I'm trying to use BouncyCastle's CMac implementation but apparently I'm doing it wrong. At least the following unit test (based on RFC 5297 test vectors) fails:
@Test
public void testCMacOfZeros() {
byte[] key = {(byte) 0xff, (byte) 0xfe, (byte) 0xfd, (byte) 0xfc, //
(byte) 0xfb, (byte) 0xfa, (byte) 0xf9, (byte) 0xf8, //
(byte) 0xf7, (byte) 0xf6, (byte) 0xf5, (byte) 0xf4, //
(byte) 0xf3, (byte) 0xf2, (byte) 0xf1, (byte) 0xf0, //
(byte) 0xf0, (byte) 0xf1, (byte) 0xf2, (byte) 0xf3, //
(byte) 0xf4, (byte) 0xf5, (byte) 0xf6, (byte) 0xf7, //
(byte) 0xf8, (byte) 0xf9, (byte) 0xfa, (byte) 0xfb, //
(byte) 0xfc, (byte) 0xfd, (byte) 0xfe, (byte) 0xff};
byte[] zeros = new byte[16];
byte[] result = new byte[16];
CipherParameters params = new KeyParameter(key);
BlockCipher aes = new AESEngine();
CMac mac = new CMac(aes);
mac.init(params);
mac.update(zeros, 0, 16);
mac.doFinal(result, 0);
byte[] expected = {(byte) 0x0e, (byte) 0x04, (byte) 0xdf, (byte) 0xaf, //
(byte) 0xc1, (byte) 0xef, (byte) 0xbf, (byte) 0x04, //
(byte) 0x01, (byte) 0x40, (byte) 0x58, (byte) 0x28, //
(byte) 0x59, (byte) 0xbf, (byte) 0x07, (byte) 0x3a};
Assert.assertArrayEquals(expected, result);
}
I assume, that the CMac implementation itself is well tested, so I must miss something.
I found my mistake:
SIV-AES uses AES in CMAC mode (S2V) and in counter mode (CTR). SIV- AES takes either a 256-, 384-, or 512-bit key (which is broken up into two equal-sized keys, one for S2V and the other for CTR)
I should have used only the first 16 bytes from the given key.
As expected, BouncyCastle works just fine.
来源:https://stackoverflow.com/questions/28516804/how-to-use-bouncycastles-cmac