5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the git commit log; for example,
7 https://github.com/openssl/openssl/commits/ and pick the appropriate
10 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
12 *) Constructed ASN.1 types with a recursive definition could exceed the stack
14 Constructed ASN.1 types with a recursive definition (such as can be found
15 in PKCS7) could eventually exceed the stack given malicious input with
16 excessive recursion. This could result in a Denial Of Service attack. There
17 are no such structures used within SSL/TLS that come from untrusted sources
18 so this is considered safe.
20 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
25 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
27 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
28 effectively reduced to only comparing the least significant bit of each
29 byte. This allows an attacker to forge messages that would be considered as
30 authenticated in an amount of tries lower than that guaranteed by the
31 security claims of the scheme. The module can only be compiled by the
32 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
34 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
39 *) Add a build target 'build_all_generated', to build all generated files
40 and only that. This can be used to prepare everything that requires
41 things like perl for a system that lacks perl and then move everything
42 to that system and do the rest of the build there.
45 *) Backport SSL_OP_NO_RENGOTIATION
47 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
48 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
49 changes this is no longer possible in 1.1.0. Therefore the new
50 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
51 1.1.0 to provide equivalent functionality.
53 Note that if an application built against 1.1.0h headers (or above) is run
54 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
55 accepted but nothing will happen, i.e. renegotiation will not be prevented.
58 *) Removed the OS390-Unix config target. It relied on a script that doesn't
62 *) rsaz_1024_mul_avx2 overflow bug on x86_64
64 There is an overflow bug in the AVX2 Montgomery multiplication procedure
65 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
66 Analysis suggests that attacks against RSA and DSA as a result of this
67 defect would be very difficult to perform and are not believed likely.
68 Attacks against DH1024 are considered just feasible, because most of the
69 work necessary to deduce information about a private key may be performed
70 offline. The amount of resources required for such an attack would be
71 significant. However, for an attack on TLS to be meaningful, the server
72 would have to share the DH1024 private key among multiple clients, which is
73 no longer an option since CVE-2016-0701.
75 This only affects processors that support the AVX2 but not ADX extensions
76 like Intel Haswell (4th generation).
78 This issue was reported to OpenSSL by David Benjamin (Google). The issue
79 was originally found via the OSS-Fuzz project.
83 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
85 *) bn_sqrx8x_internal carry bug on x86_64
87 There is a carry propagating bug in the x86_64 Montgomery squaring
88 procedure. No EC algorithms are affected. Analysis suggests that attacks
89 against RSA and DSA as a result of this defect would be very difficult to
90 perform and are not believed likely. Attacks against DH are considered just
91 feasible (although very difficult) because most of the work necessary to
92 deduce information about a private key may be performed offline. The amount
93 of resources required for such an attack would be very significant and
94 likely only accessible to a limited number of attackers. An attacker would
95 additionally need online access to an unpatched system using the target
96 private key in a scenario with persistent DH parameters and a private
97 key that is shared between multiple clients.
99 This only affects processors that support the BMI1, BMI2 and ADX extensions
100 like Intel Broadwell (5th generation) and later or AMD Ryzen.
102 This issue was reported to OpenSSL by the OSS-Fuzz project.
106 *) Malformed X.509 IPAddressFamily could cause OOB read
108 If an X.509 certificate has a malformed IPAddressFamily extension,
109 OpenSSL could do a one-byte buffer overread. The most likely result
110 would be an erroneous display of the certificate in text format.
112 This issue was reported to OpenSSL by the OSS-Fuzz project.
116 *) Ignore the '-named_curve auto' value for compatibility of applications
118 [Tomas Mraz <tmraz@fedoraproject.org>]
120 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
123 Changes between 1.1.0e and 1.1.0f [25 May 2017]
125 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
126 platform rather than 'mingw'.
129 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
130 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
131 which is the minimum version we support.
134 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
136 *) Encrypt-Then-Mac renegotiation crash
138 During a renegotiation handshake if the Encrypt-Then-Mac extension is
139 negotiated where it was not in the original handshake (or vice-versa) then
140 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
141 and servers are affected.
143 This issue was reported to OpenSSL by Joe Orton (Red Hat).
147 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
149 *) Truncated packet could crash via OOB read
151 If one side of an SSL/TLS path is running on a 32-bit host and a specific
152 cipher is being used, then a truncated packet can cause that host to
153 perform an out-of-bounds read, usually resulting in a crash.
155 This issue was reported to OpenSSL by Robert Święcki of Google.
159 *) Bad (EC)DHE parameters cause a client crash
161 If a malicious server supplies bad parameters for a DHE or ECDHE key
162 exchange then this can result in the client attempting to dereference a
163 NULL pointer leading to a client crash. This could be exploited in a Denial
166 This issue was reported to OpenSSL by Guido Vranken.
170 *) BN_mod_exp may produce incorrect results on x86_64
172 There is a carry propagating bug in the x86_64 Montgomery squaring
173 procedure. No EC algorithms are affected. Analysis suggests that attacks
174 against RSA and DSA as a result of this defect would be very difficult to
175 perform and are not believed likely. Attacks against DH are considered just
176 feasible (although very difficult) because most of the work necessary to
177 deduce information about a private key may be performed offline. The amount
178 of resources required for such an attack would be very significant and
179 likely only accessible to a limited number of attackers. An attacker would
180 additionally need online access to an unpatched system using the target
181 private key in a scenario with persistent DH parameters and a private
182 key that is shared between multiple clients. For example this can occur by
183 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
184 similar to CVE-2015-3193 but must be treated as a separate problem.
186 This issue was reported to OpenSSL by the OSS-Fuzz project.
190 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
192 *) ChaCha20/Poly1305 heap-buffer-overflow
194 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
195 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
196 crash. This issue is not considered to be exploitable beyond a DoS.
198 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
202 *) CMS Null dereference
204 Applications parsing invalid CMS structures can crash with a NULL pointer
205 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
206 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
207 structure callback if an attempt is made to free certain invalid encodings.
208 Only CHOICE structures using a callback which do not handle NULL value are
211 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
215 *) Montgomery multiplication may produce incorrect results
217 There is a carry propagating bug in the Broadwell-specific Montgomery
218 multiplication procedure that handles input lengths divisible by, but
219 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
220 and DH private keys are impossible. This is because the subroutine in
221 question is not used in operations with the private key itself and an input
222 of the attacker's direct choice. Otherwise the bug can manifest itself as
223 transient authentication and key negotiation failures or reproducible
224 erroneous outcome of public-key operations with specially crafted input.
225 Among EC algorithms only Brainpool P-512 curves are affected and one
226 presumably can attack ECDH key negotiation. Impact was not analyzed in
227 detail, because pre-requisites for attack are considered unlikely. Namely
228 multiple clients have to choose the curve in question and the server has to
229 share the private key among them, neither of which is default behaviour.
230 Even then only clients that chose the curve will be affected.
232 This issue was publicly reported as transient failures and was not
233 initially recognized as a security issue. Thanks to Richard Morgan for
234 providing reproducible case.
238 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
239 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
240 prevent issues where no progress is being made and the peer continually
241 sends unrecognised record types, using up resources processing them.
244 *) Removed automatic addition of RPATH in shared libraries and executables,
245 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
248 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
250 *) Fix Use After Free for large message sizes
252 The patch applied to address CVE-2016-6307 resulted in an issue where if a
253 message larger than approx 16k is received then the underlying buffer to
254 store the incoming message is reallocated and moved. Unfortunately a
255 dangling pointer to the old location is left which results in an attempt to
256 write to the previously freed location. This is likely to result in a
257 crash, however it could potentially lead to execution of arbitrary code.
259 This issue only affects OpenSSL 1.1.0a.
261 This issue was reported to OpenSSL by Robert Święcki.
265 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
267 *) OCSP Status Request extension unbounded memory growth
269 A malicious client can send an excessively large OCSP Status Request
270 extension. If that client continually requests renegotiation, sending a
271 large OCSP Status Request extension each time, then there will be unbounded
272 memory growth on the server. This will eventually lead to a Denial Of
273 Service attack through memory exhaustion. Servers with a default
274 configuration are vulnerable even if they do not support OCSP. Builds using
275 the "no-ocsp" build time option are not affected.
277 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
281 *) SSL_peek() hang on empty record
283 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
284 sends an empty record. This could be exploited by a malicious peer in a
285 Denial Of Service attack.
287 This issue was reported to OpenSSL by Alex Gaynor.
291 *) Excessive allocation of memory in tls_get_message_header() and
292 dtls1_preprocess_fragment()
294 A (D)TLS message includes 3 bytes for its length in the header for the
295 message. This would allow for messages up to 16Mb in length. Messages of
296 this length are excessive and OpenSSL includes a check to ensure that a
297 peer is sending reasonably sized messages in order to avoid too much memory
298 being consumed to service a connection. A flaw in the logic of version
299 1.1.0 means that memory for the message is allocated too early, prior to
300 the excessive message length check. Due to way memory is allocated in
301 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
302 to service a connection. This could lead to a Denial of Service through
303 memory exhaustion. However, the excessive message length check still takes
304 place, and this would cause the connection to immediately fail. Assuming
305 that the application calls SSL_free() on the failed connection in a timely
306 manner then the 21Mb of allocated memory will then be immediately freed
307 again. Therefore the excessive memory allocation will be transitory in
308 nature. This then means that there is only a security impact if:
310 1) The application does not call SSL_free() in a timely manner in the event
311 that the connection fails
313 2) The application is working in a constrained environment where there is
314 very little free memory
316 3) The attacker initiates multiple connection attempts such that there are
317 multiple connections in a state where memory has been allocated for the
318 connection; SSL_free() has not yet been called; and there is insufficient
319 memory to service the multiple requests.
321 Except in the instance of (1) above any Denial Of Service is likely to be
322 transitory because as soon as the connection fails the memory is
323 subsequently freed again in the SSL_free() call. However there is an
324 increased risk during this period of application crashes due to the lack of
325 memory - which would then mean a more serious Denial of Service.
327 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
328 (CVE-2016-6307 and CVE-2016-6308)
331 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
332 had to be removed. Primary reason is that vendor assembler can't
333 assemble our modules with -KPIC flag. As result it, assembly
334 support, was not even available as option. But its lack means
335 lack of side-channel resistant code, which is incompatible with
336 security by todays standards. Fortunately gcc is readily available
337 prepackaged option, which we firmly point at...
340 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
342 *) Windows command-line tool supports UTF-8 opt-in option for arguments
343 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
344 (to any value) allows Windows user to access PKCS#12 file generated
345 with Windows CryptoAPI and protected with non-ASCII password, as well
346 as files generated under UTF-8 locale on Linux also protected with
350 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
351 have been disabled by default and removed from DEFAULT, just like RC4.
352 See the RC4 item below to re-enable both.
355 *) The method for finding the storage location for the Windows RAND seed file
356 has changed. First we check %RANDFILE%. If that is not set then we check
357 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
358 all else fails we fall back to C:\.
361 *) The EVP_EncryptUpdate() function has had its return type changed from void
362 to int. A return of 0 indicates and error while a return of 1 indicates
366 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
367 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
368 off the constant time implementation for RSA, DSA and DH have been made
369 no-ops and deprecated.
372 *) Windows RAND implementation was simplified to only get entropy by
373 calling CryptGenRandom(). Various other RAND-related tickets
375 [Joseph Wylie Yandle, Rich Salz]
377 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
378 and OPENSSL_LH_, respectively. The old names are available
379 with API compatibility. They new names are now completely documented.
382 *) Unify TYPE_up_ref(obj) methods signature.
383 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
384 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
385 int (instead of void) like all others TYPE_up_ref() methods.
386 So now these methods also check the return value of CRYPTO_atomic_add(),
387 and the validity of object reference counter.
388 [fdasilvayy@gmail.com]
390 *) With Windows Visual Studio builds, the .pdb files are installed
391 alongside the installed libraries and executables. For a static
392 library installation, ossl_static.pdb is the associate compiler
393 generated .pdb file to be used when linking programs.
396 *) Remove openssl.spec. Packaging files belong with the packagers.
399 *) Automatic Darwin/OSX configuration has had a refresh, it will now
400 recognise x86_64 architectures automatically. You can still decide
401 to build for a different bitness with the environment variable
402 KERNEL_BITS (can be 32 or 64), for example:
404 KERNEL_BITS=32 ./config
408 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
409 256 bit AES and HMAC with SHA256.
412 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
415 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
418 *) To enable users to have their own config files and build file templates,
419 Configure looks in the directory indicated by the environment variable
420 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
421 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
422 name and is used as is.
425 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
426 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
427 X509_CERT_FILE_CTX was removed.
430 *) "shared" builds are now the default. To create only static libraries use
431 the "no-shared" Configure option.
434 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
435 All of these option have not worked for some while and are fundamental
439 *) Make various cleanup routines no-ops and mark them as deprecated. Most
440 global cleanup functions are no longer required because they are handled
441 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
442 Explicitly de-initing can cause problems (e.g. where a library that uses
443 OpenSSL de-inits, but an application is still using it). The affected
444 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
445 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
446 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
450 *) --strict-warnings no longer enables runtime debugging options
451 such as REF_DEBUG. Instead, debug options are automatically
452 enabled with '--debug' builds.
453 [Andy Polyakov, Emilia Käsper]
455 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
456 have been moved out of the public header files. New functions for managing
457 these have been added.
460 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
461 objects have been moved out of the public header files. New
462 functions for managing these have been added.
465 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
466 have been moved out of the public header files. New functions for managing
467 these have been added.
470 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
471 moved out of the public header files. New functions for managing these
475 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
478 *) Removed the mk1mf build scripts.
481 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
482 it is always safe to #include a header now.
485 *) Removed the aged BC-32 config and all its supporting scripts
488 *) Removed support for Ultrix, Netware, and OS/2.
491 *) Add support for HKDF.
494 *) Add support for blake2b and blake2s
497 *) Added support for "pipelining". Ciphers that have the
498 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
499 encryptions/decryptions simultaneously. There are currently no built-in
500 ciphers with this property but the expectation is that engines will be able
501 to offer it to significantly improve throughput. Support has been extended
502 into libssl so that multiple records for a single connection can be
503 processed in one go (for >=TLS 1.1).
506 *) Added the AFALG engine. This is an async capable engine which is able to
507 offload work to the Linux kernel. In this initial version it only supports
508 AES128-CBC. The kernel must be version 4.1.0 or greater.
511 *) OpenSSL now uses a new threading API. It is no longer necessary to
512 set locking callbacks to use OpenSSL in a multi-threaded environment. There
513 are two supported threading models: pthreads and windows threads. It is
514 also possible to configure OpenSSL at compile time for "no-threads". The
515 old threading API should no longer be used. The functions have been
516 replaced with "no-op" compatibility macros.
517 [Alessandro Ghedini, Matt Caswell]
519 *) Modify behavior of ALPN to invoke callback after SNI/servername
520 callback, such that updates to the SSL_CTX affect ALPN.
523 *) Add SSL_CIPHER queries for authentication and key-exchange.
526 *) Changes to the DEFAULT cipherlist:
527 - Prefer (EC)DHE handshakes over plain RSA.
528 - Prefer AEAD ciphers over legacy ciphers.
529 - Prefer ECDSA over RSA when both certificates are available.
530 - Prefer TLSv1.2 ciphers/PRF.
531 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
535 *) Change the ECC default curve list to be this, in order: x25519,
536 secp256r1, secp521r1, secp384r1.
539 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
540 disabled by default. They can be re-enabled using the
541 enable-weak-ssl-ciphers option to Configure.
544 *) If the server has ALPN configured, but supports no protocols that the
545 client advertises, send a fatal "no_application_protocol" alert.
546 This behaviour is SHALL in RFC 7301, though it isn't universally
547 implemented by other servers.
550 *) Add X25519 support.
551 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
552 for public and private key encoding using the format documented in
553 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
554 key generation and key derivation.
556 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
560 *) Deprecate SRP_VBASE_get_by_user.
561 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
562 In order to fix an unavoidable memory leak (CVE-2016-0798),
563 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
564 seed, even if the seed is configured.
566 Users should use SRP_VBASE_get1_by_user instead. Note that in
567 SRP_VBASE_get1_by_user, caller must free the returned value. Note
568 also that even though configuring the SRP seed attempts to hide
569 invalid usernames by continuing the handshake with fake
570 credentials, this behaviour is not constant time and no strong
571 guarantees are made that the handshake is indistinguishable from
572 that of a valid user.
575 *) Configuration change; it's now possible to build dynamic engines
576 without having to build shared libraries and vice versa. This
577 only applies to the engines in engines/, those in crypto/engine/
578 will always be built into libcrypto (i.e. "static").
580 Building dynamic engines is enabled by default; to disable, use
581 the configuration option "disable-dynamic-engine".
583 The only requirements for building dynamic engines are the
584 presence of the DSO module and building with position independent
585 code, so they will also automatically be disabled if configuring
586 with "disable-dso" or "disable-pic".
588 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
589 are also taken away from openssl/opensslconf.h, as they are
593 *) Configuration change; if there is a known flag to compile
594 position independent code, it will always be applied on the
595 libcrypto and libssl object files, and never on the application
596 object files. This means other libraries that use routines from
597 libcrypto / libssl can be made into shared libraries regardless
598 of how OpenSSL was configured.
600 If this isn't desirable, the configuration options "disable-pic"
601 or "no-pic" can be used to disable the use of PIC. This will
602 also disable building shared libraries and dynamic engines.
605 *) Removed JPAKE code. It was experimental and has no wide use.
608 *) The INSTALL_PREFIX Makefile variable has been renamed to
609 DESTDIR. That makes for less confusion on what this variable
610 is for. Also, the configuration option --install_prefix is
614 *) Heartbeat for TLS has been removed and is disabled by default
615 for DTLS; configure with enable-heartbeats. Code that uses the
616 old #define's might need to be updated.
617 [Emilia Käsper, Rich Salz]
619 *) Rename REF_CHECK to REF_DEBUG.
622 *) New "unified" build system
624 The "unified" build system is aimed to be a common system for all
625 platforms we support. With it comes new support for VMS.
627 This system builds supports building in a different directory tree
628 than the source tree. It produces one Makefile (for unix family
629 or lookalikes), or one descrip.mms (for VMS).
631 The source of information to make the Makefile / descrip.mms is
632 small files called 'build.info', holding the necessary
633 information for each directory with source to compile, and a
634 template in Configurations, like unix-Makefile.tmpl or
637 With this change, the library names were also renamed on Windows
638 and on VMS. They now have names that are closer to the standard
639 on Unix, and include the major version number, and in certain
640 cases, the architecture they are built for. See "Notes on shared
641 libraries" in INSTALL.
643 We rely heavily on the perl module Text::Template.
646 *) Added support for auto-initialisation and de-initialisation of the library.
647 OpenSSL no longer requires explicit init or deinit routines to be called,
648 except in certain circumstances. See the OPENSSL_init_crypto() and
649 OPENSSL_init_ssl() man pages for further information.
652 *) The arguments to the DTLSv1_listen function have changed. Specifically the
653 "peer" argument is now expected to be a BIO_ADDR object.
655 *) Rewrite of BIO networking library. The BIO library lacked consistent
656 support of IPv6, and adding it required some more extensive
657 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
658 which hold all types of addresses and chains of address information.
659 It also introduces a new API, with functions like BIO_socket,
660 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
661 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
662 have been adapted accordingly.
665 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
669 *) CRIME protection: disable compression by default, even if OpenSSL is
670 compiled with zlib enabled. Applications can still enable compression
671 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
672 using the SSL_CONF library to configure compression.
675 *) The signature of the session callback configured with
676 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
677 was explicitly marked as 'const unsigned char*' instead of
681 *) Always DPURIFY. Remove the use of uninitialized memory in the
682 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
685 *) Removed many obsolete configuration items, including
686 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
687 MD2_CHAR, MD2_INT, MD2_LONG
689 IDEA_SHORT, IDEA_LONG
690 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
691 [Rich Salz, with advice from Andy Polyakov]
693 *) Many BN internals have been moved to an internal header file.
694 [Rich Salz with help from Andy Polyakov]
696 *) Configuration and writing out the results from it has changed.
697 Files such as Makefile include/openssl/opensslconf.h and are now
698 produced through general templates, such as Makefile.in and
699 crypto/opensslconf.h.in and some help from the perl module
702 Also, the center of configuration information is no longer
703 Makefile. Instead, Configure produces a perl module in
704 configdata.pm which holds most of the config data (in the hash
705 table %config), the target data that comes from the target
706 configuration in one of the Configurations/*.conf files (in
710 *) To clarify their intended purposes, the Configure options
711 --prefix and --openssldir change their semantics, and become more
712 straightforward and less interdependent.
714 --prefix shall be used exclusively to give the location INSTALLTOP
715 where programs, scripts, libraries, include files and manuals are
716 going to be installed. The default is now /usr/local.
718 --openssldir shall be used exclusively to give the default
719 location OPENSSLDIR where certificates, private keys, CRLs are
720 managed. This is also where the default openssl.cnf gets
722 If the directory given with this option is a relative path, the
723 values of both the --prefix value and the --openssldir value will
724 be combined to become OPENSSLDIR.
725 The default for --openssldir is INSTALLTOP/ssl.
727 Anyone who uses --openssldir to specify where OpenSSL is to be
728 installed MUST change to use --prefix instead.
731 *) The GOST engine was out of date and therefore it has been removed. An up
732 to date GOST engine is now being maintained in an external repository.
733 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
734 support for GOST ciphersuites (these are only activated if a GOST engine
738 *) EGD is no longer supported by default; use enable-egd when
740 [Ben Kaduk and Rich Salz]
742 *) The distribution now has Makefile.in files, which are used to
743 create Makefile's when Configure is run. *Configure must be run
744 before trying to build now.*
747 *) The return value for SSL_CIPHER_description() for error conditions
751 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
753 Obtaining and performing DNSSEC validation of TLSA records is
754 the application's responsibility. The application provides
755 the TLSA records of its choice to OpenSSL, and these are then
756 used to authenticate the peer.
758 The TLSA records need not even come from DNS. They can, for
759 example, be used to implement local end-entity certificate or
760 trust-anchor "pinning", where the "pin" data takes the form
761 of TLSA records, which can augment or replace verification
762 based on the usual WebPKI public certification authorities.
765 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
766 continues to support deprecated interfaces in default builds.
767 However, applications are strongly advised to compile their
768 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
769 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
770 or the 1.1.0 releases.
772 In environments in which all applications have been ported to
773 not use any deprecated interfaces OpenSSL's Configure script
774 should be used with the --api=1.1.0 option to entirely remove
775 support for the deprecated features from the library and
776 unconditionally disable them in the installed headers.
777 Essentially the same effect can be achieved with the "no-deprecated"
778 argument to Configure, except that this will always restrict
779 the build to just the latest API, rather than a fixed API
782 As applications are ported to future revisions of the API,
783 they should update their compile-time OPENSSL_API_COMPAT define
784 accordingly, but in most cases should be able to continue to
785 compile with later releases.
787 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
788 0x10000000L and 0x00908000L, respectively. However those
789 versions did not support the OPENSSL_API_COMPAT feature, and
790 so applications are not typically tested for explicit support
791 of just the undeprecated features of either release.
794 *) Add support for setting the minimum and maximum supported protocol.
795 It can bet set via the SSL_set_min_proto_version() and
796 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
797 MaxProtcol. It's recommended to use the new APIs to disable
798 protocols instead of disabling individual protocols using
799 SSL_set_options() or SSL_CONF's Protocol. This change also
800 removes support for disabling TLS 1.2 in the OpenSSL TLS
801 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
804 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
807 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
808 and integrates ECDSA and ECDH functionality into EC. Implementations can
809 now redirect key generation and no longer need to convert to or from
812 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
813 include the ec.h header file instead.
816 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
817 ciphers who are no longer supported and drops support the ephemeral RSA key
818 exchange. The LOW ciphers currently doesn't have any ciphers in it.
821 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
822 opaque. For HMAC_CTX, the following constructors and destructors
825 HMAC_CTX *HMAC_CTX_new(void);
826 void HMAC_CTX_free(HMAC_CTX *ctx);
828 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
829 destroy such methods has been added. See EVP_MD_meth_new(3) and
830 EVP_CIPHER_meth_new(3) for documentation.
833 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
834 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
835 EVP_MD_CTX_reset() should be called instead to reinitialise
836 an already created structure.
837 2) For consistency with the majority of our object creators and
838 destructors, EVP_MD_CTX_(create|destroy) were renamed to
839 EVP_MD_CTX_(new|free). The old names are retained as macros
840 for deprecated builds.
843 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
844 cryptographic operations to be performed asynchronously as long as an
845 asynchronous capable engine is used. See the ASYNC_start_job() man page for
846 further details. Libssl has also had this capability integrated with the
847 introduction of the new mode SSL_MODE_ASYNC and associated error
848 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
849 pages. This work was developed in partnership with Intel Corp.
852 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
853 always enabled now. If you want to disable the support you should
854 exclude it using the list of supported ciphers. This also means that the
855 "-no_ecdhe" option has been removed from s_server.
858 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
859 SSL_{CTX_}set1_curves() which can set a list.
862 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
863 curve you want to support using SSL_{CTX_}set1_curves().
866 *) State machine rewrite. The state machine code has been significantly
867 refactored in order to remove much duplication of code and solve issues
868 with the old code (see ssl/statem/README for further details). This change
869 does have some associated API changes. Notably the SSL_state() function
870 has been removed and replaced by SSL_get_state which now returns an
871 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
872 altogether. The previous handshake states defined in ssl.h and ssl3.h have
876 *) All instances of the string "ssleay" in the public API were replaced
877 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
878 Some error codes related to internal RSA_eay API's were renamed.
881 *) The demo files in crypto/threads were moved to demo/threads.
884 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
886 [Matt Caswell, Rich Salz]
888 *) New ASN.1 embed macro.
890 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
891 structure is not allocated: it is part of the parent. That is instead of
899 This reduces memory fragmentation and make it impossible to accidentally
900 set a mandatory field to NULL.
902 This currently only works for some fields specifically a SEQUENCE, CHOICE,
903 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
904 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
908 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
911 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
912 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
913 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
914 DES and RC4 ciphersuites.
917 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
918 This changes the decoding behaviour for some invalid messages,
919 though the change is mostly in the more lenient direction, and
920 legacy behaviour is preserved as much as possible.
923 *) Fix no-stdio build.
924 [ David Woodhouse <David.Woodhouse@intel.com> and also
925 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
927 *) New testing framework
928 The testing framework has been largely rewritten and is now using
929 perl and the perl modules Test::Harness and an extended variant of
930 Test::More called OpenSSL::Test to do its work. All test scripts in
931 test/ have been rewritten into test recipes, and all direct calls to
932 executables in test/Makefile have become individual recipes using the
933 simplified testing OpenSSL::Test::Simple.
935 For documentation on our testing modules, do:
937 perldoc test/testlib/OpenSSL/Test/Simple.pm
938 perldoc test/testlib/OpenSSL/Test.pm
942 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
943 are used; the latter aborts on memory leaks (usually checked on exit).
944 Some undocumented "set malloc, etc., hooks" functions were removed
945 and others were changed. All are now documented.
948 *) In DSA_generate_parameters_ex, if the provided seed is too short,
950 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
952 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
953 from RFC4279, RFC4785, RFC5487, RFC5489.
955 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
956 original RSA_PSK patch.
959 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
960 era flag was never set throughout the codebase (only read). Also removed
961 SSL3_FLAGS_POP_BUFFER which was only used if
962 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
965 *) Changed the default name options in the "ca", "crl", "req" and "x509"
966 to be "oneline" instead of "compat".
969 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
970 not aware of clients that still exhibit this bug, and the workaround
971 hasn't been working properly for a while.
974 *) The return type of BIO_number_read() and BIO_number_written() as well as
975 the corresponding num_read and num_write members in the BIO structure has
976 changed from unsigned long to uint64_t. On platforms where an unsigned
977 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
981 *) Given the pervasive nature of TLS extensions it is inadvisable to run
982 OpenSSL without support for them. It also means that maintaining
983 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
984 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
987 *) Removed support for the two export grade static DH ciphersuites
988 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
989 were newly added (along with a number of other static DH ciphersuites) to
990 1.0.2. However the two export ones have *never* worked since they were
991 introduced. It seems strange in any case to be adding new export
992 ciphersuites, and given "logjam" it also does not seem correct to fix them.
995 *) Version negotiation has been rewritten. In particular SSLv23_method(),
996 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
997 and turned into macros which simply call the new preferred function names
998 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
999 should use the new names instead. Also as part of this change the ssl23.h
1000 header file has been removed.
1003 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1004 code and the associated standard is no longer considered fit-for-purpose.
1007 *) RT2547 was closed. When generating a private key, try to make the
1008 output file readable only by the owner. This behavior change might
1009 be noticeable when interacting with other software.
1011 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1015 *) Added HTTP GET support to the ocsp command.
1018 *) Changed default digest for the dgst and enc commands from MD5 to
1022 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1025 *) Added support for TLS extended master secret from
1026 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1027 initial patch which was a great help during development.
1030 *) All libssl internal structures have been removed from the public header
1031 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1032 now redundant). Users should not attempt to access internal structures
1033 directly. Instead they should use the provided API functions.
1036 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1037 Access to deprecated functions can be re-enabled by running config with
1038 "enable-deprecated". In addition applications wishing to use deprecated
1039 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1040 will, by default, disable some transitive includes that previously existed
1041 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1044 *) Added support for OCB mode. OpenSSL has been granted a patent license
1045 compatible with the OpenSSL license for use of OCB. Details are available
1046 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1047 for OCB can be removed by calling config with no-ocb.
1050 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1051 compatible client hello.
1054 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1055 done while fixing the error code for the key-too-small case.
1056 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1058 *) CA.sh has been removed; use CA.pl instead.
1061 *) Removed old DES API.
1064 *) Remove various unsupported platforms:
1070 Sinix/ReliantUNIX RM400
1075 16-bit platforms such as WIN16
1078 *) Clean up OPENSSL_NO_xxx #define's
1079 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1080 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1081 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1082 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1083 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1084 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1085 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1086 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1087 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1088 Remove MS_STATIC; it's a relic from platforms <32 bits.
1091 *) Cleaned up dead code
1092 Remove all but one '#ifdef undef' which is to be looked at.
1095 *) Clean up calling of xxx_free routines.
1096 Just like free(), fix most of the xxx_free routines to accept
1097 NULL. Remove the non-null checks from callers. Save much code.
1100 *) Add secure heap for storage of private keys (when possible).
1101 Add BIO_s_secmem(), CBIGNUM, etc.
1102 Contributed by Akamai Technologies under our Corporate CLA.
1105 *) Experimental support for a new, fast, unbiased prime candidate generator,
1106 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1107 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1109 *) New output format NSS in the sess_id command line tool. This allows
1110 exporting the session id and the master key in NSS keylog format.
1111 [Martin Kaiser <martin@kaiser.cx>]
1113 *) Harmonize version and its documentation. -f flag is used to display
1115 [mancha <mancha1@zoho.com>]
1117 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1118 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1119 [mancha <mancha1@zoho.com>]
1121 *) Fix some double frees. These are not thought to be exploitable.
1122 [mancha <mancha1@zoho.com>]
1124 *) A missing bounds check in the handling of the TLS heartbeat extension
1125 can be used to reveal up to 64k of memory to a connected client or
1128 Thanks for Neel Mehta of Google Security for discovering this bug and to
1129 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1130 preparing the fix (CVE-2014-0160)
1131 [Adam Langley, Bodo Moeller]
1133 *) Fix for the attack described in the paper "Recovering OpenSSL
1134 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1135 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1136 http://eprint.iacr.org/2014/140
1138 Thanks to Yuval Yarom and Naomi Benger for discovering this
1139 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1140 [Yuval Yarom and Naomi Benger]
1142 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1143 this fixes a limitation in previous versions of OpenSSL.
1146 *) Experimental encrypt-then-mac support.
1148 Experimental support for encrypt then mac from
1149 draft-gutmann-tls-encrypt-then-mac-02.txt
1151 To enable it set the appropriate extension number (0x42 for the test
1152 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1154 For non-compliant peers (i.e. just about everything) this should have no
1157 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1161 *) Add EVP support for key wrapping algorithms, to avoid problems with
1162 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1163 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1164 algorithms and include tests cases.
1167 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1171 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1172 MGF1 digest and OAEP label.
1175 *) Make openssl verify return errors.
1176 [Chris Palmer <palmer@google.com> and Ben Laurie]
1178 *) New function ASN1_TIME_diff to calculate the difference between two
1179 ASN1_TIME structures or one structure and the current time.
1182 *) Update fips_test_suite to support multiple command line options. New
1183 test to induce all self test errors in sequence and check expected
1187 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1188 sign or verify all in one operation.
1191 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1192 test programs and fips_test_suite. Includes functionality to parse
1193 the minimal script output of fipsalgest.pl directly.
1196 *) Add authorisation parameter to FIPS_module_mode_set().
1199 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1202 *) Use separate DRBG fields for internal and external flags. New function
1203 FIPS_drbg_health_check() to perform on demand health checking. Add
1204 generation tests to fips_test_suite with reduced health check interval to
1205 demonstrate periodic health checking. Add "nodh" option to
1206 fips_test_suite to skip very slow DH test.
1209 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1213 *) More extensive health check for DRBG checking many more failure modes.
1214 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1215 combination: call this in fips_test_suite.
1218 *) Add support for canonical generation of DSA parameter 'g'. See
1221 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1222 POST to handle HMAC cases.
1225 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1226 to return numerical and string versions of the FIPS module number.
1229 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1230 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1231 outside the validated module in the FIPS capable OpenSSL.
1234 *) Minor change to DRBG entropy callback semantics. In some cases
1235 there is no multiple of the block length between min_len and
1236 max_len. Allow the callback to return more than max_len bytes
1237 of entropy but discard any extra: it is the callback's responsibility
1238 to ensure that the extra data discarded does not impact the
1239 requested amount of entropy.
1242 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1243 information in FIPS186-3, SP800-57 and SP800-131A.
1246 *) CCM support via EVP. Interface is very similar to GCM case except we
1247 must supply all data in one chunk (i.e. no update, final) and the
1248 message length must be supplied if AAD is used. Add algorithm test
1252 *) Initial version of POST overhaul. Add POST callback to allow the status
1253 of POST to be monitored and/or failures induced. Modify fips_test_suite
1254 to use callback. Always run all selftests even if one fails.
1257 *) XTS support including algorithm test driver in the fips_gcmtest program.
1258 Note: this does increase the maximum key length from 32 to 64 bytes but
1259 there should be no binary compatibility issues as existing applications
1260 will never use XTS mode.
1263 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1264 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1265 performs algorithm blocking for unapproved PRNG types. Also do not
1266 set PRNG type in FIPS_mode_set(): leave this to the application.
1267 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1268 the standard OpenSSL PRNG: set additional data to a date time vector.
1271 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1272 This shouldn't present any incompatibility problems because applications
1273 shouldn't be using these directly and any that are will need to rethink
1274 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1277 *) Extensive self tests and health checking required by SP800-90 DRBG.
1278 Remove strength parameter from FIPS_drbg_instantiate and always
1279 instantiate at maximum supported strength.
1282 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1285 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1288 *) New function DH_compute_key_padded() to compute a DH key and pad with
1289 leading zeroes if needed: this complies with SP800-56A et al.
1292 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1293 anything, incomplete, subject to change and largely untested at present.
1296 *) Modify fipscanisteronly build option to only build the necessary object
1297 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1300 *) Add experimental option FIPSSYMS to give all symbols in
1301 fipscanister.o and FIPS or fips prefix. This will avoid
1302 conflicts with future versions of OpenSSL. Add perl script
1303 util/fipsas.pl to preprocess assembly language source files
1304 and rename any affected symbols.
1307 *) Add selftest checks and algorithm block of non-fips algorithms in
1308 FIPS mode. Remove DES2 from selftests.
1311 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1312 return internal method without any ENGINE dependencies. Add new
1313 tiny fips sign and verify functions.
1316 *) New build option no-ec2m to disable characteristic 2 code.
1319 *) New build option "fipscanisteronly". This only builds fipscanister.o
1320 and (currently) associated fips utilities. Uses the file Makefile.fips
1321 instead of Makefile.org as the prototype.
1324 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1325 Update fips_gcmtest to use IV generator.
1328 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1329 setting output buffer to NULL. The *Final function must be
1330 called although it will not retrieve any additional data. The tag
1331 can be set or retrieved with a ctrl. The IV length is by default 12
1332 bytes (96 bits) but can be set to an alternative value. If the IV
1333 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1337 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1338 underlying do_cipher function handles all cipher semantics itself
1339 including padding and finalisation. This is useful if (for example)
1340 an ENGINE cipher handles block padding itself. The behaviour of
1341 do_cipher is subtly changed if this flag is set: the return value
1342 is the number of characters written to the output buffer (zero is
1343 no longer an error code) or a negative error code. Also if the
1344 input buffer is NULL and length 0 finalisation should be performed.
1347 *) If a candidate issuer certificate is already part of the constructed
1348 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1351 *) Improve forward-security support: add functions
1353 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1354 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1356 for use by SSL/TLS servers; the callback function will be called whenever a
1357 new session is created, and gets to decide whether the session may be
1358 cached to make it resumable (return 0) or not (return 1). (As by the
1359 SSL/TLS protocol specifications, the session_id sent by the server will be
1360 empty to indicate that the session is not resumable; also, the server will
1361 not generate RFC 4507 (RFC 5077) session tickets.)
1363 A simple reasonable callback implementation is to return is_forward_secure.
1364 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1365 by the SSL/TLS server library, indicating whether it can provide forward
1367 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1369 *) New -verify_name option in command line utilities to set verification
1373 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1374 Add CMAC pkey methods.
1377 *) Experimental renegotiation in s_server -www mode. If the client
1378 browses /reneg connection is renegotiated. If /renegcert it is
1379 renegotiated requesting a certificate.
1382 *) Add an "external" session cache for debugging purposes to s_server. This
1383 should help trace issues which normally are only apparent in deployed
1384 multi-process servers.
1387 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1388 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1389 BIO_set_cipher() and some obscure PEM functions were changed so they
1390 can now return an error. The RAND changes required a change to the
1391 RAND_METHOD structure.
1394 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1395 a gcc attribute to warn if the result of a function is ignored. This
1396 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1397 whose return value is often ignored.
1400 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1401 These allow SCTs (signed certificate timestamps) to be requested and
1402 validated when establishing a connection.
1403 [Rob Percival <robpercival@google.com>]
1405 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1407 *) Prevent padding oracle in AES-NI CBC MAC check
1409 A MITM attacker can use a padding oracle attack to decrypt traffic
1410 when the connection uses an AES CBC cipher and the server support
1413 This issue was introduced as part of the fix for Lucky 13 padding
1414 attack (CVE-2013-0169). The padding check was rewritten to be in
1415 constant time by making sure that always the same bytes are read and
1416 compared against either the MAC or padding bytes. But it no longer
1417 checked that there was enough data to have both the MAC and padding
1420 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1424 *) Fix EVP_EncodeUpdate overflow
1426 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1427 Base64 encoding of binary data. If an attacker is able to supply very large
1428 amounts of input data then a length check can overflow resulting in a heap
1431 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1432 the PEM_write_bio* family of functions. These are mainly used within the
1433 OpenSSL command line applications, so any application which processes data
1434 from an untrusted source and outputs it as a PEM file should be considered
1435 vulnerable to this issue. User applications that call these APIs directly
1436 with large amounts of untrusted data may also be vulnerable.
1438 This issue was reported by Guido Vranken.
1442 *) Fix EVP_EncryptUpdate overflow
1444 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1445 is able to supply very large amounts of input data after a previous call to
1446 EVP_EncryptUpdate() with a partial block then a length check can overflow
1447 resulting in a heap corruption. Following an analysis of all OpenSSL
1448 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1449 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1450 the first called function after an EVP_EncryptInit(), and therefore that
1451 specific call must be safe. The second form is where the length passed to
1452 EVP_EncryptUpdate() can be seen from the code to be some small value and
1453 therefore there is no possibility of an overflow. Since all instances are
1454 one of these two forms, it is believed that there can be no overflows in
1455 internal code due to this problem. It should be noted that
1456 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1457 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1458 of these calls have also been analysed too and it is believed there are no
1459 instances in internal usage where an overflow could occur.
1461 This issue was reported by Guido Vranken.
1465 *) Prevent ASN.1 BIO excessive memory allocation
1467 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1468 a short invalid encoding can cause allocation of large amounts of memory
1469 potentially consuming excessive resources or exhausting memory.
1471 Any application parsing untrusted data through d2i BIO functions is
1472 affected. The memory based functions such as d2i_X509() are *not* affected.
1473 Since the memory based functions are used by the TLS library, TLS
1474 applications are not affected.
1476 This issue was reported by Brian Carpenter.
1482 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1483 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1484 in arbitrary stack data being returned in the buffer.
1486 This issue was reported by Guido Vranken.
1490 *) Modify behavior of ALPN to invoke callback after SNI/servername
1491 callback, such that updates to the SSL_CTX affect ALPN.
1494 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1498 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1499 methods are enabled and ssl2 is disabled the methods return NULL.
1502 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1504 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1505 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1506 provide any "EXPORT" or "LOW" strength ciphers.
1509 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1510 is by default disabled at build-time. Builds that are not configured with
1511 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1512 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1513 will need to explicitly call either of:
1515 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1517 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1519 as appropriate. Even if either of those is used, or the application
1520 explicitly uses the version-specific SSLv2_method() or its client and
1521 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1522 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1523 ciphers, and SSLv2 56-bit DES are no longer available.
1527 *) Fix a double-free in DSA code
1529 A double free bug was discovered when OpenSSL parses malformed DSA private
1530 keys and could lead to a DoS attack or memory corruption for applications
1531 that receive DSA private keys from untrusted sources. This scenario is
1534 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1539 *) Disable SRP fake user seed to address a server memory leak.
1541 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1543 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1544 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1545 was changed to ignore the "fake user" SRP seed, even if the seed
1548 Users should use SRP_VBASE_get1_by_user instead. Note that in
1549 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1550 also that even though configuring the SRP seed attempts to hide
1551 invalid usernames by continuing the handshake with fake
1552 credentials, this behaviour is not constant time and no strong
1553 guarantees are made that the handshake is indistinguishable from
1554 that of a valid user.
1558 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1560 In the BN_hex2bn function the number of hex digits is calculated using an
1561 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1562 large values of |i| this can result in |bn_expand| not allocating any
1563 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1564 field as NULL leading to a subsequent NULL ptr deref. For very large values
1565 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1566 In this case memory is allocated to the internal BIGNUM data field, but it
1567 is insufficiently sized leading to heap corruption. A similar issue exists
1568 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1569 is ever called by user applications with very large untrusted hex/dec data.
1570 This is anticipated to be a rare occurrence.
1572 All OpenSSL internal usage of these functions use data that is not expected
1573 to be untrusted, e.g. config file data or application command line
1574 arguments. If user developed applications generate config file data based
1575 on untrusted data then it is possible that this could also lead to security
1576 consequences. This is also anticipated to be rare.
1578 This issue was reported to OpenSSL by Guido Vranken.
1582 *) Fix memory issues in BIO_*printf functions
1584 The internal |fmtstr| function used in processing a "%s" format string in
1585 the BIO_*printf functions could overflow while calculating the length of a
1586 string and cause an OOB read when printing very long strings.
1588 Additionally the internal |doapr_outch| function can attempt to write to an
1589 OOB memory location (at an offset from the NULL pointer) in the event of a
1590 memory allocation failure. In 1.0.2 and below this could be caused where
1591 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1592 could be in processing a very long "%s" format string. Memory leaks can
1595 The first issue may mask the second issue dependent on compiler behaviour.
1596 These problems could enable attacks where large amounts of untrusted data
1597 is passed to the BIO_*printf functions. If applications use these functions
1598 in this way then they could be vulnerable. OpenSSL itself uses these
1599 functions when printing out human-readable dumps of ASN.1 data. Therefore
1600 applications that print this data could be vulnerable if the data is from
1601 untrusted sources. OpenSSL command line applications could also be
1602 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1603 as command line arguments.
1605 Libssl is not considered directly vulnerable. Additionally certificates etc
1606 received via remote connections via libssl are also unlikely to be able to
1607 trigger these issues because of message size limits enforced within libssl.
1609 This issue was reported to OpenSSL Guido Vranken.
1613 *) Side channel attack on modular exponentiation
1615 A side-channel attack was found which makes use of cache-bank conflicts on
1616 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1617 of RSA keys. The ability to exploit this issue is limited as it relies on
1618 an attacker who has control of code in a thread running on the same
1619 hyper-threaded core as the victim thread which is performing decryptions.
1621 This issue was reported to OpenSSL by Yuval Yarom, The University of
1622 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1623 Nadia Heninger, University of Pennsylvania with more information at
1624 http://cachebleed.info.
1628 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1629 if no keysize is specified with default_bits. This fixes an
1630 omission in an earlier change that changed all RSA/DSA key generation
1631 apps to use 2048 bits by default.
1634 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1635 *) DH small subgroups
1637 Historically OpenSSL only ever generated DH parameters based on "safe"
1638 primes. More recently (in version 1.0.2) support was provided for
1639 generating X9.42 style parameter files such as those required for RFC 5114
1640 support. The primes used in such files may not be "safe". Where an
1641 application is using DH configured with parameters based on primes that are
1642 not "safe" then an attacker could use this fact to find a peer's private
1643 DH exponent. This attack requires that the attacker complete multiple
1644 handshakes in which the peer uses the same private DH exponent. For example
1645 this could be used to discover a TLS server's private DH exponent if it's
1646 reusing the private DH exponent or it's using a static DH ciphersuite.
1648 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1649 TLS. It is not on by default. If the option is not set then the server
1650 reuses the same private DH exponent for the life of the server process and
1651 would be vulnerable to this attack. It is believed that many popular
1652 applications do set this option and would therefore not be at risk.
1654 The fix for this issue adds an additional check where a "q" parameter is
1655 available (as is the case in X9.42 based parameters). This detects the
1656 only known attack, and is the only possible defense for static DH
1657 ciphersuites. This could have some performance impact.
1659 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1660 default and cannot be disabled. This could have some performance impact.
1662 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1666 *) SSLv2 doesn't block disabled ciphers
1668 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1669 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1670 been disabled, provided that the SSLv2 protocol was not also disabled via
1673 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1674 and Sebastian Schinzel.
1678 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1680 *) BN_mod_exp may produce incorrect results on x86_64
1682 There is a carry propagating bug in the x86_64 Montgomery squaring
1683 procedure. No EC algorithms are affected. Analysis suggests that attacks
1684 against RSA and DSA as a result of this defect would be very difficult to
1685 perform and are not believed likely. Attacks against DH are considered just
1686 feasible (although very difficult) because most of the work necessary to
1687 deduce information about a private key may be performed offline. The amount
1688 of resources required for such an attack would be very significant and
1689 likely only accessible to a limited number of attackers. An attacker would
1690 additionally need online access to an unpatched system using the target
1691 private key in a scenario with persistent DH parameters and a private
1692 key that is shared between multiple clients. For example this can occur by
1693 default in OpenSSL DHE based SSL/TLS ciphersuites.
1695 This issue was reported to OpenSSL by Hanno Böck.
1699 *) Certificate verify crash with missing PSS parameter
1701 The signature verification routines will crash with a NULL pointer
1702 dereference if presented with an ASN.1 signature using the RSA PSS
1703 algorithm and absent mask generation function parameter. Since these
1704 routines are used to verify certificate signature algorithms this can be
1705 used to crash any certificate verification operation and exploited in a
1706 DoS attack. Any application which performs certificate verification is
1707 vulnerable including OpenSSL clients and servers which enable client
1710 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1714 *) X509_ATTRIBUTE memory leak
1716 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1717 memory. This structure is used by the PKCS#7 and CMS routines so any
1718 application which reads PKCS#7 or CMS data from untrusted sources is
1719 affected. SSL/TLS is not affected.
1721 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1726 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1727 This changes the decoding behaviour for some invalid messages,
1728 though the change is mostly in the more lenient direction, and
1729 legacy behaviour is preserved as much as possible.
1732 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1734 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1736 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1738 *) Alternate chains certificate forgery
1740 During certificate verification, OpenSSL will attempt to find an
1741 alternative certificate chain if the first attempt to build such a chain
1742 fails. An error in the implementation of this logic can mean that an
1743 attacker could cause certain checks on untrusted certificates to be
1744 bypassed, such as the CA flag, enabling them to use a valid leaf
1745 certificate to act as a CA and "issue" an invalid certificate.
1747 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1751 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1753 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1754 incompatibility in the handling of HMAC. The previous ABI has now been
1758 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1760 *) Malformed ECParameters causes infinite loop
1762 When processing an ECParameters structure OpenSSL enters an infinite loop
1763 if the curve specified is over a specially malformed binary polynomial
1766 This can be used to perform denial of service against any
1767 system which processes public keys, certificate requests or
1768 certificates. This includes TLS clients and TLS servers with
1769 client authentication enabled.
1771 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1775 *) Exploitable out-of-bounds read in X509_cmp_time
1777 X509_cmp_time does not properly check the length of the ASN1_TIME
1778 string and can read a few bytes out of bounds. In addition,
1779 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1782 An attacker can use this to craft malformed certificates and CRLs of
1783 various sizes and potentially cause a segmentation fault, resulting in
1784 a DoS on applications that verify certificates or CRLs. TLS clients
1785 that verify CRLs are affected. TLS clients and servers with client
1786 authentication enabled may be affected if they use custom verification
1789 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1790 independently by Hanno Böck.
1794 *) PKCS7 crash with missing EnvelopedContent
1796 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1797 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1798 with missing content and trigger a NULL pointer dereference on parsing.
1800 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1801 structures from untrusted sources are affected. OpenSSL clients and
1802 servers are not affected.
1804 This issue was reported to OpenSSL by Michal Zalewski (Google).
1808 *) CMS verify infinite loop with unknown hash function
1810 When verifying a signedData message the CMS code can enter an infinite loop
1811 if presented with an unknown hash function OID. This can be used to perform
1812 denial of service against any system which verifies signedData messages using
1814 This issue was reported to OpenSSL by Johannes Bauer.
1818 *) Race condition handling NewSessionTicket
1820 If a NewSessionTicket is received by a multi-threaded client when attempting to
1821 reuse a previous ticket then a race condition can occur potentially leading to
1822 a double free of the ticket data.
1826 *) Only support 256-bit or stronger elliptic curves with the
1827 'ecdh_auto' setting (server) or by default (client). Of supported
1828 curves, prefer P-256 (both).
1831 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1833 *) ClientHello sigalgs DoS fix
1835 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1836 invalid signature algorithms extension a NULL pointer dereference will
1837 occur. This can be exploited in a DoS attack against the server.
1839 This issue was was reported to OpenSSL by David Ramos of Stanford
1842 [Stephen Henson and Matt Caswell]
1844 *) Multiblock corrupted pointer fix
1846 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1847 feature only applies on 64 bit x86 architecture platforms that support AES
1848 NI instructions. A defect in the implementation of "multiblock" can cause
1849 OpenSSL's internal write buffer to become incorrectly set to NULL when
1850 using non-blocking IO. Typically, when the user application is using a
1851 socket BIO for writing, this will only result in a failed connection.
1852 However if some other BIO is used then it is likely that a segmentation
1853 fault will be triggered, thus enabling a potential DoS attack.
1855 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1859 *) Segmentation fault in DTLSv1_listen fix
1861 The DTLSv1_listen function is intended to be stateless and processes the
1862 initial ClientHello from many peers. It is common for user code to loop
1863 over the call to DTLSv1_listen until a valid ClientHello is received with
1864 an associated cookie. A defect in the implementation of DTLSv1_listen means
1865 that state is preserved in the SSL object from one invocation to the next
1866 that can lead to a segmentation fault. Errors processing the initial
1867 ClientHello can trigger this scenario. An example of such an error could be
1868 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1871 This issue was reported to OpenSSL by Per Allansson.
1875 *) Segmentation fault in ASN1_TYPE_cmp fix
1877 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1878 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1879 certificate signature algorithm consistency this can be used to crash any
1880 certificate verification operation and exploited in a DoS attack. Any
1881 application which performs certificate verification is vulnerable including
1882 OpenSSL clients and servers which enable client authentication.
1886 *) Segmentation fault for invalid PSS parameters fix
1888 The signature verification routines will crash with a NULL pointer
1889 dereference if presented with an ASN.1 signature using the RSA PSS
1890 algorithm and invalid parameters. Since these routines are used to verify
1891 certificate signature algorithms this can be used to crash any
1892 certificate verification operation and exploited in a DoS attack. Any
1893 application which performs certificate verification is vulnerable including
1894 OpenSSL clients and servers which enable client authentication.
1896 This issue was was reported to OpenSSL by Brian Carpenter.
1900 *) ASN.1 structure reuse memory corruption fix
1902 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1903 memory corruption via an invalid write. Such reuse is and has been
1904 strongly discouraged and is believed to be rare.
1906 Applications that parse structures containing CHOICE or ANY DEFINED BY
1907 components may be affected. Certificate parsing (d2i_X509 and related
1908 functions) are however not affected. OpenSSL clients and servers are
1913 *) PKCS7 NULL pointer dereferences fix
1915 The PKCS#7 parsing code does not handle missing outer ContentInfo
1916 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1917 missing content and trigger a NULL pointer dereference on parsing.
1919 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1920 otherwise parse PKCS#7 structures from untrusted sources are
1921 affected. OpenSSL clients and servers are not affected.
1923 This issue was reported to OpenSSL by Michal Zalewski (Google).
1927 *) DoS via reachable assert in SSLv2 servers fix
1929 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1930 servers that both support SSLv2 and enable export cipher suites by sending
1931 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1933 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1934 (OpenSSL development team).
1938 *) Empty CKE with client auth and DHE fix
1940 If client auth is used then a server can seg fault in the event of a DHE
1941 ciphersuite being selected and a zero length ClientKeyExchange message
1942 being sent by the client. This could be exploited in a DoS attack.
1946 *) Handshake with unseeded PRNG fix
1948 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1949 with an unseeded PRNG. The conditions are:
1950 - The client is on a platform where the PRNG has not been seeded
1951 automatically, and the user has not seeded manually
1952 - A protocol specific client method version has been used (i.e. not
1953 SSL_client_methodv23)
1954 - A ciphersuite is used that does not require additional random data from
1955 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1957 If the handshake succeeds then the client random that has been used will
1958 have been generated from a PRNG with insufficient entropy and therefore the
1959 output may be predictable.
1961 For example using the following command with an unseeded openssl will
1962 succeed on an unpatched platform:
1964 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1968 *) Use After Free following d2i_ECPrivatekey error fix
1970 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1971 could cause a use after free condition. This, in turn, could cause a double
1972 free in several private key parsing functions (such as d2i_PrivateKey
1973 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1974 for applications that receive EC private keys from untrusted
1975 sources. This scenario is considered rare.
1977 This issue was discovered by the BoringSSL project and fixed in their
1982 *) X509_to_X509_REQ NULL pointer deref fix
1984 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1985 the certificate key is invalid. This function is rarely used in practice.
1987 This issue was discovered by Brian Carpenter.
1991 *) Removed the export ciphers from the DEFAULT ciphers
1994 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1996 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1997 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1998 So far those who have to target multiple platforms would compromise
1999 and argue that binary targeting say ARMv5 would still execute on
2000 ARMv8. "Universal" build resolves this compromise by providing
2001 near-optimal performance even on newer platforms.
2004 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2005 (other platforms pending).
2006 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2008 *) Add support for the SignedCertificateTimestampList certificate and
2009 OCSP response extensions from RFC6962.
2012 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2013 for corner cases. (Certain input points at infinity could lead to
2014 bogus results, with non-infinity inputs mapped to infinity too.)
2017 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2018 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2019 common cases are optimized and there still is room for further
2020 improvements. Vector Permutation AES for Altivec is also added.
2023 *) Add support for little-endian ppc64 Linux target.
2024 [Marcelo Cerri (IBM)]
2026 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2027 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2028 are optimized and there still is room for further improvements.
2029 Both 32- and 64-bit modes are supported.
2030 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2032 *) Improved ARMv7 NEON support.
2035 *) Support for SPARC Architecture 2011 crypto extensions, first
2036 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2037 SHA256/512, MD5, GHASH and modular exponentiation.
2038 [Andy Polyakov, David Miller]
2040 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2042 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2044 *) Support for new and upcoming Intel processors, including AVX2,
2045 BMI and SHA ISA extensions. This includes additional "stitched"
2046 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2049 This work was sponsored by Intel Corp.
2052 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2053 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2054 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2057 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2058 this fixes a limitation in previous versions of OpenSSL.
2061 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2062 MGF1 digest and OAEP label.
2065 *) Add EVP support for key wrapping algorithms, to avoid problems with
2066 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2067 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2068 algorithms and include tests cases.
2071 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2073 [Douglas E. Engert, Steve Henson]
2075 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2076 difference in days and seconds between two tm or ASN1_TIME structures.
2079 *) Add -rev test option to s_server to just reverse order of characters
2080 received by client and send back to server. Also prints an abbreviated
2081 summary of the connection parameters.
2084 *) New option -brief for s_client and s_server to print out a brief summary
2085 of connection parameters.
2088 *) Add callbacks for arbitrary TLS extensions.
2089 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2091 *) New option -crl_download in several openssl utilities to download CRLs
2092 from CRLDP extension in certificates.
2095 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2098 *) New function X509_CRL_diff to generate a delta CRL from the difference
2099 of two full CRLs. Add support to "crl" utility.
2102 *) New functions to set lookup_crls function and to retrieve
2103 X509_STORE from X509_STORE_CTX.
2106 *) Print out deprecated issuer and subject unique ID fields in
2110 *) Extend OCSP I/O functions so they can be used for simple general purpose
2111 HTTP as well as OCSP. New wrapper function which can be used to download
2112 CRLs using the OCSP API.
2115 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2118 *) SSL_CONF* functions. These provide a common framework for application
2119 configuration using configuration files or command lines.
2122 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2123 message callback and prints the results. Needs compile time option
2124 "enable-ssl-trace". New options to s_client and s_server to enable
2128 *) New ctrl and macro to retrieve supported points extensions.
2129 Print out extension in s_server and s_client.
2132 *) New functions to retrieve certificate signature and signature
2136 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2140 *) New Suite B modes for TLS code. These use and enforce the requirements
2141 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2142 only use Suite B curves. The Suite B modes can be set by using the
2143 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2146 *) New chain verification flags for Suite B levels of security. Check
2147 algorithms are acceptable when flags are set in X509_verify_cert.
2150 *) Make tls1_check_chain return a set of flags indicating checks passed
2151 by a certificate chain. Add additional tests to handle client
2152 certificates: checks for matching certificate type and issuer name
2156 *) If an attempt is made to use a signature algorithm not in the peer
2157 preference list abort the handshake. If client has no suitable
2158 signature algorithms in response to a certificate request do not
2159 use the certificate.
2162 *) If server EC tmp key is not in client preference list abort handshake.
2165 *) Add support for certificate stores in CERT structure. This makes it
2166 possible to have different stores per SSL structure or one store in
2167 the parent SSL_CTX. Include distinct stores for certificate chain
2168 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2169 to build and store a certificate chain in CERT structure: returning
2170 an error if the chain cannot be built: this will allow applications
2171 to test if a chain is correctly configured.
2173 Note: if the CERT based stores are not set then the parent SSL_CTX
2174 store is used to retain compatibility with existing behaviour.
2178 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2179 mask based on the current session, check mask when sending client
2180 hello and checking the requested ciphersuite.
2183 *) New ctrls to retrieve and set certificate types in a certificate
2184 request message. Print out received values in s_client. If certificate
2185 types is not set with custom values set sensible values based on
2186 supported signature algorithms.
2189 *) Support for distinct client and server supported signature algorithms.
2192 *) Add certificate callback. If set this is called whenever a certificate
2193 is required by client or server. An application can decide which
2194 certificate chain to present based on arbitrary criteria: for example
2195 supported signature algorithms. Add very simple example to s_server.
2196 This fixes many of the problems and restrictions of the existing client
2197 certificate callback: for example you can now clear an existing
2198 certificate and specify the whole chain.
2201 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2202 the certificate can be used for (if anything). Set valid_flags field
2203 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2204 to have similar checks in it.
2206 Add new "cert_flags" field to CERT structure and include a "strict mode".
2207 This enforces some TLS certificate requirements (such as only permitting
2208 certificate signature algorithms contained in the supported algorithms
2209 extension) which some implementations ignore: this option should be used
2210 with caution as it could cause interoperability issues.
2213 *) Update and tidy signature algorithm extension processing. Work out
2214 shared signature algorithms based on preferences and peer algorithms
2215 and print them out in s_client and s_server. Abort handshake if no
2216 shared signature algorithms.
2219 *) Add new functions to allow customised supported signature algorithms
2220 for SSL and SSL_CTX structures. Add options to s_client and s_server
2224 *) New function SSL_certs_clear() to delete all references to certificates
2225 from an SSL structure. Before this once a certificate had been added
2226 it couldn't be removed.
2229 *) Integrate hostname, email address and IP address checking with certificate
2230 verification. New verify options supporting checking in openssl utility.
2233 *) Fixes and wildcard matching support to hostname and email checking
2234 functions. Add manual page.
2235 [Florian Weimer (Red Hat Product Security Team)]
2237 *) New functions to check a hostname email or IP address against a
2238 certificate. Add options x509 utility to print results of checks against
2242 *) Fix OCSP checking.
2243 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2245 *) Initial experimental support for explicitly trusted non-root CAs.
2246 OpenSSL still tries to build a complete chain to a root but if an
2247 intermediate CA has a trust setting included that is used. The first
2248 setting is used: whether to trust (e.g., -addtrust option to the x509
2252 *) Add -trusted_first option which attempts to find certificates in the
2253 trusted store even if an untrusted chain is also supplied.
2256 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2257 platform support for Linux and Android.
2260 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2263 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2264 When in FIPS mode the approved implementations are used as normal,
2265 when not in FIPS mode the internal unapproved versions are used instead.
2266 This means that the FIPS capable OpenSSL isn't forced to use the
2267 (often lower performance) FIPS implementations outside FIPS mode.
2270 *) Transparently support X9.42 DH parameters when calling
2271 PEM_read_bio_DHparameters. This means existing applications can handle
2272 the new parameter format automatically.
2275 *) Initial experimental support for X9.42 DH parameter format: mainly
2276 to support use of 'q' parameter for RFC5114 parameters.
2279 *) Add DH parameters from RFC5114 including test data to dhtest.
2282 *) Support for automatic EC temporary key parameter selection. If enabled
2283 the most preferred EC parameters are automatically used instead of
2284 hardcoded fixed parameters. Now a server just has to call:
2285 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2286 support ECDH and use the most appropriate parameters.
2289 *) Enhance and tidy EC curve and point format TLS extension code. Use
2290 static structures instead of allocation if default values are used.
2291 New ctrls to set curves we wish to support and to retrieve shared curves.
2292 Print out shared curves in s_server. New options to s_server and s_client
2293 to set list of supported curves.
2296 *) New ctrls to retrieve supported signature algorithms and
2297 supported curve values as an array of NIDs. Extend openssl utility
2298 to print out received values.
2301 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2302 between NIDs and the more common NIST names such as "P-256". Enhance
2303 ecparam utility and ECC method to recognise the NIST names for curves.
2306 *) Enhance SSL/TLS certificate chain handling to support different
2307 chains for each certificate instead of one chain in the parent SSL_CTX.
2310 *) Support for fixed DH ciphersuite client authentication: where both
2311 server and client use DH certificates with common parameters.
2314 *) Support for fixed DH ciphersuites: those requiring DH server
2318 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2320 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2321 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2322 X509_CINF_get_signature were reverted post internal team review.
2324 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2326 *) Build fixes for the Windows and OpenVMS platforms
2327 [Matt Caswell and Richard Levitte]
2329 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2331 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2332 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2333 dereference. This could lead to a Denial Of Service attack. Thanks to
2334 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2338 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2339 dtls1_buffer_record function under certain conditions. In particular this
2340 could occur if an attacker sent repeated DTLS records with the same
2341 sequence number but for the next epoch. The memory leak could be exploited
2342 by an attacker in a Denial of Service attack through memory exhaustion.
2343 Thanks to Chris Mueller for reporting this issue.
2347 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2348 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2349 method would be set to NULL which could later result in a NULL pointer
2350 dereference. Thanks to Frank Schmirler for reporting this issue.
2354 *) Abort handshake if server key exchange message is omitted for ephemeral
2357 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2358 reporting this issue.
2362 *) Remove non-export ephemeral RSA code on client and server. This code
2363 violated the TLS standard by allowing the use of temporary RSA keys in
2364 non-export ciphersuites and could be used by a server to effectively
2365 downgrade the RSA key length used to a value smaller than the server
2366 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2367 INRIA or reporting this issue.
2371 *) Fixed issue where DH client certificates are accepted without verification.
2372 An OpenSSL server will accept a DH certificate for client authentication
2373 without the certificate verify message. This effectively allows a client to
2374 authenticate without the use of a private key. This only affects servers
2375 which trust a client certificate authority which issues certificates
2376 containing DH keys: these are extremely rare and hardly ever encountered.
2377 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2382 *) Ensure that the session ID context of an SSL is updated when its
2383 SSL_CTX is updated via SSL_set_SSL_CTX.
2385 The session ID context is typically set from the parent SSL_CTX,
2386 and can vary with the CTX.
2389 *) Fix various certificate fingerprint issues.
2391 By using non-DER or invalid encodings outside the signed portion of a
2392 certificate the fingerprint can be changed without breaking the signature.
2393 Although no details of the signed portion of the certificate can be changed
2394 this can cause problems with some applications: e.g. those using the
2395 certificate fingerprint for blacklists.
2397 1. Reject signatures with non zero unused bits.
2399 If the BIT STRING containing the signature has non zero unused bits reject
2400 the signature. All current signature algorithms require zero unused bits.
2402 2. Check certificate algorithm consistency.
2404 Check the AlgorithmIdentifier inside TBS matches the one in the
2405 certificate signature. NB: this will result in signature failure
2406 errors for some broken certificates.
2408 Thanks to Konrad Kraszewski from Google for reporting this issue.
2410 3. Check DSA/ECDSA signatures use DER.
2412 Re-encode DSA/ECDSA signatures and compare with the original received
2413 signature. Return an error if there is a mismatch.
2415 This will reject various cases including garbage after signature
2416 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2417 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2418 (negative or with leading zeroes).
2420 Further analysis was conducted and fixes were developed by Stephen Henson
2421 of the OpenSSL core team.
2426 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2427 results on some platforms, including x86_64. This bug occurs at random
2428 with a very low probability, and is not known to be exploitable in any
2429 way, though its exact impact is difficult to determine. Thanks to Pieter
2430 Wuille (Blockstream) who reported this issue and also suggested an initial
2431 fix. Further analysis was conducted by the OpenSSL development team and
2432 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2433 the OpenSSL core team.
2437 *) Do not resume sessions on the server if the negotiated protocol
2438 version does not match the session's version. Resuming with a different
2439 version, while not strictly forbidden by the RFC, is of questionable
2440 sanity and breaks all known clients.
2441 [David Benjamin, Emilia Käsper]
2443 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2444 early CCS messages during renegotiation. (Note that because
2445 renegotiation is encrypted, this early CCS was not exploitable.)
2448 *) Tighten client-side session ticket handling during renegotiation:
2449 ensure that the client only accepts a session ticket if the server sends
2450 the extension anew in the ServerHello. Previously, a TLS client would
2451 reuse the old extension state and thus accept a session ticket if one was
2452 announced in the initial ServerHello.
2454 Similarly, ensure that the client requires a session ticket if one
2455 was advertised in the ServerHello. Previously, a TLS client would
2456 ignore a missing NewSessionTicket message.
2459 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2461 *) SRTP Memory Leak.
2463 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2464 sends a carefully crafted handshake message, to cause OpenSSL to fail
2465 to free up to 64k of memory causing a memory leak. This could be
2466 exploited in a Denial Of Service attack. This issue affects OpenSSL
2467 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2468 whether SRTP is used or configured. Implementations of OpenSSL that
2469 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2471 The fix was developed by the OpenSSL team.
2475 *) Session Ticket Memory Leak.
2477 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2478 integrity of that ticket is first verified. In the event of a session
2479 ticket integrity check failing, OpenSSL will fail to free memory
2480 causing a memory leak. By sending a large number of invalid session
2481 tickets an attacker could exploit this issue in a Denial Of Service
2486 *) Build option no-ssl3 is incomplete.
2488 When OpenSSL is configured with "no-ssl3" as a build option, servers
2489 could accept and complete a SSL 3.0 handshake, and clients could be
2490 configured to send them.
2492 [Akamai and the OpenSSL team]
2494 *) Add support for TLS_FALLBACK_SCSV.
2495 Client applications doing fallback retries should call
2496 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2498 [Adam Langley, Bodo Moeller]
2500 *) Add additional DigestInfo checks.
2502 Re-encode DigestInto in DER and check against the original when
2503 verifying RSA signature: this will reject any improperly encoded
2504 DigestInfo structures.
2506 Note: this is a precautionary measure and no attacks are currently known.
2510 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2512 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2513 SRP code can be overrun an internal buffer. Add sanity check that
2514 g, A, B < N to SRP code.
2516 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2517 Group for discovering this issue.
2521 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2522 TLS 1.0 instead of higher protocol versions when the ClientHello message
2523 is badly fragmented. This allows a man-in-the-middle attacker to force a
2524 downgrade to TLS 1.0 even if both the server and the client support a
2525 higher protocol version, by modifying the client's TLS records.
2527 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2528 researching this issue.
2532 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2533 to a denial of service attack. A malicious server can crash the client
2534 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2535 ciphersuite and sending carefully crafted handshake messages.
2537 Thanks to Felix Gröbert (Google) for discovering and researching this
2542 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2543 to leak memory. This can be exploited through a Denial of Service attack.
2544 Thanks to Adam Langley for discovering and researching this issue.
2548 *) An attacker can force openssl to consume large amounts of memory whilst
2549 processing DTLS handshake messages. This can be exploited through a
2550 Denial of Service attack.
2551 Thanks to Adam Langley for discovering and researching this issue.
2555 *) An attacker can force an error condition which causes openssl to crash
2556 whilst processing DTLS packets due to memory being freed twice. This
2557 can be exploited through a Denial of Service attack.
2558 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2563 *) If a multithreaded client connects to a malicious server using a resumed
2564 session and the server sends an ec point format extension it could write
2565 up to 255 bytes to freed memory.
2567 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2572 *) A malicious server can crash an OpenSSL client with a null pointer
2573 dereference (read) by specifying an SRP ciphersuite even though it was not
2574 properly negotiated with the client. This can be exploited through a
2575 Denial of Service attack.
2577 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2578 discovering and researching this issue.
2582 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2583 X509_name_oneline, X509_name_print_ex et al. to leak some information
2584 from the stack. Applications may be affected if they echo pretty printing
2585 output to the attacker.
2587 Thanks to Ivan Fratric (Google) for discovering this issue.
2589 [Emilia Käsper, and Steve Henson]
2591 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2592 for corner cases. (Certain input points at infinity could lead to
2593 bogus results, with non-infinity inputs mapped to infinity too.)
2596 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2598 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2599 handshake can force the use of weak keying material in OpenSSL
2600 SSL/TLS clients and servers.
2602 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2603 researching this issue. (CVE-2014-0224)
2604 [KIKUCHI Masashi, Steve Henson]
2606 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2607 OpenSSL DTLS client the code can be made to recurse eventually crashing
2610 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2612 [Imre Rad, Steve Henson]
2614 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2615 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2616 client or server. This is potentially exploitable to run arbitrary
2617 code on a vulnerable client or server.
2619 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2620 [Jüri Aedla, Steve Henson]
2622 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2623 are subject to a denial of service attack.
2625 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2626 this issue. (CVE-2014-3470)
2627 [Felix Gröbert, Ivan Fratric, Steve Henson]
2629 *) Harmonize version and its documentation. -f flag is used to display
2631 [mancha <mancha1@zoho.com>]
2633 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2634 in i2d_ECPrivateKey.
2635 [mancha <mancha1@zoho.com>]
2637 *) Fix some double frees. These are not thought to be exploitable.
2638 [mancha <mancha1@zoho.com>]
2640 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2642 *) A missing bounds check in the handling of the TLS heartbeat extension
2643 can be used to reveal up to 64k of memory to a connected client or
2646 Thanks for Neel Mehta of Google Security for discovering this bug and to
2647 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2648 preparing the fix (CVE-2014-0160)
2649 [Adam Langley, Bodo Moeller]
2651 *) Fix for the attack described in the paper "Recovering OpenSSL
2652 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2653 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2654 http://eprint.iacr.org/2014/140
2656 Thanks to Yuval Yarom and Naomi Benger for discovering this
2657 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2658 [Yuval Yarom and Naomi Benger]
2660 *) TLS pad extension: draft-agl-tls-padding-03
2662 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2663 TLS client Hello record length value would otherwise be > 255 and
2664 less that 512 pad with a dummy extension containing zeroes so it
2665 is at least 512 bytes long.
2667 [Adam Langley, Steve Henson]
2669 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2671 *) Fix for TLS record tampering bug. A carefully crafted invalid
2672 handshake could crash OpenSSL with a NULL pointer exception.
2673 Thanks to Anton Johansson for reporting this issues.
2676 *) Keep original DTLS digest and encryption contexts in retransmission
2677 structures so we can use the previous session parameters if they need
2678 to be resent. (CVE-2013-6450)
2681 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2682 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2683 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2684 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2685 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2686 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2687 [Rob Stradling, Adam Langley]
2689 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2691 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2692 supporting platforms or when small records were transferred.
2693 [Andy Polyakov, Steve Henson]
2695 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2697 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2699 This addresses the flaw in CBC record processing discovered by
2700 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2701 at: http://www.isg.rhul.ac.uk/tls/
2703 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2704 Security Group at Royal Holloway, University of London
2705 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2706 Emilia Käsper for the initial patch.
2708 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2710 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2711 ciphersuites which can be exploited in a denial of service attack.
2712 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2713 and detecting this bug and to Wolfgang Ettlinger
2714 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2718 *) Return an error when checking OCSP signatures when key is NULL.
2719 This fixes a DoS attack. (CVE-2013-0166)
2722 *) Make openssl verify return errors.
2723 [Chris Palmer <palmer@google.com> and Ben Laurie]
2725 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2726 the right response is stapled. Also change SSL_get_certificate()
2727 so it returns the certificate actually sent.
2728 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2729 [Rob Stradling <rob.stradling@comodo.com>]
2731 *) Fix possible deadlock when decoding public keys.
2734 *) Don't use TLS 1.0 record version number in initial client hello
2738 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2740 *) Sanity check record length before skipping explicit IV in TLS
2741 1.2, 1.1 and DTLS to fix DoS attack.
2743 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2744 fuzzing as a service testing platform.
2748 *) Initialise tkeylen properly when encrypting CMS messages.
2749 Thanks to Solar Designer of Openwall for reporting this issue.
2752 *) In FIPS mode don't try to use composite ciphers as they are not
2756 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2758 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2759 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2760 mean any application compiled against OpenSSL 1.0.0 headers setting
2761 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2762 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2763 0x10000000L Any application which was previously compiled against
2764 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2765 will need to be recompiled as a result. Letting be results in
2766 inability to disable specifically TLS 1.1 and in client context,
2767 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2770 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2771 disable just protocol X, but all protocols above X *if* there are
2772 protocols *below* X still enabled. In more practical terms it means
2773 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2774 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2775 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2779 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2781 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2782 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2783 in CRYPTO_realloc_clean.
2785 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2786 issue and to Adam Langley <agl@chromium.org> for fixing it.
2788 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2790 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2793 *) Workarounds for some broken servers that "hang" if a client hello
2794 record length exceeds 255 bytes.
2796 1. Do not use record version number > TLS 1.0 in initial client
2797 hello: some (but not all) hanging servers will now work.
2798 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2799 the number of ciphers sent in the client hello. This should be
2800 set to an even number, such as 50, for example by passing:
2801 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2802 Most broken servers should now work.
2803 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2804 TLS 1.2 client support entirely.
2807 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2810 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2812 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2813 STRING form instead of a DigestInfo.
2816 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2817 and the RSA_sign/RSA_verify functions. This was made more apparent when
2818 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2819 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2820 the correct format in RSA_verify so both forms transparently work.
2823 *) Some servers which support TLS 1.0 can choke if we initially indicate
2824 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2825 encrypted premaster secret. As a workaround use the maximum permitted
2826 client version in client hello, this should keep such servers happy
2827 and still work with previous versions of OpenSSL.
2830 *) Add support for TLS/DTLS heartbeats.
2831 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2833 *) Add support for SCTP.
2834 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2836 *) Improved PRNG seeding for VOS.
2837 [Paul Green <Paul.Green@stratus.com>]
2839 *) Extensive assembler packs updates, most notably:
2841 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2842 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2843 - x86_64: bit-sliced AES implementation;
2844 - ARM: NEON support, contemporary platforms optimizations;
2845 - s390x: z196 support;
2846 - *: GHASH and GF(2^m) multiplication implementations;
2850 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2851 (removal of unnecessary code)
2852 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2854 *) Add TLS key material exporter from RFC 5705.
2857 *) Add DTLS-SRTP negotiation from RFC 5764.
2860 *) Add Next Protocol Negotiation,
2861 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2862 disabled with a no-npn flag to config or Configure. Code donated
2864 [Adam Langley <agl@google.com> and Ben Laurie]
2866 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2867 NIST-P256, NIST-P521, with constant-time single point multiplication on
2868 typical inputs. Compiler support for the nonstandard type __uint128_t is
2869 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2870 Code made available under Apache License version 2.0.
2872 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2873 line to include this in your build of OpenSSL, and run "make depend" (or
2874 "make update"). This enables the following EC_METHODs:
2876 EC_GFp_nistp224_method()
2877 EC_GFp_nistp256_method()
2878 EC_GFp_nistp521_method()
2880 EC_GROUP_new_by_curve_name() will automatically use these (while
2881 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2883 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2885 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2886 all platforms. Move ssize_t definition from e_os.h to the public
2887 header file e_os2.h as it now appears in public header file cms.h
2890 *) New -sigopt option to the ca, req and x509 utilities. Additional
2891 signature parameters can be passed using this option and in
2895 *) Add RSA PSS signing function. This will generate and set the
2896 appropriate AlgorithmIdentifiers for PSS based on those in the
2897 corresponding EVP_MD_CTX structure. No application support yet.
2900 *) Support for companion algorithm specific ASN1 signing routines.
2901 New function ASN1_item_sign_ctx() signs a pre-initialised
2902 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2903 the appropriate parameters.
2906 *) Add new algorithm specific ASN1 verification initialisation function
2907 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2908 handling will be the same no matter what EVP_PKEY_METHOD is used.
2909 Add a PSS handler to support verification of PSS signatures: checked
2910 against a number of sample certificates.
2913 *) Add signature printing for PSS. Add PSS OIDs.
2914 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2916 *) Add algorithm specific signature printing. An individual ASN1 method
2917 can now print out signatures instead of the standard hex dump.
2919 More complex signatures (e.g. PSS) can print out more meaningful
2920 information. Include DSA version that prints out the signature
2924 *) Password based recipient info support for CMS library: implementing
2928 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2929 neatly separates the code into cipher and PBE sections and is required
2930 for some algorithms that split PBES2 into separate pieces (such as
2931 password based CMS).
2934 *) Session-handling fixes:
2935 - Fix handling of connections that are resuming with a session ID,
2936 but also support Session Tickets.
2937 - Fix a bug that suppressed issuing of a new ticket if the client
2938 presented a ticket with an expired session.
2939 - Try to set the ticket lifetime hint to something reasonable.
2940 - Make tickets shorter by excluding irrelevant information.
2941 - On the client side, don't ignore renewed tickets.
2942 [Adam Langley, Bodo Moeller (Google)]
2944 *) Fix PSK session representation.
2947 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2949 This work was sponsored by Intel.
2952 *) Add GCM support to TLS library. Some custom code is needed to split
2953 the IV between the fixed (from PRF) and explicit (from TLS record)
2954 portions. This adds all GCM ciphersuites supported by RFC5288 and
2955 RFC5289. Generalise some AES* cipherstrings to include GCM and
2956 add a special AESGCM string for GCM only.
2959 *) Expand range of ctrls for AES GCM. Permit setting invocation
2960 field on decrypt and retrieval of invocation field only on encrypt.
2963 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2964 As required by RFC5289 these ciphersuites cannot be used if for
2965 versions of TLS earlier than 1.2.
2968 *) For FIPS capable OpenSSL interpret a NULL default public key method
2969 as unset and return the appropriate default but do *not* set the default.
2970 This means we can return the appropriate method in applications that
2971 switch between FIPS and non-FIPS modes.
2974 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2975 ENGINE is used then we cannot handle that in the FIPS module so we
2976 keep original code iff non-FIPS operations are allowed.
2979 *) Add -attime option to openssl utilities.
2980 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2982 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2985 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2986 FIPS EC methods unconditionally for now.
2989 *) New build option no-ec2m to disable characteristic 2 code.
2992 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2993 all cases can be covered as some introduce binary incompatibilities.
2996 *) Redirect RSA operations to FIPS module including keygen,
2997 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3000 *) Add similar low level API blocking to ciphers.
3003 *) Low level digest APIs are not approved in FIPS mode: any attempt
3004 to use these will cause a fatal error. Applications that *really* want
3005 to use them can use the private_* version instead.
3008 *) Redirect cipher operations to FIPS module for FIPS builds.
3011 *) Redirect digest operations to FIPS module for FIPS builds.
3014 *) Update build system to add "fips" flag which will link in fipscanister.o
3015 for static and shared library builds embedding a signature if needed.
3018 *) Output TLS supported curves in preference order instead of numerical
3019 order. This is currently hardcoded for the highest order curves first.
3020 This should be configurable so applications can judge speed vs strength.
3023 *) Add TLS v1.2 server support for client authentication.
3026 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3030 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3031 FIPS modules versions.
3034 *) Add TLS v1.2 client side support for client authentication. Keep cache
3035 of handshake records longer as we don't know the hash algorithm to use
3036 until after the certificate request message is received.
3039 *) Initial TLS v1.2 client support. Add a default signature algorithms
3040 extension including all the algorithms we support. Parse new signature
3041 format in client key exchange. Relax some ECC signing restrictions for
3042 TLS v1.2 as indicated in RFC5246.
3045 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3046 to new signature format when needed using client digest preference.
3047 All server ciphersuites should now work correctly in TLS v1.2. No client
3048 support yet and no support for client certificates.
3051 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3052 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3053 ciphersuites. At present only RSA key exchange ciphersuites work with
3054 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3055 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3056 and version checking.
3059 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3060 with this defined it will not be affected by any changes to ssl internal
3061 structures. Add several utility functions to allow openssl application
3062 to work with OPENSSL_NO_SSL_INTERN defined.
3065 *) A long standing patch to add support for SRP from EdelWeb (Peter
3066 Sylvester and Christophe Renou) was integrated.
3067 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3068 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3071 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3074 *) Permit abbreviated handshakes when renegotiating using the function
3075 SSL_renegotiate_abbreviated().
3076 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3078 *) Add call to ENGINE_register_all_complete() to
3079 ENGINE_load_builtin_engines(), so some implementations get used
3080 automatically instead of needing explicit application support.
3083 *) Add support for TLS key exporter as described in RFC5705.
3084 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3086 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3087 a few changes are required:
3089 Add SSL_OP_NO_TLSv1_1 flag.
3090 Add TLSv1_1 methods.
3091 Update version checking logic to handle version 1.1.
3092 Add explicit IV handling (ported from DTLS code).
3093 Add command line options to s_client/s_server.
3096 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3098 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3099 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3100 content decryption and always return the same error. Note: this attack
3101 needs on average 2^20 messages so it only affects automated senders. The
3102 old behaviour can be re-enabled in the CMS code by setting the
3103 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3104 an MMA defence is not necessary.
3105 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3106 this issue. (CVE-2012-0884)
3109 *) Fix CVE-2011-4619: make sure we really are receiving a
3110 client hello before rejecting multiple SGC restarts. Thanks to
3111 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3114 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3116 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3117 Thanks to Antonio Martin, Enterprise Secure Access Research and
3118 Development, Cisco Systems, Inc. for discovering this bug and
3119 preparing a fix. (CVE-2012-0050)
3122 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3124 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3125 of the Vaudenay padding oracle attack on CBC mode encryption
3126 which enables an efficient plaintext recovery attack against
3127 the OpenSSL implementation of DTLS. Their attack exploits timing
3128 differences arising during decryption processing. A research
3129 paper describing this attack can be found at:
3130 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3131 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3132 Security Group at Royal Holloway, University of London
3133 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3134 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3135 for preparing the fix. (CVE-2011-4108)
3136 [Robin Seggelmann, Michael Tuexen]
3138 *) Clear bytes used for block padding of SSL 3.0 records.
3140 [Adam Langley (Google)]
3142 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3143 Kadianakis <desnacked@gmail.com> for discovering this issue and
3144 Adam Langley for preparing the fix. (CVE-2011-4619)
3145 [Adam Langley (Google)]
3147 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3148 [Andrey Kulikov <amdeich@gmail.com>]
3150 *) Prevent malformed RFC3779 data triggering an assertion failure.
3151 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3152 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3153 [Rob Austein <sra@hactrn.net>]
3155 *) Improved PRNG seeding for VOS.
3156 [Paul Green <Paul.Green@stratus.com>]
3158 *) Fix ssl_ciph.c set-up race.
3159 [Adam Langley (Google)]
3161 *) Fix spurious failures in ecdsatest.c.
3162 [Emilia Käsper (Google)]
3164 *) Fix the BIO_f_buffer() implementation (which was mixing different
3165 interpretations of the '..._len' fields).
3166 [Adam Langley (Google)]
3168 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3169 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3170 threads won't reuse the same blinding coefficients.
3172 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3173 lock to call BN_BLINDING_invert_ex, and avoids one use of
3174 BN_BLINDING_update for each BN_BLINDING structure (previously,
3175 the last update always remained unused).
3176 [Emilia Käsper (Google)]
3178 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3179 [Bob Buckholz (Google)]
3181 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3183 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3184 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3185 [Kaspar Brand <ossl@velox.ch>]
3187 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3188 for multi-threaded use of ECDH. (CVE-2011-3210)
3189 [Adam Langley (Google)]
3191 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3194 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3195 signature public key algorithm by using OID xref utilities instead.
3196 Before this you could only use some ECC ciphersuites with SHA1 only.
3199 *) Add protection against ECDSA timing attacks as mentioned in the paper
3200 by Billy Bob Brumley and Nicola Tuveri, see:
3202 http://eprint.iacr.org/2011/232.pdf
3204 [Billy Bob Brumley and Nicola Tuveri]
3206 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3208 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3209 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3211 *) Fix bug in string printing code: if *any* escaping is enabled we must
3212 escape the escape character (backslash) or the resulting string is
3216 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3218 *) Disable code workaround for ancient and obsolete Netscape browsers
3219 and servers: an attacker can use it in a ciphersuite downgrade attack.
3220 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3223 *) Fixed J-PAKE implementation error, originally discovered by
3224 Sebastien Martini, further info and confirmation from Stefan
3225 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3228 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3230 *) Fix extension code to avoid race conditions which can result in a buffer
3231 overrun vulnerability: resumed sessions must not be modified as they can
3232 be shared by multiple threads. CVE-2010-3864
3235 *) Fix WIN32 build system to correctly link an ENGINE directory into
3239 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3241 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3243 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3245 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3247 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3248 context. The operation can be customised via the ctrl mechanism in
3249 case ENGINEs want to include additional functionality.
3252 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3255 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3256 output hashes compatible with older versions of OpenSSL.
3257 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3259 *) Fix compression algorithm handling: if resuming a session use the
3260 compression algorithm of the resumed session instead of determining
3261 it from client hello again. Don't allow server to change algorithm.
3264 *) Add load_crls() function to apps tidying load_certs() too. Add option
3265 to verify utility to allow additional CRLs to be included.
3268 *) Update OCSP request code to permit adding custom headers to the request:
3269 some responders need this.
3272 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3274 [Julia Lawall <julia@diku.dk>]
3276 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3277 needlessly dereferenced structures, used obsolete functions and
3278 didn't handle all updated verify codes correctly.
3281 *) Disable MD2 in the default configuration.
3284 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3285 indicate the initial BIO being pushed or popped. This makes it possible
3286 to determine whether the BIO is the one explicitly called or as a result
3287 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3288 it handles reference counts correctly and doesn't zero out the I/O bio
3289 when it is not being explicitly popped. WARNING: applications which
3290 included workarounds for the old buggy behaviour will need to be modified
3291 or they could free up already freed BIOs.
3294 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3295 renaming to all platforms (within the 0.9.8 branch, this was
3296 done conditionally on Netware platforms to avoid a name clash).
3297 [Guenter <lists@gknw.net>]
3299 *) Add ECDHE and PSK support to DTLS.
3300 [Michael Tuexen <tuexen@fh-muenster.de>]
3302 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3306 *) Add "missing" function EVP_MD_flags() (without this the only way to
3307 retrieve a digest flags is by accessing the structure directly. Update
3308 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3309 or cipher is registered as in the "from" argument. Print out all
3310 registered digests in the dgst usage message instead of manually
3311 attempting to work them out.
3314 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3315 this allows the use of compression and extensions. Change default cipher
3316 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3317 by default unless an application cipher string requests it.
3320 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3321 key ids to find matching certificates and keys but some PKCS#12 files
3322 don't follow the (somewhat unwritten) rules and this strategy fails.
3323 Now just gather all certificates together and the first private key
3324 then look for the first certificate that matches the key.
3327 *) Support use of registered digest and cipher names for dgst and cipher
3328 commands instead of having to add each one as a special case. So now
3335 openssl dgst -sha256 foo
3337 and this works for ENGINE based algorithms too.
3341 *) Update Gost ENGINE to support parameter files.
3342 [Victor B. Wagner <vitus@cryptocom.ru>]
3344 *) Support GeneralizedTime in ca utility.
3345 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3347 *) Enhance the hash format used for certificate directory links. The new
3348 form uses the canonical encoding (meaning equivalent names will work
3349 even if they aren't identical) and uses SHA1 instead of MD5. This form
3350 is incompatible with the older format and as a result c_rehash should
3351 be used to rebuild symbolic links.
3354 *) Make PKCS#8 the default write format for private keys, replacing the
3355 traditional format. This form is standardised, more secure and doesn't
3356 include an implicit MD5 dependency.
3359 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3360 committed to OpenSSL should pass this lot as a minimum.
3363 *) Add session ticket override functionality for use by EAP-FAST.
3364 [Jouni Malinen <j@w1.fi>]
3366 *) Modify HMAC functions to return a value. Since these can be implemented
3367 in an ENGINE errors can occur.
3370 *) Type-checked OBJ_bsearch_ex.
3373 *) Type-checked OBJ_bsearch. Also some constification necessitated
3374 by type-checking. Still to come: TXT_DB, bsearch(?),
3375 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3379 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3380 seconds to a tm structure directly, instead of going through OS
3381 specific date routines. This avoids any issues with OS routines such
3382 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3383 and X509_time_adj_ex() to cover the extended range. The existing
3384 X509_time_adj() is still usable and will no longer have any date issues.
3387 *) Delta CRL support. New use deltas option which will attempt to locate
3388 and search any appropriate delta CRLs available.
3390 This work was sponsored by Google.
3393 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3394 code and add additional score elements. Validate alternate CRL paths
3395 as part of the CRL checking and indicate a new error "CRL path validation
3396 error" in this case. Applications wanting additional details can use
3397 the verify callback and check the new "parent" field. If this is not
3398 NULL CRL path validation is taking place. Existing applications won't
3399 see this because it requires extended CRL support which is off by
3402 This work was sponsored by Google.
3405 *) Support for freshest CRL extension.
3407 This work was sponsored by Google.
3410 *) Initial indirect CRL support. Currently only supported in the CRLs
3411 passed directly and not via lookup. Process certificate issuer
3412 CRL entry extension and lookup CRL entries by bother issuer name
3413 and serial number. Check and process CRL issuer entry in IDP extension.
3415 This work was sponsored by Google.
3418 *) Add support for distinct certificate and CRL paths. The CRL issuer
3419 certificate is validated separately in this case. Only enabled if
3420 an extended CRL support flag is set: this flag will enable additional
3421 CRL functionality in future.
3423 This work was sponsored by Google.
3426 *) Add support for policy mappings extension.
3428 This work was sponsored by Google.
3431 *) Fixes to pathlength constraint, self issued certificate handling,
3432 policy processing to align with RFC3280 and PKITS tests.
3434 This work was sponsored by Google.
3437 *) Support for name constraints certificate extension. DN, email, DNS
3438 and URI types are currently supported.
3440 This work was sponsored by Google.
3443 *) To cater for systems that provide a pointer-based thread ID rather
3444 than numeric, deprecate the current numeric thread ID mechanism and
3445 replace it with a structure and associated callback type. This
3446 mechanism allows a numeric "hash" to be extracted from a thread ID in
3447 either case, and on platforms where pointers are larger than 'long',
3448 mixing is done to help ensure the numeric 'hash' is usable even if it
3449 can't be guaranteed unique. The default mechanism is to use "&errno"
3450 as a pointer-based thread ID to distinguish between threads.
3452 Applications that want to provide their own thread IDs should now use
3453 CRYPTO_THREADID_set_callback() to register a callback that will call
3454 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3456 Note that ERR_remove_state() is now deprecated, because it is tied
3457 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3458 to free the current thread's error state should be replaced by
3459 ERR_remove_thread_state(NULL).
3461 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3462 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3463 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3464 application was previously providing a numeric thread callback that
3465 was inappropriate for distinguishing threads, then uniqueness might
3466 have been obtained with &errno that happened immediately in the
3467 intermediate development versions of OpenSSL; this is no longer the
3468 case, the numeric thread callback will now override the automatic use
3470 [Geoff Thorpe, with help from Bodo Moeller]
3472 *) Initial support for different CRL issuing certificates. This covers a
3473 simple case where the self issued certificates in the chain exist and
3474 the real CRL issuer is higher in the existing chain.
3476 This work was sponsored by Google.
3479 *) Removed effectively defunct crypto/store from the build.
3482 *) Revamp of STACK to provide stronger type-checking. Still to come:
3483 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3484 ASN1_STRING, CONF_VALUE.
3487 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3488 RAM on SSL connections. This option can save about 34k per idle SSL.
3491 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3492 STACK, TXT_DB, bsearch, qsort.
3495 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3496 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3497 support for data, signedData, compressedData, digestedData and
3498 encryptedData, envelopedData types included. Scripts to check against
3499 RFC4134 examples draft and interop and consistency checks of many
3500 content types and variants.
3503 *) Add options to enc utility to support use of zlib compression BIO.
3506 *) Extend mk1mf to support importing of options and assembly language
3507 files from Configure script, currently only included in VC-WIN32.
3508 The assembly language rules can now optionally generate the source
3509 files from the associated perl scripts.
3512 *) Implement remaining functionality needed to support GOST ciphersuites.
3513 Interop testing has been performed using CryptoPro implementations.
3514 [Victor B. Wagner <vitus@cryptocom.ru>]
3516 *) s390x assembler pack.
3519 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3523 *) Implement Opaque PRF Input TLS extension as specified in
3524 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3525 official specification yet and no extension type assignment by
3526 IANA exists, this extension (for now) will have to be explicitly
3527 enabled when building OpenSSL by providing the extension number
3528 to use. For example, specify an option
3530 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3532 to the "config" or "Configure" script to enable the extension,
3533 assuming extension number 0x9527 (which is a completely arbitrary
3534 and unofficial assignment based on the MD5 hash of the Internet
3535 Draft). Note that by doing so, you potentially lose
3536 interoperability with other TLS implementations since these might
3537 be using the same extension number for other purposes.
3539 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3540 opaque PRF input value to use in the handshake. This will create
3541 an interal copy of the length-'len' string at 'src', and will
3542 return non-zero for success.
3544 To get more control and flexibility, provide a callback function
3547 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3548 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3552 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3555 Callback function 'cb' will be called in handshakes, and is
3556 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3557 Argument 'arg' is for application purposes (the value as given to
3558 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3559 be provided to the callback function). The callback function
3560 has to return non-zero to report success: usually 1 to use opaque
3561 PRF input just if possible, or 2 to enforce use of the opaque PRF
3562 input. In the latter case, the library will abort the handshake
3563 if opaque PRF input is not successfully negotiated.
3565 Arguments 'peerinput' and 'len' given to the callback function
3566 will always be NULL and 0 in the case of a client. A server will
3567 see the client's opaque PRF input through these variables if
3568 available (NULL and 0 otherwise). Note that if the server
3569 provides an opaque PRF input, the length must be the same as the
3570 length of the client's opaque PRF input.
3572 Note that the callback function will only be called when creating
3573 a new session (session resumption can resume whatever was
3574 previously negotiated), and will not be called in SSL 2.0
3575 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3576 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3577 for applications that need to enforce opaque PRF input.
3581 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3584 [Victor B. Wagner <vitus@cryptocom.ru>]
3586 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3587 RFC4507bis. The encrypted ticket format is an encrypted encoded
3588 SSL_SESSION structure, that way new session features are automatically
3591 If a client application caches session in an SSL_SESSION structure
3592 support is transparent because tickets are now stored in the encoded
3595 The SSL_CTX structure automatically generates keys for ticket
3596 protection in servers so again support should be possible
3597 with no application modification.
3599 If a client or server wishes to disable RFC4507 support then the option
3600 SSL_OP_NO_TICKET can be set.
3602 Add a TLS extension debugging callback to allow the contents of any client
3603 or server extensions to be examined.
3605 This work was sponsored by Google.
3608 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3609 OpenSSL should now compile cleanly on gcc 4.2
3610 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3612 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3613 support including streaming MAC support: this is required for GOST
3614 ciphersuite support.
3615 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3617 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3618 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3619 to output in BER and PEM format.
3622 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3623 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3624 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3625 ENGINE support for HMAC keys which are unextractable. New -mac and
3626 -macopt options to dgst utility.
3629 *) New option -sigopt to dgst utility. Update dgst to use
3630 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3631 alternative signing parameters such as X9.31 or PSS in the dgst
3635 *) Change ssl_cipher_apply_rule(), the internal function that does
3636 the work each time a ciphersuite string requests enabling
3637 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3638 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3639 the order of disabled ciphersuites such that those ciphersuites
3640 that most recently went from enabled to disabled not only stay
3641 in order with respect to each other, but also have higher priority
3642 than other disabled ciphersuites the next time ciphersuites are
3645 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3646 the same ciphersuites as with "HIGH" alone, but in a specific
3647 order where the PSK ciphersuites come first (since they are the
3648 most recently disabled ciphersuites when "HIGH" is parsed).
3650 Also, change ssl_create_cipher_list() (using this new
3651 funcionality) such that between otherwise identical
3652 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3656 *) Change ssl_create_cipher_list() so that it automatically
3657 arranges the ciphersuites in reasonable order before starting
3658 to process the rule string. Thus, the definition for "DEFAULT"
3659 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3660 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3661 This makes it much easier to arrive at a reasonable default order
3662 in applications for which anonymous ciphers are OK (meaning
3663 that you can't actually use DEFAULT).
3664 [Bodo Moeller; suggested by Victor Duchovni]
3666 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3667 processing) into multiple integers instead of setting
3668 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3669 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3670 (These masks as well as the individual bit definitions are hidden
3671 away into the non-exported interface ssl/ssl_locl.h, so this
3672 change to the definition of the SSL_CIPHER structure shouldn't
3673 affect applications.) This give us more bits for each of these
3674 categories, so there is no longer a need to coagulate AES128 and
3675 AES256 into a single algorithm bit, and to coagulate Camellia128
3676 and Camellia256 into a single algorithm bit, which has led to all
3679 Thus, among other things, the kludge introduced in 0.9.7m and
3680 0.9.8e for masking out AES256 independently of AES128 or masking
3681 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3683 With the change, we also introduce new ciphersuite aliases that
3684 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3688 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3689 Use the leftmost N bytes of the signature input if the input is
3690 larger than the prime q (with N being the size in bytes of q).
3693 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3694 it yet and it is largely untested.
3697 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3700 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3701 some compilers (gcc 4.2 and later) reject their use. Safestack is
3702 reimplemented. Update ASN1 to avoid use of legacy functions.
3705 *) Win32/64 targets are linked with Winsock2.
3708 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3709 to external functions. This can be used to increase CRL handling
3710 efficiency especially when CRLs are very large by (for example) storing
3711 the CRL revoked certificates in a database.
3714 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3715 new CRLs added to a directory can be used. New command line option
3716 -verify_return_error to s_client and s_server. This causes real errors
3717 to be returned by the verify callback instead of carrying on no matter
3718 what. This reflects the way a "real world" verify callback would behave.
3721 *) GOST engine, supporting several GOST algorithms and public key formats.
3722 Kindly donated by Cryptocom.
3725 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3726 partitioned by DP are handled but no indirect CRL or reason partitioning
3727 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3728 selected via a scoring technique which handles IDP and AKID in CRLs.
3731 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3732 will ultimately be used for all verify operations: this will remove the
3733 X509_STORE dependency on certificate verification and allow alternative
3734 lookup methods. X509_STORE based implementations of these two callbacks.
3737 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3738 Modify get_crl() to find a valid (unexpired) CRL if possible.
3741 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3742 this would be called X509_CRL_cmp() but that name is already used by
3743 a function that just compares CRL issuer names. Cache several CRL
3744 extensions in X509_CRL structure and cache CRLDP in X509.
3747 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3748 this maps equivalent X509_NAME structures into a consistent structure.
3749 Name comparison can then be performed rapidly using memcmp().
3752 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3756 *) Allow digests to supply their own micalg string for S/MIME type using
3757 the ctrl EVP_MD_CTRL_MICALG.
3760 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3761 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3762 ctrl. It can then customise the structure before and/or after signing
3766 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3767 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3768 to free up any added signature OIDs.
3771 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3772 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3773 digest and cipher tables. New options added to openssl utility:
3774 list-message-digest-algorithms and list-cipher-algorithms.
3777 *) Change the array representation of binary polynomials: the list
3778 of degrees of non-zero coefficients is now terminated with -1.
3779 Previously it was terminated with 0, which was also part of the
3780 value; thus, the array representation was not applicable to
3781 polynomials where t^0 has coefficient zero. This change makes
3782 the array representation useful in a more general context.
3785 *) Various modifications and fixes to SSL/TLS cipher string
3786 handling. For ECC, the code now distinguishes between fixed ECDH
3787 with RSA certificates on the one hand and with ECDSA certificates
3788 on the other hand, since these are separate ciphersuites. The
3789 unused code for Fortezza ciphersuites has been removed.
3791 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3792 (not "ECDHE"). For consistency with the code for DH
3793 certificates, use of ECDH certificates is now considered ECDH
3794 authentication, not RSA or ECDSA authentication (the latter is
3795 merely the CA's signing algorithm and not actively used in the
3798 The temporary ciphersuite alias "ECCdraft" is no longer
3799 available, and ECC ciphersuites are no longer excluded from "ALL"
3800 and "DEFAULT". The following aliases now exist for RFC 4492
3801 ciphersuites, most of these by analogy with the DH case:
3803 kECDHr - ECDH cert, signed with RSA
3804 kECDHe - ECDH cert, signed with ECDSA
3805 kECDH - ECDH cert (signed with either RSA or ECDSA)
3806 kEECDH - ephemeral ECDH
3807 ECDH - ECDH cert or ephemeral ECDH
3813 AECDH - anonymous ECDH
3814 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3818 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3819 Use correct micalg parameters depending on digest(s) in signed message.
3822 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3823 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3826 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3827 an engine to register a method. Add ENGINE lookups for methods and
3828 functional reference processing.
3831 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3832 EVP_{Sign,Verify}* which allow an application to customise the signature
3836 *) New -resign option to smime utility. This adds one or more signers
3837 to an existing PKCS#7 signedData structure. Also -md option to use an
3838 alternative message digest algorithm for signing.
3841 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3842 create PKCS7 structures containing multiple signers. Update smime
3843 application to support multiple signers.
3846 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3850 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3851 Reorganize PBE internals to lookup from a static table using NIDs,
3852 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3853 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3854 PRF which will be automatically used with PBES2.
3857 *) Replace the algorithm specific calls to generate keys in "req" with the
3861 *) Update PKCS#7 enveloped data routines to use new API. This is now
3862 supported by any public key method supporting the encrypt operation. A
3863 ctrl is added to allow the public key algorithm to examine or modify
3864 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3868 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3869 a default digest type to use. In most cases this will be SHA1 but some
3870 algorithms (such as GOST) need to specify an alternative digest. The
3871 return value indicates how strong the preference is 1 means optional and
3872 2 is mandatory (that is it is the only supported type). Modify
3873 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3874 use the default md. Update openssl utilities to use the default digest
3875 type for signing if it is not explicitly indicated.
3878 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3879 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3880 signing method from the key type. This effectively removes the link
3881 between digests and public key types.
3884 *) Add an OID cross reference table and utility functions. Its purpose is to
3885 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3886 rsaEncryption. This will allow some of the algorithm specific hackery
3887 needed to use the correct OID to be removed.
3890 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3891 structures for PKCS7_sign(). They are now set up by the relevant public
3895 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3898 *) Add support for key derivation (agreement) in the API, DH method and
3902 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3903 public and private key formats. As a side effect these add additional
3904 command line functionality not previously available: DSA signatures can be
3905 generated and verified using pkeyutl and DH key support and generation in
3910 [Oliver Tappe <zooey@hirschkaefer.de>]
3912 *) New make target "install_html_docs" installs HTML renditions of the
3914 [Oliver Tappe <zooey@hirschkaefer.de>]
3916 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3917 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3918 support key and parameter generation and add initial key generation
3919 functionality for RSA.
3922 *) Add functions for main EVP_PKEY_method operations. The undocumented
3923 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3924 EVP_PKEY_{encrypt,decrypt}_old.
3927 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3928 key API, doesn't do much yet.
3931 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3932 public key algorithms. New option to openssl utility:
3933 "list-public-key-algorithms" to print out info.
3936 *) Implement the Supported Elliptic Curves Extension for
3937 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3940 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3941 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3944 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3945 utilities such as rsa, dsa, dsaparam etc except they process any key
3949 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3950 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3951 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3955 *) Initial support for pluggable public key ASN1.
3956 De-spaghettify the public key ASN1 handling. Move public and private
3957 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3958 algorithm specific handling to a single module within the relevant
3959 algorithm directory. Add functions to allow (near) opaque processing
3960 of public and private key structures.
3963 *) Implement the Supported Point Formats Extension for
3964 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3967 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3968 for the psk identity [hint] and the psk callback functions to the
3969 SSL_SESSION, SSL and SSL_CTX structure.
3972 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3976 SSL_CTX_use_psk_identity_hint
3977 SSL_get_psk_identity_hint
3978 SSL_get_psk_identity
3979 SSL_use_psk_identity_hint
3981 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3983 *) Add RFC 3161 compliant time stamp request creation, response generation
3984 and response verification functionality.
3985 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3987 *) Add initial support for TLS extensions, specifically for the server_name
3988 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3989 have new members for a host name. The SSL data structure has an
3990 additional member SSL_CTX *initial_ctx so that new sessions can be
3991 stored in that context to allow for session resumption, even after the
3992 SSL has been switched to a new SSL_CTX in reaction to a client's
3993 server_name extension.
3995 New functions (subject to change):
3997 SSL_get_servername()
3998 SSL_get_servername_type()
4001 New CTRL codes and macros (subject to change):
4003 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4004 - SSL_CTX_set_tlsext_servername_callback()
4005 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4006 - SSL_CTX_set_tlsext_servername_arg()
4007 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4009 openssl s_client has a new '-servername ...' option.
4011 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4012 '-key2 ...', '-servername_fatal' (subject to change). This allows
4013 testing the HostName extension for a specific single host name ('-cert'
4014 and '-key' remain fallbacks for handshakes without HostName
4015 negotiation). If the unrecognized_name alert has to be sent, this by
4016 default is a warning; it becomes fatal with the '-servername_fatal'
4019 [Peter Sylvester, Remy Allais, Christophe Renou]
4021 *) Whirlpool hash implementation is added.
4024 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4025 bn(64,32). Because of instruction set limitations it doesn't have
4026 any negative impact on performance. This was done mostly in order
4027 to make it possible to share assembler modules, such as bn_mul_mont
4028 implementations, between 32- and 64-bit builds without hassle.
4031 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4032 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4036 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4037 dedicated Montgomery multiplication procedure, is introduced.
4038 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4039 "64-bit" performance on certain 32-bit targets.
4042 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4043 in SSL structures. New SSL ctrl to set maximum send fragment size.
4044 Save memory by seeting the I/O buffer sizes dynamically instead of
4045 using the maximum available value.
4048 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4049 in addition to the text details.
4052 *) Very, very preliminary EXPERIMENTAL support for printing of general
4053 ASN1 structures. This currently produces rather ugly output and doesn't
4054 handle several customised structures at all.
4057 *) Integrated support for PVK file format and some related formats such
4058 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4059 these in the 'rsa' and 'dsa' utilities.
4062 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4065 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4066 place for the (very old) "NETSCAPE" format certificates which are now
4067 handled using new ASN1 code equivalents.
4070 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4071 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4072 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4075 *) Modify CRL distribution points extension code to print out previously
4076 unsupported fields. Enhance extension setting code to allow setting of
4080 *) Add print and set support for Issuing Distribution Point CRL extension.
4083 *) Change 'Configure' script to enable Camellia by default.
4086 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4088 *) When rejecting SSL/TLS records due to an incorrect version number, never
4089 update s->server with a new major version number. As of
4090 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4091 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4092 the previous behavior could result in a read attempt at NULL when
4093 receiving specific incorrect SSL/TLS records once record payload
4094 protection is active. (CVE-2010-0740)
4095 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4097 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4098 could be crashed if the relevant tables were not present (e.g. chrooted).
4099 [Tomas Hoger <thoger@redhat.com>]
4101 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4103 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
4104 [Martin Olsson, Neel Mehta]
4106 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4107 accommodate for stack sorting, always a write lock!).
4110 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4111 excessive delays in the RAND_poll(): over a minute. As a workaround
4112 include a time check in the inner Heap32Next loop too.
4115 *) The code that handled flushing of data in SSL/TLS originally used the
4116 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4117 the problem outlined in PR#1949. The fix suggested there however can
4118 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4119 of Apache). So instead simplify the code to flush unconditionally.
4120 This should be fine since flushing with no data to flush is a no op.
4123 *) Handle TLS versions 2.0 and later properly and correctly use the
4124 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4125 off ancient servers have a habit of sticking around for a while...
4128 *) Modify compression code so it frees up structures without using the
4129 ex_data callbacks. This works around a problem where some applications
4130 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4131 restarting) then use compression (e.g. SSL with compression) later.
4132 This results in significant per-connection memory leaks and
4133 has caused some security issues including CVE-2008-1678 and
4137 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4138 change when encrypting or decrypting.
4141 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4142 connect and renegotiate with servers which do not support RI.
4143 Until RI is more widely deployed this option is enabled by default.
4146 *) Add "missing" ssl ctrls to clear options and mode.
4149 *) If client attempts to renegotiate and doesn't support RI respond with
4150 a no_renegotiation alert as required by RFC5746. Some renegotiating
4151 TLS clients will continue a connection gracefully when they receive
4152 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4153 waiting for a server hello which it will never receive. Now we treat a
4154 received no_renegotiation alert as a fatal error. This is because
4155 applications requesting a renegotiation might well expect it to succeed
4156 and would have no code in place to handle the server denying it so the
4157 only safe thing to do is to terminate the connection.
4160 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4161 peer supports secure renegotiation and 0 otherwise. Print out peer
4162 renegotiation support in s_client/s_server.
4165 *) Replace the highly broken and deprecated SPKAC certification method with
4166 the updated NID creation version. This should correctly handle UTF8.
4169 *) Implement RFC5746. Re-enable renegotiation but require the extension
4170 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4171 turns out to be a bad idea. It has been replaced by
4172 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4173 SSL_CTX_set_options(). This is really not recommended unless you
4174 know what you are doing.
4175 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4177 *) Fixes to stateless session resumption handling. Use initial_ctx when
4178 issuing and attempting to decrypt tickets in case it has changed during
4179 servername handling. Use a non-zero length session ID when attempting
4180 stateless session resumption: this makes it possible to determine if
4181 a resumption has occurred immediately after receiving server hello
4182 (several places in OpenSSL subtly assume this) instead of later in
4186 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4187 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4188 fixes for a few places where the return code is not checked
4190 [Julia Lawall <julia@diku.dk>]
4192 *) Add --strict-warnings option to Configure script to include devteam
4193 warnings in other configurations.
4196 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4197 makes it possible to install openssl libraries in locations which
4198 have names other than "lib", for example "/usr/lib64" which some
4200 [Steve Henson, based on patch from Jeremy Utley]
4202 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4203 X690 8.9.12 and can produce some misleading textual output of OIDs.
4204 [Steve Henson, reported by Dan Kaminsky]
4206 *) Delete MD2 from algorithm tables. This follows the recommendation in
4207 several standards that it is not used in new applications due to
4208 several cryptographic weaknesses. For binary compatibility reasons
4209 the MD2 API is still compiled in by default.
4212 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4216 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4217 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4219 [Guenter <lists@gknw.net>]
4221 *) Fix the server certificate chain building code to use X509_verify_cert(),
4222 it used to have an ad-hoc builder which was unable to cope with anything
4223 other than a simple chain.
4224 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4226 *) Don't check self signed certificate signatures in X509_verify_cert()
4227 by default (a flag can override this): it just wastes time without
4228 adding any security. As a useful side effect self signed root CAs
4229 with non-FIPS digests are now usable in FIPS mode.
4232 *) In dtls1_process_out_of_seq_message() the check if the current message
4233 is already buffered was missing. For every new message was memory
4234 allocated, allowing an attacker to perform an denial of service attack
4235 with sending out of seq handshake messages until there is no memory
4236 left. Additionally every future messege was buffered, even if the
4237 sequence number made no sense and would be part of another handshake.
4238 So only messages with sequence numbers less than 10 in advance will be
4239 buffered. (CVE-2009-1378)
4240 [Robin Seggelmann, discovered by Daniel Mentz]
4242 *) Records are buffered if they arrive with a future epoch to be
4243 processed after finishing the corresponding handshake. There is
4244 currently no limitation to this buffer allowing an attacker to perform
4245 a DOS attack with sending records with future epochs until there is no
4246 memory left. This patch adds the pqueue_size() function to determine
4247 the size of a buffer and limits the record buffer to 100 entries.
4249 [Robin Seggelmann, discovered by Daniel Mentz]
4251 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4252 parent structure is freed. (CVE-2009-1379)
4255 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4256 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4259 [Ilya O. <vrghost@gmail.com>]
4261 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4263 *) Disable renegotiation completely - this fixes a severe security
4264 problem (CVE-2009-3555) at the cost of breaking all
4265 renegotiation. Renegotiation can be re-enabled by setting
4266 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4267 run-time. This is really not recommended unless you know what
4271 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4273 *) Don't set val to NULL when freeing up structures, it is freed up by
4274 underlying code. If sizeof(void *) > sizeof(long) this can result in
4275 zeroing past the valid field. (CVE-2009-0789)
4276 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4278 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4279 checked correctly. This would allow some invalid signed attributes to
4280 appear to verify correctly. (CVE-2009-0591)
4281 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4283 *) Reject UniversalString and BMPString types with invalid lengths. This
4284 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4285 a legal length. (CVE-2009-0590)
4288 *) Set S/MIME signing as the default purpose rather than setting it
4289 unconditionally. This allows applications to override it at the store
4293 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4294 to handle some structures.
4297 *) Improve efficiency of mem_gets: don't search whole buffer each time
4299 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4301 *) New -hex option for openssl rand.
4304 *) Print out UTF8String and NumericString when parsing ASN1.
4307 *) Support NumericString type for name components.
4310 *) Allow CC in the environment to override the automatically chosen
4311 compiler. Note that nothing is done to ensure flags work with the
4315 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4317 *) Properly check EVP_VerifyFinal() and similar return values
4319 [Ben Laurie, Bodo Moeller, Google Security Team]
4321 *) Enable TLS extensions by default.
4324 *) Allow the CHIL engine to be loaded, whether the application is
4325 multithreaded or not. (This does not release the developer from the
4326 obligation to set up the dynamic locking callbacks.)
4327 [Sander Temme <sander@temme.net>]
4329 *) Use correct exit code if there is an error in dgst command.
4330 [Steve Henson; problem pointed out by Roland Dirlewanger]
4332 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4333 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4336 *) Add experimental JPAKE support, including demo authentication in
4337 s_client and s_server.
4340 *) Set the comparison function in v3_addr_canonize().
4341 [Rob Austein <sra@hactrn.net>]
4343 *) Add support for XMPP STARTTLS in s_client.
4344 [Philip Paeps <philip@freebsd.org>]
4346 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4347 to ensure that even with this option, only ciphersuites in the
4348 server's preference list will be accepted. (Note that the option
4349 applies only when resuming a session, so the earlier behavior was
4350 just about the algorithm choice for symmetric cryptography.)
4353 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4355 *) Fix NULL pointer dereference if a DTLS server received
4356 ChangeCipherSpec as first record (CVE-2009-1386).
4359 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4360 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4363 *) The fix in 0.9.8c that supposedly got rid of unsafe
4364 double-checked locking was incomplete for RSA blinding,
4365 addressing just one layer of what turns out to have been
4366 doubly unsafe triple-checked locking.
4368 So now fix this for real by retiring the MONT_HELPER macro
4369 in crypto/rsa/rsa_eay.c.
4371 [Bodo Moeller; problem pointed out by Marius Schilder]
4373 *) Various precautionary measures:
4375 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4377 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4378 (NB: This would require knowledge of the secret session ticket key
4379 to exploit, in which case you'd be SOL either way.)
4381 - Change bn_nist.c so that it will properly handle input BIGNUMs
4382 outside the expected range.
4384 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4387 [Neel Mehta, Bodo Moeller]
4389 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4390 the load fails. Useful for distros.
4391 [Ben Laurie and the FreeBSD team]
4393 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4396 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4399 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4401 This work was sponsored by Logica.
4404 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4405 keystores. Support for SSL/TLS client authentication too.
4406 Not compiled unless enable-capieng specified to Configure.
4408 This work was sponsored by Logica.
4411 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4412 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4413 attribute creation routines such as certificate requests and PKCS#12
4417 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4419 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4420 handshake which could lead to a cilent crash as found using the
4421 Codenomicon TLS test suite (CVE-2008-1672)
4422 [Steve Henson, Mark Cox]
4424 *) Fix double free in TLS server name extensions which could lead to
4425 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4428 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4430 Clear the error queue to ensure that error entries left from
4431 older function calls do not interfere with the correct operation.
4432 [Lutz Jaenicke, Erik de Castro Lopo]
4434 *) Remove root CA certificates of commercial CAs:
4436 The OpenSSL project does not recommend any specific CA and does not
4437 have any policy with respect to including or excluding any CA.
4438 Therefore it does not make any sense to ship an arbitrary selection
4439 of root CA certificates with the OpenSSL software.
4442 *) RSA OAEP patches to fix two separate invalid memory reads.
4443 The first one involves inputs when 'lzero' is greater than
4444 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4445 before the beginning of from). The second one involves inputs where
4446 the 'db' section contains nothing but zeroes (there is a one-byte
4447 invalid read after the end of 'db').
4448 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4450 *) Partial backport from 0.9.9-dev:
4452 Introduce bn_mul_mont (dedicated Montgomery multiplication
4453 procedure) as a candidate for BIGNUM assembler implementation.
4454 While 0.9.9-dev uses assembler for various architectures, only
4455 x86_64 is available by default here in the 0.9.8 branch, and
4456 32-bit x86 is available through a compile-time setting.
4458 To try the 32-bit x86 assembler implementation, use Configure
4459 option "enable-montasm" (which exists only for this backport).
4461 As "enable-montasm" for 32-bit x86 disclaims code stability
4462 anyway, in this constellation we activate additional code
4463 backported from 0.9.9-dev for further performance improvements,
4464 namely BN_from_montgomery_word. (To enable this otherwise,
4465 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4467 [Andy Polyakov (backport partially by Bodo Moeller)]
4469 *) Add TLS session ticket callback. This allows an application to set
4470 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4471 values. This is useful for key rollover for example where several key
4472 sets may exist with different names.
4475 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4476 This was broken until now in 0.9.8 releases, such that the only way
4477 a registered ENGINE could be used (assuming it initialises
4478 successfully on the host) was to explicitly set it as the default
4479 for the relevant algorithms. This is in contradiction with 0.9.7
4480 behaviour and the documentation. With this fix, when an ENGINE is
4481 registered into a given algorithm's table of implementations, the
4482 'uptodate' flag is reset so that auto-discovery will be used next
4483 time a new context for that algorithm attempts to select an
4485 [Ian Lister (tweaked by Geoff Thorpe)]
4487 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4488 implementation in the following ways:
4490 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4493 Lack of BER streaming support means one pass streaming processing is
4494 only supported if data is detached: setting the streaming flag is
4495 ignored for embedded content.
4497 CMS support is disabled by default and must be explicitly enabled
4498 with the enable-cms configuration option.
4501 *) Update the GMP engine glue to do direct copies between BIGNUM and
4502 mpz_t when openssl and GMP use the same limb size. Otherwise the
4503 existing "conversion via a text string export" trick is still used.
4504 [Paul Sheer <paulsheer@gmail.com>]
4506 *) Zlib compression BIO. This is a filter BIO which compressed and
4507 uncompresses any data passed through it.
4510 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4511 RFC3394 compatible AES key wrapping.
4514 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4515 sets string data without copying. X509_ALGOR_set0() and
4516 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4517 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4518 from an X509_ATTRIBUTE structure optionally checking it occurs only
4519 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4523 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4524 to get the expected BN_FLG_CONSTTIME behavior.
4525 [Bodo Moeller (Google)]
4529 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4530 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4531 - added some more tests to do_tests.pl
4532 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4533 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4534 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4535 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4536 - various changes to netware.pl to enable gcc-cross builds on Win32
4538 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4539 - various changes to fix missing prototype warnings
4540 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4541 - added AES, WHIRLPOOL and CPUID assembler code to build files
4542 - added missing AES assembler make rules to mk1mf.pl
4543 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4544 [Guenter Knauf <eflash@gmx.net>]
4546 *) Implement certificate status request TLS extension defined in RFC3546.
4547 A client can set the appropriate parameters and receive the encoded
4548 OCSP response via a callback. A server can query the supplied parameters
4549 and set the encoded OCSP response in the callback. Add simplified examples
4550 to s_client and s_server.
4553 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4555 *) Fix various bugs:
4556 + Binary incompatibility of ssl_ctx_st structure
4557 + DTLS interoperation with non-compliant servers
4558 + Don't call get_session_cb() without proposed session
4559 + Fix ia64 assembler code
4560 [Andy Polyakov, Steve Henson]
4562 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4564 *) DTLS Handshake overhaul. There were longstanding issues with
4565 OpenSSL DTLS implementation, which were making it impossible for
4566 RFC 4347 compliant client to communicate with OpenSSL server.
4567 Unfortunately just fixing these incompatibilities would "cut off"
4568 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4569 server keeps tolerating non RFC compliant syntax. The opposite is
4570 not true, 0.9.8f client can not communicate with earlier server.
4571 This update even addresses CVE-2007-4995.
4574 *) Changes to avoid need for function casts in OpenSSL: some compilers
4575 (gcc 4.2 and later) reject their use.
4576 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4579 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4580 RFC4507bis. The encrypted ticket format is an encrypted encoded
4581 SSL_SESSION structure, that way new session features are automatically
4584 If a client application caches session in an SSL_SESSION structure
4585 support is transparent because tickets are now stored in the encoded
4588 The SSL_CTX structure automatically generates keys for ticket
4589 protection in servers so again support should be possible
4590 with no application modification.
4592 If a client or server wishes to disable RFC4507 support then the option
4593 SSL_OP_NO_TICKET can be set.
4595 Add a TLS extension debugging callback to allow the contents of any client
4596 or server extensions to be examined.
4598 This work was sponsored by Google.
4601 *) Add initial support for TLS extensions, specifically for the server_name
4602 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4603 have new members for a host name. The SSL data structure has an
4604 additional member SSL_CTX *initial_ctx so that new sessions can be
4605 stored in that context to allow for session resumption, even after the
4606 SSL has been switched to a new SSL_CTX in reaction to a client's
4607 server_name extension.
4609 New functions (subject to change):
4611 SSL_get_servername()
4612 SSL_get_servername_type()
4615 New CTRL codes and macros (subject to change):
4617 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4618 - SSL_CTX_set_tlsext_servername_callback()
4619 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4620 - SSL_CTX_set_tlsext_servername_arg()
4621 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4623 openssl s_client has a new '-servername ...' option.
4625 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4626 '-key2 ...', '-servername_fatal' (subject to change). This allows
4627 testing the HostName extension for a specific single host name ('-cert'
4628 and '-key' remain fallbacks for handshakes without HostName
4629 negotiation). If the unrecognized_name alert has to be sent, this by
4630 default is a warning; it becomes fatal with the '-servername_fatal'
4633 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4635 *) Add AES and SSE2 assembly language support to VC++ build.
4638 *) Mitigate attack on final subtraction in Montgomery reduction.
4641 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4642 (which previously caused an internal error).
4645 *) Squeeze another 10% out of IGE mode when in != out.
4648 *) AES IGE mode speedup.
4649 [Dean Gaudet (Google)]
4651 *) Add the Korean symmetric 128-bit cipher SEED (see
4652 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4653 add SEED ciphersuites from RFC 4162:
4655 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4656 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4657 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4658 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4660 To minimize changes between patchlevels in the OpenSSL 0.9.8
4661 series, SEED remains excluded from compilation unless OpenSSL
4662 is configured with 'enable-seed'.
4663 [KISA, Bodo Moeller]
4665 *) Mitigate branch prediction attacks, which can be practical if a
4666 single processor is shared, allowing a spy process to extract
4667 information. For detailed background information, see
4668 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4669 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4670 and Necessary Software Countermeasures"). The core of the change
4671 are new versions BN_div_no_branch() and
4672 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4673 respectively, which are slower, but avoid the security-relevant
4674 conditional branches. These are automatically called by BN_div()
4675 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4676 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4677 remove a conditional branch.
4679 BN_FLG_CONSTTIME is the new name for the previous
4680 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4681 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4682 in the exponent causes BN_mod_exp_mont() to use the alternative
4683 implementation in BN_mod_exp_mont_consttime().) The old name
4684 remains as a deprecated alias.
4686 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4687 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4688 constant-time implementations for more than just exponentiation.
4689 Here too the old name is kept as a deprecated alias.
4691 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4692 the BN_BLINDING structure gets an independent copy of the
4693 modulus. This means that the previous "BIGNUM *m" argument to
4694 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4695 essentially becomes "const BIGNUM *m", although we can't actually
4696 change this in the header file before 0.9.9. It allows
4697 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4698 enable BN_FLG_CONSTTIME.
4700 [Matthew D Wood (Intel Corp)]
4702 *) In the SSL/TLS server implementation, be strict about session ID
4703 context matching (which matters if an application uses a single
4704 external cache for different purposes). Previously,
4705 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4706 set. This did ensure strict client verification, but meant that,
4707 with applications using a single external cache for quite
4708 different requirements, clients could circumvent ciphersuite
4709 restrictions for a given session ID context by starting a session
4710 in a different context.
4713 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4714 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4715 authentication-only ciphersuites.
4718 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4719 not complete and could lead to a possible single byte overflow
4720 (CVE-2007-5135) [Ben Laurie]
4722 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4724 *) Since AES128 and AES256 (and similarly Camellia128 and
4725 Camellia256) share a single mask bit in the logic of
4726 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4727 kludge to work properly if AES128 is available and AES256 isn't
4728 (or if Camellia128 is available and Camellia256 isn't).
4731 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4732 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4733 When a point or a seed is encoded in a BIT STRING, we need to
4734 prevent the removal of trailing zero bits to get the proper DER
4735 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4736 of a NamedBitList, for which trailing 0 bits need to be removed.)
4739 *) Have SSL/TLS server implementation tolerate "mismatched" record
4740 protocol version while receiving ClientHello even if the
4741 ClientHello is fragmented. (The server can't insist on the
4742 particular protocol version it has chosen before the ServerHello
4743 message has informed the client about his choice.)
4746 *) Add RFC 3779 support.
4747 [Rob Austein for ARIN, Ben Laurie]
4749 *) Load error codes if they are not already present instead of using a
4750 static variable. This allows them to be cleanly unloaded and reloaded.
4751 Improve header file function name parsing.
4754 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4755 or CAPABILITY handshake as required by RFCs.
4758 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4760 *) Introduce limits to prevent malicious keys being able to
4761 cause a denial of service. (CVE-2006-2940)
4762 [Steve Henson, Bodo Moeller]
4764 *) Fix ASN.1 parsing of certain invalid structures that can result
4765 in a denial of service. (CVE-2006-2937) [Steve Henson]
4767 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4768 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4770 *) Fix SSL client code which could crash if connecting to a
4771 malicious SSLv2 server. (CVE-2006-4343)
4772 [Tavis Ormandy and Will Drewry, Google Security Team]
4774 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4775 match only those. Before that, "AES256-SHA" would be interpreted
4776 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4777 the same strength classification in 0.9.7h) as we currently only
4778 have a single AES bit in the ciphersuite description bitmap.
4779 That change, however, also applied to ciphersuite strings such as
4780 "RC4-MD5" that intentionally matched multiple ciphersuites --
4781 namely, SSL 2.0 ciphersuites in addition to the more common ones
4782 from SSL 3.0/TLS 1.0.
4784 So we change the selection algorithm again: Naming an explicit
4785 ciphersuite selects this one ciphersuite, and any other similar
4786 ciphersuite (same bitmap) from *other* protocol versions.
4787 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4788 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4790 Since SSL 2.0 does not have any ciphersuites for which the
4791 128/256 bit distinction would be relevant, this works for now.
4792 The proper fix will be to use different bits for AES128 and
4793 AES256, which would have avoided the problems from the beginning;
4794 however, bits are scarce, so we can only do this in a new release
4795 (not just a patchlevel) when we can change the SSL_CIPHER
4796 definition to split the single 'unsigned long mask' bitmap into
4797 multiple values to extend the available space.
4801 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4803 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4804 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4806 *) Add AES IGE and biIGE modes.
4809 *) Change the Unix randomness entropy gathering to use poll() when
4810 possible instead of select(), since the latter has some
4811 undesirable limitations.
4812 [Darryl Miles via Richard Levitte and Bodo Moeller]
4814 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4815 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4816 cannot be implicitly activated as part of, e.g., the "AES" alias.
4817 However, please upgrade to OpenSSL 0.9.9[-dev] for
4818 non-experimental use of the ECC ciphersuites to get TLS extension
4819 support, which is required for curve and point format negotiation
4820 to avoid potential handshake problems.
4823 *) Disable rogue ciphersuites:
4825 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4826 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4827 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4829 The latter two were purportedly from
4830 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4833 Also deactivate the remaining ciphersuites from
4834 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4835 unofficial, and the ID has long expired.
4838 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4839 dual-core machines) and other potential thread-safety issues.
4842 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4843 versions), which is now available for royalty-free use
4844 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4845 Also, add Camellia TLS ciphersuites from RFC 4132.
4847 To minimize changes between patchlevels in the OpenSSL 0.9.8
4848 series, Camellia remains excluded from compilation unless OpenSSL
4849 is configured with 'enable-camellia'.
4852 *) Disable the padding bug check when compression is in use. The padding
4853 bug check assumes the first packet is of even length, this is not
4854 necessarily true if compresssion is enabled and can result in false
4855 positives causing handshake failure. The actual bug test is ancient
4856 code so it is hoped that implementations will either have fixed it by
4857 now or any which still have the bug do not support compression.
4860 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4862 *) When applying a cipher rule check to see if string match is an explicit
4863 cipher suite and only match that one cipher suite if it is.
4866 *) Link in manifests for VC++ if needed.
4867 [Austin Ziegler <halostatue@gmail.com>]
4869 *) Update support for ECC-based TLS ciphersuites according to
4870 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4871 TLS extensions, which are supported starting with the 0.9.9
4872 branch, not in the OpenSSL 0.9.8 branch).
4875 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4876 opaque EVP_CIPHER_CTX handling.
4879 *) Fixes and enhancements to zlib compression code. We now only use
4880 "zlib1.dll" and use the default __cdecl calling convention on Win32
4881 to conform with the standards mentioned here:
4882 http://www.zlib.net/DLL_FAQ.txt
4883 Static zlib linking now works on Windows and the new --with-zlib-include
4884 --with-zlib-lib options to Configure can be used to supply the location
4885 of the headers and library. Gracefully handle case where zlib library
4889 *) Several fixes and enhancements to the OID generation code. The old code
4890 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4891 handle numbers larger than ULONG_MAX, truncated printing and had a
4892 non standard OBJ_obj2txt() behaviour.
4895 *) Add support for building of engines under engine/ as shared libraries
4896 under VC++ build system.
4899 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4900 Hopefully, we will not see any false combination of paths any more.
4903 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4905 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4906 (part of SSL_OP_ALL). This option used to disable the
4907 countermeasure against man-in-the-middle protocol-version
4908 rollback in the SSL 2.0 server implementation, which is a bad
4909 idea. (CVE-2005-2969)
4911 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4912 for Information Security, National Institute of Advanced Industrial
4913 Science and Technology [AIST], Japan)]
4915 *) Add two function to clear and return the verify parameter flags.
4918 *) Keep cipherlists sorted in the source instead of sorting them at
4919 runtime, thus removing the need for a lock.
4922 *) Avoid some small subgroup attacks in Diffie-Hellman.
4923 [Nick Mathewson and Ben Laurie]
4925 *) Add functions for well-known primes.
4928 *) Extended Windows CE support.
4929 [Satoshi Nakamura and Andy Polyakov]
4931 *) Initialize SSL_METHOD structures at compile time instead of during
4932 runtime, thus removing the need for a lock.
4935 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4936 attempting to decrypt each encrypted key in turn. Add support to
4940 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4942 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4945 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4948 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4949 key into the same file any more.
4952 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4955 *) Add -utf8 command line and config file option to 'ca'.
4956 [Stefan <stf@udoma.org]
4958 *) Removed the macro des_crypt(), as it seems to conflict with some
4959 libraries. Use DES_crypt().
4962 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4963 involves renaming the source and generated shared-libs for
4964 both. The engines will accept the corrected or legacy ids
4965 ('ncipher' and '4758_cca' respectively) when binding. NB,
4966 this only applies when building 'shared'.
4967 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4969 *) Add attribute functions to EVP_PKEY structure. Modify
4970 PKCS12_create() to recognize a CSP name attribute and
4971 use it. Make -CSP option work again in pkcs12 utility.
4974 *) Add new functionality to the bn blinding code:
4975 - automatic re-creation of the BN_BLINDING parameters after
4976 a fixed number of uses (currently 32)
4977 - add new function for parameter creation
4978 - introduce flags to control the update behaviour of the
4979 BN_BLINDING parameters
4980 - hide BN_BLINDING structure
4981 Add a second BN_BLINDING slot to the RSA structure to improve
4982 performance when a single RSA object is shared among several
4986 *) Add support for DTLS.
4987 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4989 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4990 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4993 *) Remove buggy and incomplete DH cert support from
4994 ssl/ssl_rsa.c and ssl/s3_both.c
4997 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4998 the apps/openssl applications.
5001 *) Compile clean with "-Wall -Wmissing-prototypes
5002 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5003 DEBUG_SAFESTACK must also be set.
5006 *) Change ./Configure so that certain algorithms can be disabled by default.
5007 The new counterpiece to "no-xxx" is "enable-xxx".
5009 The patented RC5 and MDC2 algorithms will now be disabled unless
5010 "enable-rc5" and "enable-mdc2", respectively, are specified.
5012 (IDEA remains enabled despite being patented. This is because IDEA
5013 is frequently required for interoperability, and there is no license
5014 fee for non-commercial use. As before, "no-idea" can be used to
5015 avoid this algorithm.)
5019 *) Add processing of proxy certificates (see RFC 3820). This work was
5020 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5021 EGEE (Enabling Grids for E-science in Europe).
5024 *) RC4 performance overhaul on modern architectures/implementations, such
5025 as Intel P4, IA-64 and AMD64.
5028 *) New utility extract-section.pl. This can be used specify an alternative
5029 section number in a pod file instead of having to treat each file as
5030 a separate case in Makefile. This can be done by adding two lines to the
5033 =for comment openssl_section:XXX
5035 The blank line is mandatory.
5039 *) New arguments -certform, -keyform and -pass for s_client and s_server
5040 to allow alternative format key and certificate files and passphrase
5044 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5045 update associated structures and add various utility functions.
5047 Add new policy related verify parameters, include policy checking in
5048 standard verify code. Enhance 'smime' application with extra parameters
5049 to support policy checking and print out.
5052 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5053 Nehemiah processors. These extensions support AES encryption in hardware
5054 as well as RNG (though RNG support is currently disabled).
5055 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5057 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5060 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5061 [Andy Polyakov and a number of other people]
5063 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5064 implementation contributed by IBM.
5065 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5067 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5068 exponent rather than 'unsigned long'. There is a corresponding change to
5069 the new 'rsa_keygen' element of the RSA_METHOD structure.
5070 [Jelte Jansen, Geoff Thorpe]
5072 *) Functionality for creating the initial serial number file is now
5073 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5075 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5076 number file to 1, which is bound to cause problems. To avoid
5077 the problems while respecting compatibility between different 0.9.7
5078 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5079 CA.pl for serial number initialization. With the new release 0.9.8,
5080 we can fix the problem directly in the 'ca' utility.)
5083 *) Reduced header interdepencies by declaring more opaque objects in
5084 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5085 give fewer recursive includes, which could break lazy source code - so
5086 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5087 developers should define this symbol when building and using openssl to
5088 ensure they track the recommended behaviour, interfaces, [etc], but
5089 backwards-compatible behaviour prevails when this isn't defined.
5092 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5095 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5096 This will generate a random key of the appropriate length based on the
5097 cipher context. The EVP_CIPHER can provide its own random key generation
5098 routine to support keys of a specific form. This is used in the des and
5099 3des routines to generate a key of the correct parity. Update S/MIME
5100 code to use new functions and hence generate correct parity DES keys.
5101 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5102 valid (weak or incorrect parity).
5105 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5106 as looking them up. This is useful when the verified structure may contain
5107 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5108 present unless the new PKCS7_NO_CRL flag is asserted.
5111 *) Extend ASN1 oid configuration module. It now additionally accepts the
5114 shortName = some long name, 1.2.3.4
5117 *) Reimplemented the BN_CTX implementation. There is now no more static
5118 limitation on the number of variables it can handle nor the depth of the
5119 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5120 information can now expand as required, and rather than having a single
5121 static array of bignums, BN_CTX now uses a linked-list of such arrays
5122 allowing it to expand on demand whilst maintaining the usefulness of
5123 BN_CTX's "bundling".
5126 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5127 to allow all RSA operations to function using a single BN_CTX.
5130 *) Preliminary support for certificate policy evaluation and checking. This
5131 is initially intended to pass the tests outlined in "Conformance Testing
5132 of Relying Party Client Certificate Path Processing Logic" v1.07.
5135 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5136 remained unused and not that useful. A variety of other little bignum
5137 tweaks and fixes have also been made continuing on from the audit (see
5141 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5142 associated ASN1, EVP and SSL functions and old ASN1 macros.
5145 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5146 and this should never fail. So the return value from the use of
5147 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5148 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5151 *) BN_CTX_get() should return zero-valued bignums, providing the same
5152 initialised value as BN_new().
5153 [Geoff Thorpe, suggested by Ulf Möller]
5155 *) Support for inhibitAnyPolicy certificate extension.
5158 *) An audit of the BIGNUM code is underway, for which debugging code is
5159 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5160 is considered valid when processing BIGNUMs, and causes execution to
5161 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5162 further steps are taken to deliberately pollute unused data in BIGNUM
5163 structures to try and expose faulty code further on. For now, openssl will
5164 (in its default mode of operation) continue to tolerate the inconsistent
5165 forms that it has tolerated in the past, but authors and packagers should
5166 consider trying openssl and their own applications when compiled with
5167 these debugging symbols defined. It will help highlight potential bugs in
5168 their own code, and will improve the test coverage for OpenSSL itself. At
5169 some point, these tighter rules will become openssl's default to improve
5170 maintainability, though the assert()s and other overheads will remain only
5171 in debugging configurations. See bn.h for more details.
5172 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5174 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5175 that can only be obtained through BN_CTX_new() (which implicitly
5176 initialises it). The presence of this function only made it possible
5177 to overwrite an existing structure (and cause memory leaks).
5180 *) Because of the callback-based approach for implementing LHASH as a
5181 template type, lh_insert() adds opaque objects to hash-tables and
5182 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5183 to clean up those corresponding objects before destroying the hash table
5184 (and losing the object pointers). So some over-zealous constifications in
5185 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5186 objects as "const" and the lh_doall[_arg] callback wrappers are not
5187 prototyped to have "const" restrictions on the object pointers they are
5188 given (and so aren't required to cast them away any more).
5191 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5192 (speed) prefers to use its own implementation. The two implementations
5193 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5194 its object type properly exposed (MS_TM) instead of casting to/from "char
5195 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5196 aren't necessarily the greatest nomenclatures - but this is what was used
5197 internally to the implementation so I've used that for now.
5200 *) Ensure that deprecated functions do not get compiled when
5201 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5202 the self-tests were still using deprecated key-generation functions so
5203 these have been updated also.
5206 *) Reorganise PKCS#7 code to separate the digest location functionality
5207 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5208 New function PKCS7_set_digest() to set the digest type for PKCS#7
5209 digestedData type. Add additional code to correctly generate the
5210 digestedData type and add support for this type in PKCS7 initialization
5214 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5215 structure of type "other".
5218 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5219 sure the loop does correctly stop and breaking ("division by zero")
5220 modulus operations are not performed. The (pre-generated) prime
5221 table crypto/bn/bn_prime.h was already correct, but it could not be
5222 re-generated on some platforms because of the "division by zero"
5223 situation in the script.
5224 [Ralf S. Engelschall]
5226 *) Update support for ECC-based TLS ciphersuites according to
5227 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5228 SHA-1 now is only used for "small" curves (where the
5229 representation of a field element takes up to 24 bytes); for
5230 larger curves, the field element resulting from ECDH is directly
5231 used as premaster secret.
5232 [Douglas Stebila (Sun Microsystems Laboratories)]
5234 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5235 curve secp160r1 to the tests.
5236 [Douglas Stebila (Sun Microsystems Laboratories)]
5238 *) Add the possibility to load symbols globally with DSO.
5239 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5241 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5242 control of the error stack.
5245 *) Add support for STORE in ENGINE.
5248 *) Add the STORE type. The intention is to provide a common interface
5249 to certificate and key stores, be they simple file-based stores, or
5250 HSM-type store, or LDAP stores, or...
5251 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5254 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5255 pass a list of arguments to any function as well as provide a way
5256 for a function to pass data back to the caller.
5259 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5260 works like BUF_strdup() but can be used to duplicate a portion of
5261 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5265 *) Add the function sk_find_ex() which works like sk_find(), but will
5266 return an index to an element even if an exact match couldn't be
5267 found. The index is guaranteed to point at the element where the
5268 searched-for key would be inserted to preserve sorting order.
5271 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5272 takes an extra flags argument for optional functionality. Currently,
5273 the following flags are defined:
5275 OBJ_BSEARCH_VALUE_ON_NOMATCH
5276 This one gets OBJ_bsearch_ex() to return a pointer to the first
5277 element where the comparing function returns a negative or zero
5280 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5281 This one gets OBJ_bsearch_ex() to return a pointer to the first
5282 element where the comparing function returns zero. This is useful
5283 if there are more than one element where the comparing function
5287 *) Make it possible to create self-signed certificates with 'openssl ca'
5288 in such a way that the self-signed certificate becomes part of the
5289 CA database and uses the same mechanisms for serial number generation
5290 as all other certificate signing. The new flag '-selfsign' enables
5291 this functionality. Adapt CA.sh and CA.pl.in.
5294 *) Add functionality to check the public key of a certificate request
5295 against a given private. This is useful to check that a certificate
5296 request can be signed by that key (self-signing).
5299 *) Make it possible to have multiple active certificates with the same
5300 subject in the CA index file. This is done only if the keyword
5301 'unique_subject' is set to 'no' in the main CA section (default
5302 if 'CA_default') of the configuration file. The value is saved
5303 with the database itself in a separate index attribute file,
5304 named like the index file with '.attr' appended to the name.
5307 *) Generate muti valued AVAs using '+' notation in config files for
5311 *) Support for nameConstraints certificate extension.
5314 *) Support for policyConstraints certificate extension.
5317 *) Support for policyMappings certificate extension.
5320 *) Make sure the default DSA_METHOD implementation only uses its
5321 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5322 and change its own handlers to be NULL so as to remove unnecessary
5323 indirection. This lets alternative implementations fallback to the
5324 default implementation more easily.
5327 *) Support for directoryName in GeneralName related extensions
5331 *) Make it possible to link applications using Makefile.shared.
5332 Make that possible even when linking against static libraries!
5335 *) Support for single pass processing for S/MIME signing. This now
5336 means that S/MIME signing can be done from a pipe, in addition
5337 cleartext signing (multipart/signed type) is effectively streaming
5338 and the signed data does not need to be all held in memory.
5340 This is done with a new flag PKCS7_STREAM. When this flag is set
5341 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5342 is done after the data is output (and digests calculated) in
5343 SMIME_write_PKCS7().
5346 *) Add full support for -rpath/-R, both in shared libraries and
5347 applications, at least on the platforms where it's known how
5351 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5352 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5353 will now compute a table of multiples of the generator that
5354 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5355 faster (notably in the case of a single point multiplication,
5356 scalar * generator).
5357 [Nils Larsch, Bodo Moeller]
5359 *) IPv6 support for certificate extensions. The various extensions
5360 which use the IP:a.b.c.d can now take IPv6 addresses using the
5361 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5365 *) Added an ENGINE that implements RSA by performing private key
5366 exponentiations with the GMP library. The conversions to and from
5367 GMP's mpz_t format aren't optimised nor are any montgomery forms
5368 cached, and on x86 it appears OpenSSL's own performance has caught up.
5369 However there are likely to be other architectures where GMP could
5370 provide a boost. This ENGINE is not built in by default, but it can be
5371 specified at Configure time and should be accompanied by the necessary
5372 linker additions, eg;
5373 ./config -DOPENSSL_USE_GMP -lgmp
5376 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5377 testing availability of engines with "-t" - the old behaviour is
5378 produced by increasing the feature's verbosity with "-tt".
5381 *) ECDSA routines: under certain error conditions uninitialized BN objects
5382 could be freed. Solution: make sure initialization is performed early
5383 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5387 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5388 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5389 software implementations. For DSA and DH, parameter generation can
5390 also be overridden by providing the appropriate method callbacks.
5393 *) Change the "progress" mechanism used in key-generation and
5394 primality testing to functions that take a new BN_GENCB pointer in
5395 place of callback/argument pairs. The new API functions have "_ex"
5396 postfixes and the older functions are reimplemented as wrappers for
5397 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5398 declarations of the old functions to help (graceful) attempts to
5399 migrate to the new functions. Also, the new key-generation API
5400 functions operate on a caller-supplied key-structure and return
5401 success/failure rather than returning a key or NULL - this is to
5402 help make "keygen" another member function of RSA_METHOD etc.
5404 Example for using the new callback interface:
5406 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5410 BN_GENCB_set(&my_cb, my_callback, my_arg);
5412 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5413 /* For the meaning of a, b in calls to my_callback(), see the
5414 * documentation of the function that calls the callback.
5415 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5416 * my_callback should return 1 if it wants BN_is_prime_ex()
5417 * to continue, or 0 to stop.
5422 *) Change the ZLIB compression method to be stateful, and make it
5423 available to TLS with the number defined in
5424 draft-ietf-tls-compression-04.txt.
5427 *) Add the ASN.1 structures and functions for CertificatePair, which
5428 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5430 CertificatePair ::= SEQUENCE {
5431 forward [0] Certificate OPTIONAL,
5432 reverse [1] Certificate OPTIONAL,
5433 -- at least one of the pair shall be present -- }
5435 Also implement the PEM functions to read and write certificate
5436 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5438 This needed to be defined, mostly for the sake of the LDAP
5439 attribute crossCertificatePair, but may prove useful elsewhere as
5443 *) Make it possible to inhibit symlinking of shared libraries in
5444 Makefile.shared, for Cygwin's sake.
5447 *) Extend the BIGNUM API by creating a function
5448 void BN_set_negative(BIGNUM *a, int neg);
5449 and a macro that behave like
5450 int BN_is_negative(const BIGNUM *a);
5452 to avoid the need to access 'a->neg' directly in applications.
5455 *) Implement fast modular reduction for pseudo-Mersenne primes
5456 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5457 EC_GROUP_new_curve_GFp() will now automatically use this
5459 [Nils Larsch <nla@trustcenter.de>]
5461 *) Add new lock type (CRYPTO_LOCK_BN).
5464 *) Change the ENGINE framework to automatically load engines
5465 dynamically from specific directories unless they could be
5466 found to already be built in or loaded. Move all the
5467 current engines except for the cryptodev one to a new
5469 The engines in engines/ are built as shared libraries if
5470 the "shared" options was given to ./Configure or ./config.
5471 Otherwise, they are inserted in libcrypto.a.
5472 /usr/local/ssl/engines is the default directory for dynamic
5473 engines, but that can be overridden at configure time through
5474 the usual use of --prefix and/or --openssldir, and at run
5475 time with the environment variable OPENSSL_ENGINES.
5476 [Geoff Thorpe and Richard Levitte]
5478 *) Add Makefile.shared, a helper makefile to build shared
5479 libraries. Adapt Makefile.org.
5482 *) Add version info to Win32 DLLs.
5483 [Peter 'Luna' Runestig" <peter@runestig.com>]
5485 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5486 can be added using this API to created arbitrary PKCS#12
5487 files while avoiding the low level API.
5489 New options to PKCS12_create(), key or cert can be NULL and
5490 will then be omitted from the output file. The encryption
5491 algorithm NIDs can be set to -1 for no encryption, the mac
5492 iteration count can be set to 0 to omit the mac.
5494 Enhance pkcs12 utility by making the -nokeys and -nocerts
5495 options work when creating a PKCS#12 file. New option -nomac
5496 to omit the mac, NONE can be set for an encryption algorithm.
5497 New code is modified to use the enhanced PKCS12_create()
5498 instead of the low level API.
5501 *) Extend ASN1 encoder to support indefinite length constructed
5502 encoding. This can output sequences tags and octet strings in
5503 this form. Modify pk7_asn1.c to support indefinite length
5504 encoding. This is experimental and needs additional code to
5505 be useful, such as an ASN1 bio and some enhanced streaming
5508 Extend template encode functionality so that tagging is passed
5509 down to the template encoder.
5512 *) Let 'openssl req' fail if an argument to '-newkey' is not
5513 recognized instead of using RSA as a default.
5516 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5517 As these are not official, they are not included in "ALL";
5518 the "ECCdraft" ciphersuite group alias can be used to select them.
5519 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5521 *) Add ECDH engine support.
5522 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5524 *) Add ECDH in new directory crypto/ecdh/.
5525 [Douglas Stebila (Sun Microsystems Laboratories)]
5527 *) Let BN_rand_range() abort with an error after 100 iterations
5528 without success (which indicates a broken PRNG).
5531 *) Change BN_mod_sqrt() so that it verifies that the input value
5532 is really the square of the return value. (Previously,
5533 BN_mod_sqrt would show GIGO behaviour.)
5536 *) Add named elliptic curves over binary fields from X9.62, SECG,
5537 and WAP/WTLS; add OIDs that were still missing.
5539 [Sheueling Chang Shantz and Douglas Stebila
5540 (Sun Microsystems Laboratories)]
5542 *) Extend the EC library for elliptic curves over binary fields
5543 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5546 EC_GF2m_simple_method
5550 EC_GROUP_new_curve_GF2m
5551 EC_GROUP_set_curve_GF2m
5552 EC_GROUP_get_curve_GF2m
5553 EC_POINT_set_affine_coordinates_GF2m
5554 EC_POINT_get_affine_coordinates_GF2m
5555 EC_POINT_set_compressed_coordinates_GF2m
5557 Point compression for binary fields is disabled by default for
5558 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5561 As binary polynomials are represented as BIGNUMs, various members
5562 of the EC_GROUP and EC_POINT data structures can be shared
5563 between the implementations for prime fields and binary fields;
5564 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5565 are essentially identical to their ..._GFp counterparts.
5566 (For simplicity, the '..._GFp' prefix has been dropped from
5567 various internal method names.)
5569 An internal 'field_div' method (similar to 'field_mul' and
5570 'field_sqr') has been added; this is used only for binary fields.
5572 [Sheueling Chang Shantz and Douglas Stebila
5573 (Sun Microsystems Laboratories)]
5575 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5576 through methods ('mul', 'precompute_mult').
5578 The generic implementations (now internally called 'ec_wNAF_mul'
5579 and 'ec_wNAF_precomputed_mult') remain the default if these
5580 methods are undefined.
5582 [Sheueling Chang Shantz and Douglas Stebila
5583 (Sun Microsystems Laboratories)]
5585 *) New function EC_GROUP_get_degree, which is defined through
5586 EC_METHOD. For curves over prime fields, this returns the bit
5587 length of the modulus.
5589 [Sheueling Chang Shantz and Douglas Stebila
5590 (Sun Microsystems Laboratories)]
5592 *) New functions EC_GROUP_dup, EC_POINT_dup.
5593 (These simply call ..._new and ..._copy).
5595 [Sheueling Chang Shantz and Douglas Stebila
5596 (Sun Microsystems Laboratories)]
5598 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5599 Polynomials are represented as BIGNUMs (where the sign bit is not
5600 used) in the following functions [macros]:
5603 BN_GF2m_sub [= BN_GF2m_add]
5604 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5605 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5606 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5608 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5609 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5610 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5611 BN_GF2m_cmp [= BN_ucmp]
5613 (Note that only the 'mod' functions are actually for fields GF(2^m).
5614 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5616 For some functions, an the irreducible polynomial defining a
5617 field can be given as an 'unsigned int[]' with strictly
5618 decreasing elements giving the indices of those bits that are set;
5619 i.e., p[] represents the polynomial
5620 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5622 p[0] > p[1] > ... > p[k] = 0.
5623 This applies to the following functions:
5628 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5629 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5631 BN_GF2m_mod_sqrt_arr
5632 BN_GF2m_mod_solve_quad_arr
5636 Conversion can be performed by the following functions:
5641 bntest.c has additional tests for binary polynomial arithmetic.
5643 Two implementations for BN_GF2m_mod_div() are available.
5644 The default algorithm simply uses BN_GF2m_mod_inv() and
5645 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5646 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5647 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5649 [Sheueling Chang Shantz and Douglas Stebila
5650 (Sun Microsystems Laboratories)]
5652 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5653 functionality is disabled at compile-time.
5654 [Douglas Stebila <douglas.stebila@sun.com>]
5656 *) Change default behaviour of 'openssl asn1parse' so that more
5657 information is visible when viewing, e.g., a certificate:
5659 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5660 mode the content of non-printable OCTET STRINGs is output in a
5661 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5662 avoid the appearance of a printable string.
5663 [Nils Larsch <nla@trustcenter.de>]
5665 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5667 EC_GROUP_set_asn1_flag()
5668 EC_GROUP_get_asn1_flag()
5669 EC_GROUP_set_point_conversion_form()
5670 EC_GROUP_get_point_conversion_form()
5671 These control ASN1 encoding details:
5672 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5673 has been set to OPENSSL_EC_NAMED_CURVE.
5674 - Points are encoded in uncompressed form by default; options for
5675 asn1_for are as for point2oct, namely
5676 POINT_CONVERSION_COMPRESSED
5677 POINT_CONVERSION_UNCOMPRESSED
5678 POINT_CONVERSION_HYBRID
5680 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5683 EC_GROUP_get0_seed()
5684 EC_GROUP_get_seed_len()
5685 This is used only for ASN1 purposes (so far).
5686 [Nils Larsch <nla@trustcenter.de>]
5688 *) Add 'field_type' member to EC_METHOD, which holds the NID
5689 of the appropriate field type OID. The new function
5690 EC_METHOD_get_field_type() returns this value.
5691 [Nils Larsch <nla@trustcenter.de>]
5696 EC_POINT_point2hex()
5697 EC_POINT_hex2point()
5698 providing useful interfaces to EC_POINT_point2oct() and
5699 EC_POINT_oct2point().
5700 [Nils Larsch <nla@trustcenter.de>]
5702 *) Change internals of the EC library so that the functions
5703 EC_GROUP_set_generator()
5704 EC_GROUP_get_generator()
5705 EC_GROUP_get_order()
5706 EC_GROUP_get_cofactor()
5707 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5708 to methods, which would lead to unnecessary code duplication when
5709 adding different types of curves.
5710 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5712 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5713 arithmetic, and such that modified wNAFs are generated
5714 (which avoid length expansion in many cases).
5717 *) Add a function EC_GROUP_check_discriminant() (defined via
5718 EC_METHOD) that verifies that the curve discriminant is non-zero.
5720 Add a function EC_GROUP_check() that makes some sanity tests
5721 on a EC_GROUP, its generator and order. This includes
5722 EC_GROUP_check_discriminant().
5723 [Nils Larsch <nla@trustcenter.de>]
5725 *) Add ECDSA in new directory crypto/ecdsa/.
5727 Add applications 'openssl ecparam' and 'openssl ecdsa'
5728 (these are based on 'openssl dsaparam' and 'openssl dsa').
5730 ECDSA support is also included in various other files across the
5731 library. Most notably,
5732 - 'openssl req' now has a '-newkey ecdsa:file' option;
5733 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5734 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5735 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5736 them suitable for ECDSA where domain parameters must be
5737 extracted before the specific public key;
5738 - ECDSA engine support has been added.
5739 [Nils Larsch <nla@trustcenter.de>]
5741 *) Include some named elliptic curves, and add OIDs from X9.62,
5742 SECG, and WAP/WTLS. Each curve can be obtained from the new
5744 EC_GROUP_new_by_curve_name(),
5745 and the list of available named curves can be obtained with
5746 EC_get_builtin_curves().
5747 Also add a 'curve_name' member to EC_GROUP objects, which can be
5749 EC_GROUP_set_curve_name()
5750 EC_GROUP_get_curve_name()
5751 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5753 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5754 was actually never needed) and in BN_mul(). The removal in BN_mul()
5755 required a small change in bn_mul_part_recursive() and the addition
5756 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5757 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5758 bn_sub_words() and bn_add_words() except they take arrays with
5762 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5764 *) Cleanse PEM buffers before freeing them since they may contain
5766 [Benjamin Bennett <ben@psc.edu>]
5768 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5769 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5770 authentication-only ciphersuites.
5773 *) Since AES128 and AES256 share a single mask bit in the logic of
5774 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5775 kludge to work properly if AES128 is available and AES256 isn't.
5778 *) Expand security boundary to match 1.1.1 module.
5781 *) Remove redundant features: hash file source, editing of test vectors
5782 modify fipsld to use external fips_premain.c signature.
5785 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5786 run algorithm test programs.
5789 *) Make algorithm test programs more tolerant of whitespace.
5792 *) Have SSL/TLS server implementation tolerate "mismatched" record
5793 protocol version while receiving ClientHello even if the
5794 ClientHello is fragmented. (The server can't insist on the
5795 particular protocol version it has chosen before the ServerHello
5796 message has informed the client about his choice.)
5799 *) Load error codes if they are not already present instead of using a
5800 static variable. This allows them to be cleanly unloaded and reloaded.
5803 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5805 *) Introduce limits to prevent malicious keys being able to
5806 cause a denial of service. (CVE-2006-2940)
5807 [Steve Henson, Bodo Moeller]
5809 *) Fix ASN.1 parsing of certain invalid structures that can result
5810 in a denial of service. (CVE-2006-2937) [Steve Henson]
5812 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5813 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5815 *) Fix SSL client code which could crash if connecting to a
5816 malicious SSLv2 server. (CVE-2006-4343)
5817 [Tavis Ormandy and Will Drewry, Google Security Team]
5819 *) Change ciphersuite string processing so that an explicit
5820 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5821 will no longer include "AES128-SHA"), and any other similar
5822 ciphersuite (same bitmap) from *other* protocol versions (so that
5823 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5824 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5825 changes from 0.9.8b and 0.9.8d.
5828 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5830 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5831 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5833 *) Change the Unix randomness entropy gathering to use poll() when
5834 possible instead of select(), since the latter has some
5835 undesirable limitations.
5836 [Darryl Miles via Richard Levitte and Bodo Moeller]
5838 *) Disable rogue ciphersuites:
5840 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5841 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5842 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5844 The latter two were purportedly from
5845 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5848 Also deactive the remaining ciphersuites from
5849 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5850 unofficial, and the ID has long expired.
5853 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5854 dual-core machines) and other potential thread-safety issues.
5857 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5859 *) Adapt fipsld and the build system to link against the validated FIPS
5860 module in FIPS mode.
5863 *) Fixes for VC++ 2005 build under Windows.
5866 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5867 from a Windows bash shell such as MSYS. It is autodetected from the
5868 "config" script when run from a VC++ environment. Modify standard VC++
5869 build to use fipscanister.o from the GNU make build.
5872 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5874 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5875 The value now differs depending on if you build for FIPS or not.
5876 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5877 safely run with a non-FIPSed libcrypto, as it may crash because of
5878 the difference induced by this change.
5881 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5883 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5884 (part of SSL_OP_ALL). This option used to disable the
5885 countermeasure against man-in-the-middle protocol-version
5886 rollback in the SSL 2.0 server implementation, which is a bad
5887 idea. (CVE-2005-2969)
5889 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5890 for Information Security, National Institute of Advanced Industrial
5891 Science and Technology [AIST], Japan)]
5893 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5894 mainly for FIPS compliance and not fully integrated at this stage.
5897 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5898 the exponentiation using a fixed-length exponent. (Otherwise,
5899 the information leaked through timing could expose the secret key
5900 after many signatures; cf. Bleichenbacher's attack on DSA with
5904 *) Make a new fixed-window mod_exp implementation the default for
5905 RSA, DSA, and DH private-key operations so that the sequence of
5906 squares and multiplies and the memory access pattern are
5907 independent of the particular secret key. This will mitigate
5908 cache-timing and potential related attacks.
5910 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5911 and this is automatically used by BN_mod_exp_mont() if the new flag
5912 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5913 will use this BN flag for private exponents unless the flag
5914 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5915 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5917 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5919 *) Change the client implementation for SSLv23_method() and
5920 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5921 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5922 (Previously, the SSL 2.0 backwards compatible Client Hello
5923 message format would be used even with SSL_OP_NO_SSLv2.)
5926 *) Add support for smime-type MIME parameter in S/MIME messages which some
5930 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5931 a threadsafe manner. Modify rsa code to use new function and add calls
5932 to dsa and dh code (which had race conditions before).
5935 *) Include the fixed error library code in the C error file definitions
5936 instead of fixing them up at runtime. This keeps the error code
5937 structures constant.
5940 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5942 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5945 *) Fixes for newer kerberos headers. NB: the casts are needed because
5946 the 'length' field is signed on one version and unsigned on another
5947 with no (?) obvious way to tell the difference, without these VC++
5948 complains. Also the "definition" of FAR (blank) is no longer included
5949 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5950 some needed definitions.
5953 *) Undo Cygwin change.
5956 *) Added support for proxy certificates according to RFC 3820.
5957 Because they may be a security thread to unaware applications,
5958 they must be explicitly allowed in run-time. See
5959 docs/HOWTO/proxy_certificates.txt for further information.
5962 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5964 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5965 server and client random values. Previously
5966 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5967 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5969 This change has negligible security impact because:
5971 1. Server and client random values still have 24 bytes of pseudo random
5974 2. Server and client random values are sent in the clear in the initial
5977 3. The master secret is derived using the premaster secret (48 bytes in
5978 size for static RSA ciphersuites) as well as client server and random
5981 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5984 [Stephen Henson, reported by UK NISCC]
5986 *) Use Windows randomness collection on Cygwin.
5989 *) Fix hang in EGD/PRNGD query when communication socket is closed
5990 prematurely by EGD/PRNGD.
5991 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5993 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5996 *) Back-port of selected performance improvements from development
5997 branch, as well as improved support for PowerPC platforms.
6000 *) Add lots of checks for memory allocation failure, error codes to indicate
6001 failure and freeing up memory if a failure occurs.
6002 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6004 *) Add new -passin argument to dgst.
6007 *) Perform some character comparisons of different types in X509_NAME_cmp:
6008 this is needed for some certificates that re-encode DNs into UTF8Strings
6009 (in violation of RFC3280) and can't or won't issue name rollover
6013 *) Make an explicit check during certificate validation to see that
6014 the CA setting in each certificate on the chain is correct. As a
6015 side effect always do the following basic checks on extensions,
6016 not just when there's an associated purpose to the check:
6018 - if there is an unhandled critical extension (unless the user
6019 has chosen to ignore this fault)
6020 - if the path length has been exceeded (if one is set at all)
6021 - that certain extensions fit the associated purpose (if one has
6025 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6027 *) Avoid a race condition when CRLs are checked in a multi threaded
6028 environment. This would happen due to the reordering of the revoked
6029 entries during signature checking and serial number lookup. Now the
6030 encoding is cached and the serial number sort performed under a lock.
6031 Add new STACK function sk_is_sorted().
6034 *) Add Delta CRL to the extension code.
6037 *) Various fixes to s3_pkt.c so alerts are sent properly.
6038 [David Holmes <d.holmes@f5.com>]
6040 *) Reduce the chances of duplicate issuer name and serial numbers (in
6041 violation of RFC3280) using the OpenSSL certificate creation utilities.
6042 This is done by creating a random 64 bit value for the initial serial
6043 number when a serial number file is created or when a self signed
6044 certificate is created using 'openssl req -x509'. The initial serial
6045 number file is created using 'openssl x509 -next_serial' in CA.pl
6046 rather than being initialized to 1.
6049 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6051 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6052 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6053 [Joe Orton, Steve Henson]
6055 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6057 [Joe Orton, Steve Henson]
6059 *) Make it possible to have multiple active certificates with the same
6060 subject in the CA index file. This is done only if the keyword
6061 'unique_subject' is set to 'no' in the main CA section (default
6062 if 'CA_default') of the configuration file. The value is saved
6063 with the database itself in a separate index attribute file,
6064 named like the index file with '.attr' appended to the name.
6067 *) X509 verify fixes. Disable broken certificate workarounds when
6068 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6069 keyUsage extension present. Don't accept CRLs with unhandled critical
6070 extensions: since verify currently doesn't process CRL extensions this
6071 rejects a CRL with *any* critical extensions. Add new verify error codes
6075 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6076 A clarification of RFC2560 will require the use of OCTET STRINGs and
6077 some implementations cannot handle the current raw format. Since OpenSSL
6078 copies and compares OCSP nonces as opaque blobs without any attempt at
6079 parsing them this should not create any compatibility issues.
6082 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6083 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6084 this HMAC (and other) operations are several times slower than OpenSSL
6088 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6089 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6091 *) Use the correct content when signing type "other".
6094 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6096 *) Fix various bugs revealed by running the NISCC test suite:
6098 Stop out of bounds reads in the ASN1 code when presented with
6099 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6101 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6103 If verify callback ignores invalid public key errors don't try to check
6104 certificate signature with the NULL public key.
6108 *) New -ignore_err option in ocsp application to stop the server
6109 exiting on the first error in a request.
6112 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6113 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6117 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6118 extra data after the compression methods not only for TLS 1.0
6119 but also for SSL 3.0 (as required by the specification).
6120 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6122 *) Change X509_certificate_type() to mark the key as exported/exportable
6123 when it's 512 *bits* long, not 512 bytes.
6126 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6127 blocks during encryption.
6130 *) Various fixes to base64 BIO and non blocking I/O. On write
6131 flushes were not handled properly if the BIO retried. On read
6132 data was not being buffered properly and had various logic bugs.
6133 This also affects blocking I/O when the data being decoded is a
6137 *) Various S/MIME bugfixes and compatibility changes:
6138 output correct application/pkcs7 MIME type if
6139 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6140 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6141 of files as .eml work). Correctly handle very long lines in MIME
6145 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6147 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6148 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6149 a protocol version number mismatch like a decryption error
6150 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6153 *) Turn on RSA blinding by default in the default implementation
6154 to avoid a timing attack. Applications that don't want it can call
6155 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6156 They would be ill-advised to do so in most cases.
6157 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6159 *) Change RSA blinding code so that it works when the PRNG is not
6160 seeded (in this case, the secret RSA exponent is abused as
6161 an unpredictable seed -- if it is not unpredictable, there
6162 is no point in blinding anyway). Make RSA blinding thread-safe
6163 by remembering the creator's thread ID in rsa->blinding and
6164 having all other threads use local one-time blinding factors
6165 (this requires more computation than sharing rsa->blinding, but
6166 avoids excessive locking; and if an RSA object is not shared
6167 between threads, blinding will still be very fast).
6170 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6171 ENGINE as defaults for all supported algorithms irrespective of
6172 the 'flags' parameter. 'flags' is now honoured, so applications
6173 should make sure they are passing it correctly.
6176 *) Target "mingw" now allows native Windows code to be generated in
6177 the Cygwin environment as well as with the MinGW compiler.
6180 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6182 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6183 via timing by performing a MAC computation even if incorrect
6184 block cipher padding has been found. This is a countermeasure
6185 against active attacks where the attacker has to distinguish
6186 between bad padding and a MAC verification error. (CVE-2003-0078)
6188 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6189 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6190 Martin Vuagnoux (EPFL, Ilion)]
6192 *) Make the no-err option work as intended. The intention with no-err
6193 is not to have the whole error stack handling routines removed from
6194 libcrypto, it's only intended to remove all the function name and
6195 reason texts, thereby removing some of the footprint that may not
6196 be interesting if those errors aren't displayed anyway.
6198 NOTE: it's still possible for any application or module to have it's
6199 own set of error texts inserted. The routines are there, just not
6200 used by default when no-err is given.
6203 *) Add support for FreeBSD on IA64.
6204 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6206 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6207 Kerberos function mit_des_cbc_cksum(). Before this change,
6208 the value returned by DES_cbc_cksum() was like the one from
6209 mit_des_cbc_cksum(), except the bytes were swapped.
6210 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6212 *) Allow an application to disable the automatic SSL chain building.
6213 Before this a rather primitive chain build was always performed in
6214 ssl3_output_cert_chain(): an application had no way to send the
6215 correct chain if the automatic operation produced an incorrect result.
6217 Now the chain builder is disabled if either:
6219 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6221 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6223 The reasoning behind this is that an application would not want the
6224 auto chain building to take place if extra chain certificates are
6225 present and it might also want a means of sending no additional
6226 certificates (for example the chain has two certificates and the
6230 *) Add the possibility to build without the ENGINE framework.
6231 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6233 *) Under Win32 gmtime() can return NULL: check return value in
6234 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6237 *) DSA routines: under certain error conditions uninitialized BN objects
6238 could be freed. Solution: make sure initialization is performed early
6239 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6240 Nils Larsch <nla@trustcenter.de> via PR#459)
6243 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6244 checked on reconnect on the client side, therefore session resumption
6245 could still fail with a "ssl session id is different" error. This
6246 behaviour is masked when SSL_OP_ALL is used due to
6247 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6248 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6249 followup to PR #377.
6252 *) IA-32 assembler support enhancements: unified ELF targets, support
6253 for SCO/Caldera platforms, fix for Cygwin shared build.
6256 *) Add support for FreeBSD on sparc64. As a consequence, support for
6257 FreeBSD on non-x86 processors is separate from x86 processors on
6258 the config script, much like the NetBSD support.
6259 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6261 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6263 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6266 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6267 code (06) was taken as the first octet of the session ID and the last
6268 octet was ignored consequently. As a result SSLv2 client side session
6269 caching could not have worked due to the session ID mismatch between
6271 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6275 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6276 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6280 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6281 seems that in spite of existing for more than a year, many application
6282 author have done nothing to provide the necessary callbacks, which
6283 means that this particular engine will not work properly anywhere.
6284 This is a very unfortunate situation which forces us, in the name
6285 of usability, to give the hw_ncipher.c a static lock, which is part
6287 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6288 appear in 0.9.8 or later. We EXPECT application authors to have
6289 dealt properly with this when 0.9.8 is released (unless we actually
6290 make such changes in the libcrypto locking code that changes will
6291 have to be made anyway).
6294 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6295 octets have been read, EOF or an error occurs. Without this change
6296 some truncated ASN1 structures will not produce an error.
6299 *) Disable Heimdal support, since it hasn't been fully implemented.
6300 Still give the possibility to force the use of Heimdal, but with
6301 warnings and a request that patches get sent to openssl-dev.
6304 *) Add the VC-CE target, introduce the WINCE sysname, and add
6305 INSTALL.WCE and appropriate conditionals to make it build.
6306 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6308 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6309 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6310 edit numbers of the version.
6311 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6313 *) Introduce safe string copy and catenation functions
6314 (BUF_strlcpy() and BUF_strlcat()).
6315 [Ben Laurie (CHATS) and Richard Levitte]
6317 *) Avoid using fixed-size buffers for one-line DNs.
6318 [Ben Laurie (CHATS)]
6320 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6321 resizing buffers containing secrets, and use where appropriate.
6322 [Ben Laurie (CHATS)]
6324 *) Avoid using fixed size buffers for configuration file location.
6325 [Ben Laurie (CHATS)]
6327 *) Avoid filename truncation for various CA files.
6328 [Ben Laurie (CHATS)]
6330 *) Use sizeof in preference to magic numbers.
6331 [Ben Laurie (CHATS)]
6333 *) Avoid filename truncation in cert requests.
6334 [Ben Laurie (CHATS)]
6336 *) Add assertions to check for (supposedly impossible) buffer
6338 [Ben Laurie (CHATS)]
6340 *) Don't cache truncated DNS entries in the local cache (this could
6341 potentially lead to a spoofing attack).
6342 [Ben Laurie (CHATS)]
6344 *) Fix various buffers to be large enough for hex/decimal
6345 representations in a platform independent manner.
6346 [Ben Laurie (CHATS)]
6348 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6349 resizing buffers containing secrets, and use where appropriate.
6350 [Ben Laurie (CHATS)]
6352 *) Add BIO_indent() to avoid much slightly worrying code to do
6354 [Ben Laurie (CHATS)]
6356 *) Convert sprintf()/BIO_puts() to BIO_printf().
6357 [Ben Laurie (CHATS)]
6359 *) buffer_gets() could terminate with the buffer only half
6361 [Ben Laurie (CHATS)]
6363 *) Add assertions to prevent user-supplied crypto functions from
6364 overflowing internal buffers by having large block sizes, etc.
6365 [Ben Laurie (CHATS)]
6367 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6369 [Ben Laurie (CHATS)]
6371 *) Eliminate unused copy of key in RC4.
6372 [Ben Laurie (CHATS)]
6374 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6375 [Ben Laurie (CHATS)]
6377 *) Fix off-by-one error in EGD path.
6378 [Ben Laurie (CHATS)]
6380 *) If RANDFILE path is too long, ignore instead of truncating.
6381 [Ben Laurie (CHATS)]
6383 *) Eliminate unused and incorrectly sized X.509 structure
6385 [Ben Laurie (CHATS)]
6387 *) Eliminate unused and dangerous function knumber().
6388 [Ben Laurie (CHATS)]
6390 *) Eliminate unused and dangerous structure, KSSL_ERR.
6391 [Ben Laurie (CHATS)]
6393 *) Protect against overlong session ID context length in an encoded
6394 session object. Since these are local, this does not appear to be
6396 [Ben Laurie (CHATS)]
6398 *) Change from security patch (see 0.9.6e below) that did not affect
6399 the 0.9.6 release series:
6401 Remote buffer overflow in SSL3 protocol - an attacker could
6402 supply an oversized master key in Kerberos-enabled versions.
6404 [Ben Laurie (CHATS)]
6406 *) Change the SSL kerb5 codes to match RFC 2712.
6409 *) Make -nameopt work fully for req and add -reqopt switch.
6410 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6412 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6413 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6415 *) Make sure tests can be performed even if the corresponding algorithms
6416 have been removed entirely. This was also the last step to make
6417 OpenSSL compilable with DJGPP under all reasonable conditions.
6418 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6420 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6421 to allow version independent disabling of normally unselected ciphers,
6422 which may be activated as a side-effect of selecting a single cipher.
6424 (E.g., cipher list string "RSA" enables ciphersuites that are left
6425 out of "ALL" because they do not provide symmetric encryption.
6426 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6427 [Lutz Jaenicke, Bodo Moeller]
6429 *) Add appropriate support for separate platform-dependent build
6430 directories. The recommended way to make a platform-dependent
6431 build directory is the following (tested on Linux), maybe with
6434 # Place yourself outside of the OpenSSL source tree. In
6435 # this example, the environment variable OPENSSL_SOURCE
6436 # is assumed to contain the absolute OpenSSL source directory.
6437 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6438 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6439 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6440 mkdir -p `dirname $F`
6441 ln -s $OPENSSL_SOURCE/$F $F
6444 To be absolutely sure not to disturb the source tree, a "make clean"
6445 is a good thing. If it isn't successful, don't worry about it,
6446 it probably means the source directory is very clean.
6449 *) Make sure any ENGINE control commands make local copies of string
6450 pointers passed to them whenever necessary. Otherwise it is possible
6451 the caller may have overwritten (or deallocated) the original string
6452 data when a later ENGINE operation tries to use the stored values.
6453 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6455 *) Improve diagnostics in file reading and command-line digests.
6456 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6458 *) Add AES modes CFB and OFB to the object database. Correct an
6459 error in AES-CFB decryption.
6462 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6463 allows existing EVP_CIPHER_CTX structures to be reused after
6464 calling EVP_*Final(). This behaviour is used by encryption
6465 BIOs and some applications. This has the side effect that
6466 applications must explicitly clean up cipher contexts with
6467 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6470 *) Check the values of dna and dnb in bn_mul_recursive before calling
6471 bn_mul_comba (a non zero value means the a or b arrays do not contain
6472 n2 elements) and fallback to bn_mul_normal if either is not zero.
6475 *) Fix escaping of non-ASCII characters when using the -subj option
6476 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6479 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6480 form for "surname", serialNumber has no short form.
6481 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6482 therefore remove "mail" short name for "internet 7".
6483 The OID for unique identifiers in X509 certificates is
6484 x500UniqueIdentifier, not uniqueIdentifier.
6485 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6488 *) Add an "init" command to the ENGINE config module and auto initialize
6489 ENGINEs. Without any "init" command the ENGINE will be initialized
6490 after all ctrl commands have been executed on it. If init=1 the
6491 ENGINE is initailized at that point (ctrls before that point are run
6492 on the uninitialized ENGINE and after on the initialized one). If
6493 init=0 then the ENGINE will not be iniatialized at all.
6496 *) Fix the 'app_verify_callback' interface so that the user-defined
6497 argument is actually passed to the callback: In the
6498 SSL_CTX_set_cert_verify_callback() prototype, the callback
6499 declaration has been changed from
6502 int (*cb)(X509_STORE_CTX *,void *);
6503 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6504 i=s->ctx->app_verify_callback(&ctx)
6505 has been changed into
6506 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6508 To update applications using SSL_CTX_set_cert_verify_callback(),
6509 a dummy argument can be added to their callback functions.
6510 [D. K. Smetters <smetters@parc.xerox.com>]
6512 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6513 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6515 *) Add and OPENSSL_LOAD_CONF define which will cause
6516 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6517 This allows older applications to transparently support certain
6518 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6519 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6520 load the config file and OPENSSL_add_all_algorithms_conf() which will
6521 always load it have also been added.
6524 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6525 Adjust NIDs and EVP layer.
6526 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6528 *) Config modules support in openssl utility.
6530 Most commands now load modules from the config file,
6531 though in a few (such as version) this isn't done
6532 because it couldn't be used for anything.
6534 In the case of ca and req the config file used is
6535 the same as the utility itself: that is the -config
6536 command line option can be used to specify an
6540 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6541 use "openssl_conf" if filename is NULL use default openssl config file.
6544 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6545 config section name. Add a new flag to tolerate a missing config file
6546 and move code to CONF_modules_load_file().
6549 *) Support for crypto accelerator cards from Accelerated Encryption
6550 Processing, www.aep.ie. (Use engine 'aep')
6551 The support was copied from 0.9.6c [engine] and adapted/corrected
6552 to work with the new engine framework.
6553 [AEP Inc. and Richard Levitte]
6555 *) Support for SureWare crypto accelerator cards from Baltimore
6556 Technologies. (Use engine 'sureware')
6557 The support was copied from 0.9.6c [engine] and adapted
6558 to work with the new engine framework.
6561 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6562 make the newer ENGINE framework commands for the CHIL engine work.
6563 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6565 *) Make it possible to produce shared libraries on ReliantUNIX.
6566 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6568 *) Add the configuration target debug-linux-ppro.
6569 Make 'openssl rsa' use the general key loading routines
6570 implemented in apps.c, and make those routines able to
6571 handle the key format FORMAT_NETSCAPE and the variant
6573 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6575 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6576 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6578 *) Add -keyform to rsautl, and document -engine.
6579 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6581 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6582 BIO_R_NO_SUCH_FILE error code rather than the generic
6583 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6586 *) Add new functions
6588 ERR_peek_last_error_line
6589 ERR_peek_last_error_line_data.
6590 These are similar to
6593 ERR_peek_error_line_data,
6594 but report on the latest error recorded rather than the first one
6595 still in the error queue.
6596 [Ben Laurie, Bodo Moeller]
6598 *) default_algorithms option in ENGINE config module. This allows things
6600 default_algorithms = ALL
6601 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6604 *) Preliminary ENGINE config module.
6607 *) New experimental application configuration code.
6610 *) Change the AES code to follow the same name structure as all other
6611 symmetric ciphers, and behave the same way. Move everything to
6612 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6613 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6615 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6616 [Ben Laurie and Theo de Raadt]
6618 *) Add option to output public keys in req command.
6619 [Massimiliano Pala madwolf@openca.org]
6621 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6622 (up to about 10% better than before for P-192 and P-224).
6625 *) New functions/macros
6627 SSL_CTX_set_msg_callback(ctx, cb)
6628 SSL_CTX_set_msg_callback_arg(ctx, arg)
6629 SSL_set_msg_callback(ssl, cb)
6630 SSL_set_msg_callback_arg(ssl, arg)
6632 to request calling a callback function
6634 void cb(int write_p, int version, int content_type,
6635 const void *buf, size_t len, SSL *ssl, void *arg)
6637 whenever a protocol message has been completely received
6638 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6639 protocol version according to which the SSL library interprets
6640 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6641 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6642 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6643 specification (change_cipher_spec(20), alert(21), handshake(22)).
6644 'buf' and 'len' point to the actual message, 'ssl' to the
6645 SSL object, and 'arg' is the application-defined value set by
6646 SSL[_CTX]_set_msg_callback_arg().
6648 'openssl s_client' and 'openssl s_server' have new '-msg' options
6649 to enable a callback that displays all protocol messages.
6652 *) Change the shared library support so shared libraries are built as
6653 soon as the corresponding static library is finished, and thereby get
6654 openssl and the test programs linked against the shared library.
6655 This still only happens when the keyword "shard" has been given to
6656 the configuration scripts.
6658 NOTE: shared library support is still an experimental thing, and
6659 backward binary compatibility is still not guaranteed.
6660 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6662 *) Add support for Subject Information Access extension.
6663 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6665 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6666 additional bytes when new memory had to be allocated, not just
6667 when reusing an existing buffer.
6670 *) New command line and configuration option 'utf8' for the req command.
6671 This allows field values to be specified as UTF8 strings.
6674 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6675 runs for the former and machine-readable output for the latter.
6678 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6679 of the e-mail address in the DN (i.e., it will go into a certificate
6680 extension only). The new configuration file option 'email_in_dn = no'
6681 has the same effect.
6682 [Massimiliano Pala madwolf@openca.org]
6684 *) Change all functions with names starting with des_ to be starting
6685 with DES_ instead. Add wrappers that are compatible with libdes,
6686 but are named _ossl_old_des_*. Finally, add macros that map the
6687 des_* symbols to the corresponding _ossl_old_des_* if libdes
6688 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6689 desired, the des_* symbols will be mapped to DES_*, with one
6692 Since we provide two compatibility mappings, the user needs to
6693 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6694 compatibility is desired. The default (i.e., when that macro
6695 isn't defined) is OpenSSL 0.9.6c compatibility.
6697 There are also macros that enable and disable the support of old
6698 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6699 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6700 are defined, the default will apply: to support the old des routines.
6702 In either case, one must include openssl/des.h to get the correct
6703 definitions. Do not try to just include openssl/des_old.h, that
6706 NOTE: This is a major break of an old API into a new one. Software
6707 authors are encouraged to switch to the DES_ style functions. Some
6708 time in the future, des_old.h and the libdes compatibility functions
6709 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6710 default), and then completely removed.
6713 *) Test for certificates which contain unsupported critical extensions.
6714 If such a certificate is found during a verify operation it is
6715 rejected by default: this behaviour can be overridden by either
6716 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6717 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6718 X509_supported_extension() has also been added which returns 1 if a
6719 particular extension is supported.
6722 *) Modify the behaviour of EVP cipher functions in similar way to digests
6723 to retain compatibility with existing code.
6726 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6727 compatibility with existing code. In particular the 'ctx' parameter does
6728 not have to be to be initialized before the call to EVP_DigestInit() and
6729 it is tidied up after a call to EVP_DigestFinal(). New function
6730 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6731 EVP_MD_CTX_copy() changed to not require the destination to be
6732 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6733 requires the destination to be valid.
6735 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6736 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6739 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6740 so that complete 'Handshake' protocol structures are kept in memory
6741 instead of overwriting 'msg_type' and 'length' with 'body' data.
6744 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6745 [Massimo Santin via Richard Levitte]
6747 *) Major restructuring to the underlying ENGINE code. This includes
6748 reduction of linker bloat, separation of pure "ENGINE" manipulation
6749 (initialisation, etc) from functionality dealing with implementations
6750 of specific crypto iterfaces. This change also introduces integrated
6751 support for symmetric ciphers and digest implementations - so ENGINEs
6752 can now accelerate these by providing EVP_CIPHER and EVP_MD
6753 implementations of their own. This is detailed in crypto/engine/README
6754 as it couldn't be adequately described here. However, there are a few
6755 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6756 were changed in the original introduction of ENGINE code have now
6757 reverted back - the hooking from this code to ENGINE is now a good
6758 deal more passive and at run-time, operations deal directly with
6759 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6760 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6761 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6762 they were not being used by the framework as there is no concept of a
6763 BIGNUM_METHOD and they could not be generalised to the new
6764 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6765 ENGINE_cpy() has been removed as it cannot be consistently defined in
6769 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6772 *) Change mkdef.pl to sort symbols that get the same entry number,
6773 and make sure the automatically generated functions ERR_load_*
6774 become part of libeay.num as well.
6777 *) New function SSL_renegotiate_pending(). This returns true once
6778 renegotiation has been requested (either SSL_renegotiate() call
6779 or HelloRequest/ClientHello received from the peer) and becomes
6780 false once a handshake has been completed.
6781 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6782 sends a HelloRequest, but does not ensure that a handshake takes
6783 place. SSL_renegotiate_pending() is useful for checking if the
6784 client has followed the request.)
6787 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6788 By default, clients may request session resumption even during
6789 renegotiation (if session ID contexts permit); with this option,
6790 session resumption is possible only in the first handshake.
6792 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6793 more bits available for options that should not be part of
6794 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6797 *) Add some demos for certificate and certificate request creation.
6800 *) Make maximum certificate chain size accepted from the peer application
6801 settable (SSL*_get/set_max_cert_list()), as proposed by
6802 "Douglas E. Engert" <deengert@anl.gov>.
6805 *) Add support for shared libraries for Unixware-7
6806 (Boyd Lynn Gerber <gerberb@zenez.com>).
6809 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6810 be done prior to destruction. Use this to unload error strings from
6811 ENGINEs that load their own error strings. NB: This adds two new API
6812 functions to "get" and "set" this destroy handler in an ENGINE.
6815 *) Alter all existing ENGINE implementations (except "openssl" and
6816 "openbsd") to dynamically instantiate their own error strings. This
6817 makes them more flexible to be built both as statically-linked ENGINEs
6818 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6819 Also, add stub code to each that makes building them as self-contained
6820 shared-libraries easier (see README.ENGINE).
6823 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6824 implementations into applications that are completely implemented in
6825 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6826 commands that can be used to configure what shared-library to load and
6827 to control aspects of the way it is handled. Also, made an update to
6828 the README.ENGINE file that brings its information up-to-date and
6829 provides some information and instructions on the "dynamic" ENGINE
6830 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6833 *) Make it possible to unload ranges of ERR strings with a new
6834 "ERR_unload_strings" function.
6837 *) Add a copy() function to EVP_MD.
6840 *) Make EVP_MD routines take a context pointer instead of just the
6841 md_data void pointer.
6844 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6845 that the digest can only process a single chunk of data
6846 (typically because it is provided by a piece of
6847 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6848 is only going to provide a single chunk of data, and hence the
6849 framework needn't accumulate the data for oneshot drivers.
6852 *) As with "ERR", make it possible to replace the underlying "ex_data"
6853 functions. This change also alters the storage and management of global
6854 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6855 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6856 index counters. The API functions that use this state have been changed
6857 to take a "class_index" rather than pointers to the class's local STACK
6858 and counter, and there is now an API function to dynamically create new
6859 classes. This centralisation allows us to (a) plug a lot of the
6860 thread-safety problems that existed, and (b) makes it possible to clean
6861 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6862 such data would previously have always leaked in application code and
6863 workarounds were in place to make the memory debugging turn a blind eye
6864 to it. Application code that doesn't use this new function will still
6865 leak as before, but their memory debugging output will announce it now
6866 rather than letting it slide.
6868 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6869 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6870 has a return value to indicate success or failure.
6873 *) Make it possible to replace the underlying "ERR" functions such that the
6874 global state (2 LHASH tables and 2 locks) is only used by the "default"
6875 implementation. This change also adds two functions to "get" and "set"
6876 the implementation prior to it being automatically set the first time
6877 any other ERR function takes place. Ie. an application can call "get",
6878 pass the return value to a module it has just loaded, and that module
6879 can call its own "set" function using that value. This means the
6880 module's "ERR" operations will use (and modify) the error state in the
6881 application and not in its own statically linked copy of OpenSSL code.
6884 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6885 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6886 the operation, and provides a more encapsulated way for external code
6887 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6888 to use these functions rather than manually incrementing the counts.
6890 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6893 *) Add EVP test program.
6896 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6899 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6900 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6901 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6902 These allow a CRL to be built without having to access X509_CRL fields
6903 directly. Modify 'ca' application to use new functions.
6906 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6907 bug workarounds. Rollback attack detection is a security feature.
6908 The problem will only arise on OpenSSL servers when TLSv1 is not
6909 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6910 Software authors not wanting to support TLSv1 will have special reasons
6911 for their choice and can explicitly enable this option.
6912 [Bodo Moeller, Lutz Jaenicke]
6914 *) Rationalise EVP so it can be extended: don't include a union of
6915 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6916 (similar to those existing for EVP_CIPHER_CTX).
6921 EVP_MD_CTX_init(&md); /* new function call */
6922 EVP_DigestInit(&md, EVP_sha1());
6923 EVP_DigestUpdate(&md, in, len);
6924 EVP_DigestFinal(&md, out, NULL);
6925 EVP_MD_CTX_cleanup(&md); /* new function call */
6929 *) Make DES key schedule conform to the usual scheme, as well as
6930 correcting its structure. This means that calls to DES functions
6931 now have to pass a pointer to a des_key_schedule instead of a
6932 plain des_key_schedule (which was actually always a pointer
6935 des_key_schedule ks;
6937 des_set_key_checked(..., &ks);
6938 des_ncbc_encrypt(..., &ks, ...);
6940 (Note that a later change renames 'des_...' into 'DES_...'.)
6943 *) Initial reduction of linker bloat: the use of some functions, such as
6944 PEM causes large amounts of unused functions to be linked in due to
6945 poor organisation. For example pem_all.c contains every PEM function
6946 which has a knock on effect of linking in large amounts of (unused)
6947 ASN1 code. Grouping together similar functions and splitting unrelated
6948 functions prevents this.
6951 *) Cleanup of EVP macros.
6954 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6955 correct _ecb suffix.
6958 *) Add initial OCSP responder support to ocsp application. The
6959 revocation information is handled using the text based index
6960 use by the ca application. The responder can either handle
6961 requests generated internally, supplied in files (for example
6962 via a CGI script) or using an internal minimal server.
6965 *) Add configuration choices to get zlib compression for TLS.
6968 *) Changes to Kerberos SSL for RFC 2712 compliance:
6969 1. Implemented real KerberosWrapper, instead of just using
6970 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6971 2. Implemented optional authenticator field of KerberosWrapper.
6973 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6974 and authenticator structs; see crypto/krb5/.
6976 Generalized Kerberos calls to support multiple Kerberos libraries.
6977 [Vern Staats <staatsvr@asc.hpc.mil>,
6978 Jeffrey Altman <jaltman@columbia.edu>
6979 via Richard Levitte]
6981 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6982 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6983 values for each of the key sizes rather than having just
6984 parameters (and 'speed' generating keys each time).
6987 *) Speed up EVP routines.
6990 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6991 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6992 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6993 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6995 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6996 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6997 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7000 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7002 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7005 *) Added the OS2-EMX target.
7006 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7008 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7009 to support NCONF routines in extension code. New function CONF_set_nconf()
7010 to allow functions which take an NCONF to also handle the old LHASH
7011 structure: this means that the old CONF compatible routines can be
7012 retained (in particular wrt extensions) without having to duplicate the
7013 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7016 *) Enhance the general user interface with mechanisms for inner control
7017 and with possibilities to have yes/no kind of prompts.
7020 *) Change all calls to low level digest routines in the library and
7021 applications to use EVP. Add missing calls to HMAC_cleanup() and
7022 don't assume HMAC_CTX can be copied using memcpy().
7023 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7025 *) Add the possibility to control engines through control names but with
7026 arbitrary arguments instead of just a string.
7027 Change the key loaders to take a UI_METHOD instead of a callback
7028 function pointer. NOTE: this breaks binary compatibility with earlier
7029 versions of OpenSSL [engine].
7030 Adapt the nCipher code for these new conditions and add a card insertion
7034 *) Enhance the general user interface with mechanisms to better support
7035 dialog box interfaces, application-defined prompts, the possibility
7036 to use defaults (for example default passwords from somewhere else)
7037 and interrupts/cancellations.
7040 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7041 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7044 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7045 tidy up some unnecessarily weird code in 'sk_new()').
7046 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7048 *) Change the key loading routines for ENGINEs to use the same kind
7049 callback (pem_password_cb) as all other routines that need this
7053 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7054 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7055 than this minimum value is recommended.
7058 *) New random seeder for OpenVMS, using the system process statistics
7059 that are easily reachable.
7062 *) Windows apparently can't transparently handle global
7063 variables defined in DLLs. Initialisations such as:
7065 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7067 won't compile. This is used by the any applications that need to
7068 declare their own ASN1 modules. This was fixed by adding the option
7069 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7070 needed for static libraries under Win32.
7073 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7074 setting of purpose and trust fields. New X509_STORE trust and
7075 purpose functions and tidy up setting in other SSL functions.
7078 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7079 structure. These are inherited by X509_STORE_CTX when it is
7080 initialised. This allows various defaults to be set in the
7081 X509_STORE structure (such as flags for CRL checking and custom
7082 purpose or trust settings) for functions which only use X509_STORE_CTX
7083 internally such as S/MIME.
7085 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7086 trust settings if they are not set in X509_STORE. This allows X509_STORE
7087 purposes and trust (in S/MIME for example) to override any set by default.
7089 Add command line options for CRL checking to smime, s_client and s_server
7093 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7094 are set then the CRL is looked up in the X509_STORE structure and
7095 its validity and signature checked, then if the certificate is found
7096 in the CRL the verify fails with a revoked error.
7098 Various new CRL related callbacks added to X509_STORE_CTX structure.
7100 Command line options added to 'verify' application to support this.
7102 This needs some additional work, such as being able to handle multiple
7103 CRLs with different times, extension based lookup (rather than just
7104 by subject name) and ultimately more complete V2 CRL extension
7108 *) Add a general user interface API (crypto/ui/). This is designed
7109 to replace things like des_read_password and friends (backward
7110 compatibility functions using this new API are provided).
7111 The purpose is to remove prompting functions from the DES code
7112 section as well as provide for prompting through dialog boxes in
7113 a window system and the like.
7116 *) Add "ex_data" support to ENGINE so implementations can add state at a
7117 per-structure level rather than having to store it globally.
7120 *) Make it possible for ENGINE structures to be copied when retrieved by
7121 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7122 This causes the "original" ENGINE structure to act like a template,
7123 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7124 operational state can be localised to each ENGINE structure, despite the
7125 fact they all share the same "methods". New ENGINE structures returned in
7126 this case have no functional references and the return value is the single
7127 structural reference. This matches the single structural reference returned
7128 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7132 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7133 needs to match any other type at all we need to manually clear the
7137 *) Changes to the "openssl engine" utility to include;
7138 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7139 about an ENGINE's available control commands.
7140 - executing control commands from command line arguments using the
7141 '-pre' and '-post' switches. '-post' is only used if '-t' is
7142 specified and the ENGINE is successfully initialised. The syntax for
7143 the individual commands are colon-separated, for example;
7144 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7147 *) New dynamic control command support for ENGINEs. ENGINEs can now
7148 declare their own commands (numbers), names (strings), descriptions,
7149 and input types for run-time discovery by calling applications. A
7150 subset of these commands are implicitly classed as "executable"
7151 depending on their input type, and only these can be invoked through
7152 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7153 can be based on user input, config files, etc). The distinction is
7154 that "executable" commands cannot return anything other than a boolean
7155 result and can only support numeric or string input, whereas some
7156 discoverable commands may only be for direct use through
7157 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7158 pointers, or other custom uses. The "executable" commands are to
7159 support parameterisations of ENGINE behaviour that can be
7160 unambiguously defined by ENGINEs and used consistently across any
7161 OpenSSL-based application. Commands have been added to all the
7162 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7163 control over shared-library paths without source code alterations.
7166 *) Changed all ENGINE implementations to dynamically allocate their
7167 ENGINEs rather than declaring them statically. Apart from this being
7168 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7169 this also allows the implementations to compile without using the
7170 internal engine_int.h header.
7173 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7174 'const' value. Any code that should be able to modify a RAND_METHOD
7175 should already have non-const pointers to it (ie. they should only
7176 modify their own ones).
7179 *) Made a variety of little tweaks to the ENGINE code.
7180 - "atalla" and "ubsec" string definitions were moved from header files
7181 to C code. "nuron" string definitions were placed in variables
7182 rather than hard-coded - allowing parameterisation of these values
7183 later on via ctrl() commands.
7184 - Removed unused "#if 0"'d code.
7185 - Fixed engine list iteration code so it uses ENGINE_free() to release
7186 structural references.
7187 - Constified the RAND_METHOD element of ENGINE structures.
7188 - Constified various get/set functions as appropriate and added
7189 missing functions (including a catch-all ENGINE_cpy that duplicates
7190 all ENGINE values onto a new ENGINE except reference counts/state).
7191 - Removed NULL parameter checks in get/set functions. Setting a method
7192 or function to NULL is a way of cancelling out a previously set
7193 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7194 and doesn't justify the extra error symbols and code.
7195 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7196 flags from engine_int.h to engine.h.
7197 - Changed prototypes for ENGINE handler functions (init(), finish(),
7198 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7201 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7202 to the algorithm using long division. The binary algorithm can be
7203 used only if the modulus is odd. On 32-bit systems, it is faster
7204 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7205 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7206 up to 450 bits. In 64-bit environments, the binary algorithm
7207 appears to be advantageous for much longer moduli; here we use it
7208 for moduli up to 2048 bits.
7211 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7212 could not support the combine flag in choice fields.
7215 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7216 extensions from a certificate request to the certificate.
7219 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7220 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7221 file: this allows the display of the certificate about to be
7222 signed to be customised, to allow certain fields to be included
7223 or excluded and extension details. The old system didn't display
7224 multicharacter strings properly, omitted fields not in the policy
7225 and couldn't display additional details such as extensions.
7228 *) Function EC_POINTs_mul for multiple scalar multiplication
7229 of an arbitrary number of elliptic curve points
7230 \sum scalars[i]*points[i],
7231 optionally including the generator defined for the EC_GROUP:
7232 scalar*generator + \sum scalars[i]*points[i].
7234 EC_POINT_mul is a simple wrapper function for the typical case
7235 that the point list has just one item (besides the optional
7239 *) First EC_METHODs for curves over GF(p):
7241 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7242 operations and provides various method functions that can also
7243 operate with faster implementations of modular arithmetic.
7245 EC_GFp_mont_method() reuses most functions that are part of
7246 EC_GFp_simple_method, but uses Montgomery arithmetic.
7248 [Bodo Moeller; point addition and point doubling
7249 implementation directly derived from source code provided by
7250 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7252 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7253 crypto/ec/ec_lib.c):
7255 Curves are EC_GROUP objects (with an optional group generator)
7256 based on EC_METHODs that are built into the library.
7258 Points are EC_POINT objects based on EC_GROUP objects.
7260 Most of the framework would be able to handle curves over arbitrary
7261 finite fields, but as there are no obvious types for fields other
7262 than GF(p), some functions are limited to that for now.
7265 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7266 that the file contains a complete HTTP response.
7269 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7270 change the def and num file printf format specifier from "%-40sXXX"
7271 to "%-39s XXX". The latter will always guarantee a space after the
7272 field while the former will cause them to run together if the field
7273 is 40 of more characters long.
7276 *) Constify the cipher and digest 'method' functions and structures
7277 and modify related functions to take constant EVP_MD and EVP_CIPHER
7281 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7282 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7285 *) Modify EVP_Digest*() routines so they now return values. Although the
7286 internal software routines can never fail additional hardware versions
7290 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7292 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7293 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7296 ERR_R_NESTED_ASN1_ERROR
7298 ERR_R_MISSING_ASN1_EOS
7299 were 4 .. 9, conflicting with
7300 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7302 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7303 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7305 Add new error code 'ERR_R_INTERNAL_ERROR'.
7308 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7312 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7313 sets the subject name for a new request or supersedes the
7314 subject name in a given request. Formats that can be parsed are
7315 'CN=Some Name, OU=myOU, C=IT'
7317 'CN=Some Name/OU=myOU/C=IT'.
7319 Add options '-batch' and '-verbose' to 'openssl req'.
7320 [Massimiliano Pala <madwolf@hackmasters.net>]
7322 *) Introduce the possibility to access global variables through
7323 functions on platform were that's the best way to handle exporting
7324 global variables in shared libraries. To enable this functionality,
7325 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7326 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7327 is normally done by Configure or something similar).
7329 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7330 in the source file (foo.c) like this:
7332 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7333 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7335 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7336 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7338 OPENSSL_DECLARE_GLOBAL(int,foo);
7339 #define foo OPENSSL_GLOBAL_REF(foo)
7340 OPENSSL_DECLARE_GLOBAL(double,bar);
7341 #define bar OPENSSL_GLOBAL_REF(bar)
7343 The #defines are very important, and therefore so is including the
7344 header file everywhere where the defined globals are used.
7346 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7347 of ASN.1 items, but that structure is a bit different.
7349 The largest change is in util/mkdef.pl which has been enhanced with
7350 better and easier to understand logic to choose which symbols should
7351 go into the Windows .def files as well as a number of fixes and code
7352 cleanup (among others, algorithm keywords are now sorted
7353 lexicographically to avoid constant rewrites).
7356 *) In BN_div() keep a copy of the sign of 'num' before writing the
7357 result to 'rm' because if rm==num the value will be overwritten
7358 and produce the wrong result if 'num' is negative: this caused
7359 problems with BN_mod() and BN_nnmod().
7362 *) Function OCSP_request_verify(). This checks the signature on an
7363 OCSP request and verifies the signer certificate. The signer
7364 certificate is just checked for a generic purpose and OCSP request
7368 *) Add OCSP_check_validity() function to check the validity of OCSP
7369 responses. OCSP responses are prepared in real time and may only
7370 be a few seconds old. Simply checking that the current time lies
7371 between thisUpdate and nextUpdate max reject otherwise valid responses
7372 caused by either OCSP responder or client clock inaccuracy. Instead
7373 we allow thisUpdate and nextUpdate to fall within a certain period of
7374 the current time. The age of the response can also optionally be
7375 checked. Two new options -validity_period and -status_age added to
7379 *) If signature or public key algorithm is unrecognized print out its
7380 OID rather that just UNKNOWN.
7383 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7384 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7385 ID to be generated from the issuer certificate alone which can then be
7386 passed to OCSP_id_issuer_cmp().
7389 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7390 ASN1 modules to export functions returning ASN1_ITEM pointers
7391 instead of the ASN1_ITEM structures themselves. This adds several
7392 new macros which allow the underlying ASN1 function/structure to
7393 be accessed transparently. As a result code should not use ASN1_ITEM
7394 references directly (such as &X509_it) but instead use the relevant
7395 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7396 use of the new ASN1 code on platforms where exporting structures
7397 is problematical (for example in shared libraries) but exporting
7398 functions returning pointers to structures is not.
7401 *) Add support for overriding the generation of SSL/TLS session IDs.
7402 These callbacks can be registered either in an SSL_CTX or per SSL.
7403 The purpose of this is to allow applications to control, if they wish,
7404 the arbitrary values chosen for use as session IDs, particularly as it
7405 can be useful for session caching in multiple-server environments. A
7406 command-line switch for testing this (and any client code that wishes
7407 to use such a feature) has been added to "s_server".
7408 [Geoff Thorpe, Lutz Jaenicke]
7410 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7411 of the form '#if defined(...) || defined(...) || ...' and
7412 '#if !defined(...) && !defined(...) && ...'. This also avoids
7413 the growing number of special cases it was previously handling.
7416 *) Make all configuration macros available for application by making
7417 sure they are available in opensslconf.h, by giving them names starting
7418 with "OPENSSL_" to avoid conflicts with other packages and by making
7419 sure e_os2.h will cover all platform-specific cases together with
7421 Additionally, it is now possible to define configuration/platform-
7422 specific names (called "system identities"). In the C code, these
7423 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7424 macro with the name beginning with "OPENSSL_SYS_", which is determined
7425 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7429 *) New option -set_serial to 'req' and 'x509' this allows the serial
7430 number to use to be specified on the command line. Previously self
7431 signed certificates were hard coded with serial number 0 and the
7432 CA options of 'x509' had to use a serial number in a file which was
7436 *) New options to 'ca' utility to support V2 CRL entry extensions.
7437 Currently CRL reason, invalidity date and hold instruction are
7438 supported. Add new CRL extensions to V3 code and some new objects.
7441 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7442 disable standard block padding (aka PKCS#5 padding) in the EVP
7443 API, which was previously mandatory. This means that the data is
7444 not padded in any way and so the total length much be a multiple
7445 of the block size, otherwise an error occurs.
7448 *) Initial (incomplete) OCSP SSL support.
7451 *) New function OCSP_parse_url(). This splits up a URL into its host,
7452 port and path components: primarily to parse OCSP URLs. New -url
7453 option to ocsp utility.
7456 *) New nonce behavior. The return value of OCSP_check_nonce() now
7457 reflects the various checks performed. Applications can decide
7458 whether to tolerate certain situations such as an absent nonce
7459 in a response when one was present in a request: the ocsp application
7460 just prints out a warning. New function OCSP_add1_basic_nonce()
7461 this is to allow responders to include a nonce in a response even if
7462 the request is nonce-less.
7465 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7466 skipped when using openssl x509 multiple times on a single input file,
7467 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7470 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7471 set string type: to handle setting ASN1_TIME structures. Fix ca
7472 utility to correctly initialize revocation date of CRLs.
7475 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7476 the clients preferred ciphersuites and rather use its own preferences.
7477 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7478 Internet Explorer by ensuring unchanged hash method during stepup.
7479 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7482 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7483 to aes and add a new 'exist' option to print out symbols that don't
7487 *) Additional options to ocsp utility to allow flags to be set and
7488 additional certificates supplied.
7491 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7492 OCSP client a number of certificate to only verify the response
7496 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7497 handle the new API. Currently only ECB, CBC modes supported. Add new
7500 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7501 Encryption Standard (AES) Ciphersuites for Transport Layer
7502 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7503 not enabled by default and were not part of the "ALL" ciphersuite
7504 alias because they were not yet official; they could be
7505 explicitly requested by specifying the "AESdraft" ciphersuite
7506 group alias. In the final release of OpenSSL 0.9.7, the group
7507 alias is called "AES" and is part of "ALL".)
7508 [Ben Laurie, Steve Henson, Bodo Moeller]
7510 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7511 request to response.
7514 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7515 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7516 extract information from a certificate request. OCSP_response_create()
7517 creates a response and optionally adds a basic response structure.
7518 OCSP_basic_add1_status() adds a complete single response to a basic
7519 response and returns the OCSP_SINGLERESP structure just added (to allow
7520 extensions to be included for example). OCSP_basic_add1_cert() adds a
7521 certificate to a basic response and OCSP_basic_sign() signs a basic
7522 response with various flags. New helper functions ASN1_TIME_check()
7523 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7524 (converts ASN1_TIME to GeneralizedTime).
7527 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7528 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7529 structure from a certificate. X509_pubkey_digest() digests the public_key
7530 contents: this is used in various key identifiers.
7533 *) Make sk_sort() tolerate a NULL argument.
7534 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7536 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7537 passed by the function are trusted implicitly. If any of them signed the
7538 response then it is assumed to be valid and is not verified.
7541 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7542 to data. This was previously part of the PKCS7 ASN1 code. This
7543 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7544 [Steve Henson, reported by Kenneth R. Robinette
7545 <support@securenetterm.com>]
7547 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7548 routines: without these tracing memory leaks is very painful.
7549 Fix leaks in PKCS12 and PKCS7 routines.
7552 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7553 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7554 effectively meant GeneralizedTime would never be used. Now it
7555 is initialised to -1 but X509_time_adj() now has to check the value
7556 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7557 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7558 [Steve Henson, reported by Kenneth R. Robinette
7559 <support@securenetterm.com>]
7561 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7562 result in a zero length in the ASN1_INTEGER structure which was
7563 not consistent with the structure when d2i_ASN1_INTEGER() was used
7564 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7565 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7566 where it did not print out a minus for negative ASN1_INTEGER.
7569 *) Add summary printout to ocsp utility. The various functions which
7570 convert status values to strings have been renamed to:
7571 OCSP_response_status_str(), OCSP_cert_status_str() and
7572 OCSP_crl_reason_str() and are no longer static. New options
7573 to verify nonce values and to disable verification. OCSP response
7574 printout format cleaned up.
7577 *) Add additional OCSP certificate checks. These are those specified
7578 in RFC2560. This consists of two separate checks: the CA of the
7579 certificate being checked must either be the OCSP signer certificate
7580 or the issuer of the OCSP signer certificate. In the latter case the
7581 OCSP signer certificate must contain the OCSP signing extended key
7582 usage. This check is performed by attempting to match the OCSP
7583 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7584 in the OCSP_CERTID structures of the response.
7587 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7588 and related routines. This uses the standard OpenSSL certificate
7589 verify routines to perform initial checks (just CA validity) and
7590 to obtain the certificate chain. Then additional checks will be
7591 performed on the chain. Currently the root CA is checked to see
7592 if it is explicitly trusted for OCSP signing. This is used to set
7593 a root CA as a global signing root: that is any certificate that
7594 chains to that CA is an acceptable OCSP signing certificate.
7597 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7598 extensions from a separate configuration file.
7599 As when reading extensions from the main configuration file,
7600 the '-extensions ...' option may be used for specifying the
7602 [Massimiliano Pala <madwolf@comune.modena.it>]
7604 *) New OCSP utility. Allows OCSP requests to be generated or
7605 read. The request can be sent to a responder and the output
7606 parsed, outputed or printed in text form. Not complete yet:
7607 still needs to check the OCSP response validity.
7610 *) New subcommands for 'openssl ca':
7611 'openssl ca -status <serial>' prints the status of the cert with
7612 the given serial number (according to the index file).
7613 'openssl ca -updatedb' updates the expiry status of certificates
7615 [Massimiliano Pala <madwolf@comune.modena.it>]
7617 *) New '-newreq-nodes' command option to CA.pl. This is like
7618 '-newreq', but calls 'openssl req' with the '-nodes' option
7619 so that the resulting key is not encrypted.
7620 [Damien Miller <djm@mindrot.org>]
7622 *) New configuration for the GNU Hurd.
7623 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7625 *) Initial code to implement OCSP basic response verify. This
7626 is currently incomplete. Currently just finds the signer's
7627 certificate and verifies the signature on the response.
7630 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7631 value of OPENSSLDIR. This is available via the new '-d' option
7632 to 'openssl version', and is also included in 'openssl version -a'.
7635 *) Allowing defining memory allocation callbacks that will be given
7636 file name and line number information in additional arguments
7637 (a const char* and an int). The basic functionality remains, as
7638 well as the original possibility to just replace malloc(),
7639 realloc() and free() by functions that do not know about these
7640 additional arguments. To register and find out the current
7641 settings for extended allocation functions, the following
7642 functions are provided:
7644 CRYPTO_set_mem_ex_functions
7645 CRYPTO_set_locked_mem_ex_functions
7646 CRYPTO_get_mem_ex_functions
7647 CRYPTO_get_locked_mem_ex_functions
7649 These work the same way as CRYPTO_set_mem_functions and friends.
7650 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7651 extended allocation function is enabled.
7652 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7653 a conventional allocation function is enabled.
7654 [Richard Levitte, Bodo Moeller]
7656 *) Finish off removing the remaining LHASH function pointer casts.
7657 There should no longer be any prototype-casting required when using
7658 the LHASH abstraction, and any casts that remain are "bugs". See
7659 the callback types and macros at the head of lhash.h for details
7660 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7663 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7664 If /dev/[u]random devices are not available or do not return enough
7665 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7667 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7668 /etc/entropy will be queried once each in this sequence, querying stops
7669 when enough entropy was collected without querying more sockets.
7672 *) Change the Unix RAND_poll() variant to be able to poll several
7673 random devices, as specified by DEVRANDOM, until a sufficient amount
7674 of data has been collected. We spend at most 10 ms on each file
7675 (select timeout) and read in non-blocking mode. DEVRANDOM now
7676 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7677 (previously it was just the string "/dev/urandom"), so on typical
7678 platforms the 10 ms delay will never occur.
7679 Also separate out the Unix variant to its own file, rand_unix.c.
7680 For VMS, there's a currently-empty rand_vms.c.
7683 *) Move OCSP client related routines to ocsp_cl.c. These
7684 provide utility functions which an application needing
7685 to issue a request to an OCSP responder and analyse the
7686 response will typically need: as opposed to those which an
7687 OCSP responder itself would need which will be added later.
7689 OCSP_request_sign() signs an OCSP request with an API similar
7690 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7691 response. OCSP_response_get1_basic() extracts basic response
7692 from response. OCSP_resp_find_status(): finds and extracts status
7693 information from an OCSP_CERTID structure (which will be created
7694 when the request structure is built). These are built from lower
7695 level functions which work on OCSP_SINGLERESP structures but
7696 won't normally be used unless the application wishes to examine
7697 extensions in the OCSP response for example.
7699 Replace nonce routines with a pair of functions.
7700 OCSP_request_add1_nonce() adds a nonce value and optionally
7701 generates a random value. OCSP_check_nonce() checks the
7702 validity of the nonce in an OCSP response.
7705 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7706 This doesn't copy the supplied OCSP_CERTID and avoids the
7707 need to free up the newly created id. Change return type
7708 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7709 This can then be used to add extensions to the request.
7710 Deleted OCSP_request_new(), since most of its functionality
7711 is now in OCSP_REQUEST_new() (and the case insensitive name
7712 clash) apart from the ability to set the request name which
7713 will be added elsewhere.
7716 *) Update OCSP API. Remove obsolete extensions argument from
7717 various functions. Extensions are now handled using the new
7718 OCSP extension code. New simple OCSP HTTP function which
7719 can be used to send requests and parse the response.
7722 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7723 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7724 uses the special reorder version of SET OF to sort the attributes
7725 and reorder them to match the encoded order. This resolves a long
7726 standing problem: a verify on a PKCS7 structure just after signing
7727 it used to fail because the attribute order did not match the
7728 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7729 it uses the received order. This is necessary to tolerate some broken
7730 software that does not order SET OF. This is handled by encoding
7731 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7732 to produce the required SET OF.
7735 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7736 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7737 files to get correct declarations of the ASN.1 item variables.
7740 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7741 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7742 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7743 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7744 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7745 ASN1_ITEM and no wrapper functions.
7748 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7749 replace the old function pointer based I/O routines. Change most of
7750 the *_d2i_bio() and *_d2i_fp() functions to use these.
7753 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7754 lines, recognice more "algorithms" that can be deselected, and make
7755 it complain about algorithm deselection that isn't recognised.
7758 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7759 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7760 to use new functions. Add NO_ASN1_OLD which can be set to remove
7761 some old style ASN1 functions: this can be used to determine if old
7762 code will still work when these eventually go away.
7765 *) New extension functions for OCSP structures, these follow the
7766 same conventions as certificates and CRLs.
7769 *) New function X509V3_add1_i2d(). This automatically encodes and
7770 adds an extension. Its behaviour can be customised with various
7771 flags to append, replace or delete. Various wrappers added for
7772 certificates and CRLs.
7775 *) Fix to avoid calling the underlying ASN1 print routine when
7776 an extension cannot be parsed. Correct a typo in the
7777 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7780 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7781 entries for variables.
7784 *) Add functionality to apps/openssl.c for detecting locking
7785 problems: As the program is single-threaded, all we have
7786 to do is register a locking callback using an array for
7787 storing which locks are currently held by the program.
7790 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7791 SSL_get_ex_data_X509_STORE_idx(), which is used in
7792 ssl_verify_cert_chain() and thus can be called at any time
7793 during TLS/SSL handshakes so that thread-safety is essential.
7794 Unfortunately, the ex_data design is not at all suited
7795 for multi-threaded use, so it probably should be abolished.
7798 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7799 [Broadcom, tweaked and integrated by Geoff Thorpe]
7801 *) Move common extension printing code to new function
7802 X509V3_print_extensions(). Reorganise OCSP print routines and
7803 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7806 *) New function X509_signature_print() to remove duplication in some
7810 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7811 set (this was treated exactly the same as SET OF previously). This
7812 is used to reorder the STACK representing the structure to match the
7813 encoding. This will be used to get round a problem where a PKCS7
7814 structure which was signed could not be verified because the STACK
7815 order did not reflect the encoded order.
7818 *) Reimplement the OCSP ASN1 module using the new code.
7821 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7822 for its ASN1 operations. The old style function pointers still exist
7823 for now but they will eventually go away.
7826 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7827 completely replaces the old ASN1 functionality with a table driven
7828 encoder and decoder which interprets an ASN1_ITEM structure describing
7829 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7830 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7831 has also been converted to the new form.
7834 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7835 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7836 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7837 for negative moduli.
7840 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7841 of not touching the result's sign bit.
7844 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7848 *) Changed the LHASH code to use prototypes for callbacks, and created
7849 macros to declare and implement thin (optionally static) functions
7850 that provide type-safety and avoid function pointer casting for the
7851 type-specific callbacks.
7854 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7856 [Veers Staats <staatsvr@asc.hpc.mil>,
7857 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7859 *) Reformat the FAQ so the different questions and answers can be divided
7860 in sections depending on the subject.
7863 *) Have the zlib compression code load ZLIB.DLL dynamically under
7867 *) New function BN_mod_sqrt for computing square roots modulo a prime
7868 (using the probabilistic Tonelli-Shanks algorithm unless
7869 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7870 be handled deterministically).
7871 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7873 *) Make BN_mod_inverse faster by explicitly handling small quotients
7874 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7875 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7878 *) New function BN_kronecker.
7881 *) Fix BN_gcd so that it works on negative inputs; the result is
7882 positive unless both parameters are zero.
7883 Previously something reasonably close to an infinite loop was
7884 possible because numbers could be growing instead of shrinking
7885 in the implementation of Euclid's algorithm.
7888 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7889 sign of the number in question.
7891 Fix BN_is_word(a,w) to work correctly for w == 0.
7893 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7894 because its test if the absolute value of 'a' equals 'w'.
7895 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7896 it exists mostly for use in the implementations of BN_is_zero(),
7897 BN_is_one(), and BN_is_word().
7900 *) New function BN_swap.
7903 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7904 the exponentiation functions are more likely to produce reasonable
7905 results on negative inputs.
7908 *) Change BN_mod_mul so that the result is always non-negative.
7909 Previously, it could be negative if one of the factors was negative;
7910 I don't think anyone really wanted that behaviour.
7913 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7914 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7915 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7916 and add new functions:
7925 BN_mod_lshift1_quick
7929 These functions always generate non-negative results.
7931 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7932 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7934 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7935 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7936 be reduced modulo m.
7937 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7940 The following entry accidentally appeared in the CHANGES file
7941 distributed with OpenSSL 0.9.7. The modifications described in
7942 it do *not* apply to OpenSSL 0.9.7.
7944 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7945 was actually never needed) and in BN_mul(). The removal in BN_mul()
7946 required a small change in bn_mul_part_recursive() and the addition
7947 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7948 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7949 bn_sub_words() and bn_add_words() except they take arrays with
7954 *) In 'openssl passwd', verify passwords read from the terminal
7955 unless the '-salt' option is used (which usually means that
7956 verification would just waste user's time since the resulting
7957 hash is going to be compared with some given password hash)
7958 or the new '-noverify' option is used.
7960 This is an incompatible change, but it does not affect
7961 non-interactive use of 'openssl passwd' (passwords on the command
7962 line, '-stdin' option, '-in ...' option) and thus should not
7966 *) Remove all references to RSAref, since there's no more need for it.
7969 *) Make DSO load along a path given through an environment variable
7970 (SHLIB_PATH) with shl_load().
7973 *) Constify the ENGINE code as a result of BIGNUM constification.
7974 Also constify the RSA code and most things related to it. In a
7975 few places, most notable in the depth of the ASN.1 code, ugly
7976 casts back to non-const were required (to be solved at a later
7980 *) Make it so the openssl application has all engines loaded by default.
7983 *) Constify the BIGNUM routines a little more.
7986 *) Add the following functions:
7988 ENGINE_load_cswift()
7990 ENGINE_load_atalla()
7992 ENGINE_load_builtin_engines()
7994 That way, an application can itself choose if external engines that
7995 are built-in in OpenSSL shall ever be used or not. The benefit is
7996 that applications won't have to be linked with libdl or other dso
7997 libraries unless it's really needed.
7999 Changed 'openssl engine' to load all engines on demand.
8000 Changed the engine header files to avoid the duplication of some
8001 declarations (they differed!).
8004 *) 'openssl engine' can now list capabilities.
8007 *) Better error reporting in 'openssl engine'.
8010 *) Never call load_dh_param(NULL) in s_server.
8013 *) Add engine application. It can currently list engines by name and
8014 identity, and test if they are actually available.
8017 *) Improve RPM specification file by forcing symbolic linking and making
8018 sure the installed documentation is also owned by root.root.
8019 [Damien Miller <djm@mindrot.org>]
8021 *) Give the OpenSSL applications more possibilities to make use of
8022 keys (public as well as private) handled by engines.
8025 *) Add OCSP code that comes from CertCo.
8028 *) Add VMS support for the Rijndael code.
8031 *) Added untested support for Nuron crypto accelerator.
8034 *) Add support for external cryptographic devices. This code was
8035 previously distributed separately as the "engine" branch.
8036 [Geoff Thorpe, Richard Levitte]
8038 *) Rework the filename-translation in the DSO code. It is now possible to
8039 have far greater control over how a "name" is turned into a filename
8040 depending on the operating environment and any oddities about the
8041 different shared library filenames on each system.
8044 *) Support threads on FreeBSD-elf in Configure.
8047 *) Fix for SHA1 assembly problem with MASM: it produces
8048 warnings about corrupt line number information when assembling
8049 with debugging information. This is caused by the overlapping
8051 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8054 NCONF_get_number() has no error checking at all. As a replacement,
8055 NCONF_get_number_e() is defined (_e for "error checking") and is
8056 promoted strongly. The old NCONF_get_number is kept around for
8057 binary backward compatibility.
8058 Make it possible for methods to load from something other than a BIO,
8059 by providing a function pointer that is given a name instead of a BIO.
8060 For example, this could be used to load configuration data from an
8064 *) Fix for non blocking accept BIOs. Added new I/O special reason
8065 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8066 with non blocking I/O was not possible because no retry code was
8067 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8071 *) Added the beginnings of Rijndael support.
8074 *) Fix for bug in DirectoryString mask setting. Add support for
8075 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8076 to allow certificate printing to more controllable, additional
8077 'certopt' option to 'x509' to allow new printing options to be
8081 *) Clean old EAY MD5 hack from e_os.h.
8084 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8086 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8087 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8088 [Joe Orton, Steve Henson]
8090 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8092 *) Fix additional bug revealed by the NISCC test suite:
8094 Stop bug triggering large recursion when presented with
8095 certain ASN.1 tags (CVE-2003-0851)
8098 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8100 *) Fix various bugs revealed by running the NISCC test suite:
8102 Stop out of bounds reads in the ASN1 code when presented with
8103 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8105 If verify callback ignores invalid public key errors don't try to check
8106 certificate signature with the NULL public key.
8110 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8111 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8115 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8116 extra data after the compression methods not only for TLS 1.0
8117 but also for SSL 3.0 (as required by the specification).
8118 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8120 *) Change X509_certificate_type() to mark the key as exported/exportable
8121 when it's 512 *bits* long, not 512 bytes.
8124 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8126 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8127 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8128 a protocol version number mismatch like a decryption error
8129 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8132 *) Turn on RSA blinding by default in the default implementation
8133 to avoid a timing attack. Applications that don't want it can call
8134 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8135 They would be ill-advised to do so in most cases.
8136 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8138 *) Change RSA blinding code so that it works when the PRNG is not
8139 seeded (in this case, the secret RSA exponent is abused as
8140 an unpredictable seed -- if it is not unpredictable, there
8141 is no point in blinding anyway). Make RSA blinding thread-safe
8142 by remembering the creator's thread ID in rsa->blinding and
8143 having all other threads use local one-time blinding factors
8144 (this requires more computation than sharing rsa->blinding, but
8145 avoids excessive locking; and if an RSA object is not shared
8146 between threads, blinding will still be very fast).
8149 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8151 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8152 via timing by performing a MAC computation even if incorrrect
8153 block cipher padding has been found. This is a countermeasure
8154 against active attacks where the attacker has to distinguish
8155 between bad padding and a MAC verification error. (CVE-2003-0078)
8157 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8158 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8159 Martin Vuagnoux (EPFL, Ilion)]
8161 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8163 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8164 memory from it's contents. This is done with a counter that will
8165 place alternating values in each byte. This can be used to solve
8166 two issues: 1) the removal of calls to memset() by highly optimizing
8167 compilers, and 2) cleansing with other values than 0, since those can
8168 be read through on certain media, for example a swap space on disk.
8171 *) Bugfix: client side session caching did not work with external caching,
8172 because the session->cipher setting was not restored when reloading
8173 from the external cache. This problem was masked, when
8174 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8175 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8178 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8179 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8180 [Zeev Lieber <zeev-l@yahoo.com>]
8182 *) Undo an undocumented change introduced in 0.9.6e which caused
8183 repeated calls to OpenSSL_add_all_ciphers() and
8184 OpenSSL_add_all_digests() to be ignored, even after calling
8188 *) Change the default configuration reader to deal with last line not
8189 being properly terminated.
8192 *) Change X509_NAME_cmp() so it applies the special rules on handling
8193 DN values that are of type PrintableString, as well as RDNs of type
8194 emailAddress where the value has the type ia5String.
8195 [stefank@valicert.com via Richard Levitte]
8197 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8198 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8199 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8200 the bitwise-OR of the two for use by the majority of applications
8201 wanting this behaviour, and update the docs. The documented
8202 behaviour and actual behaviour were inconsistent and had been
8203 changing anyway, so this is more a bug-fix than a behavioural
8205 [Geoff Thorpe, diagnosed by Nadav Har'El]
8207 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8208 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8211 *) Fix initialization code race conditions in
8212 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8213 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8214 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8215 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8216 ssl2_get_cipher_by_char(),
8217 ssl3_get_cipher_by_char().
8218 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8220 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8221 the cached sessions are flushed, as the remove_cb() might use ex_data
8222 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8223 (see [openssl.org #212]).
8224 [Geoff Thorpe, Lutz Jaenicke]
8226 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8227 length, instead of the encoding length to d2i_ASN1_OBJECT.
8230 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8232 *) [In 0.9.6g-engine release:]
8233 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8234 [Lynn Gazis <lgazis@rainbow.com>]
8236 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8238 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8239 and get fix the header length calculation.
8240 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8241 Alon Kantor <alonk@checkpoint.com> (and others),
8244 *) Use proper error handling instead of 'assertions' in buffer
8245 overflow checks added in 0.9.6e. This prevents DoS (the
8246 assertions could call abort()).
8247 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8249 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8251 *) Add various sanity checks to asn1_get_length() to reject
8252 the ASN1 length bytes if they exceed sizeof(long), will appear
8253 negative or the content length exceeds the length of the
8255 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8257 *) Fix cipher selection routines: ciphers without encryption had no flags
8258 for the cipher strength set and where therefore not handled correctly
8259 by the selection routines (PR #130).
8262 *) Fix EVP_dsa_sha macro.
8266 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8267 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8268 that was added in OpenSSL 0.9.6d.
8270 As the countermeasure turned out to be incompatible with some
8271 broken SSL implementations, the new option is part of SSL_OP_ALL.
8272 SSL_OP_ALL is usually employed when compatibility with weird SSL
8273 implementations is desired (e.g. '-bugs' option to 's_client' and
8274 's_server'), so the new option is automatically set in many
8278 *) Changes in security patch:
8280 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8281 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8282 Air Force Materiel Command, USAF, under agreement number
8285 *) Add various sanity checks to asn1_get_length() to reject
8286 the ASN1 length bytes if they exceed sizeof(long), will appear
8287 negative or the content length exceeds the length of the
8288 supplied buffer. (CVE-2002-0659)
8289 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8291 *) Assertions for various potential buffer overflows, not known to
8293 [Ben Laurie (CHATS)]
8295 *) Various temporary buffers to hold ASCII versions of integers were
8296 too small for 64 bit platforms. (CVE-2002-0655)
8297 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8299 *) Remote buffer overflow in SSL3 protocol - an attacker could
8300 supply an oversized session ID to a client. (CVE-2002-0656)
8301 [Ben Laurie (CHATS)]
8303 *) Remote buffer overflow in SSL2 protocol - an attacker could
8304 supply an oversized client master key. (CVE-2002-0656)
8305 [Ben Laurie (CHATS)]
8307 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8309 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8310 encoded as NULL) with id-dsa-with-sha1.
8311 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8313 *) Check various X509_...() return values in apps/req.c.
8314 [Nils Larsch <nla@trustcenter.de>]
8316 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8317 an end-of-file condition would erroneously be flagged, when the CRLF
8318 was just at the end of a processed block. The bug was discovered when
8319 processing data through a buffering memory BIO handing the data to a
8320 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8321 <ptsekov@syntrex.com> and Nedelcho Stanev.
8324 *) Implement a countermeasure against a vulnerability recently found
8325 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8326 before application data chunks to avoid the use of known IVs
8327 with data potentially chosen by the attacker.
8330 *) Fix length checks in ssl3_get_client_hello().
8333 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8334 to prevent ssl3_read_internal() from incorrectly assuming that
8335 ssl3_read_bytes() found application data while handshake
8336 processing was enabled when in fact s->s3->in_read_app_data was
8337 merely automatically cleared during the initial handshake.
8338 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8340 *) Fix object definitions for Private and Enterprise: they were not
8341 recognized in their shortname (=lowercase) representation. Extend
8342 obj_dat.pl to issue an error when using undefined keywords instead
8343 of silently ignoring the problem (Svenning Sorensen
8344 <sss@sss.dnsalias.net>).
8347 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8348 generators, i.e. generators other than 2 and 5. (Previously, the
8349 code did not properly initialise the 'add' and 'rem' values to
8350 BN_generate_prime().)
8352 In the new general case, we do not insist that 'generator' is
8353 actually a primitive root: This requirement is rather pointless;
8354 a generator of the order-q subgroup is just as good, if not
8358 *) Map new X509 verification errors to alerts. Discovered and submitted by
8359 Tom Wu <tom@arcot.com>.
8362 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8363 returning non-zero before the data has been completely received
8364 when using non-blocking I/O.
8365 [Bodo Moeller; problem pointed out by John Hughes]
8367 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8368 [Ben Laurie, Lutz Jaenicke]
8370 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8371 Yoram Zahavi <YoramZ@gilian.com>).
8374 *) Add information about CygWin 1.3 and on, and preserve proper
8375 configuration for the versions before that.
8376 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8378 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8379 check whether we deal with a copy of a session and do not delete from
8380 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8381 <izhar@checkpoint.com>.
8384 *) Do not store session data into the internal session cache, if it
8385 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8386 flag is set). Proposed by Aslam <aslam@funk.com>.
8389 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8393 *) [In 0.9.6d-engine release:]
8394 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8395 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8397 *) Add the configuration target linux-s390x.
8398 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8400 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8401 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8402 variable as an indication that a ClientHello message has been
8403 received. As the flag value will be lost between multiple
8404 invocations of ssl3_accept when using non-blocking I/O, the
8405 function may not be aware that a handshake has actually taken
8406 place, thus preventing a new session from being added to the
8409 To avoid this problem, we now set s->new_session to 2 instead of
8410 using a local variable.
8411 [Lutz Jaenicke, Bodo Moeller]
8413 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8414 if the SSL_R_LENGTH_MISMATCH error is detected.
8415 [Geoff Thorpe, Bodo Moeller]
8417 *) New 'shared_ldflag' column in Configure platform table.
8420 *) Fix EVP_CIPHER_mode macro.
8421 ["Dan S. Camper" <dan@bti.net>]
8423 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8424 type, we must throw them away by setting rr->length to 0.
8425 [D P Chang <dpc@qualys.com>]
8427 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8429 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8430 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8431 worked incorrectly for those cases where range = 10..._2 and
8432 3*range is two bits longer than range.)
8435 *) Only add signing time to PKCS7 structures if it is not already
8439 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8440 OBJ_ld_ce should be OBJ_id_ce.
8441 Also some ip-pda OIDs in crypto/objects/objects.txt were
8442 incorrect (cf. RFC 3039).
8443 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8445 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8446 returns early because it has nothing to do.
8447 [Andy Schneider <andy.schneider@bjss.co.uk>]
8449 *) [In 0.9.6c-engine release:]
8450 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8451 [Andy Schneider <andy.schneider@bjss.co.uk>]
8453 *) [In 0.9.6c-engine release:]
8454 Add support for Cryptographic Appliance's keyserver technology.
8455 (Use engine 'keyclient')
8456 [Cryptographic Appliances and Geoff Thorpe]
8458 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8459 is called via tools/c89.sh because arguments have to be
8460 rearranged (all '-L' options must appear before the first object
8462 [Richard Shapiro <rshapiro@abinitio.com>]
8464 *) [In 0.9.6c-engine release:]
8465 Add support for Broadcom crypto accelerator cards, backported
8467 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8469 *) [In 0.9.6c-engine release:]
8470 Add support for SureWare crypto accelerator cards from
8471 Baltimore Technologies. (Use engine 'sureware')
8472 [Baltimore Technologies and Mark Cox]
8474 *) [In 0.9.6c-engine release:]
8475 Add support for crypto accelerator cards from Accelerated
8476 Encryption Processing, www.aep.ie. (Use engine 'aep')
8477 [AEP Inc. and Mark Cox]
8479 *) Add a configuration entry for gcc on UnixWare.
8480 [Gary Benson <gbenson@redhat.com>]
8482 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8483 messages are stored in a single piece (fixed-length part and
8484 variable-length part combined) and fix various bugs found on the way.
8487 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8488 instead. BIO_gethostbyname() does not know what timeouts are
8489 appropriate, so entries would stay in cache even when they have
8491 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8493 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8494 faced with a pathologically small ClientHello fragment that does
8495 not contain client_version: Instead of aborting with an error,
8496 simply choose the highest available protocol version (i.e.,
8497 TLS 1.0 unless it is disabled). In practice, ClientHello
8498 messages are never sent like this, but this change gives us
8499 strictly correct behaviour at least for TLS.
8502 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8503 never resets s->method to s->ctx->method when called from within
8504 one of the SSL handshake functions.
8505 [Bodo Moeller; problem pointed out by Niko Baric]
8507 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8508 (sent using the client's version number) if client_version is
8509 smaller than the protocol version in use. Also change
8510 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8511 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8512 the client will at least see that alert.
8515 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8519 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8520 client receives HelloRequest while in a handshake.
8521 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8523 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8524 should end in 'break', not 'goto end' which circumvents various
8525 cleanups done in state SSL_ST_OK. But session related stuff
8526 must be disabled for SSL_ST_OK in the case that we just sent a
8529 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8530 before just sending a HelloRequest.
8531 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8533 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8534 reveal whether illegal block cipher padding was found or a MAC
8535 verification error occurred. (Neither SSLerr() codes nor alerts
8536 are directly visible to potential attackers, but the information
8537 may leak via logfiles.)
8539 Similar changes are not required for the SSL 2.0 implementation
8540 because the number of padding bytes is sent in clear for SSL 2.0,
8541 and the extra bytes are just ignored. However ssl/s2_pkt.c
8542 failed to verify that the purported number of padding bytes is in
8546 *) Add OpenUNIX-8 support including shared libraries
8547 (Boyd Lynn Gerber <gerberb@zenez.com>).
8550 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8551 'wristwatch attack' using huge encoding parameters (cf.
8552 James H. Manger's CRYPTO 2001 paper). Note that the
8553 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8554 encoding parameters and hence was not vulnerable.
8557 *) BN_sqr() bug fix.
8558 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8560 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8561 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8562 followed by modular reduction.
8563 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8565 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8566 equivalent based on BN_pseudo_rand() instead of BN_rand().
8569 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8570 This function was broken, as the check for a new client hello message
8571 to handle SGC did not allow these large messages.
8572 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8575 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8578 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8579 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8582 *) Rework the configuration and shared library support for Tru64 Unix.
8583 The configuration part makes use of modern compiler features and
8584 still retains old compiler behavior for those that run older versions
8585 of the OS. The shared library support part includes a variant that
8586 uses the RPATH feature, and is available through the special
8587 configuration target "alpha-cc-rpath", which will never be selected
8589 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8591 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8592 with the same message size as in ssl3_get_certificate_request().
8593 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8594 messages might inadvertently be reject as too long.
8595 [Petr Lampa <lampa@fee.vutbr.cz>]
8597 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8600 *) Modified SSL library such that the verify_callback that has been set
8601 specificly for an SSL object with SSL_set_verify() is actually being
8602 used. Before the change, a verify_callback set with this function was
8603 ignored and the verify_callback() set in the SSL_CTX at the time of
8604 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8605 to allow the necessary settings.
8608 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8609 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8610 done automatically (in contradiction to the requirements of the C
8611 standard). This made problems when used from OpenSSH.
8614 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8615 dh->length and always used
8617 BN_rand_range(priv_key, dh->p).
8619 BN_rand_range() is not necessary for Diffie-Hellman, and this
8620 specific range makes Diffie-Hellman unnecessarily inefficient if
8621 dh->length (recommended exponent length) is much smaller than the
8622 length of dh->p. We could use BN_rand_range() if the order of
8623 the subgroup was stored in the DH structure, but we only have
8628 BN_rand(priv_key, l, ...)
8630 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8636 RSA_eay_public_encrypt
8637 RSA_eay_private_decrypt
8638 RSA_eay_private_encrypt (signing)
8639 RSA_eay_public_decrypt (signature verification)
8641 (default implementations for RSA_public_encrypt,
8642 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8643 always reject numbers >= n.
8646 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8647 to synchronize access to 'locking_thread'. This is necessary on
8648 systems where access to 'locking_thread' (an 'unsigned long'
8649 variable) is not atomic.
8652 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8653 *before* setting the 'crypto_lock_rand' flag. The previous code had
8654 a race condition if 0 is a valid thread ID.
8655 [Travis Vitek <vitek@roguewave.com>]
8657 *) Add support for shared libraries under Irix.
8658 [Albert Chin-A-Young <china@thewrittenword.com>]
8660 *) Add configuration option to build on Linux on both big-endian and
8662 [Ralf Baechle <ralf@uni-koblenz.de>]
8664 *) Add the possibility to create shared libraries on HP-UX.
8667 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8669 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8670 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8671 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8672 PRNG state recovery was possible based on the output of
8673 one PRNG request appropriately sized to gain knowledge on
8674 'md' followed by enough consecutive 1-byte PRNG requests
8675 to traverse all of 'state'.
8677 1. When updating 'md_local' (the current thread's copy of 'md')
8678 during PRNG output generation, hash all of the previous
8679 'md_local' value, not just the half used for PRNG output.
8681 2. Make the number of bytes from 'state' included into the hash
8682 independent from the number of PRNG bytes requested.
8684 The first measure alone would be sufficient to avoid
8685 Markku-Juhani's attack. (Actually it had never occurred
8686 to me that the half of 'md_local' used for chaining was the
8687 half from which PRNG output bytes were taken -- I had always
8688 assumed that the secret half would be used.) The second
8689 measure makes sure that additional data from 'state' is never
8690 mixed into 'md_local' in small portions; this heuristically
8691 further strengthens the PRNG.
8694 *) Fix crypto/bn/asm/mips3.s.
8697 *) When only the key is given to "enc", the IV is undefined. Print out
8698 an error message in this case.
8701 *) Handle special case when X509_NAME is empty in X509 printing routines.
8704 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8705 positive and less than q.
8708 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8709 used: it isn't thread safe and the add_lock_callback should handle
8711 [Paul Rose <Paul.Rose@bridge.com>]
8713 *) Verify that incoming data obeys the block size in
8714 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8718 [Ulf Möller, Bodo Möller]
8720 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8721 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8722 when fixing the server behaviour for backwards-compatible 'client
8723 hello' messages. (Note that the attack is impractical against
8724 SSL 3.0 and TLS 1.0 anyway because length and version checking
8725 means that the probability of guessing a valid ciphertext is
8726 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8729 Before 0.9.5, the countermeasure (hide the error by generating a
8730 random 'decryption result') did not work properly because
8731 ERR_clear_error() was missing, meaning that SSL_get_error() would
8732 detect the supposedly ignored error.
8734 Both problems are now fixed.
8737 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8738 (previously it was 1024).
8741 *) Fix for compatibility mode trust settings: ignore trust settings
8742 unless some valid trust or reject settings are present.
8745 *) Fix for blowfish EVP: its a variable length cipher.
8748 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8749 parameters in DSA public key structures and return an error in the
8750 DSA routines if parameters are absent.
8753 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8754 in the current directory if neither $RANDFILE nor $HOME was set.
8755 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8756 caused some confusion to Windows users who haven't defined $HOME.
8757 Thus RAND_file_name() is changed again: e_os.h can define a
8758 DEFAULT_HOME, which will be used if $HOME is not set.
8759 For Windows, we use "C:"; on other platforms, we still require
8760 environment variables.
8762 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8763 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8764 having multiple threads call RAND_poll() concurrently.
8767 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8768 combination of a flag and a thread ID variable.
8769 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8770 flag), *other* threads can enter ssleay_add_bytes without obeying
8771 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8772 that they do not hold after the first thread unsets add_do_not_lock).
8775 *) Change bctest again: '-x' expressions are not available in all
8779 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8781 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8782 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8784 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8785 the default extension for executables, if any. Also, make the perl
8786 scripts that use symlink() to test if it really exists and use "cp"
8787 if it doesn't. All this made OpenSSL compilable and installable in
8791 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8792 If SEQUENCE is length is indefinite just set c->slen to the total
8793 amount of data available.
8794 [Steve Henson, reported by shige@FreeBSD.org]
8795 [This change does not apply to 0.9.7.]
8797 *) Change bctest to avoid here-documents inside command substitution
8798 (workaround for FreeBSD /bin/sh bug).
8799 For compatibility with Ultrix, avoid shell functions (introduced
8800 in the bctest version that searches along $PATH).
8803 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8804 with des_encrypt() defined on some operating systems, like Solaris
8808 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8809 On the Importance of Eliminating Errors in Cryptographic
8810 Computations, J. Cryptology 14 (2001) 2, 101-119,
8811 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8814 *) MIPS assembler BIGNUM division bug fix.
8817 *) Disabled incorrect Alpha assembler code.
8820 *) Fix PKCS#7 decode routines so they correctly update the length
8821 after reading an EOC for the EXPLICIT tag.
8823 [This change does not apply to 0.9.7.]
8825 *) Fix bug in PKCS#12 key generation routines. This was triggered
8826 if a 3DES key was generated with a 0 initial byte. Include
8827 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8828 (but broken) behaviour.
8831 *) Enhance bctest to search for a working bc along $PATH and print
8833 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8835 *) Fix memory leaks in err.c: free err_data string if necessary;
8836 don't write to the wrong index in ERR_set_error_data.
8839 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8843 *) Replace rdtsc with _emit statements for VC++ version 5.
8844 [Jeremy Cooper <jeremy@baymoo.org>]
8846 *) Make it possible to reuse SSLv2 sessions.
8849 *) In copy_email() check for >= 0 as a return value for
8850 X509_NAME_get_index_by_NID() since 0 is a valid index.
8851 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8853 *) Avoid coredump with unsupported or invalid public keys by checking if
8854 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8855 PKCS7_verify() fails with non detached data.
8858 *) Don't use getenv in library functions when run as setuid/setgid.
8859 New function OPENSSL_issetugid().
8862 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8863 due to incorrect handling of multi-threading:
8865 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8867 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8869 3. Count how many times MemCheck_off() has been called so that
8870 nested use can be treated correctly. This also avoids
8871 inband-signalling in the previous code (which relied on the
8872 assumption that thread ID 0 is impossible).
8875 *) Add "-rand" option also to s_client and s_server.
8878 *) Fix CPU detection on Irix 6.x.
8879 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8880 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8882 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8885 [This change does not apply to 0.9.7.]
8887 *) Use the cached encoding of an X509_NAME structure rather than
8888 copying it. This is apparently the reason for the libsafe "errors"
8889 but the code is actually correct.
8892 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8893 Bleichenbacher's DSA attack.
8894 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8895 to be set and top=0 forces the highest bit to be set; top=-1 is new
8896 and leaves the highest bit random.
8897 [Ulf Moeller, Bodo Moeller]
8899 *) In the NCONF_...-based implementations for CONF_... queries
8900 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8901 a temporary CONF structure with the data component set to NULL
8902 (which gives segmentation faults in lh_retrieve).
8903 Instead, use NULL for the CONF pointer in CONF_get_string and
8904 CONF_get_number (which may use environment variables) and directly
8905 return NULL from CONF_get_section.
8908 *) Fix potential buffer overrun for EBCDIC.
8911 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8912 keyUsage if basicConstraints absent for a CA.
8915 *) Make SMIME_write_PKCS7() write mail header values with a format that
8916 is more generally accepted (no spaces before the semicolon), since
8917 some programs can't parse those values properly otherwise. Also make
8918 sure BIO's that break lines after each write do not create invalid
8922 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8923 macros previously used would not encode an empty SEQUENCE OF
8924 and break the signature.
8926 [This change does not apply to 0.9.7.]
8928 *) Zero the premaster secret after deriving the master secret in
8932 *) Add some EVP_add_digest_alias registrations (as found in
8933 OpenSSL_add_all_digests()) to SSL_library_init()
8934 aka OpenSSL_add_ssl_algorithms(). This provides improved
8935 compatibility with peers using X.509 certificates
8936 with unconventional AlgorithmIdentifier OIDs.
8939 *) Fix for Irix with NO_ASM.
8940 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8942 *) ./config script fixes.
8943 [Ulf Moeller, Richard Levitte]
8945 *) Fix 'openssl passwd -1'.
8948 *) Change PKCS12_key_gen_asc() so it can cope with non null
8949 terminated strings whose length is passed in the passlen
8950 parameter, for example from PEM callbacks. This was done
8951 by adding an extra length parameter to asc2uni().
8952 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8954 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8955 call failed, free the DSA structure.
8958 *) Fix to uni2asc() to cope with zero length Unicode strings.
8959 These are present in some PKCS#12 files.
8962 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8963 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8964 when writing a 32767 byte record.
8965 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8967 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8968 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8970 (RSA objects have a reference count access to which is protected
8971 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8972 so they are meant to be shared between threads.)
8973 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8974 "Reddie, Steven" <Steven.Reddie@ca.com>]
8976 *) Fix a deadlock in CRYPTO_mem_leaks().
8979 *) Use better test patterns in bntest.
8982 *) rand_win.c fix for Borland C.
8985 *) BN_rshift bugfix for n == 0.
8988 *) Add a 'bctest' script that checks for some known 'bc' bugs
8989 so that 'make test' does not abort just because 'bc' is broken.
8992 *) Store verify_result within SSL_SESSION also for client side to
8993 avoid potential security hole. (Re-used sessions on the client side
8994 always resulted in verify_result==X509_V_OK, not using the original
8995 result of the server certificate verification.)
8998 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8999 SSL3_RT_APPLICATION_DATA, return 0.
9000 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9004 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9005 releases, have been re-implemented by renaming the previous
9006 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9007 and ssl3_read_internal, respectively, and adding 'peek' parameters
9008 to them. The new ssl[23]_{read,peek} functions are calls to
9009 ssl[23]_read_internal with the 'peek' flag set appropriately.
9010 A 'peek' parameter has also been added to ssl3_read_bytes, which
9011 does the actual work for ssl3_read_internal.
9014 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9015 the method-specific "init()" handler. Also clean up ex_data after
9016 calling the method-specific "finish()" handler. Previously, this was
9017 happening the other way round.
9020 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9021 The previous value, 12, was not always sufficient for BN_mod_exp().
9024 *) Make sure that shared libraries get the internal name engine with
9025 the full version number and not just 0. This should mark the
9026 shared libraries as not backward compatible. Of course, this should
9027 be changed again when we can guarantee backward binary compatibility.
9030 *) Fix typo in get_cert_by_subject() in by_dir.c
9031 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9033 *) Rework the system to generate shared libraries:
9035 - Make note of the expected extension for the shared libraries and
9036 if there is a need for symbolic links from for example libcrypto.so.0
9037 to libcrypto.so.0.9.7. There is extended info in Configure for
9040 - Make as few rebuilds of the shared libraries as possible.
9042 - Still avoid linking the OpenSSL programs with the shared libraries.
9044 - When installing, install the shared libraries separately from the
9048 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9050 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9051 and not in SSL_clear because the latter is also used by the
9052 accept/connect functions; previously, the settings made by
9053 SSL_set_read_ahead would be lost during the handshake.
9054 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9056 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9057 Previously, it would create entries for disabled algorithms no
9061 *) Added several new manual pages for SSL_* function.
9064 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9066 *) In ssl23_get_client_hello, generate an error message when faced
9067 with an initial SSL 3.0/TLS record that is too small to contain the
9068 first two bytes of the ClientHello message, i.e. client_version.
9069 (Note that this is a pathologic case that probably has never happened
9070 in real life.) The previous approach was to use the version number
9071 from the record header as a substitute; but our protocol choice
9072 should not depend on that one because it is not authenticated
9073 by the Finished messages.
9076 *) More robust randomness gathering functions for Windows.
9077 [Jeffrey Altman <jaltman@columbia.edu>]
9079 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9080 not set then we don't setup the error code for issuer check errors
9081 to avoid possibly overwriting other errors which the callback does
9082 handle. If an application does set the flag then we assume it knows
9083 what it is doing and can handle the new informational codes
9087 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9088 a general "ANY" type, as such it should be able to decode anything
9089 including tagged types. However it didn't check the class so it would
9090 wrongly interpret tagged types in the same way as their universal
9091 counterpart and unknown types were just rejected. Changed so that the
9092 tagged and unknown types are handled in the same way as a SEQUENCE:
9093 that is the encoding is stored intact. There is also a new type
9094 "V_ASN1_OTHER" which is used when the class is not universal, in this
9095 case we have no idea what the actual type is so we just lump them all
9099 *) On VMS, stdout may very well lead to a file that is written to
9100 in a record-oriented fashion. That means that every write() will
9101 write a separate record, which will be read separately by the
9102 programs trying to read from it. This can be very confusing.
9104 The solution is to put a BIO filter in the way that will buffer
9105 text until a linefeed is reached, and then write everything a
9106 line at a time, so every record written will be an actual line,
9107 not chunks of lines and not (usually doesn't happen, but I've
9108 seen it once) several lines in one record. BIO_f_linebuffer() is
9111 Currently, it's a VMS-only method, because that's where it has
9112 been tested well enough.
9115 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9116 it can return incorrect results.
9117 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9118 but it was in 0.9.6-beta[12].)
9121 *) Disable the check for content being present when verifying detached
9122 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9123 include zero length content when signing messages.
9126 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9127 BIO_ctrl (for BIO pairs).
9130 *) Add DSO method for VMS.
9133 *) Bug fix: Montgomery multiplication could produce results with the
9137 *) Add RPM specification openssl.spec and modify it to build three
9138 packages. The default package contains applications, application
9139 documentation and run-time libraries. The devel package contains
9140 include files, static libraries and function documentation. The
9141 doc package contains the contents of the doc directory. The original
9142 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9145 *) Add a large number of documentation files for many SSL routines.
9146 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9148 *) Add a configuration entry for Sony News 4.
9149 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9151 *) Don't set the two most significant bits to one when generating a
9152 random number < q in the DSA library.
9155 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9156 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9157 the underlying transport is blocking) if a handshake took place.
9158 (The default behaviour is needed by applications such as s_client
9159 and s_server that use select() to determine when to use SSL_read;
9160 but for applications that know in advance when to expect data, it
9161 just makes things more complicated.)
9164 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9168 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9169 work better on such systems.
9170 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9172 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9173 Update PKCS12_parse() so it copies the friendlyName and the
9174 keyid to the certificates aux info.
9177 *) Fix bug in PKCS7_verify() which caused an infinite loop
9178 if there was more than one signature.
9179 [Sven Uszpelkat <su@celocom.de>]
9181 *) Major change in util/mkdef.pl to include extra information
9182 about each symbol, as well as presenting variables as well
9183 as functions. This change means that there's n more need
9184 to rebuild the .num files when some algorithms are excluded.
9187 *) Allow the verify time to be set by an application,
9188 rather than always using the current time.
9191 *) Phase 2 verify code reorganisation. The certificate
9192 verify code now looks up an issuer certificate by a
9193 number of criteria: subject name, authority key id
9194 and key usage. It also verifies self signed certificates
9195 by the same criteria. The main comparison function is
9196 X509_check_issued() which performs these checks.
9198 Lot of changes were necessary in order to support this
9199 without completely rewriting the lookup code.
9201 Authority and subject key identifier are now cached.
9203 The LHASH 'certs' is X509_STORE has now been replaced
9204 by a STACK_OF(X509_OBJECT). This is mainly because an
9205 LHASH can't store or retrieve multiple objects with
9206 the same hash value.
9208 As a result various functions (which were all internal
9209 use only) have changed to handle the new X509_STORE
9210 structure. This will break anything that messed round
9211 with X509_STORE internally.
9213 The functions X509_STORE_add_cert() now checks for an
9214 exact match, rather than just subject name.
9216 The X509_STORE API doesn't directly support the retrieval
9217 of multiple certificates matching a given criteria, however
9218 this can be worked round by performing a lookup first
9219 (which will fill the cache with candidate certificates)
9220 and then examining the cache for matches. This is probably
9221 the best we can do without throwing out X509_LOOKUP
9222 entirely (maybe later...).
9224 The X509_VERIFY_CTX structure has been enhanced considerably.
9226 All certificate lookup operations now go via a get_issuer()
9227 callback. Although this currently uses an X509_STORE it
9228 can be replaced by custom lookups. This is a simple way
9229 to bypass the X509_STORE hackery necessary to make this
9230 work and makes it possible to use more efficient techniques
9231 in future. A very simple version which uses a simple
9232 STACK for its trusted certificate store is also provided
9233 using X509_STORE_CTX_trusted_stack().
9235 The verify_cb() and verify() callbacks now have equivalents
9236 in the X509_STORE_CTX structure.
9238 X509_STORE_CTX also has a 'flags' field which can be used
9239 to customise the verify behaviour.
9242 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9243 excludes S/MIME capabilities.
9246 *) When a certificate request is read in keep a copy of the
9247 original encoding of the signed data and use it when outputting
9248 again. Signatures then use the original encoding rather than
9249 a decoded, encoded version which may cause problems if the
9250 request is improperly encoded.
9253 *) For consistency with other BIO_puts implementations, call
9254 buffer_write(b, ...) directly in buffer_puts instead of calling
9257 In BIO_puts, increment b->num_write as in BIO_write.
9258 [Peter.Sylvester@EdelWeb.fr]
9260 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9261 BN_zero, we may not return a BIGNUM with an array consisting of
9265 *) Avoid calling abort() from within the library when problems are
9266 detected, except if preprocessor symbols have been defined
9267 (such as REF_CHECK, BN_DEBUG etc.).
9270 *) New openssl application 'rsautl'. This utility can be
9271 used for low level RSA operations. DER public key
9272 BIO/fp routines also added.
9275 *) New Configure entry and patches for compiling on QNX 4.
9276 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9278 *) A demo state-machine implementation was sponsored by
9279 Nuron (http://www.nuron.com/) and is now available in
9280 demos/state_machine.
9283 *) New options added to the 'dgst' utility for signature
9284 generation and verification.
9287 *) Unrecognized PKCS#7 content types are now handled via a
9288 catch all ASN1_TYPE structure. This allows unsupported
9289 types to be stored as a "blob" and an application can
9290 encode and decode it manually.
9293 *) Fix various signed/unsigned issues to make a_strex.c
9295 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9297 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9298 length if passed a buffer. ASN1_INTEGER_to_BN failed
9299 if passed a NULL BN and its argument was negative.
9300 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9302 *) Modification to PKCS#7 encoding routines to output definite
9303 length encoding. Since currently the whole structures are in
9304 memory there's not real point in using indefinite length
9305 constructed encoding. However if OpenSSL is compiled with
9306 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9309 *) Added BIO_vprintf() and BIO_vsnprintf().
9312 *) Added more prefixes to parse for in the the strings written
9313 through a logging bio, to cover all the levels that are available
9314 through syslog. The prefixes are now:
9316 PANIC, EMERG, EMR => LOG_EMERG
9317 ALERT, ALR => LOG_ALERT
9318 CRIT, CRI => LOG_CRIT
9319 ERROR, ERR => LOG_ERR
9320 WARNING, WARN, WAR => LOG_WARNING
9321 NOTICE, NOTE, NOT => LOG_NOTICE
9322 INFO, INF => LOG_INFO
9323 DEBUG, DBG => LOG_DEBUG
9325 and as before, if none of those prefixes are present at the
9326 beginning of the string, LOG_ERR is chosen.
9328 On Win32, the LOG_* levels are mapped according to this:
9330 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9331 LOG_WARNING => EVENTLOG_WARNING_TYPE
9332 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9336 *) Made it possible to reconfigure with just the configuration
9337 argument "reconf" or "reconfigure". The command line arguments
9338 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9339 and are retrieved from there when reconfiguring.
9343 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9345 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9348 *) The obj_dat.pl script was messing up the sorting of object
9349 names. The reason was that it compared the quoted version
9350 of strings as a result "OCSP" > "OCSP Signing" because
9351 " > SPACE. Changed script to store unquoted versions of
9352 names and add quotes on output. It was also omitting some
9353 names from the lookup table if they were given a default
9354 value (that is if SN is missing it is given the same
9355 value as LN and vice versa), these are now added on the
9356 grounds that if an object has a name we should be able to
9357 look it up. Finally added warning output when duplicate
9358 short or long names are found.
9361 *) Changes needed for Tandem NSK.
9362 [Scott Uroff <scott@xypro.com>]
9364 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9365 RSA_padding_check_SSLv23(), special padding was never detected
9366 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9367 version rollback attacks was not effective.
9369 In s23_clnt.c, don't use special rollback-attack detection padding
9370 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9371 client; similarly, in s23_srvr.c, don't do the rollback check if
9372 SSL 2.0 is the only protocol enabled in the server.
9375 *) Make it possible to get hexdumps of unprintable data with 'openssl
9376 asn1parse'. By implication, the functions ASN1_parse_dump() and
9377 BIO_dump_indent() are added.
9380 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9381 these print out strings and name structures based on various
9382 flags including RFC2253 support and proper handling of
9383 multibyte characters. Added options to the 'x509' utility
9384 to allow the various flags to be set.
9387 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9388 Also change the functions X509_cmp_current_time() and
9389 X509_gmtime_adj() work with an ASN1_TIME structure,
9390 this will enable certificates using GeneralizedTime in validity
9391 dates to be checked.
9394 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9395 negative public key encodings) on by default,
9396 NO_NEG_PUBKEY_BUG can be set to disable it.
9399 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9400 content octets. An i2c_ASN1_OBJECT is unnecessary because
9401 the encoding can be trivially obtained from the structure.
9404 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9405 not read locks (CRYPTO_r_[un]lock).
9408 *) A first attempt at creating official support for shared
9409 libraries through configuration. I've kept it so the
9410 default is static libraries only, and the OpenSSL programs
9411 are always statically linked for now, but there are
9412 preparations for dynamic linking in place.
9413 This has been tested on Linux and Tru64.
9416 *) Randomness polling function for Win9x, as described in:
9417 Peter Gutmann, Software Generation of Practically Strong
9421 *) Fix so PRNG is seeded in req if using an already existing
9425 *) New options to smime application. -inform and -outform
9426 allow alternative formats for the S/MIME message including
9427 PEM and DER. The -content option allows the content to be
9428 specified separately. This should allow things like Netscape
9429 form signing output easier to verify.
9432 *) Fix the ASN1 encoding of tags using the 'long form'.
9435 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9436 STRING types. These convert content octets to and from the
9437 underlying type. The actual tag and length octets are
9438 already assumed to have been read in and checked. These
9439 are needed because all other string types have virtually
9440 identical handling apart from the tag. By having versions
9441 of the ASN1 functions that just operate on content octets
9442 IMPLICIT tagging can be handled properly. It also allows
9443 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9444 and ASN1_INTEGER are identical apart from the tag.
9447 *) Change the handling of OID objects as follows:
9449 - New object identifiers are inserted in objects.txt, following
9450 the syntax given in objects.README.
9451 - objects.pl is used to process obj_mac.num and create a new
9453 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9456 This is currently kind of a hack, and the perl code in objects.pl
9457 isn't very elegant, but it works as I intended. The simplest way
9458 to check that it worked correctly is to look in obj_dat.h and
9459 check the array nid_objs and make sure the objects haven't moved
9460 around (this is important!). Additions are OK, as well as
9461 consistent name changes.
9464 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9467 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9468 The given file adds to whatever has already been seeded into the
9469 random pool through the RANDFILE configuration file option or
9470 environment variable, or the default random state file.
9473 *) mkstack.pl now sorts each macro group into lexical order.
9474 Previously the output order depended on the order the files
9475 appeared in the directory, resulting in needless rewriting
9479 *) Patches to make OpenSSL compile under Win32 again. Mostly
9480 work arounds for the VC++ problem that it treats func() as
9481 func(void). Also stripped out the parts of mkdef.pl that
9482 added extra typesafe functions: these no longer exist.
9485 *) Reorganisation of the stack code. The macros are now all
9486 collected in safestack.h . Each macro is defined in terms of
9487 a "stack macro" of the form SKM_<name>(type, a, b). The
9488 DEBUG_SAFESTACK is now handled in terms of function casts,
9489 this has the advantage of retaining type safety without the
9490 use of additional functions. If DEBUG_SAFESTACK is not defined
9491 then the non typesafe macros are used instead. Also modified the
9492 mkstack.pl script to handle the new form. Needs testing to see
9493 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9494 the default if no major problems. Similar behaviour for ASN1_SET_OF
9495 and PKCS12_STACK_OF.
9498 *) When some versions of IIS use the 'NET' form of private key the
9499 key derivation algorithm is different. Normally MD5(password) is
9500 used as a 128 bit RC4 key. In the modified case
9501 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9502 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9503 as the old Netscape_RSA functions except they have an additional
9504 'sgckey' parameter which uses the modified algorithm. Also added
9505 an -sgckey command line option to the rsa utility. Thanks to
9506 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9507 algorithm to openssl-dev.
9510 *) The evp_local.h macros were using 'c.##kname' which resulted in
9511 invalid expansion on some systems (SCO 5.0.5 for example).
9512 Corrected to 'c.kname'.
9513 [Phillip Porch <root@theporch.com>]
9515 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9516 a STACK of email addresses from a certificate or request, these look
9517 in the subject name and the subject alternative name extensions and
9518 omit any duplicate addresses.
9521 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9522 This makes DSA verification about 2 % faster.
9525 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9526 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9527 plus overhead for 1024 bit moduli).
9528 This makes exponentiations about 0.5 % faster for 1024 bit
9529 exponents (as measured by "openssl speed rsa2048").
9532 *) Rename memory handling macros to avoid conflicts with other
9534 Malloc => OPENSSL_malloc
9535 Malloc_locked => OPENSSL_malloc_locked
9536 Realloc => OPENSSL_realloc
9537 Free => OPENSSL_free
9540 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9541 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9544 *) CygWin32 support.
9545 [John Jarvie <jjarvie@newsguy.com>]
9547 *) The type-safe stack code has been rejigged. It is now only compiled
9548 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9549 by default all type-specific stack functions are "#define"d back to
9550 standard stack functions. This results in more streamlined output
9551 but retains the type-safety checking possibilities of the original
9555 *) The STACK code has been cleaned up, and certain type declarations
9556 that didn't make a lot of sense have been brought in line. This has
9557 also involved a cleanup of sorts in safestack.h to more correctly
9558 map type-safe stack functions onto their plain stack counterparts.
9559 This work has also resulted in a variety of "const"ifications of
9560 lots of the code, especially "_cmp" operations which should normally
9561 be prototyped with "const" parameters anyway.
9564 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9565 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9566 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9567 where all of 'md' is used each time the PRNG is used, but 'state'
9568 is used only indexed by a cyclic counter. As entropy may not be
9569 well distributed from the beginning, 'md' is important as a
9570 chaining variable. However, the output function chains only half
9571 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9572 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9573 in all of 'state' being rewritten, with the new values depending
9574 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9577 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9578 the handshake is continued after ssl_verify_cert_chain();
9579 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9580 can lead to 'unexplainable' connection aborts later.
9581 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9583 *) Major EVP API cipher revision.
9584 Add hooks for extra EVP features. This allows various cipher
9585 parameters to be set in the EVP interface. Support added for variable
9586 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9587 setting of RC2 and RC5 parameters.
9589 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9592 Remove lots of duplicated code from the EVP library. For example *every*
9593 cipher init() function handles the 'iv' in the same way according to the
9594 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9595 for CFB and OFB modes they zero ctx->num.
9597 New functionality allows removal of S/MIME code RC2 hack.
9599 Most of the routines have the same form and so can be declared in terms
9602 By shifting this to the top level EVP_CipherInit() it can be removed from
9603 all individual ciphers. If the cipher wants to handle IVs or keys
9604 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9607 Change lots of functions like EVP_EncryptUpdate() to now return a
9608 value: although software versions of the algorithms cannot fail
9609 any installed hardware versions can.
9612 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9613 this option is set, tolerate broken clients that send the negotiated
9614 protocol version number instead of the requested protocol version
9618 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9619 i.e. non-zero for export ciphersuites, zero otherwise.
9620 Previous versions had this flag inverted, inconsistent with
9621 rsa_tmp_cb (..._TMP_RSA_CB).
9622 [Bodo Moeller; problem reported by Amit Chopra]
9624 *) Add missing DSA library text string. Work around for some IIS
9625 key files with invalid SEQUENCE encoding.
9628 *) Add a document (doc/standards.txt) that list all kinds of standards
9629 and so on that are implemented in OpenSSL.
9632 *) Enhance c_rehash script. Old version would mishandle certificates
9633 with the same subject name hash and wouldn't handle CRLs at all.
9634 Added -fingerprint option to crl utility, to support new c_rehash
9638 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9641 *) Fix for SSL server purpose checking. Server checking was
9642 rejecting certificates which had extended key usage present
9643 but no ssl client purpose.
9644 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9646 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9647 is a little unclear about how a blank password is handled.
9648 Since the password in encoded as a BMPString with terminating
9649 double NULL a zero length password would end up as just the
9650 double NULL. However no password at all is different and is
9651 handled differently in the PKCS#12 key generation code. NS
9652 treats a blank password as zero length. MSIE treats it as no
9653 password on export: but it will try both on import. We now do
9654 the same: PKCS12_parse() tries zero length and no password if
9655 the password is set to "" or NULL (NULL is now a valid password:
9656 it wasn't before) as does the pkcs12 application.
9659 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9660 perror when PEM_read_bio_X509_REQ fails, the error message must
9661 be obtained from the error queue.
9664 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9665 it in ERR_remove_state if appropriate, and change ERR_get_state
9666 accordingly to avoid race conditions (this is necessary because
9667 thread_hash is no longer constant once set).
9670 *) Bugfix for linux-elf makefile.one.
9673 *) RSA_get_default_method() will now cause a default
9674 RSA_METHOD to be chosen if one doesn't exist already.
9675 Previously this was only set during a call to RSA_new()
9676 or RSA_new_method(NULL) meaning it was possible for
9677 RSA_get_default_method() to return NULL.
9680 *) Added native name translation to the existing DSO code
9681 that will convert (if the flag to do so is set) filenames
9682 that are sufficiently small and have no path information
9683 into a canonical native form. Eg. "blah" converted to
9684 "libblah.so" or "blah.dll" etc.
9687 *) New function ERR_error_string_n(e, buf, len) which is like
9688 ERR_error_string(e, buf), but writes at most 'len' bytes
9689 including the 0 terminator. For ERR_error_string_n, 'buf'
9691 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9693 *) CONF library reworked to become more general. A new CONF
9694 configuration file reader "class" is implemented as well as a
9695 new functions (NCONF_*, for "New CONF") to handle it. The now
9696 old CONF_* functions are still there, but are reimplemented to
9697 work in terms of the new functions. Also, a set of functions
9698 to handle the internal storage of the configuration data is
9699 provided to make it easier to write new configuration file
9700 reader "classes" (I can definitely see something reading a
9701 configuration file in XML format, for example), called _CONF_*,
9702 or "the configuration storage API"...
9704 The new configuration file reading functions are:
9706 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9707 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9709 NCONF_default, NCONF_WIN32
9711 NCONF_dump_fp, NCONF_dump_bio
9713 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9714 NCONF_new creates a new CONF object. This works in the same way
9715 as other interfaces in OpenSSL, like the BIO interface.
9716 NCONF_dump_* dump the internal storage of the configuration file,
9717 which is useful for debugging. All other functions take the same
9718 arguments as the old CONF_* functions wth the exception of the
9719 first that must be a `CONF *' instead of a `LHASH *'.
9721 To make it easer to use the new classes with the old CONF_* functions,
9722 the function CONF_set_default_method is provided.
9725 *) Add '-tls1' option to 'openssl ciphers', which was already
9726 mentioned in the documentation but had not been implemented.
9727 (This option is not yet really useful because even the additional
9728 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9731 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9732 OpenSSL-based applications) load shared libraries and bind to
9733 them in a portable way.
9734 [Geoff Thorpe, with contributions from Richard Levitte]
9736 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9738 *) Make sure _lrotl and _lrotr are only used with MSVC.
9740 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9741 (the default implementation of RAND_status).
9743 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9744 to '-clrext' (= clear extensions), as intended and documented.
9745 [Bodo Moeller; inconsistency pointed out by Michael Attili
9748 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9749 was larger than the MD block size.
9750 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9752 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9753 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9754 using the passed key: if the passed key was a private key the result
9755 of X509_print(), for example, would be to print out all the private key
9759 *) des_quad_cksum() byte order bug fix.
9760 [Ulf Möller, using the problem description in krb4-0.9.7, where
9761 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9763 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9765 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9767 *) For easily testing in shell scripts whether some command
9768 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9769 returns with exit code 0 iff no command of the given name is available.
9770 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9771 the output goes to stdout and nothing is printed to stderr.
9772 Additional arguments are always ignored.
9774 Since for each cipher there is a command of the same name,
9775 the 'no-cipher' compilation switches can be tested this way.
9777 ('openssl no-XXX' is not able to detect pseudo-commands such
9778 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9781 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9784 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9785 is set; it will be thrown away anyway because each handshake creates
9787 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9788 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9789 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9790 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9793 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9794 'Q' and 'R' lose their special meanings (quit/renegotiate).
9795 This is part of what -quiet does; unlike -quiet, -ign_eof
9796 does not suppress any output.
9799 *) Add compatibility options to the purpose and trust code. The
9800 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9801 accepts a certificate or CA, this was the previous behaviour,
9802 with all the associated security issues.
9804 X509_TRUST_COMPAT is the old trust behaviour: only and
9805 automatically trust self signed roots in certificate store. A
9806 new trust setting X509_TRUST_DEFAULT is used to specify that
9807 a purpose has no associated trust setting and it should instead
9808 use the value in the default purpose.
9811 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9812 and fix a memory leak.
9815 *) In util/mkerr.pl (which implements 'make errors'), preserve
9816 reason strings from the previous version of the .c file, as
9817 the default to have only downcase letters (and digits) in
9818 automatically generated reasons codes is not always appropriate.
9821 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9822 using strerror. Previously, ERR_reason_error_string() returned
9823 library names as reason strings for SYSerr; but SYSerr is a special
9824 case where small numbers are errno values, not library numbers.
9827 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9828 converts DSA parameters into DH parameters. (When creating parameters,
9829 DSA_generate_parameters is used.)
9832 *) Include 'length' (recommended exponent length) in C code generated
9833 by 'openssl dhparam -C'.
9836 *) The second argument to set_label in perlasm was already being used
9837 so couldn't be used as a "file scope" flag. Moved to third argument
9841 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9842 instead of RAND_bytes for encryption IVs and salts.
9845 *) Include RAND_status() into RAND_METHOD instead of implementing
9846 it only for md_rand.c Otherwise replacing the PRNG by calling
9847 RAND_set_rand_method would be impossible.
9850 *) Don't let DSA_generate_key() enter an infinite loop if the random
9851 number generation fails.
9854 *) New 'rand' application for creating pseudo-random output.
9857 *) Added configuration support for Linux/IA64
9858 [Rolf Haberrecker <rolf@suse.de>]
9860 *) Assembler module support for Mingw32.
9863 *) Shared library support for HPUX (in shlib/).
9864 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9866 *) Shared library support for Solaris gcc.
9867 [Lutz Behnke <behnke@trustcenter.de>]
9869 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9871 *) PKCS7_encrypt() was adding text MIME headers twice because they
9872 were added manually and by SMIME_crlf_copy().
9875 *) In bntest.c don't call BN_rand with zero bits argument.
9876 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9878 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9879 case was implemented. This caused BN_div_recp() to fail occasionally.
9882 *) Add an optional second argument to the set_label() in the perl
9883 assembly language builder. If this argument exists and is set
9884 to 1 it signals that the assembler should use a symbol whose
9885 scope is the entire file, not just the current function. This
9886 is needed with MASM which uses the format label:: for this scope.
9887 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9889 *) Change the ASN1 types so they are typedefs by default. Before
9890 almost all types were #define'd to ASN1_STRING which was causing
9891 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9895 *) Change names of new functions to the new get1/get0 naming
9896 convention: After 'get1', the caller owns a reference count
9897 and has to call ..._free; 'get0' returns a pointer to some
9898 data structure without incrementing reference counters.
9899 (Some of the existing 'get' functions increment a reference
9900 counter, some don't.)
9901 Similarly, 'set1' and 'add1' functions increase reference
9902 counters or duplicate objects.
9905 *) Allow for the possibility of temp RSA key generation failure:
9906 the code used to assume it always worked and crashed on failure.
9909 *) Fix potential buffer overrun problem in BIO_printf().
9910 [Ulf Möller, using public domain code by Patrick Powell; problem
9911 pointed out by David Sacerdote <das33@cornell.edu>]
9913 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9914 RAND_egd() and RAND_status(). In the command line application,
9915 the EGD socket can be specified like a seed file using RANDFILE
9919 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9920 Some CAs (e.g. Verisign) distribute certificates in this form.
9923 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9924 list to exclude them. This means that no special compilation option
9925 is needed to use anonymous DH: it just needs to be included in the
9929 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9930 EVP_MD_type. The old functionality is available in a new macro called
9931 EVP_MD_md(). Change code that uses it and update docs.
9934 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9935 where the 'void *' argument is replaced by a function pointer argument.
9936 Previously 'void *' was abused to point to functions, which works on
9937 many platforms, but is not correct. As these functions are usually
9938 called by macros defined in OpenSSL header files, most source code
9939 should work without changes.
9942 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9943 sections with information on -D... compiler switches used for
9944 compiling the library so that applications can see them. To enable
9945 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9946 must be defined. E.g.,
9947 #define OPENSSL_ALGORITHM_DEFINES
9948 #include <openssl/opensslconf.h>
9949 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9950 [Richard Levitte, Ulf and Bodo Möller]
9952 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9956 *) Change the 'other' type in certificate aux info to a STACK_OF
9957 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9958 the required ASN1 format: arbitrary types determined by an OID.
9961 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9962 argument to 'req'. This is not because the function is newer or
9963 better than others it just uses the work 'NEW' in the certificate
9964 request header lines. Some software needs this.
9967 *) Reorganise password command line arguments: now passwords can be
9968 obtained from various sources. Delete the PEM_cb function and make
9969 it the default behaviour: i.e. if the callback is NULL and the
9970 usrdata argument is not NULL interpret it as a null terminated pass
9971 phrase. If usrdata and the callback are NULL then the pass phrase
9972 is prompted for as usual.
9975 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9976 the support is automatically enabled. The resulting binaries will
9977 autodetect the card and use it if present.
9978 [Ben Laurie and Compaq Inc.]
9980 *) Work around for Netscape hang bug. This sends certificate request
9981 and server done in one record. Since this is perfectly legal in the
9982 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9983 the bugs/SSLv3 entry for more info.
9986 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9989 *) Add -rand argument to smime and pkcs12 applications and read/write
9993 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9996 *) Add command line password options to the remaining applications.
9999 *) Bug fix for BN_div_recp() for numerators with an even number of
10003 *) More tests in bntest.c, and changed test_bn output.
10006 *) ./config recognizes MacOS X now.
10009 *) Bug fix for BN_div() when the first words of num and divsor are
10010 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10013 *) Add support for various broken PKCS#8 formats, and command line
10014 options to produce them.
10017 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10018 get temporary BIGNUMs from a BN_CTX.
10021 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10025 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10026 include a #define from the old name to the new. The original intent
10027 was that statically linked binaries could for example just call
10028 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10029 link with digests. This never worked because SSLeay_add_all_digests()
10030 and SSLeay_add_all_ciphers() were in the same source file so calling
10031 one would link with the other. They are now in separate source files.
10034 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10037 *) Use a less unusual form of the Miller-Rabin primality test (it used
10038 a binary algorithm for exponentiation integrated into the Miller-Rabin
10039 loop, our standard modexp algorithms are faster).
10042 *) Support for the EBCDIC character set completed.
10043 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10045 *) Source code cleanups: use const where appropriate, eliminate casts,
10046 use void * instead of char * in lhash.
10049 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10050 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10051 this the server could overwrite ephemeral keys that the client
10055 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10056 using 50 iterations of the Rabin-Miller test.
10058 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10059 iterations of the Rabin-Miller test as required by the appendix
10060 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10061 As BN_is_prime_fasttest includes trial division, DSA parameter
10062 generation becomes much faster.
10064 This implies a change for the callback functions in DSA_is_prime
10065 and DSA_generate_parameters: The callback function is called once
10066 for each positive witness in the Rabin-Miller test, not just
10067 occasionally in the inner loop; and the parameters to the
10068 callback function now provide an iteration count for the outer
10069 loop rather than for the current invocation of the inner loop.
10070 DSA_generate_parameters additionally can call the callback
10071 function with an 'iteration count' of -1, meaning that a
10072 candidate has passed the trial division test (when q is generated
10073 from an application-provided seed, trial division is skipped).
10076 *) New function BN_is_prime_fasttest that optionally does trial
10077 division before starting the Rabin-Miller test and has
10078 an additional BN_CTX * argument (whereas BN_is_prime always
10079 has to allocate at least one BN_CTX).
10080 'callback(1, -1, cb_arg)' is called when a number has passed the
10081 trial division stage.
10084 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10088 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10091 *) New function BN_pseudo_rand().
10094 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10095 bignum version of BN_from_montgomery() with the working code from
10096 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10100 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10101 made it impossible to use the same SSL_SESSION data structure in
10102 SSL2 clients in multiple threads.
10105 *) The return value of RAND_load_file() no longer counts bytes obtained
10106 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10107 to seed the PRNG (previously an explicit byte count was required).
10108 [Ulf Möller, Bodo Möller]
10110 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10111 used (char *) instead of (void *) and had casts all over the place.
10114 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10117 *) Retain source code compatibility for BN_prime_checks macro:
10118 BN_is_prime(..., BN_prime_checks, ...) now uses
10119 BN_prime_checks_for_size to determine the appropriate number of
10120 Rabin-Miller iterations.
10123 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10124 DH_CHECK_P_NOT_SAFE_PRIME.
10125 (Check if this is true? OpenPGP calls them "strong".)
10128 *) Merge the functionality of "dh" and "gendh" programs into a new program
10129 "dhparam". The old programs are retained for now but will handle DH keys
10130 (instead of parameters) in future.
10133 *) Make the ciphers, s_server and s_client programs check the return values
10134 when a new cipher list is set.
10137 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10138 ciphers. Before when the 56bit ciphers were enabled the sorting was
10141 The syntax for the cipher sorting has been extended to support sorting by
10142 cipher-strength (using the strength_bits hard coded in the tables).
10143 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10145 Fix a bug in the cipher-command parser: when supplying a cipher command
10146 string with an "undefined" symbol (neither command nor alphanumeric
10147 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10148 an error is flagged.
10150 Due to the strength-sorting extension, the code of the
10151 ssl_create_cipher_list() function was completely rearranged. I hope that
10152 the readability was also increased :-)
10153 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10155 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10156 for the first serial number and places 2 in the serial number file. This
10157 avoids problems when the root CA is created with serial number zero and
10158 the first user certificate has the same issuer name and serial number
10162 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10163 the new code. Add documentation for this stuff.
10166 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10167 X509_*() to X509at_*() on the grounds that they don't handle X509
10168 structures and behave in an analogous way to the X509v3 functions:
10169 they shouldn't be called directly but wrapper functions should be used
10172 So we also now have some wrapper functions that call the X509at functions
10173 when passed certificate requests. (TO DO: similar things can be done with
10174 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10175 things. Some of these need some d2i or i2d and print functionality
10176 because they handle more complex structures.)
10179 *) Add missing #ifndefs that caused missing symbols when building libssl
10180 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10181 NO_RSA in ssl/s2*.c.
10182 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10184 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10185 has a return value which indicates the quality of the random data
10186 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10187 error queue. New function RAND_pseudo_bytes() generates output that is
10188 guaranteed to be unique but not unpredictable. RAND_add is like
10189 RAND_seed, but takes an extra argument for an entropy estimate
10190 (RAND_seed always assumes full entropy).
10193 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10194 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10195 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10196 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10197 false-positive rate of at most 2^-80 for random input.
10200 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10203 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10204 in the 0.9.5 release), this returns the chain
10205 from an X509_CTX structure with a dup of the stack and all
10206 the X509 reference counts upped: so the stack will exist
10207 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10210 Also make SSL_SESSION_print() print out the verify return
10214 *) Add manpage for the pkcs12 command. Also change the default
10215 behaviour so MAC iteration counts are used unless the new
10216 -nomaciter option is used. This improves file security and
10217 only older versions of MSIE (4.0 for example) need it.
10220 *) Honor the no-xxx Configure options when creating .DEF files.
10223 *) Add PKCS#10 attributes to field table: challengePassword,
10224 unstructuredName and unstructuredAddress. These are taken from
10225 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10226 international characters are used.
10228 More changes to X509_ATTRIBUTE code: allow the setting of types
10229 based on strings. Remove the 'loc' parameter when adding
10230 attributes because these will be a SET OF encoding which is sorted
10234 *) Initial changes to the 'req' utility to allow request generation
10235 automation. This will allow an application to just generate a template
10236 file containing all the field values and have req construct the
10239 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10240 used all over the place including certificate requests and PKCS#7
10241 structures. They are currently handled manually where necessary with
10242 some primitive wrappers for PKCS#7. The new functions behave in a
10243 manner analogous to the X509 extension functions: they allow
10244 attributes to be looked up by NID and added.
10246 Later something similar to the X509V3 code would be desirable to
10247 automatically handle the encoding, decoding and printing of the
10248 more complex types. The string types like challengePassword can
10249 be handled by the string table functions.
10251 Also modified the multi byte string table handling. Now there is
10252 a 'global mask' which masks out certain types. The table itself
10253 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10254 is useful when for example there is only one permissible type
10255 (as in countryName) and using the mask might result in no valid
10259 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10260 SSL_get_peer_finished to allow applications to obtain the latest
10261 Finished messages sent to the peer or expected from the peer,
10262 respectively. (SSL_get_peer_finished is usually the Finished message
10263 actually received from the peer, otherwise the protocol will be aborted.)
10265 As the Finished message are message digests of the complete handshake
10266 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10267 be used for external authentication procedures when the authentication
10268 provided by SSL/TLS is not desired or is not enough.
10271 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10272 the host supports BWX extension and if Compaq C is present on the
10273 $PATH. Just exploiting of the BWX extension results in 20-30%
10274 performance kick for some algorithms, e.g. DES and RC4 to mention
10275 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10279 *) Add support for MS "fast SGC". This is arguably a violation of the
10280 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10281 weak crypto and after checking the certificate is SGC a second one
10282 with strong crypto. MS SGC stops the first handshake after receiving
10283 the server certificate message and sends a second client hello. Since
10284 a server will typically do all the time consuming operations before
10285 expecting any further messages from the client (server key exchange
10286 is the most expensive) there is little difference between the two.
10288 To get OpenSSL to support MS SGC we have to permit a second client
10289 hello message after we have sent server done. In addition we have to
10290 reset the MAC if we do get this second client hello.
10293 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10294 if a DER encoded private key is RSA or DSA traditional format. Changed
10295 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10296 format DER encoded private key. Newer code should use PKCS#8 format which
10297 has the key type encoded in the ASN1 structure. Added DER private key
10298 support to pkcs8 application.
10301 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10302 ciphersuites has been selected (as required by the SSL 3/TLS 1
10303 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10304 is set, we interpret this as a request to violate the specification
10305 (the worst that can happen is a handshake failure, and 'correct'
10306 behaviour would result in a handshake failure anyway).
10309 *) In SSL_CTX_add_session, take into account that there might be multiple
10310 SSL_SESSION structures with the same session ID (e.g. when two threads
10311 concurrently obtain them from an external cache).
10312 The internal cache can handle only one SSL_SESSION with a given ID,
10313 so if there's a conflict, we now throw out the old one to achieve
10317 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10318 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10319 some routines that use cipher OIDs: some ciphers do not have OIDs
10320 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10324 *) Simplify the trust setting structure and code. Now we just have
10325 two sequences of OIDs for trusted and rejected settings. These will
10326 typically have values the same as the extended key usage extension
10327 and any application specific purposes.
10329 The trust checking code now has a default behaviour: it will just
10330 check for an object with the same NID as the passed id. Functions can
10331 be provided to override either the default behaviour or the behaviour
10332 for a given id. SSL client, server and email already have functions
10333 in place for compatibility: they check the NID and also return "trusted"
10334 if the certificate is self signed.
10337 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10338 traditional format into an EVP_PKEY structure.
10341 *) Add a password callback function PEM_cb() which either prompts for
10342 a password if usr_data is NULL or otherwise assumes it is a null
10343 terminated password. Allow passwords to be passed on command line
10344 environment or config files in a few more utilities.
10347 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10348 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10349 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10350 Update documentation.
10353 *) Support for ASN1 "NULL" type. This could be handled before by using
10354 ASN1_TYPE but there wasn't any function that would try to read a NULL
10355 and produce an error if it couldn't. For compatibility we also have
10356 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10357 don't allocate anything because they don't need to.
10360 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10362 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10364 *) Rebuild of the memory allocation routines used by OpenSSL code and
10365 possibly others as well. The purpose is to make an interface that
10366 provide hooks so anyone can build a separate set of allocation and
10367 deallocation routines to be used by OpenSSL, for example memory
10368 pool implementations, or something else, which was previously hard
10369 since Malloc(), Realloc() and Free() were defined as macros having
10370 the values malloc, realloc and free, respectively (except for Win32
10371 compilations). The same is provided for memory debugging code.
10372 OpenSSL already comes with functionality to find memory leaks, but
10373 this gives people a chance to debug other memory problems.
10375 With these changes, a new set of functions and macros have appeared:
10377 CRYPTO_set_mem_debug_functions() [F]
10378 CRYPTO_get_mem_debug_functions() [F]
10379 CRYPTO_dbg_set_options() [F]
10380 CRYPTO_dbg_get_options() [F]
10381 CRYPTO_malloc_debug_init() [M]
10383 The memory debug functions are NULL by default, unless the library
10384 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10385 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10386 gives the standard debugging functions that come with OpenSSL) or
10387 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10388 provided by the library user) must be used. When the standard
10389 debugging functions are used, CRYPTO_dbg_set_options can be used to
10390 request additional information:
10391 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10392 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10394 Also, things like CRYPTO_set_mem_functions will always give the
10395 expected result (the new set of functions is used for allocation
10396 and deallocation) at all times, regardless of platform and compiler
10399 To finish it up, some functions that were never use in any other
10400 way than through macros have a new API and new semantic:
10402 CRYPTO_dbg_malloc()
10403 CRYPTO_dbg_realloc()
10406 All macros of value have retained their old syntax.
10407 [Richard Levitte and Bodo Moeller]
10409 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10410 ordering of SMIMECapabilities wasn't in "strength order" and there
10411 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10415 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10416 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10417 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10419 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10420 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10421 functionality to handle multipart/signed properly) and a utility
10422 called 'smime' to call all this stuff. This is based on code I
10423 originally wrote for Celo who have kindly allowed it to be
10424 included in OpenSSL.
10427 *) Add variants des_set_key_checked and des_set_key_unchecked of
10428 des_set_key (aka des_key_sched). Global variable des_check_key
10429 decides which of these is called by des_set_key; this way
10430 des_check_key behaves as it always did, but applications and
10431 the library itself, which was buggy for des_check_key == 1,
10432 have a cleaner way to pick the version they need.
10435 *) New function PKCS12_newpass() which changes the password of a
10439 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10440 dynamic mix. In both cases the ids can be used as an index into the
10441 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10442 functions so they accept a list of the field values and the
10443 application doesn't need to directly manipulate the X509_TRUST
10447 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10451 *) Modify the way the V3 extension code looks up extensions. This now
10452 works in a similar way to the object code: we have some "standard"
10453 extensions in a static table which is searched with OBJ_bsearch()
10454 and the application can add dynamic ones if needed. The file
10455 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10456 updated whenever a new extension is added to the core code and kept
10457 in ext_nid order. There is a simple program 'tabtest.c' which checks
10458 this. New extensions are not added too often so this file can readily
10459 be maintained manually.
10461 There are two big advantages in doing things this way. The extensions
10462 can be looked up immediately and no longer need to be "added" using
10463 X509V3_add_standard_extensions(): this function now does nothing.
10464 [Side note: I get *lots* of email saying the extension code doesn't
10465 work because people forget to call this function]
10466 Also no dynamic allocation is done unless new extensions are added:
10467 so if we don't add custom extensions there is no need to call
10468 X509V3_EXT_cleanup().
10471 *) Modify enc utility's salting as follows: make salting the default. Add a
10472 magic header, so unsalted files fail gracefully instead of just decrypting
10473 to garbage. This is because not salting is a big security hole, so people
10474 should be discouraged from doing it.
10477 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10478 digest to be passed on the command line but it only used this
10479 parameter when signing a certificate. Modified so all relevant
10480 operations are affected by the digest parameter including the
10481 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10482 DSA key was used because it didn't fix the digest.
10485 *) Initial certificate chain verify code. Currently tests the untrusted
10486 certificates for consistency with the verify purpose (which is set
10487 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10489 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10490 this is because it will reject chains with invalid extensions whereas
10491 every previous version of OpenSSL and SSLeay made no checks at all.
10493 Trust code: checks the root CA for the relevant trust settings. Trust
10494 settings have an initial value consistent with the verify purpose: e.g.
10495 if the verify purpose is for SSL client use it expects the CA to be
10496 trusted for SSL client use. However the default value can be changed to
10497 permit custom trust settings: one example of this would be to only trust
10498 certificates from a specific "secure" set of CAs.
10500 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10501 which should be used for version portability: especially since the
10502 verify structure is likely to change more often now.
10504 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10505 to set them. If not set then assume SSL clients will verify SSL servers
10508 Two new options to the verify program: -untrusted allows a set of
10509 untrusted certificates to be passed in and -purpose which sets the
10510 intended purpose of the certificate. If a purpose is set then the
10511 new chain verify code is used to check extension consistency.
10514 *) Support for the authority information access extension.
10517 *) Modify RSA and DSA PEM read routines to transparently handle
10518 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10519 public keys in a format compatible with certificate
10520 SubjectPublicKeyInfo structures. Unfortunately there were already
10521 functions called *_PublicKey_* which used various odd formats so
10522 these are retained for compatibility: however the DSA variants were
10523 never in a public release so they have been deleted. Changed dsa/rsa
10524 utilities to handle the new format: note no releases ever handled public
10525 keys so we should be OK.
10527 The primary motivation for this change is to avoid the same fiasco
10528 that dogs private keys: there are several incompatible private key
10529 formats some of which are standard and some OpenSSL specific and
10530 require various evil hacks to allow partial transparent handling and
10531 even then it doesn't work with DER formats. Given the option anything
10532 other than PKCS#8 should be dumped: but the other formats have to
10533 stay in the name of compatibility.
10535 With public keys and the benefit of hindsight one standard format
10536 is used which works with EVP_PKEY, RSA or DSA structures: though
10537 it clearly returns an error if you try to read the wrong kind of key.
10539 Added a -pubkey option to the 'x509' utility to output the public key.
10540 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10541 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10542 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10543 that do the same as the EVP_PKEY_assign_*() except they up the
10544 reference count of the added key (they don't "swallow" the
10548 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10549 CRLs would fail if the file contained no certificates or no CRLs:
10550 added a new function to read in both types and return the number
10551 read: this means that if none are read it will be an error. The
10552 DER versions of the certificate and CRL reader would always fail
10553 because it isn't possible to mix certificates and CRLs in DER format
10554 without choking one or the other routine. Changed this to just read
10555 a certificate: this is the best we can do. Also modified the code
10556 in apps/verify.c to take notice of return codes: it was previously
10557 attempting to read in certificates from NULL pointers and ignoring
10558 any errors: this is one reason why the cert and CRL reader seemed
10559 to work. It doesn't check return codes from the default certificate
10560 routines: these may well fail if the certificates aren't installed.
10563 *) Code to support otherName option in GeneralName.
10566 *) First update to verify code. Change the verify utility
10567 so it warns if it is passed a self signed certificate:
10568 for consistency with the normal behaviour. X509_verify
10569 has been modified to it will now verify a self signed
10570 certificate if *exactly* the same certificate appears
10571 in the store: it was previously impossible to trust a
10572 single self signed certificate. This means that:
10573 openssl verify ss.pem
10574 now gives a warning about a self signed certificate but
10575 openssl verify -CAfile ss.pem ss.pem
10579 *) For servers, store verify_result in SSL_SESSION data structure
10580 (and add it to external session representation).
10581 This is needed when client certificate verifications fails,
10582 but an application-provided verification callback (set by
10583 SSL_CTX_set_cert_verify_callback) allows accepting the session
10584 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10585 but returns 1): When the session is reused, we have to set
10586 ssl->verify_result to the appropriate error code to avoid
10588 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10590 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10591 case in PKCS7_dataInit() where the signed PKCS7 structure
10592 didn't contain any existing data because it was being created.
10593 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10595 *) Add a salt to the key derivation routines in enc.c. This
10596 forms the first 8 bytes of the encrypted file. Also add a
10597 -S option to allow a salt to be input on the command line.
10600 *) New function X509_cmp(). Oddly enough there wasn't a function
10601 to compare two certificates. We do this by working out the SHA1
10602 hash and comparing that. X509_cmp() will be needed by the trust
10606 *) SSL_get1_session() is like SSL_get_session(), but increments
10607 the reference count in the SSL_SESSION returned.
10608 [Geoff Thorpe <geoff@eu.c2.net>]
10610 *) Fix for 'req': it was adding a null to request attributes.
10611 Also change the X509_LOOKUP and X509_INFO code to handle
10612 certificate auxiliary information.
10615 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10619 *) Add the possibility to add extra information to the memory leak
10620 detecting output, to form tracebacks, showing from where each
10621 allocation was originated: CRYPTO_push_info("constant string") adds
10622 the string plus current file name and line number to a per-thread
10623 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10624 is like calling CYRPTO_pop_info() until the stack is empty.
10625 Also updated memory leak detection code to be multi-thread-safe.
10628 *) Add options -text and -noout to pkcs7 utility and delete the
10629 encryption options which never did anything. Update docs.
10632 *) Add options to some of the utilities to allow the pass phrase
10633 to be included on either the command line (not recommended on
10634 OSes like Unix) or read from the environment. Update the
10635 manpages and fix a few bugs.
10638 *) Add a few manpages for some of the openssl commands.
10641 *) Fix the -revoke option in ca. It was freeing up memory twice,
10642 leaking and not finding already revoked certificates.
10645 *) Extensive changes to support certificate auxiliary information.
10646 This involves the use of X509_CERT_AUX structure and X509_AUX
10647 functions. An X509_AUX function such as PEM_read_X509_AUX()
10648 can still read in a certificate file in the usual way but it
10649 will also read in any additional "auxiliary information". By
10650 doing things this way a fair degree of compatibility can be
10651 retained: existing certificates can have this information added
10652 using the new 'x509' options.
10654 Current auxiliary information includes an "alias" and some trust
10655 settings. The trust settings will ultimately be used in enhanced
10656 certificate chain verification routines: currently a certificate
10657 can only be trusted if it is self signed and then it is trusted
10661 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10662 The problem was that one of the replacement routines had not been working
10663 since SSLeay releases. For now the offending routine has been replaced
10664 with non-optimised assembler. Even so, this now gives around 95%
10665 performance improvement for 1024 bit RSA signs.
10668 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10669 handling. Most clients have the effective key size in bits equal to
10670 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10671 A few however don't do this and instead use the size of the decrypted key
10672 to determine the RC2 key length and the AlgorithmIdentifier to determine
10673 the effective key length. In this case the effective key length can still
10674 be 40 bits but the key length can be 168 bits for example. This is fixed
10675 by manually forcing an RC2 key into the EVP_PKEY structure because the
10676 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10677 the key length and effective key length are equal.
10680 *) Add a bunch of functions that should simplify the creation of
10681 X509_NAME structures. Now you should be able to do:
10682 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10683 and have it automatically work out the correct field type and fill in
10684 the structures. The more adventurous can try:
10685 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10686 and it will (hopefully) work out the correct multibyte encoding.
10689 *) Change the 'req' utility to use the new field handling and multibyte
10690 copy routines. Before the DN field creation was handled in an ad hoc
10691 way in req, ca, and x509 which was rather broken and didn't support
10692 BMPStrings or UTF8Strings. Since some software doesn't implement
10693 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10694 using the dirstring_type option. See the new comment in the default
10695 openssl.cnf for more info.
10698 *) Make crypto/rand/md_rand.c more robust:
10699 - Assure unique random numbers after fork().
10700 - Make sure that concurrent threads access the global counter and
10701 md serializably so that we never lose entropy in them
10702 or use exactly the same state in multiple threads.
10703 Access to the large state is not always serializable because
10704 the additional locking could be a performance killer, and
10705 md should be large enough anyway.
10708 *) New file apps/app_rand.c with commonly needed functionality
10709 for handling the random seed file.
10711 Use the random seed file in some applications that previously did not:
10713 dsaparam -genkey (which also ignored its '-rand' option),
10716 x509 (when signing).
10717 Except on systems with /dev/urandom, it is crucial to have a random
10718 seed file at least for key creation, DSA signing, and for DH exchanges;
10719 for RSA signatures we could do without one.
10721 gendh and gendsa (unlike genrsa) used to read only the first byte
10722 of each file listed in the '-rand' option. The function as previously
10723 found in genrsa is now in app_rand.c and is used by all programs
10724 that support '-rand'.
10727 *) In RAND_write_file, use mode 0600 for creating files;
10728 don't just chmod when it may be too late.
10731 *) Report an error from X509_STORE_load_locations
10732 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10735 *) New function ASN1_mbstring_copy() this copies a string in either
10736 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10737 into an ASN1_STRING type. A mask of permissible types is passed
10738 and it chooses the "minimal" type to use or an error if not type
10742 *) Add function equivalents to the various macros in asn1.h. The old
10743 macros are retained with an M_ prefix. Code inside the library can
10744 use the M_ macros. External code (including the openssl utility)
10745 should *NOT* in order to be "shared library friendly".
10748 *) Add various functions that can check a certificate's extensions
10749 to see if it usable for various purposes such as SSL client,
10750 server or S/MIME and CAs of these types. This is currently
10751 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10752 verification. Also added a -purpose flag to x509 utility to
10753 print out all the purposes.
10756 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10760 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10761 for, obtain and decode and extension and obtain its critical flag.
10762 This allows all the necessary extension code to be handled in a
10763 single function call.
10766 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10767 platforms. See crypto/rc4/rc4_enc.c for further details.
10770 *) New -noout option to asn1parse. This causes no output to be produced
10771 its main use is when combined with -strparse and -out to extract data
10772 from a file (which may not be in ASN.1 format).
10775 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10776 when producing the local key id.
10777 [Richard Levitte <levitte@stacken.kth.se>]
10779 *) New option -dhparam in s_server. This allows a DH parameter file to be
10780 stated explicitly. If it is not stated then it tries the first server
10781 certificate file. The previous behaviour hard coded the filename
10785 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10786 a public key to be input or output. For example:
10787 openssl rsa -in key.pem -pubout -out pubkey.pem
10788 Also added necessary DSA public key functions to handle this.
10791 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10792 in the message. This was handled by allowing
10793 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10794 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10796 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10797 to the end of the strings whereas this didn't. This would cause problems
10798 if strings read with d2i_ASN1_bytes() were later modified.
10799 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10801 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10802 data and it contains EOF it will end up returning an error. This is
10803 caused by input 46 bytes long. The cause is due to the way base64
10804 BIOs find the start of base64 encoded data. They do this by trying a
10805 trial decode on each line until they find one that works. When they
10806 do a flag is set and it starts again knowing it can pass all the
10807 data directly through the decoder. Unfortunately it doesn't reset
10808 the context it uses. This means that if EOF is reached an attempt
10809 is made to pass two EOFs through the context and this causes the
10810 resulting error. This can also cause other problems as well. As is
10811 usual with these problems it takes *ages* to find and the fix is
10812 trivial: move one line.
10813 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10815 *) Ugly workaround to get s_client and s_server working under Windows. The
10816 old code wouldn't work because it needed to select() on sockets and the
10817 tty (for keypresses and to see if data could be written). Win32 only
10818 supports select() on sockets so we select() with a 1s timeout on the
10819 sockets and then see if any characters are waiting to be read, if none
10820 are present then we retry, we also assume we can always write data to
10821 the tty. This isn't nice because the code then blocks until we've
10822 received a complete line of data and it is effectively polling the
10823 keyboard at 1s intervals: however it's quite a bit better than not
10824 working at all :-) A dedicated Windows application might handle this
10825 with an event loop for example.
10828 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10829 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10830 will be called when RSA_sign() and RSA_verify() are used. This is useful
10831 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10832 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10833 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10834 This necessitated the support of an extra signature type NID_md5_sha1
10835 for SSL signatures and modifications to the SSL library to use it instead
10836 of calling RSA_public_decrypt() and RSA_private_encrypt().
10839 *) Add new -verify -CAfile and -CApath options to the crl program, these
10840 will lookup a CRL issuers certificate and verify the signature in a
10841 similar way to the verify program. Tidy up the crl program so it
10842 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10843 less strict. It will now permit CRL extensions even if it is not
10844 a V2 CRL: this will allow it to tolerate some broken CRLs.
10847 *) Initialize all non-automatic variables each time one of the openssl
10848 sub-programs is started (this is necessary as they may be started
10849 multiple times from the "OpenSSL>" prompt).
10850 [Lennart Bang, Bodo Moeller]
10852 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10853 removing all other RSA functionality (this is what NO_RSA does). This
10854 is so (for example) those in the US can disable those operations covered
10855 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10859 *) Non-copying interface to BIO pairs.
10860 (still largely untested)
10863 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10864 ASCII string. This was handled independently in various places before.
10867 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10868 UTF8 strings a character at a time.
10871 *) Use client_version from client hello to select the protocol
10872 (s23_srvr.c) and for RSA client key exchange verification
10873 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10876 *) Add various utility functions to handle SPKACs, these were previously
10877 handled by poking round in the structure internals. Added new function
10878 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10879 print, verify and generate SPKACs. Based on an original idea from
10880 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10883 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10886 *) Allow the config file extension section to be overwritten on the
10887 command line. Based on an original idea from Massimiliano Pala
10888 <madwolf@comune.modena.it>. The new option is called -extensions
10889 and can be applied to ca, req and x509. Also -reqexts to override
10890 the request extensions in req and -crlexts to override the crl extensions
10894 *) Add new feature to the SPKAC handling in ca. Now you can include
10895 the same field multiple times by preceding it by "XXXX." for example:
10898 this is the same syntax as used in the req config file.
10901 *) Allow certificate extensions to be added to certificate requests. These
10902 are specified in a 'req_extensions' option of the req section of the
10903 config file. They can be printed out with the -text option to req but
10904 are otherwise ignored at present.
10907 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10908 data read consists of only the final block it would not decrypted because
10909 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10910 A misplaced 'break' also meant the decrypted final block might not be
10911 copied until the next read.
10914 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10915 a few extra parameters to the DH structure: these will be useful if
10916 for example we want the value of 'q' or implement X9.42 DH.
10919 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10920 provides hooks that allow the default DSA functions or functions on a
10921 "per key" basis to be replaced. This allows hardware acceleration and
10922 hardware key storage to be handled without major modification to the
10923 library. Also added low level modexp hooks and CRYPTO_EX structure and
10924 associated functions.
10927 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10928 as "read only": it can't be written to and the buffer it points to will
10929 not be freed. Reading from a read only BIO is much more efficient than
10930 a normal memory BIO. This was added because there are several times when
10931 an area of memory needs to be read from a BIO. The previous method was
10932 to create a memory BIO and write the data to it, this results in two
10933 copies of the data and an O(n^2) reading algorithm. There is a new
10934 function BIO_new_mem_buf() which creates a read only memory BIO from
10935 an area of memory. Also modified the PKCS#7 routines to use read only
10939 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10940 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10941 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10942 but a retry condition occurred while trying to read the rest.
10945 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10946 NID_pkcs7_encrypted by default: this was wrong since this should almost
10947 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10948 the encrypted data type: this is a more sensible place to put it and it
10949 allows the PKCS#12 code to be tidied up that duplicated this
10953 *) Changed obj_dat.pl script so it takes its input and output files on
10954 the command line. This should avoid shell escape redirection problems
10958 *) Initial support for certificate extension requests, these are included
10959 in things like Xenroll certificate requests. Included functions to allow
10960 extensions to be obtained and added.
10963 *) -crlf option to s_client and s_server for sending newlines as
10964 CRLF (as required by many protocols).
10967 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10969 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10970 [Ralf S. Engelschall]
10972 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10973 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10975 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10979 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10980 DH parameters/keys (q is lost during that conversion, but the resulting
10981 DH parameters contain its length).
10983 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10984 much faster than DH_generate_parameters (which creates parameters
10985 where p = 2*q + 1), and also the smaller q makes DH computations
10986 much more efficient (160-bit exponentiation instead of 1024-bit
10987 exponentiation); so this provides a convenient way to support DHE
10988 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10989 utter importance to use
10990 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10992 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10993 when such DH parameters are used, because otherwise small subgroup
10994 attacks may become possible!
10997 *) Avoid memory leak in i2d_DHparams.
11000 *) Allow the -k option to be used more than once in the enc program:
11001 this allows the same encrypted message to be read by multiple recipients.
11004 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11005 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11006 it will always use the numerical form of the OID, even if it has a short
11010 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11011 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11012 otherwise bn_mod_exp was called. In the case of hardware keys for example
11013 no private key components need be present and it might store extra data
11014 in the RSA structure, which cannot be accessed from bn_mod_exp.
11015 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11016 private key operations.
11019 *) Added support for SPARC Linux.
11022 *) pem_password_cb function type incompatibly changed from
11023 typedef int pem_password_cb(char *buf, int size, int rwflag);
11025 ....(char *buf, int size, int rwflag, void *userdata);
11026 so that applications can pass data to their callbacks:
11027 The PEM[_ASN1]_{read,write}... functions and macros now take an
11028 additional void * argument, which is just handed through whenever
11029 the password callback is called.
11030 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11032 New function SSL_CTX_set_default_passwd_cb_userdata.
11034 Compatibility note: As many C implementations push function arguments
11035 onto the stack in reverse order, the new library version is likely to
11036 interoperate with programs that have been compiled with the old
11037 pem_password_cb definition (PEM_whatever takes some data that
11038 happens to be on the stack as its last argument, and the callback
11039 just ignores this garbage); but there is no guarantee whatsoever that
11042 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11043 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11044 problems not only on Windows, but also on some Unix platforms.
11045 To avoid problematic command lines, these definitions are now in an
11046 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11047 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11050 *) MIPS III/IV assembler module is reimplemented.
11053 *) More DES library cleanups: remove references to srand/rand and
11054 delete an unused file.
11057 *) Add support for the the free Netwide assembler (NASM) under Win32,
11058 since not many people have MASM (ml) and it can be hard to obtain.
11059 This is currently experimental but it seems to work OK and pass all
11060 the tests. Check out INSTALL.W32 for info.
11063 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11064 without temporary keys kept an extra copy of the server key,
11065 and connections with temporary keys did not free everything in case
11069 *) New function RSA_check_key and new openssl rsa option -check
11070 for verifying the consistency of RSA keys.
11071 [Ulf Moeller, Bodo Moeller]
11073 *) Various changes to make Win32 compile work:
11074 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11075 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11076 comparison" warnings.
11077 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11080 *) Add a debugging option to PKCS#5 v2 key generation function: when
11081 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11082 derived keys are printed to stderr.
11085 *) Copy the flags in ASN1_STRING_dup().
11086 [Roman E. Pavlov <pre@mo.msk.ru>]
11088 *) The x509 application mishandled signing requests containing DSA
11089 keys when the signing key was also DSA and the parameters didn't match.
11091 It was supposed to omit the parameters when they matched the signing key:
11092 the verifying software was then supposed to automatically use the CA's
11093 parameters if they were absent from the end user certificate.
11095 Omitting parameters is no longer recommended. The test was also
11096 the wrong way round! This was probably due to unusual behaviour in
11097 EVP_cmp_parameters() which returns 1 if the parameters match.
11098 This meant that parameters were omitted when they *didn't* match and
11099 the certificate was useless. Certificates signed with 'ca' didn't have
11101 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11103 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11104 The interface is as follows:
11105 Applications can use
11106 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11107 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11108 "off" is now the default.
11109 The library internally uses
11110 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11111 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11112 to disable memory-checking temporarily.
11114 Some inconsistent states that previously were possible (and were
11115 even the default) are now avoided.
11117 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11118 with each memory chunk allocated; this is occasionally more helpful
11119 than just having a counter.
11121 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11123 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11127 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11128 which largely parallels "options", but is for changing API behaviour,
11129 whereas "options" are about protocol behaviour.
11130 Initial "mode" flags are:
11132 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11133 a single record has been written.
11134 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11135 retries use the same buffer location.
11136 (But all of the contents must be
11140 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11143 *) Fix problems with no-hmac etc.
11144 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11146 *) New functions RSA_get_default_method(), RSA_set_method() and
11147 RSA_get_method(). These allows replacement of RSA_METHODs without having
11148 to mess around with the internals of an RSA structure.
11151 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11152 Also really enable memory leak checks in openssl.c and in some
11154 [Chad C. Mulligan, Bodo Moeller]
11156 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11157 up the length of negative integers. This has now been simplified to just
11158 store the length when it is first determined and use it later, rather
11159 than trying to keep track of where data is copied and updating it to
11161 [Steve Henson, reported by Brien Wheeler
11162 <bwheeler@authentica-security.com>]
11164 *) Add a new function PKCS7_signatureVerify. This allows the verification
11165 of a PKCS#7 signature but with the signing certificate passed to the
11166 function itself. This contrasts with PKCS7_dataVerify which assumes the
11167 certificate is present in the PKCS#7 structure. This isn't always the
11168 case: certificates can be omitted from a PKCS#7 structure and be
11169 distributed by "out of band" means (such as a certificate database).
11172 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11173 function prototypes in pem.h, also change util/mkdef.pl to add the
11174 necessary function names.
11177 *) mk1mf.pl (used by Windows builds) did not properly read the
11178 options set by Configure in the top level Makefile, and Configure
11179 was not even able to write more than one option correctly.
11180 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11183 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11184 file to be loaded from a BIO or FILE pointer. The BIO version will
11185 for example allow memory BIOs to contain config info.
11188 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11189 Whoever hopes to achieve shared-library compatibility across versions
11190 must use this, not the compile-time macro.
11191 (Exercise 0.9.4: Which is the minimum library version required by
11193 Note: All this applies only to multi-threaded programs, others don't
11197 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11198 through a BIO pair triggered the default case, i.e.
11199 SSLerr(...,SSL_R_UNKNOWN_STATE).
11202 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11203 can use the SSL library even if none of the specific BIOs is
11207 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11208 for the encoded length.
11209 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11211 *) Add initial documentation of the X509V3 functions.
11214 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11215 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11216 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11217 secure PKCS#8 private key format with a high iteration count.
11220 *) Fix determination of Perl interpreter: A perl or perl5
11221 _directory_ in $PATH was also accepted as the interpreter.
11222 [Ralf S. Engelschall]
11224 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11225 wrong with it but it was very old and did things like calling
11226 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11227 unusual formatting.
11230 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11231 to use the new extension code.
11234 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11235 with macros. This should make it easier to change their form, add extra
11236 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11240 *) Add to configuration table a new entry that can specify an alternative
11241 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11242 according to Mark Crispin <MRC@Panda.COM>.
11246 *) DES CBC did not update the IV. Weird.
11249 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11250 Changing the behaviour of the former might break existing programs --
11251 where IV updating is needed, des_ncbc_encrypt can be used.
11254 *) When bntest is run from "make test" it drives bc to check its
11255 calculations, as well as internally checking them. If an internal check
11256 fails, it needs to cause bc to give a non-zero result or make test carries
11257 on without noticing the failure. Fixed.
11260 *) DES library cleanups.
11263 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11264 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11265 ciphers. NOTE: although the key derivation function has been verified
11266 against some published test vectors it has not been extensively tested
11267 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11271 *) Instead of "mkdir -p", which is not fully portable, use new
11272 Perl script "util/mkdir-p.pl".
11275 *) Rewrite the way password based encryption (PBE) is handled. It used to
11276 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11277 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11278 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11279 the 'parameter' field of the AlgorithmIdentifier is passed to the
11280 underlying key generation function so it must do its own ASN1 parsing.
11281 This has also changed the EVP_PBE_CipherInit() function which now has a
11282 'parameter' argument instead of literal salt and iteration count values
11283 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11286 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11287 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11288 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11289 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11290 value was just used as a "magic string" and not used directly its
11291 value doesn't matter.
11294 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11298 *) "linux-sparc64" configuration (ultrapenguin).
11299 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11300 "linux-sparc" configuration.
11301 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11303 *) config now generates no-xxx options for missing ciphers.
11306 *) Support the EBCDIC character set (work in progress).
11307 File ebcdic.c not yet included because it has a different license.
11308 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11310 *) Support BS2000/OSD-POSIX.
11311 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11313 *) Make callbacks for key generation use void * instead of char *.
11316 *) Make S/MIME samples compile (not yet tested).
11319 *) Additional typesafe stacks.
11322 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11326 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11328 *) New configuration variant "sco5-gcc".
11330 *) Updated some demos.
11331 [Sean O Riordain, Wade Scholine]
11333 *) Add missing BIO_free at exit of pkcs12 application.
11336 *) Fix memory leak in conf.c.
11339 *) Updates for Win32 to assembler version of MD5.
11342 *) Set #! path to perl in apps/der_chop to where we found it
11343 instead of using a fixed path.
11346 *) SHA library changes for irix64-mips4-cc.
11349 *) Improvements for VMS support.
11353 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11355 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11356 This also avoids the problems with SC4.2 and unpatched SC5.
11357 [Andy Polyakov <appro@fy.chalmers.se>]
11359 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11360 These are required because of the typesafe stack would otherwise break
11361 existing code. If old code used a structure member which used to be STACK
11362 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11363 sk_num or sk_value it would produce an error because the num, data members
11364 are not present in STACK_OF. Now it just produces a warning. sk_set
11365 replaces the old method of assigning a value to sk_value
11366 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11367 that does this will no longer work (and should use sk_set instead) but
11368 this could be regarded as a "questionable" behaviour anyway.
11371 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11372 correctly handle encrypted S/MIME data.
11375 *) Change type of various DES function arguments from des_cblock
11376 (which means, in function argument declarations, pointer to char)
11377 to des_cblock * (meaning pointer to array with 8 char elements),
11378 which allows the compiler to do more typechecking; it was like
11379 that back in SSLeay, but with lots of ugly casts.
11381 Introduce new type const_des_cblock.
11384 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11385 problems: find RecipientInfo structure that matches recipient certificate
11386 and initialise the ASN1 structures properly based on passed cipher.
11389 *) Belatedly make the BN tests actually check the results.
11392 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11393 to and from BNs: it was completely broken. New compilation option
11394 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11395 key elements as negative integers.
11398 *) Reorganize and speed up MD5.
11399 [Andy Polyakov <appro@fy.chalmers.se>]
11402 [Richard Levitte <richard@levitte.org>]
11404 *) New option -out to asn1parse to allow the parsed structure to be
11405 output to a file. This is most useful when combined with the -strparse
11406 option to examine the output of things like OCTET STRINGS.
11409 *) Make SSL library a little more fool-proof by not requiring any longer
11410 that SSL_set_{accept,connect}_state be called before
11411 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11412 in many applications because usually everything *appeared* to work as
11413 intended anyway -- now it really works as intended).
11416 *) Move openssl.cnf out of lib/.
11419 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11420 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11421 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11422 [Ralf S. Engelschall]
11424 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11425 handle PKCS#7 enveloped data properly.
11426 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11428 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11429 copying pointers. The cert_st handling is changed by this in
11430 various ways (and thus what used to be known as ctx->default_cert
11431 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11432 any longer when s->cert does not give us what we need).
11433 ssl_cert_instantiate becomes obsolete by this change.
11434 As soon as we've got the new code right (possibly it already is?),
11435 we have solved a couple of bugs of the earlier code where s->cert
11436 was used as if it could not have been shared with other SSL structures.
11438 Note that using the SSL API in certain dirty ways now will result
11439 in different behaviour than observed with earlier library versions:
11440 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11441 does not influence s as it used to.
11443 In order to clean up things more thoroughly, inside SSL_SESSION
11444 we don't use CERT any longer, but a new structure SESS_CERT
11445 that holds per-session data (if available); currently, this is
11446 the peer's certificate chain and, for clients, the server's certificate
11447 and temporary key. CERT holds only those values that can have
11448 meaningful defaults in an SSL_CTX.
11451 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11452 from the internal representation. Various PKCS#7 fixes: remove some
11453 evil casts and set the enc_dig_alg field properly based on the signing
11457 *) Allow PKCS#12 password to be set from the command line or the
11458 environment. Let 'ca' get its config file name from the environment
11459 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11463 *) Allow certificate policies extension to use an IA5STRING for the
11464 organization field. This is contrary to the PKIX definition but
11465 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11469 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11470 without disallowing inline assembler and the like for non-pedantic builds.
11473 *) Support Borland C++ builder.
11474 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11476 *) Support Mingw32.
11479 *) SHA-1 cleanups and performance enhancements.
11480 [Andy Polyakov <appro@fy.chalmers.se>]
11482 *) Sparc v8plus assembler for the bignum library.
11483 [Andy Polyakov <appro@fy.chalmers.se>]
11485 *) Accept any -xxx and +xxx compiler options in Configure.
11488 *) Update HPUX configuration.
11491 *) Add missing sk_<type>_unshift() function to safestack.h
11492 [Ralf S. Engelschall]
11494 *) New function SSL_CTX_use_certificate_chain_file that sets the
11495 "extra_cert"s in addition to the certificate. (This makes sense
11496 only for "PEM" format files, as chains as a whole are not
11500 *) Support verify_depth from the SSL API.
11501 x509_vfy.c had what can be considered an off-by-one-error:
11502 Its depth (which was not part of the external interface)
11503 was actually counting the number of certificates in a chain;
11504 now it really counts the depth.
11507 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11508 instead of X509err, which often resulted in confusing error
11509 messages since the error codes are not globally unique
11510 (e.g. an alleged error in ssl3_accept when a certificate
11511 didn't match the private key).
11513 *) New function SSL_CTX_set_session_id_context that allows to set a default
11514 value (so that you don't need SSL_set_session_id_context for each
11515 connection using the SSL_CTX).
11518 *) OAEP decoding bug fix.
11521 *) Support INSTALL_PREFIX for package builders, as proposed by
11525 *) New Configure options "threads" and "no-threads". For systems
11526 where the proper compiler options are known (currently Solaris
11527 and Linux), "threads" is the default.
11530 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11533 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11534 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11535 such as /usr/local/bin.
11538 *) "make linux-shared" to build shared libraries.
11539 [Niels Poppe <niels@netbox.org>]
11541 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11544 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11545 extension adding in x509 utility.
11548 *) Remove NOPROTO sections and error code comments.
11551 *) Partial rewrite of the DEF file generator to now parse the ANSI
11555 *) New Configure options --prefix=DIR and --openssldir=DIR.
11558 *) Complete rewrite of the error code script(s). It is all now handled
11559 by one script at the top level which handles error code gathering,
11560 header rewriting and C source file generation. It should be much better
11561 than the old method: it now uses a modified version of Ulf's parser to
11562 read the ANSI prototypes in all header files (thus the old K&R definitions
11563 aren't needed for error creation any more) and do a better job of
11564 translating function codes into names. The old 'ASN1 error code imbedded
11565 in a comment' is no longer necessary and it doesn't use .err files which
11566 have now been deleted. Also the error code call doesn't have to appear all
11567 on one line (which resulted in some large lines...).
11570 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11573 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11574 0 (which usually indicates a closed connection), but continue reading.
11577 *) Fix some race conditions.
11580 *) Add support for CRL distribution points extension. Add Certificate
11581 Policies and CRL distribution points documentation.
11584 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11587 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11588 8 of keying material. Merlin has also confirmed interop with this fix
11589 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11590 [Merlin Hughes <merlin@baltimore.ie>]
11592 *) Fix lots of warnings.
11593 [Richard Levitte <levitte@stacken.kth.se>]
11595 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11596 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11597 [Richard Levitte <levitte@stacken.kth.se>]
11599 *) Fix problems with sizeof(long) == 8.
11600 [Andy Polyakov <appro@fy.chalmers.se>]
11602 *) Change functions to ANSI C.
11605 *) Fix typos in error codes.
11606 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11608 *) Remove defunct assembler files from Configure.
11611 *) SPARC v8 assembler BIGNUM implementation.
11612 [Andy Polyakov <appro@fy.chalmers.se>]
11614 *) Support for Certificate Policies extension: both print and set.
11615 Various additions to support the r2i method this uses.
11618 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11619 return a const string when you are expecting an allocated buffer.
11622 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11623 types DirectoryString and DisplayText.
11626 *) Add code to allow r2i extensions to access the configuration database,
11627 add an LHASH database driver and add several ctx helper functions.
11630 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11631 fail when they extended the size of a BIGNUM.
11634 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11635 support typesafe stack.
11638 *) Fix typo in SSL_[gs]et_options().
11639 [Nils Frostberg <nils@medcom.se>]
11641 *) Delete various functions and files that belonged to the (now obsolete)
11642 old X509V3 handling code.
11645 *) New Configure option "rsaref".
11648 *) Don't auto-generate pem.h.
11651 *) Introduce type-safe ASN.1 SETs.
11654 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11655 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11657 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11658 that links with OpenSSL (well at least cause lots of warnings), but fear
11659 not: the conversion is trivial, and it eliminates loads of evil casts. A
11660 few STACKed things have been converted already. Feel free to convert more.
11661 In the fullness of time, I'll do away with the STACK type altogether.
11664 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11665 specified in <certfile> by updating the entry in the index.txt file.
11666 This way one no longer has to edit the index.txt file manually for
11667 revoking a certificate. The -revoke option does the gory details now.
11668 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11670 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11671 `-text' option at all and this way the `-noout -text' combination was
11672 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11673 [Ralf S. Engelschall]
11675 *) Make sure a corresponding plain text error message exists for the
11676 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11677 verify callback function determined that a certificate was revoked.
11678 [Ralf S. Engelschall]
11680 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11681 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11682 all available cipers including rc5, which was forgotten until now.
11683 In order to let the testing shell script know which algorithms
11684 are available, a new (up to now undocumented) command
11685 "openssl list-cipher-commands" is used.
11688 *) Bugfix: s_client occasionally would sleep in select() when
11689 it should have checked SSL_pending() first.
11692 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11693 the raw DSA values prior to ASN.1 encoding.
11696 *) Tweaks to Configure
11697 [Niels Poppe <niels@netbox.org>]
11699 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11703 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11706 *) New config option to avoid instructions that are illegal on the 80386.
11707 The default code is faster, but requires at least a 486.
11710 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11711 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11712 same as SSL2_VERSION anyway.
11715 *) New "-showcerts" option for s_client.
11718 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11719 application. Various cleanups and fixes.
11722 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11723 modify error routines to work internally. Add error codes and PBE init
11724 to library startup routines.
11727 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11728 packing functions to asn1 and evp. Changed function names and error
11729 codes along the way.
11732 *) PKCS12 integration: and so it begins... First of several patches to
11733 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11734 objects to objects.h
11737 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11738 and display support for Thawte strong extranet extension.
11741 *) Add LinuxPPC support.
11742 [Jeff Dubrule <igor@pobox.org>]
11744 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11745 bn_div_words in alpha.s.
11746 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11748 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11749 OAEP isn't supported when OpenSSL is built with RSAref.
11750 [Ulf Moeller <ulf@fitug.de>]
11752 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11753 so they no longer are missing under -DNOPROTO.
11754 [Soren S. Jorvang <soren@t.dk>]
11757 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11759 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11760 doesn't work when the session is reused. Coming soon!
11763 *) Fix a security hole, that allows sessions to be reused in the wrong
11764 context thus bypassing client cert protection! All software that uses
11765 client certs and session caches in multiple contexts NEEDS PATCHING to
11766 allow session reuse! A fuller solution is in the works.
11767 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11769 *) Some more source tree cleanups (removed obsolete files
11770 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11771 permission on "config" script to be executable) and a fix for the INSTALL
11773 [Ulf Moeller <ulf@fitug.de>]
11775 *) Remove some legacy and erroneous uses of malloc, free instead of
11777 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11779 *) Make rsa_oaep_test return non-zero on error.
11780 [Ulf Moeller <ulf@fitug.de>]
11782 *) Add support for native Solaris shared libraries. Configure
11783 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11784 if someone would make that last step automatic.
11785 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11787 *) ctx_size was not built with the right compiler during "make links". Fixed.
11790 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11791 except NULL ciphers". This means the default cipher list will no longer
11792 enable NULL ciphers. They need to be specifically enabled e.g. with
11793 the string "DEFAULT:eNULL".
11796 *) Fix to RSA private encryption routines: if p < q then it would
11797 occasionally produce an invalid result. This will only happen with
11798 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11801 *) Be less restrictive and allow also `perl util/perlpath.pl
11802 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11803 because this way one can also use an interpreter named `perl5' (which is
11804 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11805 installed as `perl').
11806 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11808 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11809 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11811 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11812 advapi32.lib to Win32 build and change the pem test comparison
11813 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11814 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11815 and crypto/des/ede_cbcm_enc.c.
11818 *) DES quad checksum was broken on big-endian architectures. Fixed.
11821 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11822 Win32 test batch file so it (might) work again. The Win32 test batch file
11823 is horrible: I feel ill....
11826 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11827 in e_os.h. Audit of header files to check ANSI and non ANSI
11828 sections: 10 functions were absent from non ANSI section and not exported
11829 from Windows DLLs. Fixed up libeay.num for new functions.
11832 *) Make `openssl version' output lines consistent.
11833 [Ralf S. Engelschall]
11835 *) Fix Win32 symbol export lists for BIO functions: Added
11836 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11837 to ms/libeay{16,32}.def.
11838 [Ralf S. Engelschall]
11840 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11841 fine under Unix and passes some trivial tests I've now added. But the
11842 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11843 added to make sure no one expects that this stuff really works in the
11844 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11845 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11847 [Ralf S. Engelschall]
11849 *) Fix the generation of two part addresses in perl.
11850 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11852 *) Add config entry for Linux on MIPS.
11853 [John Tobey <jtobey@channel1.com>]
11855 *) Make links whenever Configure is run, unless we are on Windoze.
11858 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11859 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11863 *) Add a useful kludge to allow package maintainers to specify compiler and
11864 other platforms details on the command line without having to patch the
11865 Configure script everytime: One now can use ``perl Configure
11866 <id>:<details>'', i.e. platform ids are allowed to have details appended
11867 to them (separated by colons). This is treated as there would be a static
11868 pre-configured entry in Configure's %table under key <id> with value
11869 <details> and ``perl Configure <id>'' is called. So, when you want to
11870 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11871 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11872 now, which overrides the FreeBSD-elf entry on-the-fly.
11873 [Ralf S. Engelschall]
11875 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11878 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11879 on the `perl Configure ...' command line. This way one can compile
11880 OpenSSL libraries with Position Independent Code (PIC) which is needed
11881 for linking it into DSOs.
11882 [Ralf S. Engelschall]
11884 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11888 *) Cleaned up the LICENSE document: The official contact for any license
11889 questions now is the OpenSSL core team under openssl-core@openssl.org.
11890 And add a paragraph about the dual-license situation to make sure people
11891 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11892 to the OpenSSL toolkit.
11893 [Ralf S. Engelschall]
11895 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11896 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11897 Additionally cleaned up the `make links' target: Remove unnecessary
11898 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11899 to speed processing and no longer clutter the display with confusing
11900 stuff. Instead only the actually done links are displayed.
11901 [Ralf S. Engelschall]
11903 *) Permit null encryption ciphersuites, used for authentication only. It used
11904 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11905 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11909 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11910 signed attributes when verifying signatures (this would break them),
11911 the detached data encoding was wrong and public keys obtained using
11912 X509_get_pubkey() weren't freed.
11915 *) Add text documentation for the BUFFER functions. Also added a work around
11916 to a Win95 console bug. This was triggered by the password read stuff: the
11917 last character typed gets carried over to the next fread(). If you were
11918 generating a new cert request using 'req' for example then the last
11919 character of the passphrase would be CR which would then enter the first
11923 *) Added the new `Includes OpenSSL Cryptography Software' button as
11924 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11925 button and can be used by applications based on OpenSSL to show the
11926 relationship to the OpenSSL project.
11927 [Ralf S. Engelschall]
11929 *) Remove confusing variables in function signatures in files
11930 ssl/ssl_lib.c and ssl/ssl.h.
11931 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11933 *) Don't install bss_file.c under PREFIX/include/
11934 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11936 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11937 functions that return function pointers and has support for NT specific
11938 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11939 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11940 unsigned to signed types: this was killing the Win32 compile.
11943 *) Add new certificate file to stack functions,
11944 SSL_add_dir_cert_subjects_to_stack() and
11945 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11946 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11947 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11948 This means that Apache-SSL and similar packages don't have to mess around
11949 to add as many CAs as they want to the preferred list.
11952 *) Experiment with doxygen documentation. Currently only partially applied to
11954 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11955 openssl.doxy as the configuration file.
11958 *) Get rid of remaining C++-style comments which strict C compilers hate.
11959 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11961 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11962 compiled in by default: it has problems with large keys.
11965 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11966 DH private keys and/or callback functions which directly correspond to
11967 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11968 is needed for applications which have to configure certificates on a
11969 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11971 For the RSA certificate situation is makes no difference, but
11972 for the DSA certificate situation this fixes the "no shared cipher"
11973 problem where the OpenSSL cipher selection procedure failed because the
11974 temporary keys were not overtaken from the context and the API provided
11975 no way to reconfigure them.
11976 The new functions now let applications reconfigure the stuff and they
11977 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11978 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11979 non-public-API function ssl_cert_instantiate() is used as a helper
11980 function and also to reduce code redundancy inside ssl_rsa.c.
11981 [Ralf S. Engelschall]
11983 *) Move s_server -dcert and -dkey options out of the undocumented feature
11984 area because they are useful for the DSA situation and should be
11985 recognized by the users.
11986 [Ralf S. Engelschall]
11988 *) Fix the cipher decision scheme for export ciphers: the export bits are
11989 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11990 SSL_EXP_MASK. So, the original variable has to be used instead of the
11991 already masked variable.
11992 [Richard Levitte <levitte@stacken.kth.se>]
11994 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11995 [Richard Levitte <levitte@stacken.kth.se>]
11997 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11998 from `int' to `unsigned int' because it's a length and initialized by
11999 EVP_DigestFinal() which expects an `unsigned int *'.
12000 [Richard Levitte <levitte@stacken.kth.se>]
12002 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12003 script. Instead use the usual Shell->Perl transition trick.
12004 [Ralf S. Engelschall]
12006 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12007 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12008 -noout -modulus' as it's already the case for `openssl rsa -noout
12009 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12010 currently the public key is printed (a decision which was already done by
12011 `openssl dsa -modulus' in the past) which serves a similar purpose.
12012 Additionally the NO_RSA no longer completely removes the whole -modulus
12013 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12015 [Ralf S. Engelschall]
12017 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12018 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12019 [Arne Ansper <arne@ats.cyber.ee>]
12021 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12022 to be added. Now both 'req' and 'ca' can use new objects defined in the
12026 *) Add cool BIO that does syslog (or event log on NT).
12027 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12029 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12030 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12031 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12032 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12035 *) Add preliminary config info for new extension code.
12038 *) Make RSA_NO_PADDING really use no padding.
12039 [Ulf Moeller <ulf@fitug.de>]
12041 *) Generate errors when private/public key check is done.
12044 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12045 for some CRL extensions and new objects added.
12048 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12049 key usage extension and fuller support for authority key id.
12052 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12053 padding method for RSA, which is recommended for new applications in PKCS
12054 #1 v2.0 (RFC 2437, October 1998).
12055 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12056 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12057 against Bleichbacher's attack on RSA.
12058 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12061 *) Updates to the new SSL compression code
12062 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12064 *) Fix so that the version number in the master secret, when passed
12065 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12066 (because the server will not accept higher), that the version number
12067 is 0x03,0x01, not 0x03,0x00
12068 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12070 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12071 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12072 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12075 *) Support for RAW extensions where an arbitrary extension can be
12076 created by including its DER encoding. See apps/openssl.cnf for
12080 *) Make sure latest Perl versions don't interpret some generated C array
12081 code as Perl array code in the crypto/err/err_genc.pl script.
12082 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12084 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12085 not many people have the assembler. Various Win32 compilation fixes and
12086 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12087 build instructions.
12090 *) Modify configure script 'Configure' to automatically create crypto/date.h
12091 file under Win32 and also build pem.h from pem.org. New script
12092 util/mkfiles.pl to create the MINFO file on environments that can't do a
12093 'make files': perl util/mkfiles.pl >MINFO should work.
12096 *) Major rework of DES function declarations, in the pursuit of correctness
12097 and purity. As a result, many evil casts evaporated, and some weirdness,
12098 too. You may find this causes warnings in your code. Zapping your evil
12099 casts will probably fix them. Mostly.
12102 *) Fix for a typo in asn1.h. Bug fix to object creation script
12103 obj_dat.pl. It considered a zero in an object definition to mean
12104 "end of object": none of the objects in objects.h have any zeros
12105 so it wasn't spotted.
12106 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12108 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12109 Masking (CBCM). In the absence of test vectors, the best I have been able
12110 to do is check that the decrypt undoes the encrypt, so far. Send me test
12111 vectors if you have them.
12114 *) Correct calculation of key length for export ciphers (too much space was
12115 allocated for null ciphers). This has not been tested!
12118 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12119 message is now correct (it understands "crypto" and "ssl" on its
12120 command line). There is also now an "update" option. This will update
12121 the util/ssleay.num and util/libeay.num files with any new functions.
12123 perl util/mkdef.pl crypto ssl update
12124 it will update them.
12127 *) Overhauled the Perl interface (perl/*):
12128 - ported BN stuff to OpenSSL's different BN library
12129 - made the perl/ source tree CVS-aware
12130 - renamed the package from SSLeay to OpenSSL (the files still contain
12131 their history because I've copied them in the repository)
12132 - removed obsolete files (the test scripts will be replaced
12133 by better Test::Harness variants in the future)
12134 [Ralf S. Engelschall]
12136 *) First cut for a very conservative source tree cleanup:
12137 1. merge various obsolete readme texts into doc/ssleay.txt
12138 where we collect the old documents and readme texts.
12139 2. remove the first part of files where I'm already sure that we no
12140 longer need them because of three reasons: either they are just temporary
12141 files which were left by Eric or they are preserved original files where
12142 I've verified that the diff is also available in the CVS via "cvs diff
12143 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12144 the crypto/md/ stuff).
12145 [Ralf S. Engelschall]
12147 *) More extension code. Incomplete support for subject and issuer alt
12148 name, issuer and authority key id. Change the i2v function parameters
12149 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12150 what that's for :-) Fix to ASN1 macro which messed up
12151 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12154 *) Preliminary support for ENUMERATED type. This is largely copied from the
12158 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12159 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12161 *) Make sure `make rehash' target really finds the `openssl' program.
12162 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12164 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12165 like to hear about it if this slows down other processors.
12168 *) Add CygWin32 platform information to Configure script.
12169 [Alan Batie <batie@aahz.jf.intel.com>]
12171 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12172 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12174 *) New program nseq to manipulate netscape certificate sequences
12177 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12181 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12182 but the BN code had some problems that would cause failures when
12183 doing certificate verification and some other functions.
12184 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12186 *) Add ASN1 and PEM code to support netscape certificate sequences.
12189 *) Add ASN1 and PEM code to support netscape certificate sequences.
12192 *) Add several PKIX and private extended key usage OIDs.
12195 *) Modify the 'ca' program to handle the new extension code. Modify
12196 openssl.cnf for new extension format, add comments.
12199 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12200 and add a sample to openssl.cnf so req -x509 now adds appropriate
12204 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12205 error code, add initial support to X509_print() and x509 application.
12208 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12209 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12210 stuff is currently isolated and isn't even compiled yet.
12213 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12214 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12215 Removed the versions check from X509 routines when loading extensions:
12216 this allows certain broken certificates that don't set the version
12217 properly to be processed.
12220 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12221 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12222 can still be regenerated with "make depend".
12225 *) Spelling mistake in C version of CAST-128.
12226 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12228 *) Changes to the error generation code. The perl script err-code.pl
12229 now reads in the old error codes and retains the old numbers, only
12230 adding new ones if necessary. It also only changes the .err files if new
12231 codes are added. The makefiles have been modified to only insert errors
12232 when needed (to avoid needlessly modifying header files). This is done
12233 by only inserting errors if the .err file is newer than the auto generated
12234 C file. To rebuild all the error codes from scratch (the old behaviour)
12235 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12236 or delete all the .err files.
12239 *) CAST-128 was incorrectly implemented for short keys. The C version has
12240 been fixed, but is untested. The assembler versions are also fixed, but
12241 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12242 to regenerate it if needed.
12243 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12244 Hagino <itojun@kame.net>]
12246 *) File was opened incorrectly in randfile.c.
12247 [Ulf Möller <ulf@fitug.de>]
12249 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12250 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12251 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12252 al: it's just almost always a UTCTime. Note this patch adds new error
12253 codes so do a "make errors" if there are problems.
12256 *) Correct Linux 1 recognition in config.
12257 [Ulf Möller <ulf@fitug.de>]
12259 *) Remove pointless MD5 hash when using DSA keys in ca.
12260 [Anonymous <nobody@replay.com>]
12262 *) Generate an error if given an empty string as a cert directory. Also
12263 generate an error if handed NULL (previously returned 0 to indicate an
12264 error, but didn't set one).
12265 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12267 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12270 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12271 parameters. This was causing a warning which killed off the Win32 compile.
12274 *) Remove C++ style comments from crypto/bn/bn_local.h.
12275 [Neil Costigan <neil.costigan@celocom.com>]
12277 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12278 based on a text string, looking up short and long names and finally
12279 "dot" format. The "dot" format stuff didn't work. Added new function
12280 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12281 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12282 OID is not part of the table.
12285 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12286 X509_LOOKUP_by_alias().
12289 *) Sort openssl functions by name.
12292 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12293 encryption from sample DSA keys (in case anyone is interested the password
12297 *) Make _all_ *_free functions accept a NULL pointer.
12298 [Frans Heymans <fheymans@isaserver.be>]
12300 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12302 [Anonymous <nobody@replay.com>]
12304 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12305 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12307 *) Don't blow it for numeric -newkey arguments to apps/req.
12308 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12310 *) Temp key "for export" tests were wrong in s3_srvr.c.
12311 [Anonymous <nobody@replay.com>]
12313 *) Add prototype for temp key callback functions
12314 SSL_CTX_set_tmp_{rsa,dh}_callback().
12317 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12318 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12321 *) X509_name_add_entry() freed the wrong thing after an error.
12322 [Arne Ansper <arne@ats.cyber.ee>]
12324 *) rsa_eay.c would attempt to free a NULL context.
12325 [Arne Ansper <arne@ats.cyber.ee>]
12327 *) BIO_s_socket() had a broken should_retry() on Windoze.
12328 [Arne Ansper <arne@ats.cyber.ee>]
12330 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12331 [Arne Ansper <arne@ats.cyber.ee>]
12333 *) Make sure the already existing X509_STORE->depth variable is initialized
12334 in X509_STORE_new(), but document the fact that this variable is still
12335 unused in the certificate verification process.
12336 [Ralf S. Engelschall]
12338 *) Fix the various library and apps files to free up pkeys obtained from
12339 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12342 *) Fix reference counting in X509_PUBKEY_get(). This makes
12343 demos/maurice/example2.c work, amongst others, probably.
12344 [Steve Henson and Ben Laurie]
12346 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12347 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12348 are no longer created. This way we have a single and consistent command
12349 line interface `openssl <command>', similar to `cvs <command>'.
12350 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12352 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12353 BIT STRING wrapper always have zero unused bits.
12356 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12359 *) Make the top-level INSTALL documentation easier to understand.
12362 *) Makefiles updated to exit if an error occurs in a sub-directory
12363 make (including if user presses ^C) [Paul Sutton]
12365 *) Make Montgomery context stuff explicit in RSA data structure.
12368 *) Fix build order of pem and err to allow for generated pem.h.
12371 *) Fix renumbering bug in X509_NAME_delete_entry().
12374 *) Enhanced the err-ins.pl script so it makes the error library number
12375 global and can add a library name. This is needed for external ASN1 and
12376 other error libraries.
12379 *) Fixed sk_insert which never worked properly.
12382 *) Fix ASN1 macros so they can handle indefinite length constructed
12383 EXPLICIT tags. Some non standard certificates use these: they can now
12387 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12388 into a single doc/ssleay.txt bundle. This way the information is still
12389 preserved but no longer messes up this directory. Now it's new room for
12390 the new set of documentation files.
12391 [Ralf S. Engelschall]
12393 *) SETs were incorrectly DER encoded. This was a major pain, because they
12394 shared code with SEQUENCEs, which aren't coded the same. This means that
12395 almost everything to do with SETs or SEQUENCEs has either changed name or
12396 number of arguments.
12397 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12399 *) Fix test data to work with the above.
12402 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12403 was already fixed by Eric for 0.9.1 it seems.
12404 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12406 *) Autodetect FreeBSD3.
12409 *) Fix various bugs in Configure. This affects the following platforms:
12413 unixware-2.0-pentium
12417 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12418 before they are needed.
12421 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12425 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12427 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12428 changed SSLeay to OpenSSL in version strings.
12429 [Ralf S. Engelschall]
12431 *) Some fixups to the top-level documents.
12434 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12435 because the symlink to include/ was missing.
12436 [Ralf S. Engelschall]
12438 *) Incorporated the popular no-RSA/DSA-only patches
12439 which allow to compile a RSA-free SSLeay.
12440 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12442 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12443 when "ssleay" is still not found.
12444 [Ralf S. Engelschall]
12446 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12447 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12449 *) Updated the README file.
12450 [Ralf S. Engelschall]
12452 *) Added various .cvsignore files in the CVS repository subdirs
12453 to make a "cvs update" really silent.
12454 [Ralf S. Engelschall]
12456 *) Recompiled the error-definition header files and added
12457 missing symbols to the Win32 linker tables.
12458 [Ralf S. Engelschall]
12460 *) Cleaned up the top-level documents;
12461 o new files: CHANGES and LICENSE
12462 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12463 o merged COPYRIGHT into LICENSE
12464 o removed obsolete TODO file
12465 o renamed MICROSOFT to INSTALL.W32
12466 [Ralf S. Engelschall]
12468 *) Removed dummy files from the 0.9.1b source tree:
12469 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12470 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12471 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12472 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12473 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12474 [Ralf S. Engelschall]
12476 *) Added various platform portability fixes.
12479 *) The Genesis of the OpenSSL rpject:
12480 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12481 Young and Tim J. Hudson created while they were working for C2Net until
12483 [The OpenSSL Project]
12486 Changes between 0.9.0b and 0.9.1b [not released]
12488 *) Updated a few CA certificates under certs/
12491 *) Changed some BIGNUM api stuff.
12494 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12495 DGUX x86, Linux Alpha, etc.
12498 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12499 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12503 *) Add -strparse option to asn1pars program which parses nested
12505 [Dr Stephen Henson <shenson@bigfoot.com>]
12507 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12510 *) DSA fix for "ca" program.
12513 *) Added "-genkey" option to "dsaparam" program.
12516 *) Added RIPE MD160 (rmd160) message digest.
12519 *) Added -a (all) option to "ssleay version" command.
12522 *) Added PLATFORM define which is the id given to Configure.
12525 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12528 *) Extended the ASN.1 parser routines.
12531 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12534 *) Added a BN_CTX to the BN library.
12537 *) Fixed the weak key values in DES library
12540 *) Changed API in EVP library for cipher aliases.
12543 *) Added support for RC2/64bit cipher.
12546 *) Converted the lhash library to the crypto/mem.c functions.
12549 *) Added more recognized ASN.1 object ids.
12552 *) Added more RSA padding checks for SSL/TLS.
12555 *) Added BIO proxy/filter functionality.
12558 *) Added extra_certs to SSL_CTX which can be used
12559 send extra CA certificates to the client in the CA cert chain sending
12560 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12563 *) Now Fortezza is denied in the authentication phase because
12564 this is key exchange mechanism is not supported by SSLeay at all.
12567 *) Additional PKCS1 checks.
12570 *) Support the string "TLSv1" for all TLS v1 ciphers.
12573 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12574 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12577 *) Fixed a few memory leaks.
12580 *) Fixed various code and comment typos.
12583 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12584 bytes sent in the client random.
12585 [Edward Bishop <ebishop@spyglass.com>]