generate_pkcs7_tests.py (7066B)
1 #!/usr/bin/env python3 2 # 3 # Copyright The Mbed TLS Contributors 4 # SPDX-License-Identifier: Apache-2.0 OR GPL-2.0-or-later 5 # 6 7 """ 8 Make fuzz like testing for pkcs7 tests 9 Given a valid DER pkcs7 file add tests to the test_suite_pkcs7.data file 10 - It is expected that the pkcs7_asn1_fail( data_t *pkcs7_buf ) 11 function is defined in test_suite_pkcs7.function 12 - This is not meant to be portable code, if anything it is meant to serve as 13 documentation for showing how those ugly tests in test_suite_pkcs7.data were created 14 """ 15 16 17 import sys 18 from os.path import exists 19 from mbedtls_framework import test_case 20 21 PKCS7_TEST_FILE = "../suites/test_suite_pkcs7.data" 22 23 class Test: # pylint: disable=too-few-public-methods 24 """ 25 A instance of a test in test_suite_pkcs7.data 26 """ 27 def __init__(self, name, depends, func_call): 28 self.name = name 29 self.depends = depends 30 self.func_call = func_call 31 32 # pylint: disable=no-self-use 33 def to_string(self): 34 return "\n" + self.name + "\n" + self.depends + "\n" + self.func_call + "\n" 35 36 class TestData: 37 """ 38 Take in test_suite_pkcs7.data file. 39 Allow for new tests to be added. 40 """ 41 mandatory_dep = test_case.psa_or_3_6_feature_macro("PSA_ALG_SHA_256", 42 test_case.Domain36.USE_PSA) 43 44 test_name = "PKCS7 Parse Failure Invalid ASN1" 45 test_function = "pkcs7_asn1_fail:" 46 def __init__(self, file_name): 47 self.file_name = file_name 48 self.last_test_num, self.old_tests = self.read_test_file(file_name) 49 self.new_tests = [] 50 51 # pylint: disable=no-self-use 52 def read_test_file(self, file): 53 """ 54 Parse the test_suite_pkcs7.data file. 55 """ 56 tests = [] 57 if not exists(file): 58 print(file + " Does not exist") 59 sys.exit() 60 with open(file, "r", encoding='UTF-8') as fp: 61 data = fp.read() 62 lines = [line.strip() for line in data.split('\n') if len(line.strip()) > 1] 63 i = 0 64 while i < len(lines): 65 if "depends" in lines[i+1]: 66 tests.append(Test(lines[i], lines[i+1], lines[i+2])) 67 i += 3 68 else: 69 tests.append(Test(lines[i], None, lines[i+1])) 70 i += 2 71 latest_test_num = float(tests[-1].name.split('#')[1]) 72 return latest_test_num, tests 73 74 def add(self, name, func_call): 75 self.last_test_num += 1 76 self.new_tests.append(Test(self.test_name + ": " + name + " #" + \ 77 str(self.last_test_num), "depends_on:" + self.mandatory_dep, \ 78 self.test_function + '"' + func_call + '"')) 79 80 def write_changes(self): 81 with open(self.file_name, 'a', encoding='UTF-8') as fw: 82 fw.write("\n") 83 for t in self.new_tests: 84 fw.write(t.to_string()) 85 86 87 def asn1_mutate(data): 88 """ 89 We have been given an asn1 structure representing a pkcs7. 90 We want to return an array of slightly modified versions of this data 91 they should be modified in a way which makes the structure invalid 92 93 We know that asn1 structures are: 94 |---1 byte showing data type---|----byte(s) for length of data---|---data content--| 95 We know that some data types can contain other data types. 96 Return a dictionary of reasons and mutated data types. 97 """ 98 99 # off the bat just add bytes to start and end of the buffer 100 mutations = [] 101 reasons = [] 102 mutations.append(["00"] + data) 103 reasons.append("Add null byte to start") 104 mutations.append(data + ["00"]) 105 reasons.append("Add null byte to end") 106 # for every asn1 entry we should attempt to: 107 # - change the data type tag 108 # - make the length longer than actual 109 # - make the length shorter than actual 110 i = 0 111 while i < len(data): 112 tag_i = i 113 leng_i = tag_i + 1 114 data_i = leng_i + 1 + (int(data[leng_i][1], 16) if data[leng_i][0] == '8' else 0) 115 if data[leng_i][0] == '8': 116 length = int(''.join(data[leng_i + 1: data_i]), 16) 117 else: 118 length = int(data[leng_i], 16) 119 120 tag = data[tag_i] 121 print("Looking at ans1: offset " + str(i) + " tag = " + tag + \ 122 ", length = " + str(length)+ ":") 123 print(''.join(data[data_i:data_i+length])) 124 # change tag to something else 125 if tag == "02": 126 # turn integers into octet strings 127 new_tag = "04" 128 else: 129 # turn everything else into an integer 130 new_tag = "02" 131 mutations.append(data[:tag_i] + [new_tag] + data[leng_i:]) 132 reasons.append("Change tag " + tag + " to " + new_tag) 133 134 # change lengths to too big 135 # skip any edge cases which would cause carry over 136 if int(data[data_i - 1], 16) < 255: 137 new_length = str(hex(int(data[data_i - 1], 16) + 1))[2:] 138 if len(new_length) == 1: 139 new_length = "0"+new_length 140 mutations.append(data[:data_i -1] + [new_length] + data[data_i:]) 141 reasons.append("Change length from " + str(length) + " to " \ 142 + str(length + 1)) 143 # we can add another test here for tags that contain other tags \ 144 # where they have more data than there containing tags account for 145 if tag in ["30", "a0", "31"]: 146 mutations.append(data[:data_i -1] + [new_length] + \ 147 data[data_i:data_i + length] + ["00"] + \ 148 data[data_i + length:]) 149 reasons.append("Change contents of tag " + tag + " to contain \ 150 one unaccounted extra byte") 151 # change lengths to too small 152 if int(data[data_i - 1], 16) > 0: 153 new_length = str(hex(int(data[data_i - 1], 16) - 1))[2:] 154 if len(new_length) == 1: 155 new_length = "0"+new_length 156 mutations.append(data[:data_i -1] + [new_length] + data[data_i:]) 157 reasons.append("Change length from " + str(length) + " to " + str(length - 1)) 158 159 # some tag types contain other tag types so we should iterate into the data 160 if tag in ["30", "a0", "31"]: 161 i = data_i 162 else: 163 i = data_i + length 164 165 return list(zip(reasons, mutations)) 166 167 if __name__ == "__main__": 168 if len(sys.argv) < 2: 169 print("USAGE: " + sys.argv[0] + " <pkcs7_der_file>") 170 sys.exit() 171 172 DATA_FILE = sys.argv[1] 173 TEST_DATA = TestData(PKCS7_TEST_FILE) 174 with open(DATA_FILE, 'rb') as f: 175 DATA_STR = f.read().hex() 176 # make data an array of byte strings eg ['de','ad','be','ef'] 177 HEX_DATA = list(map(''.join, [[DATA_STR[i], DATA_STR[i+1]] for i in range(0, len(DATA_STR), \ 178 2)])) 179 # returns tuples of test_names and modified data buffers 180 MUT_ARR = asn1_mutate(HEX_DATA) 181 182 print("made " + str(len(MUT_ARR)) + " new tests") 183 for new_test in MUT_ARR: 184 TEST_DATA.add(new_test[0], ''.join(new_test[1])) 185 186 TEST_DATA.write_changes()