s3api: Add support for crc64nvme checksum calculation

Add anycrc as a soft dependency in case ISA-L isn't available.
Plus we'll want it later: when we start writing down checksums,
we'll need it to combine per-part checksums for MPUs.

Like with crc32c, we won't provide any pure-python version as the
CPU-intensiveness could present a DoS vector. Worst case, we 501
as before.

Co-Authored-By: Tim Burke <tim.burke@gmail.com>
Signed-off-by: Tim Burke <tim.burke@gmail.com>
Change-Id: Ia05e5677a8ca89a62b142078abfb7371b1badd3f
Signed-off-by: Alistair Coles <alistairncoles@gmail.com>

test/s3api/test_object_checksums.py

@@ -191,7 +191,11 @@ class ObjectChecksumMixin(object):
             'ChecksumAlgorithm': self.ALGORITHM,
         }
         if boto_at_least(1, 36):
-            checksum_kwargs['ChecksumType'] = 'COMPOSITE'
+            if self.ALGORITHM == 'CRC64NVME':
+                # crc64nvme only allows full-object
+                checksum_kwargs['ChecksumType'] = 'FULL_OBJECT'
+            else:
+                checksum_kwargs['ChecksumType'] = 'COMPOSITE'
 
         obj_name = self.create_name(self.ALGORITHM + '-mpu-complete-good')
         create_mpu_resp = self.client.create_multipart_upload(
@@ -247,6 +251,24 @@ class TestObjectChecksumCRC32C(ObjectChecksumMixin, BaseS3TestCaseWithBucket):
         super().setUpClass()
 
 
+class TestObjectChecksumCRC64NVME(ObjectChecksumMixin,
+                                  BaseS3TestCaseWithBucket):
+    ALGORITHM = 'CRC64NVME'
+    EXPECTED = 'rosUhgp5mIg='
+    INVALID = 'rosUhgp5mIh='
+    BAD = 'sosUhgp5mIg='
+
+    @classmethod
+    def setUpClass(cls):
+        if [int(x) for x in botocore.__version__.split('.')] < [1, 36]:
+            raise SkipTest('botocore cannot crc64nvme (run '
+                           '`pip install -U boto3 botocore`)')
+        if not botocore.httpchecksum.HAS_CRT:
+            raise SkipTest('botocore cannot crc64nvme (run '
+                           '`pip install awscrt`)')
+        super().setUpClass()
+
+
 class TestObjectChecksumSHA1(ObjectChecksumMixin, BaseS3TestCaseWithBucket):
     ALGORITHM = 'SHA1'
     EXPECTED = '98O8HYCOBHMq32eZZczDTKeuNEE='

test/unit/__init__.py

@@ -1096,6 +1096,17 @@ def requires_crc32c(func):
     return wrapper
 
 
+def requires_crc64nvme(func):
+    @functools.wraps(func)
+    def wrapper(*args, **kwargs):
+        try:
+            checksum.crc64nvme()
+        except NotImplementedError as e:
+            raise SkipTest(str(e))
+        return func(*args, **kwargs)
+    return wrapper
+
+
 class StubResponse(object):
 
     def __init__(self, status, body=b'', headers=None, frag_index=None,

test/unit/common/middleware/s3api/test_s3api.py

@@ -248,11 +248,15 @@ class TestS3ApiMiddleware(S3ApiTestCase):
         with mock.patch('swift.common.middleware.s3api.s3api.get_logger',
                         return_value=self.logger), \
                 mock.patch('swift.common.utils.checksum.crc32c_isal') \
-                as mock_crc32c:
+                as mock_crc32c, \
+                mock.patch('swift.common.utils.checksum.crc64nvme_isal') \
+                as mock_crc64nvme:
             mock_crc32c.__name__ = 'crc32c_isal'
+            mock_crc64nvme.__name__ = 'crc64nvme_isal'
             S3ApiMiddleware(None, {})
         self.assertEqual(
-            {'info': ['Using crc32c_isal implementation for CRC32C.']},
+            {'info': ['Using crc32c_isal implementation for CRC32C.',
+                      'Using crc64nvme_isal implementation for CRC64NVME.']},
             self.logger.all_log_lines())
 
     def test_non_s3_request_passthrough(self):

test/unit/common/middleware/s3api/test_s3request.py

@@ -41,7 +41,7 @@ from swift.common.middleware.s3api.s3response import InvalidArgument, \
     XAmzContentSHA256Mismatch, ErrorResponse, S3NotImplemented
 from swift.common.utils import checksum
 from test.debug_logger import debug_logger
-from test.unit import requires_crc32c
+from test.unit import requires_crc32c, requires_crc64nvme
 from test.unit.common.middleware.s3api.test_s3api import S3ApiTestCase
 
 Fake_ACL_MAP = {
@@ -2001,40 +2001,38 @@ class TestRequest(S3ApiTestCase):
         with self.assertRaises(S3InputChecksumMismatch):
             sigv4_req.environ['wsgi.input'].read()
 
+    @requires_crc64nvme
     @patch.object(S3Request, '_validate_dates', lambda *a: None)
-    def test_sig_v4_strm_unsgnd_pyld_trl_checksum_hdr_crc64nvme_valid(self):
-        # apparently valid value provokes the not implemented error
+    def test_sig_v4_strm_unsgnd_pyld_trl_checksum_hdr_crc64nvme_ok(self):
         body = 'a\r\nabcdefghij\r\n' \
                'a\r\nklmnopqrst\r\n' \
                '7\r\nuvwxyz\n\r\n' \
                '0\r\n'
-        crc = base64.b64encode(b'12345678')
+        crc = base64.b64encode(
+            checksum.crc64nvme(b'abcdefghijklmnopqrstuvwxyz\n').digest())
         req = self._make_sig_v4_streaming_unsigned_payload_trailer_req(
             body=body,
             extra_headers={'x-amz-checksum-crc64nvme': crc}
         )
-        with self.assertRaises(S3NotImplemented) as cm:
-            SigV4Request(req.environ)
-        self.assertIn(
-            b'The x-amz-checksum-crc64nvme algorithm is not supported.',
-            cm.exception.body)
+        sigv4_req = SigV4Request(req.environ)
+        self.assertEqual(b'abcdefghijklmnopqrstuvwxyz\n',
+                         sigv4_req.environ['wsgi.input'].read())
 
+    @requires_crc64nvme
     @patch.object(S3Request, '_validate_dates', lambda *a: None)
     def test_sig_v4_strm_unsgnd_pyld_trl_checksum_hdr_crc64nvme_invalid(self):
-        # the not implemented error is raised before the value is validated
         body = 'a\r\nabcdefghij\r\n' \
                'a\r\nklmnopqrst\r\n' \
                '7\r\nuvwxyz\n\r\n' \
                '0\r\n'
+        crc = base64.b64encode(checksum.crc64nvme(b'not-the-body').digest())
         req = self._make_sig_v4_streaming_unsigned_payload_trailer_req(
             body=body,
-            extra_headers={'x-amz-checksum-crc64nvme': 'not-a-valid-crc'}
+            extra_headers={'x-amz-checksum-crc64nvme': crc}
         )
-        with self.assertRaises(S3NotImplemented) as cm:
-            SigV4Request(req.environ)
-        self.assertIn(
-            b'The x-amz-checksum-crc64nvme algorithm is not supported.',
-            cm.exception.body)
+        sigv4_req = SigV4Request(req.environ)
+        with self.assertRaises(S3InputChecksumMismatch):
+            sigv4_req.environ['wsgi.input'].read()
 
     @patch.object(S3Request, '_validate_dates', lambda *a: None)
     def test_sig_v4_strm_unsgnd_pyld_trl_checksum_hdr_sha1_ok(self):
@@ -2896,13 +2894,21 @@ class TestModuleFunctions(unittest.TestCase):
         do_test('crc32c')
         do_test('sha1')
         do_test('sha256')
+        try:
+            checksum._select_crc64nvme_impl()
+        except NotImplementedError:
+            pass
+        else:
+            do_test('crc64nvme')
 
     def test_get_checksum_hasher_invalid(self):
         def do_test(crc):
             with self.assertRaises(s3response.S3NotImplemented):
                 _get_checksum_hasher('x-amz-checksum-%s' % crc)
 
-        do_test('crc64nvme')
+        with mock.patch.object(checksum, '_select_crc64nvme_impl',
+                               side_effect=NotImplementedError):
+            do_test('crc64nvme')
         do_test('nonsense')
         do_test('')
 

test/unit/common/utils/test_checksum.py

@@ -20,7 +20,7 @@ import zlib
 
 from swift.common.utils import checksum
 from test.debug_logger import debug_logger
-from test.unit import requires_crc32c
+from test.unit import requires_crc32c, requires_crc64nvme
 
 
 # If you're curious about the 0xe3069283, see "check" at
@@ -32,10 +32,18 @@ class TestCRC32C(unittest.TestCase):
         partial = impl(b"12345")
         self.assertEqual(impl(b"6789", partial), 0xe3069283)
 
+    @unittest.skipIf(checksum.crc32c_anycrc is None, 'No anycrc CRC32C')
+    def test_anycrc(self):
+        self.check_crc_func(checksum.crc32c_anycrc)
+        # Check preferences -- beats out reference, but not kernel or ISA-L
+        if checksum.crc32c_isal is None and checksum.crc32c_kern is None:
+            self.assertIs(checksum._select_crc32c_impl(),
+                          checksum.crc32c_anycrc)
+
     @unittest.skipIf(checksum.crc32c_kern is None, 'No kernel CRC32C')
     def test_kern(self):
         self.check_crc_func(checksum.crc32c_kern)
-        # Check preferences -- beats out reference, but not ISA-L
+        # Check preferences -- beats out reference and anycrc, but not ISA-L
         if checksum.crc32c_isal is None:
             self.assertIs(checksum._select_crc32c_impl(), checksum.crc32c_kern)
 
@@ -128,6 +136,28 @@ class TestCRC32C(unittest.TestCase):
         self.assertIs(checksum._select_crc32c_impl(), checksum.crc32c_isal)
 
 
+class TestCRC64NVME(unittest.TestCase):
+    def check_crc_func(self, impl):
+        self.assertEqual(impl(b"123456789"), 0xae8b14860a799888)
+        # Check that we can save/continue
+        partial = impl(b"12345")
+        self.assertEqual(impl(b"6789", partial), 0xae8b14860a799888)
+
+    @unittest.skipIf(checksum.crc64nvme_anycrc is None, 'No anycrc CRC64NVME')
+    def test_anycrc(self):
+        self.check_crc_func(checksum.crc64nvme_anycrc)
+        if checksum.crc64nvme_isal is None:
+            self.assertIs(checksum._select_crc64nvme_impl(),
+                          checksum.crc64nvme_anycrc)
+
+    @unittest.skipIf(checksum.crc64nvme_isal is None, 'No ISA-L CRC64NVME')
+    def test_isal(self):
+        self.check_crc_func(checksum.crc64nvme_isal)
+        # Check preferences -- ISA-L always wins
+        self.assertIs(checksum._select_crc64nvme_impl(),
+                      checksum.crc64nvme_isal)
+
+
 class TestCRCHasher(unittest.TestCase):
     def setUp(self):
         self.logger = debug_logger()
@@ -238,21 +268,101 @@ class TestCRCHasher(unittest.TestCase):
         self.assertEqual('6b2fc5b0', hasher_copy.hexdigest())
 
     def test_crc32c_hasher_selects_kern_impl(self):
-        with mock.patch('swift.common.utils.checksum.crc32c_isal', None), \
-                mock.patch(
-                    'swift.common.utils.checksum.crc32c_kern') as mock_kern:
+        scuc = 'swift.common.utils.checksum'
+        with mock.patch(scuc + '.crc32c_isal', None), \
+                mock.patch(scuc + '.crc32c_kern') as mock_kern, \
+                mock.patch(scuc + '.crc32c_anycrc', None):
             mock_kern.__name__ = 'crc32c_kern'
             self.assertIs(mock_kern, checksum.crc32c().crc_func)
             checksum.log_selected_implementation(self.logger)
         self.assertIn('Using crc32c_kern implementation for CRC32C.',
                       self.logger.get_lines_for_level('info'))
 
+    def test_crc32c_hasher_selects_anycrc_impl(self):
+        scuc = 'swift.common.utils.checksum'
+        with mock.patch(scuc + '.crc32c_isal', None), \
+                mock.patch(scuc + '.crc32c_kern', None), \
+                mock.patch(scuc + '.crc32c_anycrc') as mock_anycrc:
+            mock_anycrc.__name__ = 'crc32c_anycrc'
+            self.assertIs(mock_anycrc, checksum.crc32c().crc_func)
+            checksum.log_selected_implementation(self.logger)
+        self.assertIn('Using crc32c_anycrc implementation for CRC32C.',
+                      self.logger.get_lines_for_level('info'))
+
     def test_crc32c_hasher_selects_isal_impl(self):
-        with mock.patch(
-                'swift.common.utils.checksum.crc32c_isal') as mock_isal, \
-                mock.patch('swift.common.utils.checksum.crc32c_kern'):
+        scuc = 'swift.common.utils.checksum'
+        with mock.patch(scuc + '.crc32c_isal') as mock_isal, \
+                mock.patch(scuc + '.crc32c_kern'), \
+                mock.patch(scuc + '.crc32c_anycrc'):
             mock_isal.__name__ = 'crc32c_isal'
             self.assertIs(mock_isal, checksum.crc32c().crc_func)
             checksum.log_selected_implementation(self.logger)
         self.assertIn('Using crc32c_isal implementation for CRC32C.',
                       self.logger.get_lines_for_level('info'))
+
+    @requires_crc64nvme
+    def test_crc64nvme_hasher(self):
+        # See CRC-64/NVME at
+        # https://reveng.sourceforge.io/crc-catalogue/17plus.htm
+        hasher = checksum.crc64nvme()
+        self.assertEqual('crc64nvme', hasher.name)
+        self.assertEqual(8, hasher.digest_size)
+        self.assertEqual(64, hasher.width)
+        self.assertEqual(0, hasher.crc)
+        self.assertEqual(b'\x00\x00\x00\x00\x00\x00\x00\x00', hasher.digest())
+        self.assertEqual('0000000000000000', hasher.hexdigest())
+
+        hasher.update(b'123456789')
+        self.assertEqual(0xae8b14860a799888, hasher.crc)
+        self.assertEqual(b'\xae\x8b\x14\x86\x0a\x79\x98\x88', hasher.digest())
+        self.assertEqual('ae8b14860a799888', hasher.hexdigest())
+
+    @requires_crc64nvme
+    def test_crc64nvme_hasher_constructed_with_data(self):
+        hasher = checksum.crc64nvme(b'123456789')
+        self.assertEqual(b'\xae\x8b\x14\x86\x0a\x79\x98\x88', hasher.digest())
+        self.assertEqual('ae8b14860a799888', hasher.hexdigest())
+
+    @requires_crc64nvme
+    def test_crc64nvme_hasher_initial_value(self):
+        hasher = checksum.crc64nvme(initial_value=0xae8b14860a799888)
+        self.assertEqual(b'\xae\x8b\x14\x86\x0a\x79\x98\x88', hasher.digest())
+        self.assertEqual('ae8b14860a799888', hasher.hexdigest())
+
+    @requires_crc64nvme
+    def test_crc64nvme_hasher_copy(self):
+        hasher = checksum.crc64nvme(b'123456789')
+        self.assertEqual('ae8b14860a799888', hasher.hexdigest())
+        hasher_copy = hasher.copy()
+        self.assertEqual('crc64nvme', hasher_copy.name)
+        self.assertIs(hasher.crc_func, hasher_copy.crc_func)
+        self.assertEqual('ae8b14860a799888', hasher_copy.hexdigest())
+        hasher_copy.update(b'foo')
+        self.assertEqual('ae8b14860a799888', hasher.hexdigest())
+        self.assertEqual('673ece0d56523f46', hasher_copy.hexdigest())
+        hasher.update(b'bar')
+        self.assertEqual('0991d5edf1b0062e', hasher.hexdigest())
+        self.assertEqual('673ece0d56523f46', hasher_copy.hexdigest())
+
+    def test_crc64nvme_hasher_selects_anycrc_impl(self):
+        scuc = 'swift.common.utils.checksum'
+        with mock.patch(scuc + '.crc64nvme_isal', None), \
+                mock.patch(scuc + '.crc64nvme_anycrc') as mock_anycrc:
+            mock_anycrc.__name__ = 'crc64nvme_anycrc'
+            self.assertIs(mock_anycrc,
+                          checksum.crc64nvme().crc_func)
+            checksum.log_selected_implementation(self.logger)
+        self.assertIn(
+            'Using crc64nvme_anycrc implementation for CRC64NVME.',
+            self.logger.get_lines_for_level('info'))
+
+    def test_crc64nvme_hasher_selects_isal_impl(self):
+        scuc = 'swift.common.utils.checksum'
+        with mock.patch(scuc + '.crc64nvme_isal') as mock_isal, \
+                mock.patch(scuc + '.crc64nvme_anycrc'):
+            mock_isal.__name__ = 'crc64nvme_isal'
+            self.assertIs(mock_isal, checksum.crc64nvme().crc_func)
+            checksum.log_selected_implementation(self.logger)
+        self.assertIn(
+            'Using crc64nvme_isal implementation for CRC64NVME.',
+            self.logger.get_lines_for_level('info'))