tests/scripts/audit-validity-dates.py

#!/usr/bin/env python3
#
# Copyright The Mbed TLS Contributors
# SPDX-License-Identifier: Apache-2.0
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

"""Audit validity date of X509 crt/crl/csr.

This script is used to audit the validity date of crt/crl/csr used for testing.
It prints the information about X.509 objects excluding the objects that
are valid throughout the desired validity period. The data are collected
from tests/data_files/ and tests/suites/*.data files by default.
"""

import os
import sys
import re
import typing
import argparse
import datetime
import glob
import logging
import hashlib
from enum import Enum

# The script requires cryptography >= 35.0.0 which is only available
# for Python >= 3.6.
import cryptography
from cryptography import x509

from generate_test_code import FileWrapper

import scripts_path # pylint: disable=unused-import
from mbedtls_dev import build_tree

def check_cryptography_version():
    match = re.match(r'^[0-9]+', cryptography.__version__)
    if match is None or int(match.group(0)) < 35:
        raise Exception("audit-validity-dates requires cryptography >= 35.0.0"
                        + "({} is too old)".format(cryptography.__version__))

class DataType(Enum):
    CRT = 1 # Certificate
    CRL = 2 # Certificate Revocation List
    CSR = 3 # Certificate Signing Request


class DataFormat(Enum):
    PEM = 1 # Privacy-Enhanced Mail
    DER = 2 # Distinguished Encoding Rules


class AuditData:
    """Store data location, type and validity period of X.509 objects."""
    #pylint: disable=too-few-public-methods
    def __init__(self, data_type: DataType, x509_obj):
        self.data_type = data_type
        # the locations that the x509 object could be found
        self.locations = [] # type: typing.List[str]
        self.fill_validity_duration(x509_obj)
        self._obj = x509_obj
        encoding = cryptography.hazmat.primitives.serialization.Encoding.DER
        self._identifier = hashlib.sha1(self._obj.public_bytes(encoding)).hexdigest()

    def __eq__(self, __value) -> bool:
        return self._obj == __value._obj

    @property
    def identifier(self):
        """
        Identifier of the underlying X.509 object, which is consistent across
        different runs.
        """
        return self._identifier

    def fill_validity_duration(self, x509_obj):
        """Read validity period from an X.509 object."""
        # Certificate expires after "not_valid_after"
        # Certificate is invalid before "not_valid_before"
        if self.data_type == DataType.CRT:
            self.not_valid_after = x509_obj.not_valid_after
            self.not_valid_before = x509_obj.not_valid_before
        # CertificateRevocationList expires after "next_update"
        # CertificateRevocationList is invalid before "last_update"
        elif self.data_type == DataType.CRL:
            self.not_valid_after = x509_obj.next_update
            self.not_valid_before = x509_obj.last_update
        # CertificateSigningRequest is always valid.
        elif self.data_type == DataType.CSR:
            self.not_valid_after = datetime.datetime.max
            self.not_valid_before = datetime.datetime.min
        else:
            raise ValueError("Unsupported file_type: {}".format(self.data_type))


class X509Parser:
    """A parser class to parse crt/crl/csr file or data in PEM/DER format."""
    PEM_REGEX = br'-{5}BEGIN (?P<type>.*?)-{5}(?P<data>.*?)-{5}END (?P=type)-{5}'
    PEM_TAG_REGEX = br'-{5}BEGIN (?P<type>.*?)-{5}\n'
    PEM_TAGS = {
        DataType.CRT: 'CERTIFICATE',
        DataType.CRL: 'X509 CRL',
        DataType.CSR: 'CERTIFICATE REQUEST'
    }

    def __init__(self,
                 backends:
                 typing.Dict[DataType,
                             typing.Dict[DataFormat,
                                         typing.Callable[[bytes], object]]]) \
    -> None:
        self.backends = backends
        self.__generate_parsers()

    def __generate_parser(self, data_type: DataType):
        """Parser generator for a specific DataType"""
        tag = self.PEM_TAGS[data_type]
        pem_loader = self.backends[data_type][DataFormat.PEM]
        der_loader = self.backends[data_type][DataFormat.DER]
        def wrapper(data: bytes):
            pem_type = X509Parser.pem_data_type(data)
            # It is in PEM format with target tag
            if pem_type == tag:
                return pem_loader(data)
            # It is in PEM format without target tag
            if pem_type:
                return None
            # It might be in DER format
            try:
                result = der_loader(data)
            except ValueError:
                result = None
            return result
        wrapper.__name__ = "{}.parser[{}]".format(type(self).__name__, tag)
        return wrapper

    def __generate_parsers(self):
        """Generate parsers for all support DataType"""
        self.parsers = {}
        for data_type, _ in self.PEM_TAGS.items():
            self.parsers[data_type] = self.__generate_parser(data_type)

    def __getitem__(self, item):
        return self.parsers[item]

    @staticmethod
    def pem_data_type(data: bytes) -> typing.Optional[str]:
        """Get the tag from the data in PEM format

        :param data: data to be checked in binary mode.
        :return: PEM tag or "" when no tag detected.
        """
        m = re.search(X509Parser.PEM_TAG_REGEX, data)
        if m is not None:
            return m.group('type').decode('UTF-8')
        else:
            return None

    @staticmethod
    def check_hex_string(hex_str: str) -> bool:
        """Check if the hex string is possibly DER data."""
        hex_len = len(hex_str)
        # At least 6 hex char for 3 bytes: Type + Length + Content
        if hex_len < 6:
            return False
        # Check if Type (1 byte) is SEQUENCE.
        if hex_str[0:2] != '30':
            return False
        # Check LENGTH (1 byte) value
        content_len = int(hex_str[2:4], base=16)
        consumed = 4
        if content_len in (128, 255):
            # Indefinite or Reserved
            return False
        elif content_len > 127:
            # Definite, Long
            length_len = (content_len - 128) * 2
            content_len = int(hex_str[consumed:consumed+length_len], base=16)
            consumed += length_len
        # Check LENGTH
        if hex_len != content_len * 2 + consumed:
            return False
        return True


class Auditor:
    """
    A base class that uses X509Parser to parse files to a list of AuditData.

    A subclass must implement the following methods:
      - collect_default_files: Return a list of file names that are defaultly
        used for parsing (auditing). The list will be stored in
        Auditor.default_files.
      - parse_file: Method that parses a single file to a list of AuditData.

    A subclass may override the following methods:
      - parse_bytes: Defaultly, it parses `bytes` that contains only one valid
        X.509 data(DER/PEM format) to an X.509 object.
      - walk_all: Defaultly, it iterates over all the files in the provided
        file name list, calls `parse_file` for each file and stores the results
        by extending the `results` passed to the function.
    """
    def __init__(self, logger):
        self.logger = logger
        self.default_files = self.collect_default_files()
        self.parser = X509Parser({
            DataType.CRT: {
                DataFormat.PEM: x509.load_pem_x509_certificate,
                DataFormat.DER: x509.load_der_x509_certificate
            },
            DataType.CRL: {
                DataFormat.PEM: x509.load_pem_x509_crl,
                DataFormat.DER: x509.load_der_x509_crl
            },
            DataType.CSR: {
                DataFormat.PEM: x509.load_pem_x509_csr,
                DataFormat.DER: x509.load_der_x509_csr
            },
        })

    def collect_default_files(self) -> typing.List[str]:
        """Collect the default files for parsing."""
        raise NotImplementedError

    def parse_file(self, filename: str) -> typing.List[AuditData]:
        """
        Parse a list of AuditData from file.

        :param filename: name of the file to parse.
        :return list of AuditData parsed from the file.
        """
        raise NotImplementedError

    def parse_bytes(self, data: bytes):
        """Parse AuditData from bytes."""
        for data_type in list(DataType):
            try:
                result = self.parser[data_type](data)
            except ValueError as val_error:
                result = None
                self.logger.warning(val_error)
            if result is not None:
                audit_data = AuditData(data_type, result)
                return audit_data
        return None

    def walk_all(self,
                 results: typing.Dict[str, AuditData],
                 file_list: typing.Optional[typing.List[str]] = None) \
        -> None:
        """
        Iterate over all the files in the list and get audit data. The
        results will be written to `results` passed to this function.

        :param results: The dictionary used to store the parsed 
                        AuditData. The keys of this dictionary should
                        be the identifier of the AuditData.
        """
        if file_list is None:
            file_list = self.default_files
        for filename in file_list:
            data_list = self.parse_file(filename)
            for d in data_list:
                if d.identifier in results:
                    results[d.identifier].locations.extend(d.locations)
                else:
                    results[d.identifier] = d

    @staticmethod
    def find_test_dir():
        """Get the relative path for the MbedTLS test directory."""
        return os.path.relpath(build_tree.guess_mbedtls_root() + '/tests')


class TestDataAuditor(Auditor):
    """Class for auditing files in `tests/data_files/`"""

    def collect_default_files(self):
        """Collect all files in `tests/data_files/`"""
        test_dir = self.find_test_dir()
        test_data_glob = os.path.join(test_dir, 'data_files/**')
        data_files = [f for f in glob.glob(test_data_glob, recursive=True)
                      if os.path.isfile(f)]
        return data_files

    def parse_file(self, filename: str) -> typing.List[AuditData]:
        """
        Parse a list of AuditData from data file.

        :param filename: name of the file to parse.
        :return list of AuditData parsed from the file.
        """
        with open(filename, 'rb') as f:
            data = f.read()

        results = []
        for idx, m in enumerate(re.finditer(X509Parser.PEM_REGEX, data, flags=re.S), 1):
            result = self.parse_bytes(data[m.start():m.end()])
            if result is not None:
                result.locations.append("{}#{}".format(filename, idx))
                results.append(result)

        return results


def parse_suite_data(data_f):
    """
    Parses .data file for test arguments that possiblly have a
    valid X.509 data. If you need a more precise parser, please
    use generate_test_code.parse_test_data instead.

    :param data_f: file object of the data file.
    :return: Generator that yields test function argument list.
    """
    for line in data_f:
        line = line.strip()
        # Skip comments
        if line.startswith('#'):
            continue

        # Check parameters line
        match = re.search(r'\A\w+(.*:)?\"', line)
        if match:
            # Read test vectors
            parts = re.split(r'(?<!\\):', line)
            parts = [x for x in parts if x]
            args = parts[1:]
            yield args


class SuiteDataAuditor(Auditor):
    """Class for auditing files in `tests/suites/*.data`"""

    def collect_default_files(self):
        """Collect all files in `tests/suites/*.data`"""
        test_dir = self.find_test_dir()
        suites_data_folder = os.path.join(test_dir, 'suites')
        data_files = glob.glob(os.path.join(suites_data_folder, '*.data'))
        return data_files

    def parse_file(self, filename: str):
        """
        Parse a list of AuditData from test suite data file.

        :param filename: name of the file to parse.
        :return list of AuditData parsed from the file.
        """
        audit_data_list = []
        data_f = FileWrapper(filename)
        for test_args in parse_suite_data(data_f):
            for idx, test_arg in enumerate(test_args):
                match = re.match(r'"(?P<data>[0-9a-fA-F]+)"', test_arg)
                if not match:
                    continue
                if not X509Parser.check_hex_string(match.group('data')):
                    continue
                audit_data = self.parse_bytes(bytes.fromhex(match.group('data')))
                if audit_data is None:
                    continue
                audit_data.locations.append("{}:{}:#{}".format(filename,
                                                               data_f.line_no,
                                                               idx + 1))
                audit_data_list.append(audit_data)

        return audit_data_list


def merge_auditdata(original: typing.List[AuditData]) \
    -> typing.List[AuditData]:
    """
    Multiple AuditData might be extracted from different locations for
    an identical X.509 object. Merge them into one entry in the list.
    """
    results = []
    for x in original:
        if x not in results:
            results.append(x)
        else:
            idx = results.index(x)
            results[idx].locations.extend(x.locations)
    return results


def list_all(audit_data: AuditData):
    for loc in audit_data.locations:
        print("{}\t{:20}\t{:20}\t{:3}\t{}".format(
            audit_data.identifier,
            audit_data.not_valid_before.isoformat(timespec='seconds'),
            audit_data.not_valid_after.isoformat(timespec='seconds'),
            audit_data.data_type.name,
            loc))


def configure_logger(logger: logging.Logger) -> None:
    """
    Configure the logging.Logger instance so that:
        - Format is set to "[%(levelname)s]: %(message)s".
        - loglevel >= WARNING are printed to stderr.
        - loglevel <  WARNING are printed to stdout.
    """
    class MaxLevelFilter(logging.Filter):
        # pylint: disable=too-few-public-methods
        def __init__(self, max_level, name=''):
            super().__init__(name)
            self.max_level = max_level

        def filter(self, record: logging.LogRecord) -> bool:
            return record.levelno <= self.max_level

    log_formatter = logging.Formatter("[%(levelname)s]: %(message)s")

    # set loglevel >= WARNING to be printed to stderr
    stderr_hdlr = logging.StreamHandler(sys.stderr)
    stderr_hdlr.setLevel(logging.WARNING)
    stderr_hdlr.setFormatter(log_formatter)

    # set loglevel <= INFO to be printed to stdout
    stdout_hdlr = logging.StreamHandler(sys.stdout)
    stdout_hdlr.addFilter(MaxLevelFilter(logging.INFO))
    stdout_hdlr.setFormatter(log_formatter)

    logger.addHandler(stderr_hdlr)
    logger.addHandler(stdout_hdlr)


def main():
    """
    Perform argument parsing.
    """
    parser = argparse.ArgumentParser(description=__doc__)

    parser.add_argument('-a', '--all',
                        action='store_true',
                        help='list the information of all the files')
    parser.add_argument('-v', '--verbose',
                        action='store_true', dest='verbose',
                        help='show logs')
    parser.add_argument('--from', dest='start_date',
                        help=('Start of desired validity period (UTC, YYYY-MM-DD). '
                              'Default: today'),
                        metavar='DATE')
    parser.add_argument('--to', dest='end_date',
                        help=('End of desired validity period (UTC, YYYY-MM-DD). '
                              'Default: --from'),
                        metavar='DATE')
    parser.add_argument('--data-files', action='append', nargs='*',
                        help='data files to audit',
                        metavar='FILE')
    parser.add_argument('--suite-data-files', action='append', nargs='*',
                        help='suite data files to audit',
                        metavar='FILE')

    args = parser.parse_args()

    # start main routine
    # setup logger
    logger = logging.getLogger()
    configure_logger(logger)
    logger.setLevel(logging.DEBUG if args.verbose else logging.ERROR)

    td_auditor = TestDataAuditor(logger)
    sd_auditor = SuiteDataAuditor(logger)

    data_files = []
    suite_data_files = []
    if args.data_files is None and args.suite_data_files is None:
        data_files = td_auditor.default_files
        suite_data_files = sd_auditor.default_files
    else:
        if args.data_files is not None:
            data_files = [x for l in args.data_files for x in l]
        if args.suite_data_files is not None:
            suite_data_files = [x for l in args.suite_data_files for x in l]

    # validity period start date
    if args.start_date:
        start_date = datetime.datetime.fromisoformat(args.start_date)
    else:
        start_date = datetime.datetime.today()
    # validity period end date
    if args.end_date:
        end_date = datetime.datetime.fromisoformat(args.end_date)
    else:
        end_date = start_date

    # go through all the files
    audit_results = {}
    td_auditor.walk_all(audit_results, data_files)
    sd_auditor.walk_all(audit_results, suite_data_files)

    logger.info("Total: {} objects found!".format(len(audit_results)))

    # we filter out the files whose validity duration covers the provided
    # duration.
    filter_func = lambda d: (start_date < d.not_valid_before) or \
                            (d.not_valid_after < end_date)

    sortby_end = lambda d: d.not_valid_after

    if args.all:
        filter_func = None

    # filter and output the results
    for d in sorted(filter(filter_func, audit_results.values()), key=sortby_end):
        list_all(d)

    logger.debug("Done!")

check_cryptography_version()
if __name__ == "__main__":
    main()