CHANGES

     This addresses the flaw in CBC record processing discovered by 
     Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
     at: http://www.isg.rhul.ac.uk/tls/     

     Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
     Security Group at Royal Holloway, University of London
     (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
     Emilia Käsper for the initial patch.
     (CVE-2013-0169)
     [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]

  *) Return an error when checking OCSP signatures when key is NULL.
     This fixes a DoS attack. (CVE-2013-0166)
     [Steve Henson]

  *) Call OCSP Stapling callback after ciphersuite has been chosen, so
     the right response is stapled. Also change SSL_get_certificate()
     so it returns the certificate actually sent.
     See http://rt.openssl.org/Ticket/Display.html?id=2836.
     (This is a backport)
     [Rob Stradling <rob.stradling@comodo.com>]

  *) Fix possible deadlock when decoding public keys.
     [Steve Henson]

 Changes between 0.9.8w and 0.9.8x [10 May 2012]

  *) Sanity check record length before skipping explicit IV in DTLS
     to fix DoS attack.

     Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
     fuzzing as a service testing platform.
     (CVE-2012-2333)
     [Steve Henson]

  *) Initialise tkeylen properly when encrypting CMS messages.
     Thanks to Solar Designer of Openwall for reporting this issue.
     [Steve Henson]

 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]

  *) The fix for CVE-2012-2110 did not take into account that the 
     'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
     int in OpenSSL 0.9.8, making it still vulnerable. Fix by 
     rejecting negative len parameter. (CVE-2012-2131)
     [Tomas Hoger <thoger@redhat.com>]

 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]

  *) Check for potentially exploitable overflows in asn1_d2i_read_bio
     BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
     in CRYPTO_realloc_clean.

     Thanks to Tavis Ormandy, Google Security Team, for discovering this
     issue and to Adam Langley <agl@chromium.org> for fixing it.
     (CVE-2012-2110)
     [Adam Langley (Google), Tavis Ormandy, Google Security Team]

 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]

  *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
     in CMS and PKCS7 code. When RSA decryption fails use a random key for
     content decryption and always return the same error. Note: this attack
     needs on average 2^20 messages so it only affects automated senders. The
     old behaviour can be reenabled in the CMS code by setting the
     CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
     an MMA defence is not necessary.
     Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
     this issue. (CVE-2012-0884)
     [Steve Henson]

  *) Fix CVE-2011-4619: make sure we really are receiving a 
     client hello before rejecting multiple SGC restarts. Thanks to
     Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
     [Steve Henson]

 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]

  *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
     Thanks to Antonio Martin, Enterprise Secure Access Research and
     Development, Cisco Systems, Inc. for discovering this bug and
     preparing a fix. (CVE-2012-0050)
     [Antonio Martin]

 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]

  *) Nadhem Alfardan and Kenny Paterson have discovered an extension
     of the Vaudenay padding oracle attack on CBC mode encryption
     which enables an efficient plaintext recovery attack against
     the OpenSSL implementation of DTLS. Their attack exploits timing
     differences arising during decryption processing. A research
     paper describing this attack can be found at:
                  http://www.isg.rhul.ac.uk/~kp/dtls.pdf
     Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
     Security Group at Royal Holloway, University of London
     (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
     <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
     for preparing the fix. (CVE-2011-4108)
     [Robin Seggelmann, Michael Tuexen]

  *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
     [Ben Laurie, Kasper <ekasper@google.com>]

  *) Clear bytes used for block padding of SSL 3.0 records.
     (CVE-2011-4576)
     [Adam Langley (Google)]

  *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
     Kadianakis <desnacked@gmail.com> for discovering this issue and
     Adam Langley for preparing the fix. (CVE-2011-4619)
     [Adam Langley (Google)]
 
  *) Prevent malformed RFC3779 data triggering an assertion failure.
     Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
     and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
     [Rob Austein <sra@hactrn.net>]

  *) Fix ssl_ciph.c set-up race.
     [Adam Langley (Google)]

  *) Fix spurious failures in ecdsatest.c.
     [Emilia Käsper (Google)]

  *) Fix the BIO_f_buffer() implementation (which was mixing different
     interpretations of the '..._len' fields).
     [Adam Langley (Google)]

  *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
     BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
     threads won't reuse the same blinding coefficients.

     This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
     lock to call BN_BLINDING_invert_ex, and avoids one use of
     BN_BLINDING_update for each BN_BLINDING structure (previously,
     the last update always remained unused).
     [Emilia Käsper (Google)]

  *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
     for multi-threaded use of ECDH.
     [Adam Langley (Google)]

  *) Fix x509_name_ex_d2i memory leak on bad inputs.
     [Bodo Moeller]

  *) Add protection against ECDSA timing attacks as mentioned in the paper
     by Billy Bob Brumley and Nicola Tuveri, see:

	http://eprint.iacr.org/2011/232.pdf

     [Billy Bob Brumley and Nicola Tuveri]

 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]

  *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
     [Neel Mehta, Adam Langley, Bodo Moeller (Google)]

  *) Fix bug in string printing code: if *any* escaping is enabled we must
     escape the escape character (backslash) or the resulting string is
     ambiguous.
     [Steve Henson]

 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]

  *) Disable code workaround for ancient and obsolete Netscape browsers
     and servers: an attacker can use it in a ciphersuite downgrade attack.
     Thanks to Martin Rex for discovering this bug. CVE-2010-4180
     [Steve Henson]

  *) Fixed J-PAKE implementation error, originally discovered by
     Sebastien Martini, further info and confirmation from Stefan
     Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
     [Ben Laurie]

 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]

  *) Fix extension code to avoid race conditions which can result in a buffer
     overrun vulnerability: resumed sessions must not be modified as they can
     be shared by multiple threads. CVE-2010-3864
     [Steve Henson]

  *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
     [Steve Henson]

  *) Don't reencode certificate when calculating signature: cache and use
     the original encoding instead. This makes signature verification of
     some broken encodings work correctly.
     [Steve Henson]

  *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
     is also one of the inputs.
     [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]

  *) Don't repeatedly append PBE algorithms to table if they already exist.
     Sort table on each new add. This effectively makes the table read only
     after all algorithms are added and subsequent calls to PKCS12_pbe_add
     etc are non-op.
     [Steve Henson]

 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]

  [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
  OpenSSL 1.0.0.]

  *) Correct a typo in the CMS ASN1 module which can result in invalid memory
     access or freeing data twice (CVE-2010-0742)
     [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]

  *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
     common in certificates and some applications which only call
     SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
     [Steve Henson]

  *) VMS fixes: 
     Reduce copying into .apps and .test in makevms.com
     Don't try to use blank CA certificate in CA.com
     Allow use of C files from original directories in maketests.com
     [Steven M. Schweda" <sms@antinode.info>]

 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]

  *) When rejecting SSL/TLS records due to an incorrect version number, never
     update s->server with a new major version number.  As of
     - OpenSSL 0.9.8m if 'short' is a 16-bit type,
     - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
     the previous behavior could result in a read attempt at NULL when
     receiving specific incorrect SSL/TLS records once record payload
     protection is active.  (CVE-2010-0740)
     [Bodo Moeller, Adam Langley <agl@chromium.org>]

  *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 
     could be crashed if the relevant tables were not present (e.g. chrooted).
     [Tomas Hoger <thoger@redhat.com>]

 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]

  *) Always check bn_wexpend() return values for failure.  (CVE-2009-3245)
     [Martin Olsson, Neel Mehta]

  *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
     accommodate for stack sorting, always a write lock!).
     [Bodo Moeller]

  *) On some versions of WIN32 Heap32Next is very slow. This can cause
     excessive delays in the RAND_poll(): over a minute. As a workaround
     include a time check in the inner Heap32Next loop too.
     [Steve Henson]

  *) The code that handled flushing of data in SSL/TLS originally used the
     BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
     the problem outlined in PR#1949. The fix suggested there however can
     trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
     of Apache). So instead simplify the code to flush unconditionally.
     This should be fine since flushing with no data to flush is a no op.
     [Steve Henson]

  *) Handle TLS versions 2.0 and later properly and correctly use the
     highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
     off ancient servers have a habit of sticking around for a while...
     [Steve Henson]

  *) Modify compression code so it frees up structures without using the
     ex_data callbacks. This works around a problem where some applications
     call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
     restarting) then use compression (e.g. SSL with compression) later.
     This results in significant per-connection memory leaks and
     has caused some security issues including CVE-2008-1678 and
     CVE-2009-4355.
     [Steve Henson]

  *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
     change when encrypting or decrypting.
     [Bodo Moeller]

  *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
     connect and renegotiate with servers which do not support RI.
     Until RI is more widely deployed this option is enabled by default.
     [Steve Henson]

  *) Add "missing" ssl ctrls to clear options and mode.
     [Steve Henson]

  *) If client attempts to renegotiate and doesn't support RI respond with
     a no_renegotiation alert as required by RFC5746.  Some renegotiating
     TLS clients will continue a connection gracefully when they receive
     the alert. Unfortunately OpenSSL mishandled this alert and would hang
     waiting for a server hello which it will never receive. Now we treat a
     received no_renegotiation alert as a fatal error. This is because
     applications requesting a renegotiation might well expect it to succeed
     and would have no code in place to handle the server denying it so the
     only safe thing to do is to terminate the connection.
     [Steve Henson]

  *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
     peer supports secure renegotiation and 0 otherwise. Print out peer
     renegotiation support in s_client/s_server.
     [Steve Henson]

  *) Replace the highly broken and deprecated SPKAC certification method with
     the updated NID creation version. This should correctly handle UTF8.
     [Steve Henson]

  *) Implement RFC5746. Re-enable renegotiation but require the extension
     as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
     turns out to be a bad idea. It has been replaced by
     SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
     SSL_CTX_set_options(). This is really not recommended unless you
     know what you are doing.
     [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]

  *) Fixes to stateless session resumption handling. Use initial_ctx when
     issuing and attempting to decrypt tickets in case it has changed during
     servername handling. Use a non-zero length session ID when attempting
     stateless session resumption: this makes it possible to determine if
     a resumption has occurred immediately after receiving server hello
     (several places in OpenSSL subtly assume this) instead of later in
     the handshake.
     [Steve Henson]

  *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
     CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
     fixes for a few places where the return code is not checked
     correctly.
     [Julia Lawall <julia@diku.dk>]

  *) Add --strict-warnings option to Configure script to include devteam
     warnings in other configurations.
     [Steve Henson]

  *) Add support for --libdir option and LIBDIR variable in makefiles. This
     makes it possible to install openssl libraries in locations which
     have names other than "lib", for example "/usr/lib64" which some
     systems need.
     [Steve Henson, based on patch from Jeremy Utley]

  *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
     X690 8.9.12 and can produce some misleading textual output of OIDs.
     [Steve Henson, reported by Dan Kaminsky]

  *) Delete MD2 from algorithm tables. This follows the recommendation in
     several standards that it is not used in new applications due to
     several cryptographic weaknesses. For binary compatibility reasons
     the MD2 API is still compiled in by default.
     [Steve Henson]

  *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
     and restored.
     [Steve Henson]

  *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
     OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
     clash.
     [Guenter <lists@gknw.net>]

  *) Fix the server certificate chain building code to use X509_verify_cert(),
     it used to have an ad-hoc builder which was unable to cope with anything
     other than a simple chain.
     [David Woodhouse <dwmw2@infradead.org>, Steve Henson]

  *) Don't check self signed certificate signatures in X509_verify_cert()
     by default (a flag can override this): it just wastes time without
     adding any security. As a useful side effect self signed root CAs
     with non-FIPS digests are now usable in FIPS mode.
     [Steve Henson]

  *) In dtls1_process_out_of_seq_message() the check if the current message
     is already buffered was missing. For every new message was memory
     allocated, allowing an attacker to perform an denial of service attack
     with sending out of seq handshake messages until there is no memory
     left. Additionally every future messege was buffered, even if the
     sequence number made no sense and would be part of another handshake.
     So only messages with sequence numbers less than 10 in advance will be
     buffered.  (CVE-2009-1378)
     [Robin Seggelmann, discovered by Daniel Mentz] 	

  *) Records are buffered if they arrive with a future epoch to be
     processed after finishing the corresponding handshake. There is
     currently no limitation to this buffer allowing an attacker to perform
     a DOS attack with sending records with future epochs until there is no
     memory left. This patch adds the pqueue_size() function to detemine
     the size of a buffer and limits the record buffer to 100 entries.
     (CVE-2009-1377)
     [Robin Seggelmann, discovered by Daniel Mentz] 	

  *) Keep a copy of frag->msg_header.frag_len so it can be used after the
     parent structure is freed.  (CVE-2009-1379)
     [Daniel Mentz] 	

  *) Handle non-blocking I/O properly in SSL_shutdown() call.
     [Darryl Miles <darryl-mailinglists@netbauds.net>]

  *) Add 2.5.4.* OIDs
     [Ilya O. <vrghost@gmail.com>]

 Changes between 0.9.8k and 0.9.8l  [5 Nov 2009]

  *) Disable renegotiation completely - this fixes a severe security
     problem (CVE-2009-3555) at the cost of breaking all
     renegotiation. Renegotiation can be re-enabled by setting
     SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
     run-time. This is really not recommended unless you know what
     you're doing.
     [Ben Laurie]

 Changes between 0.9.8j and 0.9.8k  [25 Mar 2009]

  *) Don't set val to NULL when freeing up structures, it is freed up by
     underlying code. If sizeof(void *) > sizeof(long) this can result in
     zeroing past the valid field. (CVE-2009-0789)
     [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]

  *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
     checked correctly. This would allow some invalid signed attributes to
     appear to verify correctly. (CVE-2009-0591)
     [Ivan Nestlerode <inestlerode@us.ibm.com>]

  *) Reject UniversalString and BMPString types with invalid lengths. This
     prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
     a legal length. (CVE-2009-0590)
     [Steve Henson]

  *) Set S/MIME signing as the default purpose rather than setting it 
     unconditionally. This allows applications to override it at the store
     level.
     [Steve Henson]

  *) Permit restricted recursion of ASN1 strings. This is needed in practice
     to handle some structures.
     [Steve Henson]

  *) Improve efficiency of mem_gets: don't search whole buffer each time
     for a '\n'
     [Jeremy Shapiro <jnshapir@us.ibm.com>]

  *) New -hex option for openssl rand.
     [Matthieu Herrb]

  *) Print out UTF8String and NumericString when parsing ASN1.
     [Steve Henson]

  *) Support NumericString type for name components.
     [Steve Henson]

  *) Allow CC in the environment to override the automatically chosen
     compiler. Note that nothing is done to ensure flags work with the
     chosen compiler.
     [Ben Laurie]

 Changes between 0.9.8i and 0.9.8j  [07 Jan 2009]

  *) Properly check EVP_VerifyFinal() and similar return values
     (CVE-2008-5077).
     [Ben Laurie, Bodo Moeller, Google Security Team]

  *) Enable TLS extensions by default.
     [Ben Laurie]

  *) Allow the CHIL engine to be loaded, whether the application is
     multithreaded or not. (This does not release the developer from the
     obligation to set up the dynamic locking callbacks.)
     [Sander Temme <sander@temme.net>]

  *) Use correct exit code if there is an error in dgst command.
     [Steve Henson; problem pointed out by Roland Dirlewanger]

  *) Tweak Configure so that you need to say "experimental-jpake" to enable
     JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
     [Bodo Moeller]

  *) Add experimental JPAKE support, including demo authentication in
     s_client and s_server.
     [Ben Laurie]

  *) Set the comparison function in v3_addr_canonize().
     [Rob Austein <sra@hactrn.net>]

  *) Add support for XMPP STARTTLS in s_client.
     [Philip Paeps <philip@freebsd.org>]

  *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
     to ensure that even with this option, only ciphersuites in the
     server's preference list will be accepted.  (Note that the option
     applies only when resuming a session, so the earlier behavior was
     just about the algorithm choice for symmetric cryptography.)
     [Bodo Moeller]

 Changes between 0.9.8h and 0.9.8i  [15 Sep 2008]

  *) Fix NULL pointer dereference if a DTLS server received
     ChangeCipherSpec as first record (CVE-2009-1386).
     [PR #1679]

  *) Fix a state transitition in s3_srvr.c and d1_srvr.c
     (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
     [Nagendra Modadugu]

  *) The fix in 0.9.8c that supposedly got rid of unsafe
     double-checked locking was incomplete for RSA blinding,
     addressing just one layer of what turns out to have been
     doubly unsafe triple-checked locking.

     So now fix this for real by retiring the MONT_HELPER macro
     in crypto/rsa/rsa_eay.c.

     [Bodo Moeller; problem pointed out by Marius Schilder]

  *) Various precautionary measures:

     - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).

     - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
       (NB: This would require knowledge of the secret session ticket key
       to exploit, in which case you'd be SOL either way.)

     - Change bn_nist.c so that it will properly handle input BIGNUMs
       outside the expected range.

     - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
       builds.

     [Neel Mehta, Bodo Moeller]

  *) Allow engines to be "soft loaded" - i.e. optionally don't die if
     the load fails. Useful for distros.
     [Ben Laurie and the FreeBSD team]

  *) Add support for Local Machine Keyset attribute in PKCS#12 files.
     [Steve Henson]

  *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
     [Huang Ying]

  *) Expand ENGINE to support engine supplied SSL client certificate functions.

     This work was sponsored by Logica.
     [Steve Henson]

  *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
     keystores. Support for SSL/TLS client authentication too.
     Not compiled unless enable-capieng specified to Configure.

     This work was sponsored by Logica.
     [Steve Henson]

  *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
     ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
     attribute creation routines such as certifcate requests and PKCS#12
     files.
     [Steve Henson]

 Changes between 0.9.8g and 0.9.8h  [28 May 2008]

  *) 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) 
     [Steve Henson, Mark Cox]

  *) Fix double free in TLS server name extensions which could lead to
     a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 
     [Joe Orton]

  *) Clear error queue in SSL_CTX_use_certificate_chain_file()

     Clear the error queue to ensure that error entries left from
     older function calls do not interfere with the correct operation.
     [Lutz Jaenicke, Erik de Castro Lopo]

  *) Remove root CA certificates of commercial CAs:

     The OpenSSL project does not recommend any specific CA and does not
     have any policy with respect to including or excluding any CA.
     Therefore it does not make any sense to ship an arbitrary selection
     of root CA certificates with the OpenSSL software.
     [Lutz Jaenicke]

  *) RSA OAEP patches to fix two separate invalid memory reads.
     The first one involves inputs when 'lzero' is greater than
     'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
     before the beginning of from). The second one involves inputs where
     the 'db' section contains nothing but zeroes (there is a one-byte
     invalid read after the end of 'db').
     [Ivan Nestlerode <inestlerode@us.ibm.com>]

  *) Partial backport from 0.9.9-dev:

     Introduce bn_mul_mont (dedicated Montgomery multiplication
     procedure) as a candidate for BIGNUM assembler implementation.
     While 0.9.9-dev uses assembler for various architectures, only
     x86_64 is available by default here in the 0.9.8 branch, and
     32-bit x86 is available through a compile-time setting.

     To try the 32-bit x86 assembler implementation, use Configure
     option "enable-montasm" (which exists only for this backport).

     As "enable-montasm" for 32-bit x86 disclaims code stability
     anyway, in this constellation we activate additional code
     backported from 0.9.9-dev for further performance improvements,
     namely BN_from_montgomery_word.  (To enable this otherwise,
     e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)

     [Andy Polyakov (backport partially by Bodo Moeller)]

  *) Add TLS session ticket callback. This allows an application to set
     TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
     values. This is useful for key rollover for example where several key
     sets may exist with different names.
     [Steve Henson]

  *) Reverse ENGINE-internal logic for caching default ENGINE handles.
     This was broken until now in 0.9.8 releases, such that the only way
     a registered ENGINE could be used (assuming it initialises
     successfully on the host) was to explicitly set it as the default
     for the relevant algorithms. This is in contradiction with 0.9.7
     behaviour and the documentation. With this fix, when an ENGINE is
     registered into a given algorithm's table of implementations, the
     'uptodate' flag is reset so that auto-discovery will be used next
     time a new context for that algorithm attempts to select an
     implementation.
     [Ian Lister (tweaked by Geoff Thorpe)]

  *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
     implemention in the following ways:

     Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
     hard coded.

     Lack of BER streaming support means one pass streaming processing is
     only supported if data is detached: setting the streaming flag is
     ignored for embedded content.

     CMS support is disabled by default and must be explicitly enabled
     with the enable-cms configuration option.
     [Steve Henson]

  *) Update the GMP engine glue to do direct copies between BIGNUM and
     mpz_t when openssl and GMP use the same limb size. Otherwise the
     existing "conversion via a text string export" trick is still used.
     [Paul Sheer <paulsheer@gmail.com>]

  *) Zlib compression BIO. This is a filter BIO which compressed and
     uncompresses any data passed through it.
     [Steve Henson]

  *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
     RFC3394 compatible AES key wrapping.
     [Steve Henson]

  *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
     sets string data without copying. X509_ALGOR_set0() and
     X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
     data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
     from an X509_ATTRIBUTE structure optionally checking it occurs only
     once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
     data.
     [Steve Henson]

  *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
     to get the expected BN_FLG_CONSTTIME behavior.
     [Bodo Moeller (Google)]
  
  *) Netware support:

     - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
     - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
     - added some more tests to do_tests.pl
     - fixed RunningProcess usage so that it works with newer LIBC NDKs too
     - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
     - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
       netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
     - various changes to netware.pl to enable gcc-cross builds on Win32
       platform
     - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
     - various changes to fix missing prototype warnings
     - fixed x86nasm.pl to create correct asm files for NASM COFF output
     - added AES, WHIRLPOOL and CPUID assembler code to build files
     - added missing AES assembler make rules to mk1mf.pl
     - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
     [Guenter Knauf <eflash@gmx.net>]

  *) Implement certificate status request TLS extension defined in RFC3546.
     A client can set the appropriate parameters and receive the encoded
     OCSP response via a callback. A server can query the supplied parameters
     and set the encoded OCSP response in the callback. Add simplified examples
     to s_client and s_server.
     [Steve Henson]

 Changes between 0.9.8f and 0.9.8g  [19 Oct 2007]

  *) Fix various bugs:
     + Binary incompatibility of ssl_ctx_st structure
     + DTLS interoperation with non-compliant servers
     + Don't call get_session_cb() without proposed session
     + Fix ia64 assembler code
     [Andy Polyakov, Steve Henson]

 Changes between 0.9.8e and 0.9.8f  [11 Oct 2007]

  *) DTLS Handshake overhaul. There were longstanding issues with
     OpenSSL DTLS implementation, which were making it impossible for
     RFC 4347 compliant client to communicate with OpenSSL server.
     Unfortunately just fixing these incompatibilities would "cut off"
     pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
     server keeps tolerating non RFC compliant syntax. The opposite is
     not true, 0.9.8f client can not communicate with earlier server.
     This update even addresses CVE-2007-4995.
     [Andy Polyakov]

  *) Changes to avoid need for function casts in OpenSSL: some compilers
     (gcc 4.2 and later) reject their use.
     [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
      Steve Henson]
  
  *) Add RFC4507 support to OpenSSL. This includes the corrections in
     RFC4507bis. The encrypted ticket format is an encrypted encoded
     SSL_SESSION structure, that way new session features are automatically
     supported.

     If a client application caches session in an SSL_SESSION structure
     support is transparent because tickets are now stored in the encoded
     SSL_SESSION.
     
     The SSL_CTX structure automatically generates keys for ticket
     protection in servers so again support should be possible
     with no application modification.

     If a client or server wishes to disable RFC4507 support then the option
     SSL_OP_NO_TICKET can be set.

     Add a TLS extension debugging callback to allow the contents of any client
     or server extensions to be examined.

     This work was sponsored by Google.
     [Steve Henson]

  *) Add initial support for TLS extensions, specifically for the server_name
     extension so far.  The SSL_SESSION, SSL_CTX, and SSL data structures now
     have new members for a host name.  The SSL data structure has an
     additional member SSL_CTX *initial_ctx so that new sessions can be
     stored in that context to allow for session resumption, even after the
     SSL has been switched to a new SSL_CTX in reaction to a client's
     server_name extension.

     New functions (subject to change):

         SSL_get_servername()
         SSL_get_servername_type()
         SSL_set_SSL_CTX()

     New CTRL codes and macros (subject to change):

         SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
                                 - SSL_CTX_set_tlsext_servername_callback()
         SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
                                      - SSL_CTX_set_tlsext_servername_arg()
         SSL_CTRL_SET_TLSEXT_HOSTNAME           - SSL_set_tlsext_host_name()

     openssl s_client has a new '-servername ...' option.

     openssl s_server has new options '-servername_host ...', '-cert2 ...',
     '-key2 ...', '-servername_fatal' (subject to change).  This allows
     testing the HostName extension for a specific single host name ('-cert'
     and '-key' remain fallbacks for handshakes without HostName
     negotiation).  If the unrecogninzed_name alert has to be sent, this by
     default is a warning; it becomes fatal with the '-servername_fatal'
     option.

     [Peter Sylvester,  Remy Allais, Christophe Renou, Steve Henson]

  *) Add AES and SSE2 assembly language support to VC++ build.
     [Steve Henson]

  *) Mitigate attack on final subtraction in Montgomery reduction.
     [Andy Polyakov]

  *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
     (which previously caused an internal error).
     [Bodo Moeller]

  *) Squeeze another 10% out of IGE mode when in != out.
     [Ben Laurie]

  *) AES IGE mode speedup.
     [Dean Gaudet (Google)]

  *) Add the Korean symmetric 128-bit cipher SEED (see
     http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
     add SEED ciphersuites from RFC 4162:

        TLS_RSA_WITH_SEED_CBC_SHA      =  "SEED-SHA"
        TLS_DHE_DSS_WITH_SEED_CBC_SHA  =  "DHE-DSS-SEED-SHA"
        TLS_DHE_RSA_WITH_SEED_CBC_SHA  =  "DHE-RSA-SEED-SHA"
        TLS_DH_anon_WITH_SEED_CBC_SHA  =  "ADH-SEED-SHA"

     To minimize changes between patchlevels in the OpenSSL 0.9.8
     series, SEED remains excluded from compilation unless OpenSSL
     is configured with 'enable-seed'.
     [KISA, Bodo Moeller]

  *) Mitigate branch prediction attacks, which can be practical if a
     single processor is shared, allowing a spy process to extract
     information.  For detailed background information, see
     http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
     J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
     and Necessary Software Countermeasures").  The core of the change
     are new versions BN_div_no_branch() and
     BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
     respectively, which are slower, but avoid the security-relevant
     conditional branches.  These are automatically called by BN_div()
     and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
     of the input BIGNUMs.  Also, BN_is_bit_set() has been changed to
     remove a conditional branch.

     BN_FLG_CONSTTIME is the new name for the previous
     BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
     modular exponentiation.  (Since OpenSSL 0.9.7h, setting this flag
     in the exponent causes BN_mod_exp_mont() to use the alternative
     implementation in BN_mod_exp_mont_consttime().)  The old name
     remains as a deprecated alias.

     Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
     RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
     constant-time implementations for more than just exponentiation.
     Here too the old name is kept as a deprecated alias.

     BN_BLINDING_new() will now use BN_dup() for the modulus so that
     the BN_BLINDING structure gets an independent copy of the
     modulus.  This means that the previous "BIGNUM *m" argument to
     BN_BLINDING_new() and to BN_BLINDING_create_param() now
     essentially becomes "const BIGNUM *m", although we can't actually
     change this in the header file before 0.9.9.  It allows
     RSA_setup_blinding() to use BN_with_flags() on the modulus to
     enable BN_FLG_CONSTTIME.

     [Matthew D Wood (Intel Corp)]

  *) In the SSL/TLS server implementation, be strict about session ID
     context matching (which matters if an application uses a single
     external cache for different purposes).  Previously,
     out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
     set.  This did ensure strict client verification, but meant that,
     with applications using a single external cache for quite
     different requirements, clients could circumvent ciphersuite
     restrictions for a given session ID context by starting a session
     in a different context.
     [Bodo Moeller]

  *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
     a ciphersuite string such as "DEFAULT:RSA" cannot enable
     authentication-only ciphersuites.
     [Bodo Moeller]

  *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
     not complete and could lead to a possible single byte overflow
     (CVE-2007-5135) [Ben Laurie]

 Changes between 0.9.8d and 0.9.8e  [23 Feb 2007]

  *) Since AES128 and AES256 (and similarly Camellia128 and
     Camellia256) share a single mask bit in the logic of
     ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
     kludge to work properly if AES128 is available and AES256 isn't
     (or if Camellia128 is available and Camellia256 isn't).
     [Victor Duchovni]

  *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
     (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
     When a point or a seed is encoded in a BIT STRING, we need to
     prevent the removal of trailing zero bits to get the proper DER
     encoding.  (By default, crypto/asn1/a_bitstr.c assumes the case
     of a NamedBitList, for which trailing 0 bits need to be removed.)
     [Bodo Moeller]

  *) Have SSL/TLS server implementation tolerate "mismatched" record
     protocol version while receiving ClientHello even if the
     ClientHello is fragmented.  (The server can't insist on the
     particular protocol version it has chosen before the ServerHello
     message has informed the client about his choice.)
     [Bodo Moeller]

  *) Add RFC 3779 support.
     [Rob Austein for ARIN, Ben Laurie]

  *) Load error codes if they are not already present instead of using a
     static variable. This allows them to be cleanly unloaded and reloaded.
     Improve header file function name parsing.
     [Steve Henson]

  *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
     or CAPABILITY handshake as required by RFCs.
     [Goetz Babin-Ebell]

 Changes between 0.9.8c and 0.9.8d  [28 Sep 2006]

  *) Introduce limits to prevent malicious keys being able to
     cause a denial of service.  (CVE-2006-2940)
     [Steve Henson, Bodo Moeller]

  *) Fix ASN.1 parsing of certain invalid structures that can result
     in a denial of service.  (CVE-2006-2937)  [Steve Henson]

  *) Fix buffer overflow in SSL_get_shared_ciphers() function. 
     (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]

  *) Fix SSL client code which could crash if connecting to a
     malicious SSLv2 server.  (CVE-2006-4343)
     [Tavis Ormandy and Will Drewry, Google Security Team]

  *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
     match only those.  Before that, "AES256-SHA" would be interpreted
     as a pattern and match "AES128-SHA" too (since AES128-SHA got
     the same strength classification in 0.9.7h) as we currently only
     have a single AES bit in the ciphersuite description bitmap.
     That change, however, also applied to ciphersuite strings such as
     "RC4-MD5" that intentionally matched multiple ciphersuites --
     namely, SSL 2.0 ciphersuites in addition to the more common ones
     from SSL 3.0/TLS 1.0.

     So we change the selection algorithm again: Naming an explicit
     ciphersuite selects this one ciphersuite, and any other similar
     ciphersuite (same bitmap) from *other* protocol versions.
     Thus, "RC4-MD5" again will properly select both the SSL 2.0
     ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.

     Since SSL 2.0 does not have any ciphersuites for which the
     128/256 bit distinction would be relevant, this works for now.
     The proper fix will be to use different bits for AES128 and
     AES256, which would have avoided the problems from the beginning;
     however, bits are scarce, so we can only do this in a new release
     (not just a patchlevel) when we can change the SSL_CIPHER
     definition to split the single 'unsigned long mask' bitmap into
     multiple values to extend the available space.

     [Bodo Moeller]

 Changes between 0.9.8b and 0.9.8c  [05 Sep 2006]

  *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
     (CVE-2006-4339)  [Ben Laurie and Google Security Team]

  *) Add AES IGE and biIGE modes.
     [Ben Laurie]

  *) Change the Unix randomness entropy gathering to use poll() when
     possible instead of select(), since the latter has some
     undesirable limitations.
     [Darryl Miles via Richard Levitte and Bodo Moeller]

  *) Disable "ECCdraft" ciphersuites more thoroughly.  Now special
     treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
     cannot be implicitly activated as part of, e.g., the "AES" alias.
     However, please upgrade to OpenSSL 0.9.9[-dev] for
     non-experimental use of the ECC ciphersuites to get TLS extension
     support, which is required for curve and point format negotiation
     to avoid potential handshake problems.
     [Bodo Moeller]

  *) Disable rogue ciphersuites:

      - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
      - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
      - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")

     The latter two were purportedly from
     draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
     appear there.

     Also deactivate the remaining ciphersuites from
     draft-ietf-tls-56-bit-ciphersuites-01.txt.  These are just as
     unofficial, and the ID has long expired.
     [Bodo Moeller]

  *) Fix RSA blinding Heisenbug (problems sometimes occured on
     dual-core machines) and other potential thread-safety issues.
     [Bodo Moeller]

  *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
     versions), which is now available for royalty-free use
     (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
     Also, add Camellia TLS ciphersuites from RFC 4132.

     To minimize changes between patchlevels in the OpenSSL 0.9.8
     series, Camellia remains excluded from compilation unless OpenSSL
     is configured with 'enable-camellia'.
     [NTT]

  *) Disable the padding bug check when compression is in use. The padding
     bug check assumes the first packet is of even length, this is not
     necessarily true if compresssion is enabled and can result in false
     positives causing handshake failure. The actual bug test is ancient
     code so it is hoped that implementations will either have fixed it by
     now or any which still have the bug do not support compression.
     [Steve Henson]

 Changes between 0.9.8a and 0.9.8b  [04 May 2006]

  *) When applying a cipher rule check to see if string match is an explicit
     cipher suite and only match that one cipher suite if it is.
     [Steve Henson]

  *) Link in manifests for VC++ if needed.