Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 03 Sep 2019 14:07:45 -0000
From:      John Baldwin <jhb@FreeBSD.org>
To:        src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org
Subject:   svn commit: r346617 - head/tests/sys/opencrypto
Message-ID:  <201904240023.x3O0N7VT029391@repo.freebsd.org>

next in thread | raw e-mail | index | archive | help
Author: jhb
Date: Wed Apr 24 00:23:06 2019
New Revision: 346617
URL: https://svnweb.freebsd.org/changeset/base/346617

Log:
  Test the AES-CCM test vectors from the NIST Known Answer Tests.
  
  The CCM test vectors use a slightly different file format in that
  there are global key-value pairs as well as section key-value pairs
  that need to be used in each test.  In addition, the sections can set
  multiple key-value pairs in the section name.  The CCM KAT parser
  class is an iterator that returns a dictionary once per test where the
  dictionary contains all of the relevant key-value pairs for a given
  test (global, section name, section, test-specific).
  
  Note that all of the CCM decrypt tests use nonce and tag lengths that
  are not supported by OCF (OCF only supports a 12 byte nonce and 16
  byte tag), so none of the decryption vectors are actually tested.
  
  Reviewed by:	ngie
  MFC after:	1 month
  Sponsored by:	Chelsio Communications
  Differential Revision:	https://reviews.freebsd.org/D19978

Modified:
  head/tests/sys/opencrypto/cryptodev.py
  head/tests/sys/opencrypto/cryptotest.py

Modified: head/tests/sys/opencrypto/cryptodev.py
==============================================================================
--- head/tests/sys/opencrypto/cryptodev.py	Wed Apr 24 00:16:39 2019	(r346616)
+++ head/tests/sys/opencrypto/cryptodev.py	Wed Apr 24 00:23:06 2019	(r346617)
@@ -381,6 +381,112 @@ class KATParser:
 
 			yield values
 
+# The CCM files use a bit of a different syntax that doesn't quite fit
+# the generic KATParser.  In particular, some keys are set globally at
+# the start of the file, and some are set globally at the start of a
+# section.
+class KATCCMParser:
+	def __init__(self, fname):
+		self.fp = open(fname)
+		self._pending = None
+		self.read_globals()
+
+	def read_globals(self):
+		self.global_values = {}
+		while True:
+			line = self.fp.readline()
+			if not line:
+				return
+			if line[0] == '#' or not line.strip():
+				continue
+			if line[0] == '[':
+				self._pending = line
+				return
+
+			try:
+				f, v = line.split(' =')
+			except:
+				print('line:', repr(line))
+				raise
+
+			v = v.strip()
+
+			if f in self.global_values:
+				raise ValueError('already present: %r' % repr(f))
+			self.global_values[f] = v
+
+	def read_section_values(self, kwpairs):
+		self.section_values = self.global_values.copy()
+		for pair in kwpairs.split(', '):
+			f, v = pair.split(' = ')
+			if f in self.section_values:
+				raise ValueError('already present: %r' % repr(f))
+			self.section_values[f] = v
+
+		while True:
+			line = self.fp.readline()
+			if not line:
+				return
+			if line[0] == '#' or not line.strip():
+				continue
+			if line[0] == '[':
+				self._pending = line
+				return
+
+			try:
+				f, v = line.split(' =')
+			except:
+				print('line:', repr(line))
+				raise
+
+			if f == 'Count':
+				self._pending = line
+				return
+
+			v = v.strip()
+
+			if f in self.section_values:
+				raise ValueError('already present: %r' % repr(f))
+			self.section_values[f] = v
+
+	def __iter__(self):
+		while True:
+			if self._pending:
+				line = self._pending
+				self._pending = None
+			else:
+				line = self.fp.readline()
+				if not line:
+					return
+
+			if (line and line[0] == '#') or not line.strip():
+				continue
+
+			if line[0] == '[':
+				section = line[1:].split(']', 1)[0]
+				self.read_section_values(section)
+				continue
+
+			values = self.section_values.copy()
+
+			while True:
+				try:
+					f, v = line.split(' =')
+				except:
+					print('line:', repr(line))
+					raise
+				v = v.strip()
+
+				if f in values:
+					raise ValueError('already present: %r' % repr(f))
+				values[f] = v
+				line = self.fp.readline().strip()
+				if not line:
+					break
+
+			yield values
+
+
 def _spdechex(s):
 	return ''.join(s.split()).decode('hex')
 

Modified: head/tests/sys/opencrypto/cryptotest.py
==============================================================================
--- head/tests/sys/opencrypto/cryptotest.py	Wed Apr 24 00:16:39 2019	(r346616)
+++ head/tests/sys/opencrypto/cryptotest.py	Wed Apr 24 00:23:06 2019	(r346617)
@@ -71,6 +71,14 @@ def GenTestCase(cname):
 			for i in katg('KAT_AES', 'CBC[GKV]*.rsp'):
 				self.runCBC(i)
 
+		@unittest.skipIf(cname not in aesmodules, 'skipping AES-CCM on %s' % (cname))
+		def test_ccm(self):
+			for i in katg('ccmtestvectors', 'V*.rsp'):
+				self.runCCMEncrypt(i)
+
+			for i in katg('ccmtestvectors', 'D*.rsp'):
+				self.runCCMDecrypt(i)
+
 		@unittest.skipIf(cname not in aesmodules, 'skipping AES-GCM on %s' % (cname))
 		def test_gcm(self):
 			for i in katg('gcmtestvectors', 'gcmEncrypt*'):
@@ -219,6 +227,93 @@ def GenTestCase(cname):
 							raise
 						continue
 					self.assertEqual(r, ct)
+
+		def runCCMEncrypt(self, fname):
+			for data in cryptodev.KATCCMParser(fname):
+				Nlen = int(data['Nlen'])
+				if Nlen != 12:
+					# OCF only supports 12 byte IVs
+					continue
+				key = data['Key'].decode('hex')
+				nonce = data['Nonce'].decode('hex')
+				Alen = int(data['Alen'])
+				if Alen != 0:
+					aad = data['Adata'].decode('hex')
+				else:
+					aad = None
+				payload = data['Payload'].decode('hex')
+				ct = data['CT'].decode('hex')
+
+				try:
+					c = Crypto(crid=crid,
+					    cipher=cryptodev.CRYPTO_AES_CCM_16,
+					    key=key,
+					    mac=cryptodev.CRYPTO_AES_CCM_CBC_MAC,
+					    mackey=key, maclen=16)
+					r, tag = Crypto.encrypt(c, payload,
+					    nonce, aad)
+				except EnvironmentError, e:
+					if e.errno != errno.EOPNOTSUPP:
+						raise
+					continue
+
+				out = r + tag
+				self.assertEqual(out, ct,
+				    "Count " + data['Count'] + " Actual: " + \
+				    repr(out.encode("hex")) + " Expected: " + \
+				    repr(data) + " on " + cname)
+
+		def runCCMDecrypt(self, fname):
+			# XXX: Note that all of the current CCM
+			# decryption test vectors use IV and tag sizes
+			# that aren't supported by OCF none of the
+			# tests are actually ran.
+			for data in cryptodev.KATCCMParser(fname):
+				Nlen = int(data['Nlen'])
+				if Nlen != 12:
+					# OCF only supports 12 byte IVs
+					continue
+				Tlen = int(data['Tlen'])
+				if Tlen != 16:
+					# OCF only supports 16 byte tags
+					continue
+				key = data['Key'].decode('hex')
+				nonce = data['Nonce'].decode('hex')
+				Alen = int(data['Alen'])
+				if Alen != 0:
+					aad = data['Adata'].decode('hex')
+				else:
+					aad = None
+				ct = data['CT'].decode('hex')
+				tag = ct[-16:]
+				ct = ct[:-16]
+
+				try:
+					c = Crypto(crid=crid,
+					    cipher=cryptodev.CRYPTO_AES_CCM_16,
+					    key=key,
+					    mac=cryptodev.CRYPTO_AES_CCM_CBC_MAC,
+					    mackey=key, maclen=16)
+				except EnvironmentError, e:
+					if e.errno != errno.EOPNOTSUPP:
+						raise
+					continue
+
+				if data['Result'] == 'Fail':
+					self.assertRaises(IOError,
+					    c.decrypt, payload, nonce, aad, tag)
+				else:
+					r = Crypto.decrypt(c, payload, nonce,
+					    aad, tag)
+
+					payload = data['Payload'].decode('hex')
+					Plen = int(data('Plen'))
+					payload = payload[:plen]
+					self.assertEqual(r, payload,
+					    "Count " + data['Count'] + \
+					    " Actual: " + repr(r.encode("hex")) + \
+					    " Expected: " + repr(data) + \
+					    " on " + cname)
 
 		###############
 		##### DES #####





Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201904240023.x3O0N7VT029391>