_diff = [] allowlisted_diff = []
for filename, sha1_target_map in artifact_sha1_target_map.items(): for filename, sha1_target_map in artifact_sha1_target_map.items():
if is_common(sha1_target_map): if is_common(sha1_target_map):
for sha1, targets in sha1_target_map.items(): for sha1, targets in sha1_target_map.items():
common.append(pretty_print(sha1, filename, targets, list_only)) common.append(pretty_print(sha1, filename, targets, list_only))
else: else:
if whitelist_filter(filename): if allowlist_filter(filename):
for sha1, targets in sha1_target_map.items(): for sha1, targets in sha1_target_map.items():
whitelisted_diff.append(pretty_print(sha1, filename, targets, list_only)) allowlisted_diff.append(
pretty_print(sha1, filename, targets, list_only))
else: else:
for sha1, targets in sha1_target_map.items(): for sha1, targets in sha1_target_map.items():
diff.append(pretty_print(sha1, filename, targets, list_only)) diff.append(pretty_print(sha1, filename, targets, list_only))
common = sorted(common) common = sorted(common)
diff = sorted(diff) diff = sorted(diff)
whitelisted_diff = sorted(whitelisted_diff) allowlisted_diff = sorted(allowlisted_diff)
header = "filename, sha1sum, targets\n" header = "filename, sha1sum, targets\n"
if list_only: if list_only:
header = "filename, targets\n" header = "filename, targets\n"
with open("common.csv", "w") as fout:
with open("common.csv", 'w') as fout:
fout.write(header) fout.write(header)
fout.writelines(common) fout.writelines(common)
with open("diff.csv", 'w') as fout: with open("diff.csv", "w") as fout:
fout.write(header) fout.write(header)
fout.writelines(diff) fout.writelines(diff)
with open("whitelisted_diff.csv", 'w') as fout: with open("allowlisted_diff.csv", "w") as fout:
fout.write(header) fout.write(header)
fout.writelines(whitelisted_diff) fout.writelines(allowlisted_diff)
def main_with_zip(extracted_paths, main_args):
def main_with_zip(extracted_paths, args): for origin_path, tmp_path in zip(main_args.target, extracted_paths):
for origin_path, tmp_path in zip(args.target, extracted_paths):
unzip_cmd = ["unzip", "-qd", tmp_path, os.path.join(origin_path, "*.zip")] unzip_cmd = ["unzip", "-qd", tmp_path, os.path.join(origin_path, "*.zip")]
unzip_cmd.extend([os.path.join(s, "*") for s in args.search_path]) unzip_cmd.extend([os.path.join(s, "*") for s in main_args.search_path])
subprocess.call(unzip_cmd) subprocess.call(unzip_cmd)
main(extracted_paths, args.search_path, args.whitelist, args.ignore_signing_key, list_only=args.list_only) main(
extracted_paths,
main_args.search_path,
main_args.allowlist,
main_args.ignore_signing_key,
list_only=main_args.list_only)
if __name__ == "__main__": if __name__ == "__main__":
parser = argparse.ArgumentParser(prog="compare_images", usage="compare_images -t model1 model2 [model...] -s dir1 [dir...] [-i] [-u] [-p] [-w whitelist1] [-w whitelist2]") parser = argparse.ArgumentParser(
parser.add_argument("-t", "--target", nargs='+', required=True) prog="compare_images",
parser.add_argument("-s", "--search_path", nargs='+', required=True) usage="compare_images -t model1 model2 [model...] -s dir1 [dir...] [-i] [-u] [-p] [-w allowlist1] [-w allowlist2]"
parser.add_argument("-i", "--ignore_signing_key", action='store_true') )
parser.add_argument("-l", "--list_only", action='store_true', help='Compare file list only and ignore SHA-1 diff') parser.add_argument("-t", "--target", nargs="+", required=True)
parser.add_argument("-u", "--unzip", action='store_true') parser.add_argument("-s", "--search_path", nargs="+", required=True)
parser.add_argument("-p", "--preserve_extracted_files", action='store_true') parser.add_argument("-i", "--ignore_signing_key", action="store_true")
parser.add_argument("-w", "--whitelist", action="append", default=[]) parser.add_argument(
"-l",
"--list_only",
action="store_true",
help="Compare file list only and ignore SHA-1 diff")
parser.add_argument("-u", "--unzip", action="store_true")
parser.add_argument("-p", "--preserve_extracted_files", action="store_true")
parser.add_argument("-w", "--allowlist", action="append", default=[])
args = parser.parse_args() args = parser.parse_args()
if len(args.target) < 2: if len(args.target) < 2:
parser.error("The number of targets has to be at least two.") parser.error("The number of targets has to be at least two.")
@@ -247,4 +232,9 @@ if __name__ == "__main__":
os.makedirs(p) os.makedirs(p)
main_with_zip(target_in_tmp, args) main_with_zip(target_in_tmp, args)
else: else:
main(args.target, args.search_path, args.whitelist, args.ignore_signing_key, list_only=args.list_only) main(
args.target,
args.search_path,
args.allowlist,
args.ignore_signing_key,
list_only=args.list_only)