Merge pull request #8222 from tgonzalezorlandoarm/tg/backport-psa-low-hash-mac-size

Backport 2.28: Start testing the PSA built-in drivers: hashes

scripts/mbedtls_dev/crypto_data_tests.py

@@ -0,0 +1,112 @@
+"""Generate test data for cryptographic mechanisms.
+This module is a work in progress, only implementing a few cases for now.
+"""
+
+# Copyright The Mbed TLS Contributors
+# SPDX-License-Identifier: Apache-2.0 OR GPL-2.0-or-later
+
+
+import hashlib
+from typing import Callable, Dict, Iterator, List, Optional #pylint: disable=unused-import
+
+from . import crypto_knowledge
+from . import psa_information
+from . import test_case
+
+
+def psa_low_level_dependencies(*expressions: str) -> List[str]:
+    """Infer dependencies of a PSA low-level test case by looking for PSA_xxx symbols.
+    This function generates MBEDTLS_PSA_BUILTIN_xxx symbols.
+    """
+    high_level = psa_information.automatic_dependencies(*expressions)
+    for dep in high_level:
+        assert dep.startswith('PSA_WANT_')
+    return ['MBEDTLS_PSA_BUILTIN_' + dep[9:] for dep in high_level]
+
+
+class HashPSALowLevel:
+    """Generate test cases for the PSA low-level hash interface."""
+
+    def __init__(self, info: psa_information.Information) -> None:
+        self.info = info
+        base_algorithms = sorted(info.constructors.algorithms)
+        all_algorithms = \
+            [crypto_knowledge.Algorithm(expr)
+             for expr in info.constructors.generate_expressions(base_algorithms)]
+        self.algorithms = \
+            [alg
+             for alg in all_algorithms
+             if (not alg.is_wildcard and
+                 alg.can_do(crypto_knowledge.AlgorithmCategory.HASH))]
+
+    # CALCULATE[alg] = function to return the hash of its argument in hex
+    # TO-DO: implement the None entries with a third-party library, because
+    # hashlib might not have everything, depending on the Python version and
+    # the underlying OpenSSL. On Ubuntu 16.04, truncated sha512 and sha3/shake
+    # are not available. On Ubuntu 22.04, md2, md4 and ripemd160 are not
+    # available.
+    CALCULATE = {
+        'PSA_ALG_MD2': None,
+        'PSA_ALG_MD4': None,
+        'PSA_ALG_MD5': lambda data: hashlib.md5(data).hexdigest(),
+        'PSA_ALG_RIPEMD160': None, #lambda data: hashlib.new('ripdemd160').hexdigest()
+        'PSA_ALG_SHA_1': lambda data: hashlib.sha1(data).hexdigest(),
+        'PSA_ALG_SHA_224': lambda data: hashlib.sha224(data).hexdigest(),
+        'PSA_ALG_SHA_256': lambda data: hashlib.sha256(data).hexdigest(),
+        'PSA_ALG_SHA_384': lambda data: hashlib.sha384(data).hexdigest(),
+        'PSA_ALG_SHA_512': lambda data: hashlib.sha512(data).hexdigest(),
+        'PSA_ALG_SHA_512_224': None, #lambda data: hashlib.new('sha512_224').hexdigest()
+        'PSA_ALG_SHA_512_256': None, #lambda data: hashlib.new('sha512_256').hexdigest()
+        'PSA_ALG_SHA3_224': None, #lambda data: hashlib.sha3_224(data).hexdigest(),
+        'PSA_ALG_SHA3_256': None, #lambda data: hashlib.sha3_256(data).hexdigest(),
+        'PSA_ALG_SHA3_384': None, #lambda data: hashlib.sha3_384(data).hexdigest(),
+        'PSA_ALG_SHA3_512': None, #lambda data: hashlib.sha3_512(data).hexdigest(),
+        'PSA_ALG_SHAKE256_512': None, #lambda data: hashlib.shake_256(data).hexdigest(64),
+    } #type: Dict[str, Optional[Callable[[bytes], str]]]
+
+    @staticmethod
+    def one_test_case(alg: crypto_knowledge.Algorithm,
+                      function: str, note: str,
+                      arguments: List[str]) -> test_case.TestCase:
+        """Construct one test case involving a hash."""
+        tc = test_case.TestCase()
+        tc.set_description('{}{} {}'
+                           .format(function,
+                                   ' ' + note if note else '',
+                                   alg.short_expression()))
+        tc.set_dependencies(psa_low_level_dependencies(alg.expression))
+        tc.set_function(function)
+        tc.set_arguments([alg.expression] +
+                         ['"{}"'.format(arg) for arg in arguments])
+        return tc
+
+    def test_cases_for_hash(self,
+                            alg: crypto_knowledge.Algorithm
+                            ) -> Iterator[test_case.TestCase]:
+        """Enumerate all test cases for one hash algorithm."""
+        calc = self.CALCULATE[alg.expression]
+        if calc is None:
+            return # not implemented yet
+
+        short = b'abc'
+        hash_short = calc(short)
+        long = (b'Hello, world. Here are 16 unprintable bytes: ['
+                b'\x00\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0a'
+                b'\x80\x81\x82\x83\xfe\xff]. '
+                b' This message was brought to you by a natural intelligence. '
+                b' If you can read this, good luck with your debugging!')
+        hash_long = calc(long)
+
+        yield self.one_test_case(alg, 'hash_empty', '', [calc(b'')])
+        yield self.one_test_case(alg, 'hash_valid_one_shot', '',
+                                 [short.hex(), hash_short])
+        for n in [0, 1, 64, len(long) - 1, len(long)]:
+            yield self.one_test_case(alg, 'hash_valid_multipart',
+                                     '{} + {}'.format(n, len(long) - n),
+                                     [long[:n].hex(), calc(long[:n]),
+                                      long[n:].hex(), hash_long])
+
+    def all_test_cases(self) -> Iterator[test_case.TestCase]:
+        """Enumerate all test cases for all hash algorithms."""
+        for alg in self.algorithms:
+            yield from self.test_cases_for_hash(alg)

scripts/mbedtls_dev/psa_information.py

@@ -0,0 +1,106 @@
+"""Collect information about PSA cryptographic mechanisms.
+"""
+
+# Copyright The Mbed TLS Contributors
+# SPDX-License-Identifier: Apache-2.0 OR GPL-2.0-or-later
+
+
+import re
+from typing import Dict, FrozenSet, List, Optional
+
+from . import macro_collector
+
+
+def psa_want_symbol(name: str) -> str:
+    """Return the PSA_WANT_xxx symbol associated with a PSA crypto feature."""
+    if name.startswith('PSA_'):
+        return name[:4] + 'WANT_' + name[4:]
+    else:
+        raise ValueError('Unable to determine the PSA_WANT_ symbol for ' + name)
+
+def finish_family_dependency(dep: str, bits: int) -> str:
+    """Finish dep if it's a family dependency symbol prefix.
+    A family dependency symbol prefix is a PSA_WANT_ symbol that needs to be
+    qualified by the key size. If dep is such a symbol, finish it by adjusting
+    the prefix and appending the key size. Other symbols are left unchanged.
+    """
+    return re.sub(r'_FAMILY_(.*)', r'_\1_' + str(bits), dep)
+
+def finish_family_dependencies(dependencies: List[str], bits: int) -> List[str]:
+    """Finish any family dependency symbol prefixes.
+    Apply `finish_family_dependency` to each element of `dependencies`.
+    """
+    return [finish_family_dependency(dep, bits) for dep in dependencies]
+
+SYMBOLS_WITHOUT_DEPENDENCY = frozenset([
+    'PSA_ALG_AEAD_WITH_AT_LEAST_THIS_LENGTH_TAG', # modifier, only in policies
+    'PSA_ALG_AEAD_WITH_SHORTENED_TAG', # modifier
+    'PSA_ALG_ANY_HASH', # only in policies
+    'PSA_ALG_AT_LEAST_THIS_LENGTH_MAC', # modifier, only in policies
+    'PSA_ALG_KEY_AGREEMENT', # chaining
+    'PSA_ALG_TRUNCATED_MAC', # modifier
+])
+
+def automatic_dependencies(*expressions: str) -> List[str]:
+    """Infer dependencies of a test case by looking for PSA_xxx symbols.
+    The arguments are strings which should be C expressions. Do not use
+    string literals or comments as this function is not smart enough to
+    skip them.
+    """
+    used = set()
+    for expr in expressions:
+        used.update(re.findall(r'PSA_(?:ALG|ECC_FAMILY|KEY_TYPE)_\w+', expr))
+    used.difference_update(SYMBOLS_WITHOUT_DEPENDENCY)
+    return sorted(psa_want_symbol(name) for name in used)
+
+# A temporary hack: at the time of writing, not all dependency symbols
+# are implemented yet. Skip test cases for which the dependency symbols are
+# not available. Once all dependency symbols are available, this hack must
+# be removed so that a bug in the dependency symbols properly leads to a test
+# failure.
+def read_implemented_dependencies(filename: str) -> FrozenSet[str]:
+    return frozenset(symbol
+                     for line in open(filename)
+                     for symbol in re.findall(r'\bPSA_WANT_\w+\b', line))
+_implemented_dependencies = None #type: Optional[FrozenSet[str]] #pylint: disable=invalid-name
+
+def hack_dependencies_not_implemented(dependencies: List[str]) -> None:
+    global _implemented_dependencies #pylint: disable=global-statement,invalid-name
+    if _implemented_dependencies is None:
+        _implemented_dependencies = \
+            read_implemented_dependencies('include/psa/crypto_config.h')
+    if not all((dep.lstrip('!') in _implemented_dependencies or
+                not dep.lstrip('!').startswith('PSA_WANT'))
+               for dep in dependencies):
+        dependencies.append('DEPENDENCY_NOT_IMPLEMENTED_YET')
+
+class Information:
+    """Gather information about PSA constructors."""
+
+    def __init__(self) -> None:
+        self.constructors = self.read_psa_interface()
+
+    @staticmethod
+    def remove_unwanted_macros(
+            constructors: macro_collector.PSAMacroEnumerator
+    ) -> None:
+        # Mbed TLS doesn't support finite-field DH yet and will not support
+        # finite-field DSA. Don't attempt to generate any related test case.
+        constructors.key_types.discard('PSA_KEY_TYPE_DH_KEY_PAIR')
+        constructors.key_types.discard('PSA_KEY_TYPE_DH_PUBLIC_KEY')
+        constructors.key_types.discard('PSA_KEY_TYPE_DSA_KEY_PAIR')
+        constructors.key_types.discard('PSA_KEY_TYPE_DSA_PUBLIC_KEY')
+
+    def read_psa_interface(self) -> macro_collector.PSAMacroEnumerator:
+        """Return the list of known key types, algorithms, etc."""
+        constructors = macro_collector.InputsForTest()
+        header_file_names = ['include/psa/crypto_values.h',
+                             'include/psa/crypto_extra.h']
+        test_suites = ['tests/suites/test_suite_psa_crypto_metadata.data']
+        for header_file_name in header_file_names:
+            constructors.parse_header(header_file_name)
+        for test_cases in test_suites:
+            constructors.parse_test_cases(test_cases)
+        self.remove_unwanted_macros(constructors)
+        constructors.gather_arguments()
+        return constructors