• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1# Diffie-Hellman
2
3## Subgroup confinement attacks
4
5The papers by van Oorshot and Wiener [OW96] rsp. Lim and Lee [LL98] show that
6Diffie-Hellman keys can be found much faster if the short exponents are used and
7if the multiplicative group modulo p contains small subgroups. In particular an
8attacker can try to send a public key that is an element of a small subgroup. If
9the receiver does not check for such elements then may be possible to find the
10private key modulo the order of the small subgroup. Several countermeasures
11against such attacks have been proposed: For example IKE uses fields of order p
12where p is a safe prime (i.e. $$q=(p-1)/2),$$ hence the only elements of small
13order are 1 and p-1.
14
15[NIST SP 800-56A] rev. 2, Section 5.5.1.1 only requires that the size of the
16subgroup generated by the generator g is big enough to prevent the baby-step
17giant-step algorithm. I.e. for 80-bit security p must be at least 1024 bits long
18and the prime q must be at least 160 bits long. A 2048 bit prime p and a 224 bit
19prime q are sufficient for 112 bit security. To avoid subgroup confinment
20attacks NIST requires that public keys are validated, i.e. by checking that a
21public key y satisfies the conditions $$2 \leq y \leq p-2$$ and $$y^q \mod p =
221$$ (Section 5.6.2.3.1). Further, after generating the shared secret $$z =
23y_a^{x_b} \mod p$$ each party should check that $$z \neq 1.$$ RFC 2785 contains
24similar recommendations. The public key validation described by NIST requires
25that the order q of the generator g is known to the verifier. Unfortunately, the
26order q is missing in [PKCS #3]. [PKCS #3] describes the Diffie-Hellman
27parameters only by the values p, g and optionally the key size in bits.
28
29The class DHParameterSpec that defines the Diffie-Hellman parameters in JCE
30contains the same values as [PKCS #3]. In particular, it does not contain the
31order of the subgroup q. Moreover, the SUN provider uses the minimal sizes
32specified by NIST for q. Essentially the provider reuses the parameters for DSA.
33
34Therefore, there is no guarantee that an implementation of Diffie-Hellman is secure against
35subgroup confinement attacks. Without a key validation it is insecure to use the key-pair
36generation from [NIST SP 800-56A] Section 5.6.1.1 (The key-pair generation there only requires that
37static and ephemeral private keys are randomly chosen in the range \\(1..q-1)\\).
38
39To avoid big disasters the tests below require that key sizes are not minimal. I.e., currently
40the tests require at least 512 bit keys for 1024 bit fields. We use this lower limit because that
41is what the SUN provider is currently doing.
42
43TODO(bleichen): Find a reference supporting or disproving that decision.
44
45## Weak parameters
46
47The DH parameters must be carefully chosen to avoid security issues. A panel at
48Eurocrypt'92 discussed the possiblity of trapdoors in DL based primitives
49[Eurocrypt92 panel]. A. Lenstra pointed out that the primes chould be chosen
50such that the special number field sieve can be used to compute discrete
51logarithms. Gordon has analyzed methods to generate and detect weak parameters
52[G92]. Section 4 of Gordons paper describes a method that can detect some
53special cases, but no general method was given. Recently Fried et al. showed
54that 1024 bit discrete logarithms with the special number field sieve are
55feasible [FGHT16]. Moreover some libraries use primes that are susceptible to
56this attack [FGHT16].
57
58TODO(bleichen): So far not test for weak DH parameters has been implemented.
59Possibly we should at least implement a test that detects special cases, so
60that weak primes (such as the one used in libtomcrypt) are detected.
61
62DH implementations are sometimes misconfigured. Adrian et al. [WeakDh] analyzed
63various implementations and found for example the following problems in the
64parameters: p is sometimes composite, p-1 contains no large prime factor, q is
65used instead of the generator g.
66
67## References
68[Eurocrypt92 panel]: "The Eurocrypt'92 Controversial Issue Trapdoor Primes and Moduli",
69EUROCRYPT '92, LNCS 658, pp. 194-199.
70
71[G92]: D. M. Gordon. "Designing and detecting trapdoors for discrete log
72cryptosystems." CRYPTO’92, pp. 66–75.
73
74\[FGHT16]: J. Fried, P. Gaudry, N. Heininger, E. Thome. "A kilobit hidden SNFS
75discrete logarithm computation". http://eprint.iacr.org/2016/961.pdf
76
77[OW96]: P. C. van Oorschot, M. J. Wiener, "On Diffie-Hellman key agreement with short exponents",
78Eurocrypt 96, pp 332–343.
79
80[LL98]: C.H. Lim and P.J. Lee,
81"A key recovery attack on discrete log-based schemes using a prime order subgroup",
82CRYPTO' 98, pp 249–263.
83
84[WeakDh]: D. Adrian, K. Bhargavan, Z. Durumeric, P. Gaudry, M. Green,
85J. A. Halderman, N. Heninger, D. Springall, E. Thomé, Luke Valenta,
86B. VanderSloot, E. Wustrow, S. Zanella-Béguelink, P. Zimmermann,
87"Imperfect Forward Secrecy: How Diffie-Hellman Fails in Practice"
88https://weakdh.org/imperfect-forward-secrecy-ccs15.pdf
89
90[NIST SP 800-56A], revision 2, May 2013
91http://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.800-56Ar2.pdf
92
93[PKCS #3]: "Diffie–Hellman Key Agreement",
94http://uk.emc.com/emc-plus/rsa-labs/standards-initiatives/pkcs-3-diffie-hellman-key-agreement-standar.htm
95
96[RFC 2785]:  R. Zuccherato,
97"Methods for Avoiding 'Small-Subgroup' Attacks on the Diffie-Hellman Key Agreement Method for S/MIME",
98March 2000
99https://www.ietf.org/rfc/rfc2785.txt
100
101<!--
102## Sources that might be used for additional tests:
103
104CVE-2015-3193: The Montgomery squaring implementation in crypto/bn/asm/x86_64-mont5.pl
105in OpenSSL 1.0.2 before 1.0.2e on the x86_64 platform, as used by the BN_mod_exp function,
106mishandles carry propagation
107https://blog.fuzzing-project.org/31-Fuzzing-Math-miscalculations-in-OpenSSLs-BN_mod_exp-CVE-2015-3193.html
108
109CVE-2016-0739: libssh before 0.7.3 improperly truncates ephemeral secrets generated for the
110(1) diffie-hellman-group1 and (2) diffie-hellman-group14 key exchange methods to 128 bits ...
111
112CVE-2015-1787 The ssl3_get_client_key_exchange function in s3_srvr.c in OpenSSL 1.0.2 before
1131.0.2a, when client authentication and an ephemeral Diffie-Hellman ciphersuite are enabled,
114allows remote attackers to cause a denial of service (daemon crash) via a ClientKeyExchange
115message with a length of zero.
116
117CVE-2015-0205 The ssl3_get_cert_verify function in s3_srvr.c in OpenSSL 1.0.0 before 1.0.0p
118and 1.0.1 before 1.0.1k accepts client authentication with a Diffie-Hellman (DH) certificate
119without requiring a CertificateVerify message, which allows remote attackers to obtain access
120without knowledge of a private key via crafted TLS Handshake Protocol traffic to a server that
121recognizes a Certification Authority with DH support.
122
123CVE-2016-0701 The DH_check_pub_key function in crypto/dh/dh_check.c in OpenSSL 1.0.2 before
1241.0.2f does not ensure that prime numbers are appropriate for Diffie-Hellman (DH) key exchange,
125which makes it easier for remote attackers to discover a private DH exponent by making multiple
126handshakes with a peer that chose an inappropriate number, as demonstrated by a number in an
127X9.42 file.
128
129CVE-2006-1115 nCipher HSM before 2.22.6, when generating a Diffie-Hellman public/private key
130pair without any specified DiscreteLogGroup parameters, chooses random parameters that could
131allow an attacker to crack the private key in significantly less time than a brute force attack.
132
133CVE-2015-1716 Schannel in Microsoft Windows Server 2003 SP2, Windows Vista SP2, Windows Server
1342008 SP2 and R2 SP1, Windows 7 SP1, Windows 8, Windows 8.1, Windows Server 2012 Gold and R2, and
135Windows RT Gold and 8.1 does not properly restrict Diffie-Hellman Ephemeral (DHE) key lengths,
136which makes it easier for remote attackers to defeat cryptographic protection mechanisms via
137unspecified vectors, aka "Schannel Information Disclosure Vulnerability.
138
139CVE-2015-2419: Random generation of the prime p allows Pohlig-Hellman and probably other
140stuff.
141-->
142