Fix flaw if 'Server Key exchange message' is omitted from a TLS
handshake which could lead to a cilent crash as found using the
Codenomicon TLS test suite (CVE-2008-1672)

Reviewed by: openssl-security@openssl.org

Obtained from: mark@awe.com

of handshake failure

2. Changes to x509_certificate_type function (crypto/x509/x509type.c) to
make it recognize GOST certificates as EVP_PKT_SIGN|EVP_PKT_EXCH
(required for s3_srvr to accept GOST client certificates).

3. Changes to EVP
	- adding of function EVP_PKEY_CTX_get0_peerkey
	- Make function EVP_PKEY_derive_set_peerkey work for context with
	  ENCRYPT operation, because we use peerkey field in the context to
	  pass non-ephemeral secret key to GOST encrypt operation.
	- added EVP_PKEY_CTRL_SET_IV control command. It is really
	  GOST-specific, but it is used in SSL code, so it has to go
	  in some header file, available during libssl compilation

4. Fix to HMAC to avoid call of OPENSSL_cleanse on undefined data

5. Include des.h if KSSL_DEBUG is defined into some libssl files, to
  make debugging output which depends on constants defined there, work
  and other KSSL_DEBUG output fixes

6. Declaration of real GOST ciphersuites, two authentication methods
   SSL_aGOST94 and SSL_aGOST2001 and one key exchange method SSL_kGOST

7. Implementation  of these methods.

8. Support for sending unsolicited serverhello extension if GOST
  ciphersuite is selected. It is require for interoperability with
  CryptoPro CSP 3.0 and 3.6 and controlled by
  SSL_OP_CRYPTOPRO_TLSEXT_BUG constant.
  This constant is added to SSL_OP_ALL, because it does nothing, if
  non-GOST ciphersuite is selected, and all implementation of GOST
  include compatibility with CryptoPro.

9. Support for CertificateVerify message without length field. It is
   another CryptoPro bug, but support is made unconditional, because it
   does no harm for draft-conforming implementation.

10. In tls1_mac extra copy of stream mac context is no more done.
  When I've written currently commited code I haven't read
  EVP_DigestSignFinal manual carefully enough and haven't noticed that
  it does an internal digest ctx copying.

This implementation was tested against
1. CryptoPro CSP 3.6 client and server
2. Cryptopro CSP 3.0 server

(draft-rescorla-tls-opaque-prf-input-00.txt), and do some cleanups and
bugfixes on the way.  In particular, this fixes the buffer bounds
checks in ssl_add_clienthello_tlsext() and in ssl_add_serverhello_tlsext().

Note that the opaque PRF Input TLS extension is not compiled by default;
see CHANGES.

Submitted by: Victor B. Wagner <vitus@cryptocom.ru>

for OpenSSL.

ciphersuite selection code in HEAD

Submitted by: Victor Duchovni

This change resolves a number of problems and obviates multiple kludges.
A new feature is that you can now say "AES256" or "AES128" (not just
"AES", which enables both).

In some cases the ciphersuite list generated from a given string is
affected by this change.  I hope this is just in those cases where the
previous behaviour did not make sense.

  1) Certificate Message with no certs

  OpenSSL implementation sends the Certificate message during SSL
  handshake, however as per the specification, these have been omitted.

  -- RFC 2712 --
     CertificateRequest, and the ServerKeyExchange shown in Figure 1
     will be omitted since authentication and the establishment of a
     master secret will be done using the client's Kerberos credentials
     for the TLS server.  The client's certificate will be omitted for
     the same reason.
  -- RFC 2712 --

  3) Pre-master secret Protocol version

  The pre-master secret generated by OpenSSL does not have the correct
  client version.

  RFC 2712 says, if the Kerberos option is selected, the pre-master
  secret structure is the same as that used in the RSA case.

  TLS specification defines pre-master secret as:
         struct {
             ProtocolVersion client_version;
             opaque random[46];
         } PreMasterSecret;

  where client_version is the latest protocol version supported by the
  client

  The pre-master secret generated by OpenSSL does not have the correct
  client version. The implementation does not update the first 2 bytes
  of random secret for Kerberos Cipher suites. At the server-end, the
  client version from the pre-master secret is not validated.

PR: 1336

Also, change the default ciphersuite to give some prefererence to
ciphersuites with forwared secrecy (rather than using a random order).

Submitted by: Douglas Stebila

Submitted by: Douglas Stebila

PR: 1191
Submitted by: Mika Kousa and Pasi Eronen of Nokia Corporation
Reviewed by: Nils Larsch

Submitted by: Peter Sylvester

Submitted by: Peter Sylvester

 - fix indentation
 - rename some functions and macros
 - fix up confusion between SSL_ERROR_... and SSL_AD_... values

Submitted by: Peter Sylvester

Submitted by: Douglas Stebila

maximum send fragment size. Allocate I/O buffers accordingly.

pointer and make the SSL_METHOD parameter in SSL_CTX_new,
SSL_CTX_set_ssl_version and SSL_set_ssl_method const.

for locking code. The CRYPTO_LOCK_SSL_METHOD lock is now no longer used.

    	./configure no-deprecated [no-dsa] [no-dh] [no-ec] [no-rsa]
    	make depend all test
work again

PR: 1159

- hide the EC_KEY structure definition in ec_lcl.c + add
  some functions to use/access the EC_KEY fields
- change the way how method specific data (ecdsa/ecdh) is
  attached to a EC_KEY
- add ECDSA_sign_ex and ECDSA_do_sign_ex functions with
  additional parameters for pre-computed values
- rebuild libeay.num from 0.9.7

with the SSL_OP_NO_SSLv2 option.

EC_GROUP_new_by_nid -> EC_GROUP_new_by_curve_name

("perl util/ck_errf.pl */*.c */*/*.c" still reports many more.)

client random values.