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 1.1.0f and 1.1.1 [xx XXX xxxx]
12 *) The last traces of Netware support, first removed in 1.1.0, have
16 *) Get rid of Makefile.shared, and in the process, make the processing
17 of certain files (rc.obj, or the .def/.map/.opt files produced from
18 the ordinal files) more visible and hopefully easier to trace and
19 debug (or make silent).
22 *) Make it possible to have environment variable assignments as
23 arguments to config / Configure.
26 *) Add multi-prime RSA (RFC 8017) support.
29 *) Add SM3 implemented according to GB/T 32905-2016
30 [ Jack Lloyd <jack.lloyd@ribose.com>,
31 Ronald Tse <ronald.tse@ribose.com>,
32 Erick Borsboom <erick.borsboom@ribose.com> ]
34 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
35 as documented in RFC6066.
36 Based on a patch from Tomasz Moń
37 [Filipe Raimundo da Silva]
39 *) Add SM4 implemented according to GB/T 32907-2016.
40 [ Jack Lloyd <jack.lloyd@ribose.com>,
41 Ronald Tse <ronald.tse@ribose.com>,
42 Erick Borsboom <erick.borsboom@ribose.com> ]
44 *) Reimplement -newreq-nodes and ERR_error_string_n; the
45 original author does not agree with the license change.
48 *) Add ARIA AEAD TLS support.
51 *) Some macro definitions to support VS6 have been removed. Visual
52 Studio 6 has not worked since 1.1.0
55 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
56 without clearing the errors.
59 *) Add "atfork" functions. If building on a system that without
60 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
61 requirements. The RAND facility now uses/requires this.
67 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
68 not possible to disable entirely. However, it's still possible to
69 disable the console reading UI method, UI_OpenSSL() (use UI_null()
72 To disable, configure with 'no-ui-console'. 'no-ui' is still
73 possible to use as an alias. Check at compile time with the
74 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
75 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
78 *) Add a STORE module, which implements a uniform and URI based reader of
79 stores that can contain keys, certificates, CRLs and numerous other
80 objects. The main API is loosely based on a few stdio functions,
81 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
82 OSSL_STORE_error and OSSL_STORE_close.
83 The implementation uses backends called "loaders" to implement arbitrary
84 URI schemes. There is one built in "loader" for the 'file' scheme.
87 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
88 then adjusted to work on FreeBSD 8.4 as well.
89 Enable by configuring with 'enable-devcryptoeng'. This is done by default
90 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
93 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
94 util/mkerr.pl, which is adapted to allow those prefixes, leading to
95 error code calls like this:
97 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
99 With this change, we claim the namespaces OSSL and OPENSSL in a manner
100 that can be encoded in C. For the foreseeable future, this will only
102 [Richard Levitte and Tim Hudson]
104 *) Removed BSD cryptodev engine.
107 *) Add a build target 'build_all_generated', to build all generated files
108 and only that. This can be used to prepare everything that requires
109 things like perl for a system that lacks perl and then move everything
110 to that system and do the rest of the build there.
113 *) In the UI interface, make it possible to duplicate the user data. This
114 can be used by engines that need to retain the data for a longer time
115 than just the call where this user data is passed.
118 *) Ignore the '-named_curve auto' value for compatibility of applications
120 [Tomas Mraz <tmraz@fedoraproject.org>]
122 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
123 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
124 alerts across multiple records (some of which could be empty). In practice
125 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
126 prohibts this altogether and other libraries (BoringSSL, NSS) do not
127 support this at all. Supporting it adds significant complexity to the
128 record layer, and its removal is unlikely to cause inter-operability
132 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
133 with Z. These are meant to replace LONG and ZLONG and to be size safe.
134 The use of LONG and ZLONG is discouraged and scheduled for deprecation
138 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
139 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
140 [Richard Levitte, Andy Polyakov]
142 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
146 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
147 platform rather than 'mingw'.
150 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
151 success if they are asked to add an object which already exists
152 in the store. This change cascades to other functions which load
153 certificates and CRLs.
156 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
157 facilitate stack unwinding even from assembly subroutines.
160 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
161 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
164 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
165 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
166 which is the minimum version we support.
169 *) Certificate time validation (X509_cmp_time) enforces stricter
170 compliance with RFC 5280. Fractional seconds and timezone offsets
171 are no longer allowed.
174 *) Add support for ARIA
177 *) s_client will now send the Server Name Indication (SNI) extension by
178 default unless the new "-noservername" option is used. The server name is
179 based on the host provided to the "-connect" option unless overridden by
183 *) Add support for SipHash
186 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
187 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
188 prevent issues where no progress is being made and the peer continually
189 sends unrecognised record types, using up resources processing them.
192 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
193 using the algorithm defined in
194 https://www.akkadia.org/drepper/SHA-crypt.txt
197 *) Heartbeat support has been removed; the ABI is changed for now.
198 [Richard Levitte, Rich Salz]
200 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
203 *) The RSA "null" method, which was partially supported to avoid patent
204 issues, has been replaced to always returns NULL.
207 Changes between 1.1.0g and 1.1.0h [xx XXX xxxx]
209 *) rsaz_1024_mul_avx2 overflow bug on x86_64
211 There is an overflow bug in the AVX2 Montgomery multiplication procedure
212 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
213 Analysis suggests that attacks against RSA and DSA as a result of this
214 defect would be very difficult to perform and are not believed likely.
215 Attacks against DH1024 are considered just feasible, because most of the
216 work necessary to deduce information about a private key may be performed
217 offline. The amount of resources required for such an attack would be
218 significant. However, for an attack on TLS to be meaningful, the server
219 would have to share the DH1024 private key among multiple clients, which is
220 no longer an option since CVE-2016-0701.
222 This only affects processors that support the AVX2 but not ADX extensions
223 like Intel Haswell (4th generation).
225 This issue was reported to OpenSSL by David Benjamin (Google). The issue
226 was originally found via the OSS-Fuzz project.
230 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
232 *) bn_sqrx8x_internal carry bug on x86_64
234 There is a carry propagating bug in the x86_64 Montgomery squaring
235 procedure. No EC algorithms are affected. Analysis suggests that attacks
236 against RSA and DSA as a result of this defect would be very difficult to
237 perform and are not believed likely. Attacks against DH are considered just
238 feasible (although very difficult) because most of the work necessary to
239 deduce information about a private key may be performed offline. The amount
240 of resources required for such an attack would be very significant and
241 likely only accessible to a limited number of attackers. An attacker would
242 additionally need online access to an unpatched system using the target
243 private key in a scenario with persistent DH parameters and a private
244 key that is shared between multiple clients.
246 This only affects processors that support the BMI1, BMI2 and ADX extensions
247 like Intel Broadwell (5th generation) and later or AMD Ryzen.
249 This issue was reported to OpenSSL by the OSS-Fuzz project.
253 *) Malformed X.509 IPAddressFamily could cause OOB read
255 If an X.509 certificate has a malformed IPAddressFamily extension,
256 OpenSSL could do a one-byte buffer overread. The most likely result
257 would be an erroneous display of the certificate in text format.
259 This issue was reported to OpenSSL by the OSS-Fuzz project.
263 Changes between 1.1.0e and 1.1.0f [25 May 2017]
265 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
266 platform rather than 'mingw'.
269 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
270 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
271 which is the minimum version we support.
274 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
276 *) Encrypt-Then-Mac renegotiation crash
278 During a renegotiation handshake if the Encrypt-Then-Mac extension is
279 negotiated where it was not in the original handshake (or vice-versa) then
280 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
281 and servers are affected.
283 This issue was reported to OpenSSL by Joe Orton (Red Hat).
287 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
289 *) Truncated packet could crash via OOB read
291 If one side of an SSL/TLS path is running on a 32-bit host and a specific
292 cipher is being used, then a truncated packet can cause that host to
293 perform an out-of-bounds read, usually resulting in a crash.
295 This issue was reported to OpenSSL by Robert Święcki of Google.
299 *) Bad (EC)DHE parameters cause a client crash
301 If a malicious server supplies bad parameters for a DHE or ECDHE key
302 exchange then this can result in the client attempting to dereference a
303 NULL pointer leading to a client crash. This could be exploited in a Denial
306 This issue was reported to OpenSSL by Guido Vranken.
310 *) BN_mod_exp may produce incorrect results on x86_64
312 There is a carry propagating bug in the x86_64 Montgomery squaring
313 procedure. No EC algorithms are affected. Analysis suggests that attacks
314 against RSA and DSA as a result of this defect would be very difficult to
315 perform and are not believed likely. Attacks against DH are considered just
316 feasible (although very difficult) because most of the work necessary to
317 deduce information about a private key may be performed offline. The amount
318 of resources required for such an attack would be very significant and
319 likely only accessible to a limited number of attackers. An attacker would
320 additionally need online access to an unpatched system using the target
321 private key in a scenario with persistent DH parameters and a private
322 key that is shared between multiple clients. For example this can occur by
323 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
324 similar to CVE-2015-3193 but must be treated as a separate problem.
326 This issue was reported to OpenSSL by the OSS-Fuzz project.
330 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
332 *) ChaCha20/Poly1305 heap-buffer-overflow
334 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
335 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
336 crash. This issue is not considered to be exploitable beyond a DoS.
338 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
342 *) CMS Null dereference
344 Applications parsing invalid CMS structures can crash with a NULL pointer
345 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
346 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
347 structure callback if an attempt is made to free certain invalid encodings.
348 Only CHOICE structures using a callback which do not handle NULL value are
351 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
355 *) Montgomery multiplication may produce incorrect results
357 There is a carry propagating bug in the Broadwell-specific Montgomery
358 multiplication procedure that handles input lengths divisible by, but
359 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
360 and DH private keys are impossible. This is because the subroutine in
361 question is not used in operations with the private key itself and an input
362 of the attacker's direct choice. Otherwise the bug can manifest itself as
363 transient authentication and key negotiation failures or reproducible
364 erroneous outcome of public-key operations with specially crafted input.
365 Among EC algorithms only Brainpool P-512 curves are affected and one
366 presumably can attack ECDH key negotiation. Impact was not analyzed in
367 detail, because pre-requisites for attack are considered unlikely. Namely
368 multiple clients have to choose the curve in question and the server has to
369 share the private key among them, neither of which is default behaviour.
370 Even then only clients that chose the curve will be affected.
372 This issue was publicly reported as transient failures and was not
373 initially recognized as a security issue. Thanks to Richard Morgan for
374 providing reproducible case.
378 *) Removed automatic addition of RPATH in shared libraries and executables,
379 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
382 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
384 *) Fix Use After Free for large message sizes
386 The patch applied to address CVE-2016-6307 resulted in an issue where if a
387 message larger than approx 16k is received then the underlying buffer to
388 store the incoming message is reallocated and moved. Unfortunately a
389 dangling pointer to the old location is left which results in an attempt to
390 write to the previously freed location. This is likely to result in a
391 crash, however it could potentially lead to execution of arbitrary code.
393 This issue only affects OpenSSL 1.1.0a.
395 This issue was reported to OpenSSL by Robert Święcki.
399 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
401 *) OCSP Status Request extension unbounded memory growth
403 A malicious client can send an excessively large OCSP Status Request
404 extension. If that client continually requests renegotiation, sending a
405 large OCSP Status Request extension each time, then there will be unbounded
406 memory growth on the server. This will eventually lead to a Denial Of
407 Service attack through memory exhaustion. Servers with a default
408 configuration are vulnerable even if they do not support OCSP. Builds using
409 the "no-ocsp" build time option are not affected.
411 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
415 *) SSL_peek() hang on empty record
417 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
418 sends an empty record. This could be exploited by a malicious peer in a
419 Denial Of Service attack.
421 This issue was reported to OpenSSL by Alex Gaynor.
425 *) Excessive allocation of memory in tls_get_message_header() and
426 dtls1_preprocess_fragment()
428 A (D)TLS message includes 3 bytes for its length in the header for the
429 message. This would allow for messages up to 16Mb in length. Messages of
430 this length are excessive and OpenSSL includes a check to ensure that a
431 peer is sending reasonably sized messages in order to avoid too much memory
432 being consumed to service a connection. A flaw in the logic of version
433 1.1.0 means that memory for the message is allocated too early, prior to
434 the excessive message length check. Due to way memory is allocated in
435 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
436 to service a connection. This could lead to a Denial of Service through
437 memory exhaustion. However, the excessive message length check still takes
438 place, and this would cause the connection to immediately fail. Assuming
439 that the application calls SSL_free() on the failed connection in a timely
440 manner then the 21Mb of allocated memory will then be immediately freed
441 again. Therefore the excessive memory allocation will be transitory in
442 nature. This then means that there is only a security impact if:
444 1) The application does not call SSL_free() in a timely manner in the event
445 that the connection fails
447 2) The application is working in a constrained environment where there is
448 very little free memory
450 3) The attacker initiates multiple connection attempts such that there are
451 multiple connections in a state where memory has been allocated for the
452 connection; SSL_free() has not yet been called; and there is insufficient
453 memory to service the multiple requests.
455 Except in the instance of (1) above any Denial Of Service is likely to be
456 transitory because as soon as the connection fails the memory is
457 subsequently freed again in the SSL_free() call. However there is an
458 increased risk during this period of application crashes due to the lack of
459 memory - which would then mean a more serious Denial of Service.
461 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
462 (CVE-2016-6307 and CVE-2016-6308)
465 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
466 had to be removed. Primary reason is that vendor assembler can't
467 assemble our modules with -KPIC flag. As result it, assembly
468 support, was not even available as option. But its lack means
469 lack of side-channel resistant code, which is incompatible with
470 security by todays standards. Fortunately gcc is readily available
471 prepackaged option, which we firmly point at...
474 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
476 *) Windows command-line tool supports UTF-8 opt-in option for arguments
477 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
478 (to any value) allows Windows user to access PKCS#12 file generated
479 with Windows CryptoAPI and protected with non-ASCII password, as well
480 as files generated under UTF-8 locale on Linux also protected with
484 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
485 have been disabled by default and removed from DEFAULT, just like RC4.
486 See the RC4 item below to re-enable both.
489 *) The method for finding the storage location for the Windows RAND seed file
490 has changed. First we check %RANDFILE%. If that is not set then we check
491 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
492 all else fails we fall back to C:\.
495 *) The EVP_EncryptUpdate() function has had its return type changed from void
496 to int. A return of 0 indicates and error while a return of 1 indicates
500 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
501 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
502 off the constant time implementation for RSA, DSA and DH have been made
503 no-ops and deprecated.
506 *) Windows RAND implementation was simplified to only get entropy by
507 calling CryptGenRandom(). Various other RAND-related tickets
509 [Joseph Wylie Yandle, Rich Salz]
511 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
512 and OPENSSL_LH_, respectively. The old names are available
513 with API compatibility. They new names are now completely documented.
516 *) Unify TYPE_up_ref(obj) methods signature.
517 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
518 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
519 int (instead of void) like all others TYPE_up_ref() methods.
520 So now these methods also check the return value of CRYPTO_atomic_add(),
521 and the validity of object reference counter.
522 [fdasilvayy@gmail.com]
524 *) With Windows Visual Studio builds, the .pdb files are installed
525 alongside the installed libraries and executables. For a static
526 library installation, ossl_static.pdb is the associate compiler
527 generated .pdb file to be used when linking programs.
530 *) Remove openssl.spec. Packaging files belong with the packagers.
533 *) Automatic Darwin/OSX configuration has had a refresh, it will now
534 recognise x86_64 architectures automatically. You can still decide
535 to build for a different bitness with the environment variable
536 KERNEL_BITS (can be 32 or 64), for example:
538 KERNEL_BITS=32 ./config
542 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
543 256 bit AES and HMAC with SHA256.
546 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
549 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
552 *) To enable users to have their own config files and build file templates,
553 Configure looks in the directory indicated by the environment variable
554 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
555 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
556 name and is used as is.
559 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
560 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
561 X509_CERT_FILE_CTX was removed.
564 *) "shared" builds are now the default. To create only static libraries use
565 the "no-shared" Configure option.
568 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
569 All of these option have not worked for some while and are fundamental
573 *) Make various cleanup routines no-ops and mark them as deprecated. Most
574 global cleanup functions are no longer required because they are handled
575 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
576 Explicitly de-initing can cause problems (e.g. where a library that uses
577 OpenSSL de-inits, but an application is still using it). The affected
578 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
579 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
580 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
584 *) --strict-warnings no longer enables runtime debugging options
585 such as REF_DEBUG. Instead, debug options are automatically
586 enabled with '--debug' builds.
587 [Andy Polyakov, Emilia Käsper]
589 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
590 have been moved out of the public header files. New functions for managing
591 these have been added.
594 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
595 objects have been moved out of the public header files. New
596 functions for managing these have been added.
599 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
600 have been moved out of the public header files. New functions for managing
601 these have been added.
604 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
605 moved out of the public header files. New functions for managing these
609 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
612 *) Removed the mk1mf build scripts.
615 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
616 it is always safe to #include a header now.
619 *) Removed the aged BC-32 config and all its supporting scripts
622 *) Removed support for Ultrix, Netware, and OS/2.
625 *) Add support for HKDF.
628 *) Add support for blake2b and blake2s
631 *) Added support for "pipelining". Ciphers that have the
632 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
633 encryptions/decryptions simultaneously. There are currently no built-in
634 ciphers with this property but the expectation is that engines will be able
635 to offer it to significantly improve throughput. Support has been extended
636 into libssl so that multiple records for a single connection can be
637 processed in one go (for >=TLS 1.1).
640 *) Added the AFALG engine. This is an async capable engine which is able to
641 offload work to the Linux kernel. In this initial version it only supports
642 AES128-CBC. The kernel must be version 4.1.0 or greater.
645 *) OpenSSL now uses a new threading API. It is no longer necessary to
646 set locking callbacks to use OpenSSL in a multi-threaded environment. There
647 are two supported threading models: pthreads and windows threads. It is
648 also possible to configure OpenSSL at compile time for "no-threads". The
649 old threading API should no longer be used. The functions have been
650 replaced with "no-op" compatibility macros.
651 [Alessandro Ghedini, Matt Caswell]
653 *) Modify behavior of ALPN to invoke callback after SNI/servername
654 callback, such that updates to the SSL_CTX affect ALPN.
657 *) Add SSL_CIPHER queries for authentication and key-exchange.
660 *) Changes to the DEFAULT cipherlist:
661 - Prefer (EC)DHE handshakes over plain RSA.
662 - Prefer AEAD ciphers over legacy ciphers.
663 - Prefer ECDSA over RSA when both certificates are available.
664 - Prefer TLSv1.2 ciphers/PRF.
665 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
669 *) Change the ECC default curve list to be this, in order: x25519,
670 secp256r1, secp521r1, secp384r1.
673 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
674 disabled by default. They can be re-enabled using the
675 enable-weak-ssl-ciphers option to Configure.
678 *) If the server has ALPN configured, but supports no protocols that the
679 client advertises, send a fatal "no_application_protocol" alert.
680 This behaviour is SHALL in RFC 7301, though it isn't universally
681 implemented by other servers.
684 *) Add X25519 support.
685 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
686 for public and private key encoding using the format documented in
687 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
688 key generation and key derivation.
690 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
694 *) Deprecate SRP_VBASE_get_by_user.
695 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
696 In order to fix an unavoidable memory leak (CVE-2016-0798),
697 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
698 seed, even if the seed is configured.
700 Users should use SRP_VBASE_get1_by_user instead. Note that in
701 SRP_VBASE_get1_by_user, caller must free the returned value. Note
702 also that even though configuring the SRP seed attempts to hide
703 invalid usernames by continuing the handshake with fake
704 credentials, this behaviour is not constant time and no strong
705 guarantees are made that the handshake is indistinguishable from
706 that of a valid user.
709 *) Configuration change; it's now possible to build dynamic engines
710 without having to build shared libraries and vice versa. This
711 only applies to the engines in engines/, those in crypto/engine/
712 will always be built into libcrypto (i.e. "static").
714 Building dynamic engines is enabled by default; to disable, use
715 the configuration option "disable-dynamic-engine".
717 The only requirements for building dynamic engines are the
718 presence of the DSO module and building with position independent
719 code, so they will also automatically be disabled if configuring
720 with "disable-dso" or "disable-pic".
722 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
723 are also taken away from openssl/opensslconf.h, as they are
727 *) Configuration change; if there is a known flag to compile
728 position independent code, it will always be applied on the
729 libcrypto and libssl object files, and never on the application
730 object files. This means other libraries that use routines from
731 libcrypto / libssl can be made into shared libraries regardless
732 of how OpenSSL was configured.
734 If this isn't desirable, the configuration options "disable-pic"
735 or "no-pic" can be used to disable the use of PIC. This will
736 also disable building shared libraries and dynamic engines.
739 *) Removed JPAKE code. It was experimental and has no wide use.
742 *) The INSTALL_PREFIX Makefile variable has been renamed to
743 DESTDIR. That makes for less confusion on what this variable
744 is for. Also, the configuration option --install_prefix is
748 *) Heartbeat for TLS has been removed and is disabled by default
749 for DTLS; configure with enable-heartbeats. Code that uses the
750 old #define's might need to be updated.
751 [Emilia Käsper, Rich Salz]
753 *) Rename REF_CHECK to REF_DEBUG.
756 *) New "unified" build system
758 The "unified" build system is aimed to be a common system for all
759 platforms we support. With it comes new support for VMS.
761 This system builds supports building in a different directory tree
762 than the source tree. It produces one Makefile (for unix family
763 or lookalikes), or one descrip.mms (for VMS).
765 The source of information to make the Makefile / descrip.mms is
766 small files called 'build.info', holding the necessary
767 information for each directory with source to compile, and a
768 template in Configurations, like unix-Makefile.tmpl or
771 With this change, the library names were also renamed on Windows
772 and on VMS. They now have names that are closer to the standard
773 on Unix, and include the major version number, and in certain
774 cases, the architecture they are built for. See "Notes on shared
775 libraries" in INSTALL.
777 We rely heavily on the perl module Text::Template.
780 *) Added support for auto-initialisation and de-initialisation of the library.
781 OpenSSL no longer requires explicit init or deinit routines to be called,
782 except in certain circumstances. See the OPENSSL_init_crypto() and
783 OPENSSL_init_ssl() man pages for further information.
786 *) The arguments to the DTLSv1_listen function have changed. Specifically the
787 "peer" argument is now expected to be a BIO_ADDR object.
789 *) Rewrite of BIO networking library. The BIO library lacked consistent
790 support of IPv6, and adding it required some more extensive
791 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
792 which hold all types of addresses and chains of address information.
793 It also introduces a new API, with functions like BIO_socket,
794 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
795 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
796 have been adapted accordingly.
799 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
803 *) CRIME protection: disable compression by default, even if OpenSSL is
804 compiled with zlib enabled. Applications can still enable compression
805 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
806 using the SSL_CONF library to configure compression.
809 *) The signature of the session callback configured with
810 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
811 was explicitly marked as 'const unsigned char*' instead of
815 *) Always DPURIFY. Remove the use of uninitialized memory in the
816 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
819 *) Removed many obsolete configuration items, including
820 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
821 MD2_CHAR, MD2_INT, MD2_LONG
823 IDEA_SHORT, IDEA_LONG
824 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
825 [Rich Salz, with advice from Andy Polyakov]
827 *) Many BN internals have been moved to an internal header file.
828 [Rich Salz with help from Andy Polyakov]
830 *) Configuration and writing out the results from it has changed.
831 Files such as Makefile include/openssl/opensslconf.h and are now
832 produced through general templates, such as Makefile.in and
833 crypto/opensslconf.h.in and some help from the perl module
836 Also, the center of configuration information is no longer
837 Makefile. Instead, Configure produces a perl module in
838 configdata.pm which holds most of the config data (in the hash
839 table %config), the target data that comes from the target
840 configuration in one of the Configurations/*.conf files (in
844 *) To clarify their intended purposes, the Configure options
845 --prefix and --openssldir change their semantics, and become more
846 straightforward and less interdependent.
848 --prefix shall be used exclusively to give the location INSTALLTOP
849 where programs, scripts, libraries, include files and manuals are
850 going to be installed. The default is now /usr/local.
852 --openssldir shall be used exclusively to give the default
853 location OPENSSLDIR where certificates, private keys, CRLs are
854 managed. This is also where the default openssl.cnf gets
856 If the directory given with this option is a relative path, the
857 values of both the --prefix value and the --openssldir value will
858 be combined to become OPENSSLDIR.
859 The default for --openssldir is INSTALLTOP/ssl.
861 Anyone who uses --openssldir to specify where OpenSSL is to be
862 installed MUST change to use --prefix instead.
865 *) The GOST engine was out of date and therefore it has been removed. An up
866 to date GOST engine is now being maintained in an external repository.
867 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
868 support for GOST ciphersuites (these are only activated if a GOST engine
872 *) EGD is no longer supported by default; use enable-egd when
874 [Ben Kaduk and Rich Salz]
876 *) The distribution now has Makefile.in files, which are used to
877 create Makefile's when Configure is run. *Configure must be run
878 before trying to build now.*
881 *) The return value for SSL_CIPHER_description() for error conditions
885 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
887 Obtaining and performing DNSSEC validation of TLSA records is
888 the application's responsibility. The application provides
889 the TLSA records of its choice to OpenSSL, and these are then
890 used to authenticate the peer.
892 The TLSA records need not even come from DNS. They can, for
893 example, be used to implement local end-entity certificate or
894 trust-anchor "pinning", where the "pin" data takes the form
895 of TLSA records, which can augment or replace verification
896 based on the usual WebPKI public certification authorities.
899 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
900 continues to support deprecated interfaces in default builds.
901 However, applications are strongly advised to compile their
902 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
903 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
904 or the 1.1.0 releases.
906 In environments in which all applications have been ported to
907 not use any deprecated interfaces OpenSSL's Configure script
908 should be used with the --api=1.1.0 option to entirely remove
909 support for the deprecated features from the library and
910 unconditionally disable them in the installed headers.
911 Essentially the same effect can be achieved with the "no-deprecated"
912 argument to Configure, except that this will always restrict
913 the build to just the latest API, rather than a fixed API
916 As applications are ported to future revisions of the API,
917 they should update their compile-time OPENSSL_API_COMPAT define
918 accordingly, but in most cases should be able to continue to
919 compile with later releases.
921 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
922 0x10000000L and 0x00908000L, respectively. However those
923 versions did not support the OPENSSL_API_COMPAT feature, and
924 so applications are not typically tested for explicit support
925 of just the undeprecated features of either release.
928 *) Add support for setting the minimum and maximum supported protocol.
929 It can bet set via the SSL_set_min_proto_version() and
930 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
931 MaxProtocol. It's recommended to use the new APIs to disable
932 protocols instead of disabling individual protocols using
933 SSL_set_options() or SSL_CONF's Protocol. This change also
934 removes support for disabling TLS 1.2 in the OpenSSL TLS
935 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
938 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
941 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
942 and integrates ECDSA and ECDH functionality into EC. Implementations can
943 now redirect key generation and no longer need to convert to or from
946 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
947 include the ec.h header file instead.
950 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
951 ciphers who are no longer supported and drops support the ephemeral RSA key
952 exchange. The LOW ciphers currently doesn't have any ciphers in it.
955 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
956 opaque. For HMAC_CTX, the following constructors and destructors
959 HMAC_CTX *HMAC_CTX_new(void);
960 void HMAC_CTX_free(HMAC_CTX *ctx);
962 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
963 destroy such methods has been added. See EVP_MD_meth_new(3) and
964 EVP_CIPHER_meth_new(3) for documentation.
967 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
968 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
969 EVP_MD_CTX_reset() should be called instead to reinitialise
970 an already created structure.
971 2) For consistency with the majority of our object creators and
972 destructors, EVP_MD_CTX_(create|destroy) were renamed to
973 EVP_MD_CTX_(new|free). The old names are retained as macros
974 for deprecated builds.
977 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
978 cryptographic operations to be performed asynchronously as long as an
979 asynchronous capable engine is used. See the ASYNC_start_job() man page for
980 further details. Libssl has also had this capability integrated with the
981 introduction of the new mode SSL_MODE_ASYNC and associated error
982 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
983 pages. This work was developed in partnership with Intel Corp.
986 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
987 always enabled now. If you want to disable the support you should
988 exclude it using the list of supported ciphers. This also means that the
989 "-no_ecdhe" option has been removed from s_server.
992 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
993 SSL_{CTX_}set1_curves() which can set a list.
996 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
997 curve you want to support using SSL_{CTX_}set1_curves().
1000 *) State machine rewrite. The state machine code has been significantly
1001 refactored in order to remove much duplication of code and solve issues
1002 with the old code (see ssl/statem/README for further details). This change
1003 does have some associated API changes. Notably the SSL_state() function
1004 has been removed and replaced by SSL_get_state which now returns an
1005 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1006 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1010 *) All instances of the string "ssleay" in the public API were replaced
1011 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1012 Some error codes related to internal RSA_eay API's were renamed.
1015 *) The demo files in crypto/threads were moved to demo/threads.
1018 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1020 [Matt Caswell, Rich Salz]
1022 *) New ASN.1 embed macro.
1024 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1025 structure is not allocated: it is part of the parent. That is instead of
1033 This reduces memory fragmentation and make it impossible to accidentally
1034 set a mandatory field to NULL.
1036 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1037 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1038 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1042 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1045 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1046 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1047 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1048 DES and RC4 ciphersuites.
1051 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1052 This changes the decoding behaviour for some invalid messages,
1053 though the change is mostly in the more lenient direction, and
1054 legacy behaviour is preserved as much as possible.
1057 *) Fix no-stdio build.
1058 [ David Woodhouse <David.Woodhouse@intel.com> and also
1059 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1061 *) New testing framework
1062 The testing framework has been largely rewritten and is now using
1063 perl and the perl modules Test::Harness and an extended variant of
1064 Test::More called OpenSSL::Test to do its work. All test scripts in
1065 test/ have been rewritten into test recipes, and all direct calls to
1066 executables in test/Makefile have become individual recipes using the
1067 simplified testing OpenSSL::Test::Simple.
1069 For documentation on our testing modules, do:
1071 perldoc test/testlib/OpenSSL/Test/Simple.pm
1072 perldoc test/testlib/OpenSSL/Test.pm
1076 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1077 are used; the latter aborts on memory leaks (usually checked on exit).
1078 Some undocumented "set malloc, etc., hooks" functions were removed
1079 and others were changed. All are now documented.
1082 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1084 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1086 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1087 from RFC4279, RFC4785, RFC5487, RFC5489.
1089 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1090 original RSA_PSK patch.
1093 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1094 era flag was never set throughout the codebase (only read). Also removed
1095 SSL3_FLAGS_POP_BUFFER which was only used if
1096 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1099 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1100 to be "oneline" instead of "compat".
1103 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1104 not aware of clients that still exhibit this bug, and the workaround
1105 hasn't been working properly for a while.
1108 *) The return type of BIO_number_read() and BIO_number_written() as well as
1109 the corresponding num_read and num_write members in the BIO structure has
1110 changed from unsigned long to uint64_t. On platforms where an unsigned
1111 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1115 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1116 OpenSSL without support for them. It also means that maintaining
1117 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1118 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1121 *) Removed support for the two export grade static DH ciphersuites
1122 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1123 were newly added (along with a number of other static DH ciphersuites) to
1124 1.0.2. However the two export ones have *never* worked since they were
1125 introduced. It seems strange in any case to be adding new export
1126 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1129 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1130 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1131 and turned into macros which simply call the new preferred function names
1132 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1133 should use the new names instead. Also as part of this change the ssl23.h
1134 header file has been removed.
1137 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1138 code and the associated standard is no longer considered fit-for-purpose.
1141 *) RT2547 was closed. When generating a private key, try to make the
1142 output file readable only by the owner. This behavior change might
1143 be noticeable when interacting with other software.
1145 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1149 *) Added HTTP GET support to the ocsp command.
1152 *) Changed default digest for the dgst and enc commands from MD5 to
1156 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1159 *) Added support for TLS extended master secret from
1160 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1161 initial patch which was a great help during development.
1164 *) All libssl internal structures have been removed from the public header
1165 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1166 now redundant). Users should not attempt to access internal structures
1167 directly. Instead they should use the provided API functions.
1170 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1171 Access to deprecated functions can be re-enabled by running config with
1172 "enable-deprecated". In addition applications wishing to use deprecated
1173 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1174 will, by default, disable some transitive includes that previously existed
1175 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1178 *) Added support for OCB mode. OpenSSL has been granted a patent license
1179 compatible with the OpenSSL license for use of OCB. Details are available
1180 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1181 for OCB can be removed by calling config with no-ocb.
1184 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1185 compatible client hello.
1188 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1189 done while fixing the error code for the key-too-small case.
1190 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1192 *) CA.sh has been removed; use CA.pl instead.
1195 *) Removed old DES API.
1198 *) Remove various unsupported platforms:
1204 Sinix/ReliantUNIX RM400
1209 16-bit platforms such as WIN16
1212 *) Clean up OPENSSL_NO_xxx #define's
1213 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1214 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1215 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1216 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1217 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1218 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1219 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1220 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1221 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1222 Remove MS_STATIC; it's a relic from platforms <32 bits.
1225 *) Cleaned up dead code
1226 Remove all but one '#ifdef undef' which is to be looked at.
1229 *) Clean up calling of xxx_free routines.
1230 Just like free(), fix most of the xxx_free routines to accept
1231 NULL. Remove the non-null checks from callers. Save much code.
1234 *) Add secure heap for storage of private keys (when possible).
1235 Add BIO_s_secmem(), CBIGNUM, etc.
1236 Contributed by Akamai Technologies under our Corporate CLA.
1239 *) Experimental support for a new, fast, unbiased prime candidate generator,
1240 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1241 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1243 *) New output format NSS in the sess_id command line tool. This allows
1244 exporting the session id and the master key in NSS keylog format.
1245 [Martin Kaiser <martin@kaiser.cx>]
1247 *) Harmonize version and its documentation. -f flag is used to display
1249 [mancha <mancha1@zoho.com>]
1251 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1252 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1253 [mancha <mancha1@zoho.com>]
1255 *) Fix some double frees. These are not thought to be exploitable.
1256 [mancha <mancha1@zoho.com>]
1258 *) A missing bounds check in the handling of the TLS heartbeat extension
1259 can be used to reveal up to 64k of memory to a connected client or
1262 Thanks for Neel Mehta of Google Security for discovering this bug and to
1263 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1264 preparing the fix (CVE-2014-0160)
1265 [Adam Langley, Bodo Moeller]
1267 *) Fix for the attack described in the paper "Recovering OpenSSL
1268 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1269 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1270 http://eprint.iacr.org/2014/140
1272 Thanks to Yuval Yarom and Naomi Benger for discovering this
1273 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1274 [Yuval Yarom and Naomi Benger]
1276 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1277 this fixes a limitation in previous versions of OpenSSL.
1280 *) Experimental encrypt-then-mac support.
1282 Experimental support for encrypt then mac from
1283 draft-gutmann-tls-encrypt-then-mac-02.txt
1285 To enable it set the appropriate extension number (0x42 for the test
1286 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1288 For non-compliant peers (i.e. just about everything) this should have no
1291 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1295 *) Add EVP support for key wrapping algorithms, to avoid problems with
1296 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1297 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1298 algorithms and include tests cases.
1301 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1305 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1306 MGF1 digest and OAEP label.
1309 *) Make openssl verify return errors.
1310 [Chris Palmer <palmer@google.com> and Ben Laurie]
1312 *) New function ASN1_TIME_diff to calculate the difference between two
1313 ASN1_TIME structures or one structure and the current time.
1316 *) Update fips_test_suite to support multiple command line options. New
1317 test to induce all self test errors in sequence and check expected
1321 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1322 sign or verify all in one operation.
1325 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1326 test programs and fips_test_suite. Includes functionality to parse
1327 the minimal script output of fipsalgest.pl directly.
1330 *) Add authorisation parameter to FIPS_module_mode_set().
1333 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1336 *) Use separate DRBG fields for internal and external flags. New function
1337 FIPS_drbg_health_check() to perform on demand health checking. Add
1338 generation tests to fips_test_suite with reduced health check interval to
1339 demonstrate periodic health checking. Add "nodh" option to
1340 fips_test_suite to skip very slow DH test.
1343 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1347 *) More extensive health check for DRBG checking many more failure modes.
1348 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1349 combination: call this in fips_test_suite.
1352 *) Add support for canonical generation of DSA parameter 'g'. See
1355 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1356 POST to handle HMAC cases.
1359 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1360 to return numerical and string versions of the FIPS module number.
1363 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1364 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1365 outside the validated module in the FIPS capable OpenSSL.
1368 *) Minor change to DRBG entropy callback semantics. In some cases
1369 there is no multiple of the block length between min_len and
1370 max_len. Allow the callback to return more than max_len bytes
1371 of entropy but discard any extra: it is the callback's responsibility
1372 to ensure that the extra data discarded does not impact the
1373 requested amount of entropy.
1376 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1377 information in FIPS186-3, SP800-57 and SP800-131A.
1380 *) CCM support via EVP. Interface is very similar to GCM case except we
1381 must supply all data in one chunk (i.e. no update, final) and the
1382 message length must be supplied if AAD is used. Add algorithm test
1386 *) Initial version of POST overhaul. Add POST callback to allow the status
1387 of POST to be monitored and/or failures induced. Modify fips_test_suite
1388 to use callback. Always run all selftests even if one fails.
1391 *) XTS support including algorithm test driver in the fips_gcmtest program.
1392 Note: this does increase the maximum key length from 32 to 64 bytes but
1393 there should be no binary compatibility issues as existing applications
1394 will never use XTS mode.
1397 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1398 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1399 performs algorithm blocking for unapproved PRNG types. Also do not
1400 set PRNG type in FIPS_mode_set(): leave this to the application.
1401 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1402 the standard OpenSSL PRNG: set additional data to a date time vector.
1405 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1406 This shouldn't present any incompatibility problems because applications
1407 shouldn't be using these directly and any that are will need to rethink
1408 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1411 *) Extensive self tests and health checking required by SP800-90 DRBG.
1412 Remove strength parameter from FIPS_drbg_instantiate and always
1413 instantiate at maximum supported strength.
1416 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1419 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1422 *) New function DH_compute_key_padded() to compute a DH key and pad with
1423 leading zeroes if needed: this complies with SP800-56A et al.
1426 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1427 anything, incomplete, subject to change and largely untested at present.
1430 *) Modify fipscanisteronly build option to only build the necessary object
1431 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1434 *) Add experimental option FIPSSYMS to give all symbols in
1435 fipscanister.o and FIPS or fips prefix. This will avoid
1436 conflicts with future versions of OpenSSL. Add perl script
1437 util/fipsas.pl to preprocess assembly language source files
1438 and rename any affected symbols.
1441 *) Add selftest checks and algorithm block of non-fips algorithms in
1442 FIPS mode. Remove DES2 from selftests.
1445 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1446 return internal method without any ENGINE dependencies. Add new
1447 tiny fips sign and verify functions.
1450 *) New build option no-ec2m to disable characteristic 2 code.
1453 *) New build option "fipscanisteronly". This only builds fipscanister.o
1454 and (currently) associated fips utilities. Uses the file Makefile.fips
1455 instead of Makefile.org as the prototype.
1458 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1459 Update fips_gcmtest to use IV generator.
1462 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1463 setting output buffer to NULL. The *Final function must be
1464 called although it will not retrieve any additional data. The tag
1465 can be set or retrieved with a ctrl. The IV length is by default 12
1466 bytes (96 bits) but can be set to an alternative value. If the IV
1467 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1471 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1472 underlying do_cipher function handles all cipher semantics itself
1473 including padding and finalisation. This is useful if (for example)
1474 an ENGINE cipher handles block padding itself. The behaviour of
1475 do_cipher is subtly changed if this flag is set: the return value
1476 is the number of characters written to the output buffer (zero is
1477 no longer an error code) or a negative error code. Also if the
1478 input buffer is NULL and length 0 finalisation should be performed.
1481 *) If a candidate issuer certificate is already part of the constructed
1482 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1485 *) Improve forward-security support: add functions
1487 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1488 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1490 for use by SSL/TLS servers; the callback function will be called whenever a
1491 new session is created, and gets to decide whether the session may be
1492 cached to make it resumable (return 0) or not (return 1). (As by the
1493 SSL/TLS protocol specifications, the session_id sent by the server will be
1494 empty to indicate that the session is not resumable; also, the server will
1495 not generate RFC 4507 (RFC 5077) session tickets.)
1497 A simple reasonable callback implementation is to return is_forward_secure.
1498 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1499 by the SSL/TLS server library, indicating whether it can provide forward
1501 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1503 *) New -verify_name option in command line utilities to set verification
1507 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1508 Add CMAC pkey methods.
1511 *) Experimental renegotiation in s_server -www mode. If the client
1512 browses /reneg connection is renegotiated. If /renegcert it is
1513 renegotiated requesting a certificate.
1516 *) Add an "external" session cache for debugging purposes to s_server. This
1517 should help trace issues which normally are only apparent in deployed
1518 multi-process servers.
1521 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1522 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1523 BIO_set_cipher() and some obscure PEM functions were changed so they
1524 can now return an error. The RAND changes required a change to the
1525 RAND_METHOD structure.
1528 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1529 a gcc attribute to warn if the result of a function is ignored. This
1530 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1531 whose return value is often ignored.
1534 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1535 These allow SCTs (signed certificate timestamps) to be requested and
1536 validated when establishing a connection.
1537 [Rob Percival <robpercival@google.com>]
1539 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1541 *) Prevent padding oracle in AES-NI CBC MAC check
1543 A MITM attacker can use a padding oracle attack to decrypt traffic
1544 when the connection uses an AES CBC cipher and the server support
1547 This issue was introduced as part of the fix for Lucky 13 padding
1548 attack (CVE-2013-0169). The padding check was rewritten to be in
1549 constant time by making sure that always the same bytes are read and
1550 compared against either the MAC or padding bytes. But it no longer
1551 checked that there was enough data to have both the MAC and padding
1554 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1558 *) Fix EVP_EncodeUpdate overflow
1560 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1561 Base64 encoding of binary data. If an attacker is able to supply very large
1562 amounts of input data then a length check can overflow resulting in a heap
1565 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1566 the PEM_write_bio* family of functions. These are mainly used within the
1567 OpenSSL command line applications, so any application which processes data
1568 from an untrusted source and outputs it as a PEM file should be considered
1569 vulnerable to this issue. User applications that call these APIs directly
1570 with large amounts of untrusted data may also be vulnerable.
1572 This issue was reported by Guido Vranken.
1576 *) Fix EVP_EncryptUpdate overflow
1578 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1579 is able to supply very large amounts of input data after a previous call to
1580 EVP_EncryptUpdate() with a partial block then a length check can overflow
1581 resulting in a heap corruption. Following an analysis of all OpenSSL
1582 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1583 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1584 the first called function after an EVP_EncryptInit(), and therefore that
1585 specific call must be safe. The second form is where the length passed to
1586 EVP_EncryptUpdate() can be seen from the code to be some small value and
1587 therefore there is no possibility of an overflow. Since all instances are
1588 one of these two forms, it is believed that there can be no overflows in
1589 internal code due to this problem. It should be noted that
1590 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1591 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1592 of these calls have also been analysed too and it is believed there are no
1593 instances in internal usage where an overflow could occur.
1595 This issue was reported by Guido Vranken.
1599 *) Prevent ASN.1 BIO excessive memory allocation
1601 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1602 a short invalid encoding can cause allocation of large amounts of memory
1603 potentially consuming excessive resources or exhausting memory.
1605 Any application parsing untrusted data through d2i BIO functions is
1606 affected. The memory based functions such as d2i_X509() are *not* affected.
1607 Since the memory based functions are used by the TLS library, TLS
1608 applications are not affected.
1610 This issue was reported by Brian Carpenter.
1616 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1617 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1618 in arbitrary stack data being returned in the buffer.
1620 This issue was reported by Guido Vranken.
1624 *) Modify behavior of ALPN to invoke callback after SNI/servername
1625 callback, such that updates to the SSL_CTX affect ALPN.
1628 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1632 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1633 methods are enabled and ssl2 is disabled the methods return NULL.
1636 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1638 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1639 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1640 provide any "EXPORT" or "LOW" strength ciphers.
1643 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1644 is by default disabled at build-time. Builds that are not configured with
1645 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1646 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1647 will need to explicitly call either of:
1649 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1651 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1653 as appropriate. Even if either of those is used, or the application
1654 explicitly uses the version-specific SSLv2_method() or its client and
1655 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1656 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1657 ciphers, and SSLv2 56-bit DES are no longer available.
1661 *) Fix a double-free in DSA code
1663 A double free bug was discovered when OpenSSL parses malformed DSA private
1664 keys and could lead to a DoS attack or memory corruption for applications
1665 that receive DSA private keys from untrusted sources. This scenario is
1668 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1673 *) Disable SRP fake user seed to address a server memory leak.
1675 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1677 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1678 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1679 was changed to ignore the "fake user" SRP seed, even if the seed
1682 Users should use SRP_VBASE_get1_by_user instead. Note that in
1683 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1684 also that even though configuring the SRP seed attempts to hide
1685 invalid usernames by continuing the handshake with fake
1686 credentials, this behaviour is not constant time and no strong
1687 guarantees are made that the handshake is indistinguishable from
1688 that of a valid user.
1692 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1694 In the BN_hex2bn function the number of hex digits is calculated using an
1695 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1696 large values of |i| this can result in |bn_expand| not allocating any
1697 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1698 field as NULL leading to a subsequent NULL ptr deref. For very large values
1699 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1700 In this case memory is allocated to the internal BIGNUM data field, but it
1701 is insufficiently sized leading to heap corruption. A similar issue exists
1702 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1703 is ever called by user applications with very large untrusted hex/dec data.
1704 This is anticipated to be a rare occurrence.
1706 All OpenSSL internal usage of these functions use data that is not expected
1707 to be untrusted, e.g. config file data or application command line
1708 arguments. If user developed applications generate config file data based
1709 on untrusted data then it is possible that this could also lead to security
1710 consequences. This is also anticipated to be rare.
1712 This issue was reported to OpenSSL by Guido Vranken.
1716 *) Fix memory issues in BIO_*printf functions
1718 The internal |fmtstr| function used in processing a "%s" format string in
1719 the BIO_*printf functions could overflow while calculating the length of a
1720 string and cause an OOB read when printing very long strings.
1722 Additionally the internal |doapr_outch| function can attempt to write to an
1723 OOB memory location (at an offset from the NULL pointer) in the event of a
1724 memory allocation failure. In 1.0.2 and below this could be caused where
1725 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1726 could be in processing a very long "%s" format string. Memory leaks can
1729 The first issue may mask the second issue dependent on compiler behaviour.
1730 These problems could enable attacks where large amounts of untrusted data
1731 is passed to the BIO_*printf functions. If applications use these functions
1732 in this way then they could be vulnerable. OpenSSL itself uses these
1733 functions when printing out human-readable dumps of ASN.1 data. Therefore
1734 applications that print this data could be vulnerable if the data is from
1735 untrusted sources. OpenSSL command line applications could also be
1736 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1737 as command line arguments.
1739 Libssl is not considered directly vulnerable. Additionally certificates etc
1740 received via remote connections via libssl are also unlikely to be able to
1741 trigger these issues because of message size limits enforced within libssl.
1743 This issue was reported to OpenSSL Guido Vranken.
1747 *) Side channel attack on modular exponentiation
1749 A side-channel attack was found which makes use of cache-bank conflicts on
1750 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1751 of RSA keys. The ability to exploit this issue is limited as it relies on
1752 an attacker who has control of code in a thread running on the same
1753 hyper-threaded core as the victim thread which is performing decryptions.
1755 This issue was reported to OpenSSL by Yuval Yarom, The University of
1756 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1757 Nadia Heninger, University of Pennsylvania with more information at
1758 http://cachebleed.info.
1762 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1763 if no keysize is specified with default_bits. This fixes an
1764 omission in an earlier change that changed all RSA/DSA key generation
1765 apps to use 2048 bits by default.
1768 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1769 *) DH small subgroups
1771 Historically OpenSSL only ever generated DH parameters based on "safe"
1772 primes. More recently (in version 1.0.2) support was provided for
1773 generating X9.42 style parameter files such as those required for RFC 5114
1774 support. The primes used in such files may not be "safe". Where an
1775 application is using DH configured with parameters based on primes that are
1776 not "safe" then an attacker could use this fact to find a peer's private
1777 DH exponent. This attack requires that the attacker complete multiple
1778 handshakes in which the peer uses the same private DH exponent. For example
1779 this could be used to discover a TLS server's private DH exponent if it's
1780 reusing the private DH exponent or it's using a static DH ciphersuite.
1782 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1783 TLS. It is not on by default. If the option is not set then the server
1784 reuses the same private DH exponent for the life of the server process and
1785 would be vulnerable to this attack. It is believed that many popular
1786 applications do set this option and would therefore not be at risk.
1788 The fix for this issue adds an additional check where a "q" parameter is
1789 available (as is the case in X9.42 based parameters). This detects the
1790 only known attack, and is the only possible defense for static DH
1791 ciphersuites. This could have some performance impact.
1793 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1794 default and cannot be disabled. This could have some performance impact.
1796 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1800 *) SSLv2 doesn't block disabled ciphers
1802 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1803 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1804 been disabled, provided that the SSLv2 protocol was not also disabled via
1807 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1808 and Sebastian Schinzel.
1812 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1814 *) BN_mod_exp may produce incorrect results on x86_64
1816 There is a carry propagating bug in the x86_64 Montgomery squaring
1817 procedure. No EC algorithms are affected. Analysis suggests that attacks
1818 against RSA and DSA as a result of this defect would be very difficult to
1819 perform and are not believed likely. Attacks against DH are considered just
1820 feasible (although very difficult) because most of the work necessary to
1821 deduce information about a private key may be performed offline. The amount
1822 of resources required for such an attack would be very significant and
1823 likely only accessible to a limited number of attackers. An attacker would
1824 additionally need online access to an unpatched system using the target
1825 private key in a scenario with persistent DH parameters and a private
1826 key that is shared between multiple clients. For example this can occur by
1827 default in OpenSSL DHE based SSL/TLS ciphersuites.
1829 This issue was reported to OpenSSL by Hanno Böck.
1833 *) Certificate verify crash with missing PSS parameter
1835 The signature verification routines will crash with a NULL pointer
1836 dereference if presented with an ASN.1 signature using the RSA PSS
1837 algorithm and absent mask generation function parameter. Since these
1838 routines are used to verify certificate signature algorithms this can be
1839 used to crash any certificate verification operation and exploited in a
1840 DoS attack. Any application which performs certificate verification is
1841 vulnerable including OpenSSL clients and servers which enable client
1844 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1848 *) X509_ATTRIBUTE memory leak
1850 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1851 memory. This structure is used by the PKCS#7 and CMS routines so any
1852 application which reads PKCS#7 or CMS data from untrusted sources is
1853 affected. SSL/TLS is not affected.
1855 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1860 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1861 This changes the decoding behaviour for some invalid messages,
1862 though the change is mostly in the more lenient direction, and
1863 legacy behaviour is preserved as much as possible.
1866 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1868 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1870 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1872 *) Alternate chains certificate forgery
1874 During certificate verification, OpenSSL will attempt to find an
1875 alternative certificate chain if the first attempt to build such a chain
1876 fails. An error in the implementation of this logic can mean that an
1877 attacker could cause certain checks on untrusted certificates to be
1878 bypassed, such as the CA flag, enabling them to use a valid leaf
1879 certificate to act as a CA and "issue" an invalid certificate.
1881 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1885 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1887 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1888 incompatibility in the handling of HMAC. The previous ABI has now been
1892 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1894 *) Malformed ECParameters causes infinite loop
1896 When processing an ECParameters structure OpenSSL enters an infinite loop
1897 if the curve specified is over a specially malformed binary polynomial
1900 This can be used to perform denial of service against any
1901 system which processes public keys, certificate requests or
1902 certificates. This includes TLS clients and TLS servers with
1903 client authentication enabled.
1905 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1909 *) Exploitable out-of-bounds read in X509_cmp_time
1911 X509_cmp_time does not properly check the length of the ASN1_TIME
1912 string and can read a few bytes out of bounds. In addition,
1913 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1916 An attacker can use this to craft malformed certificates and CRLs of
1917 various sizes and potentially cause a segmentation fault, resulting in
1918 a DoS on applications that verify certificates or CRLs. TLS clients
1919 that verify CRLs are affected. TLS clients and servers with client
1920 authentication enabled may be affected if they use custom verification
1923 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1924 independently by Hanno Böck.
1928 *) PKCS7 crash with missing EnvelopedContent
1930 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1931 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1932 with missing content and trigger a NULL pointer dereference on parsing.
1934 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1935 structures from untrusted sources are affected. OpenSSL clients and
1936 servers are not affected.
1938 This issue was reported to OpenSSL by Michal Zalewski (Google).
1942 *) CMS verify infinite loop with unknown hash function
1944 When verifying a signedData message the CMS code can enter an infinite loop
1945 if presented with an unknown hash function OID. This can be used to perform
1946 denial of service against any system which verifies signedData messages using
1948 This issue was reported to OpenSSL by Johannes Bauer.
1952 *) Race condition handling NewSessionTicket
1954 If a NewSessionTicket is received by a multi-threaded client when attempting to
1955 reuse a previous ticket then a race condition can occur potentially leading to
1956 a double free of the ticket data.
1960 *) Only support 256-bit or stronger elliptic curves with the
1961 'ecdh_auto' setting (server) or by default (client). Of supported
1962 curves, prefer P-256 (both).
1965 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1967 *) ClientHello sigalgs DoS fix
1969 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1970 invalid signature algorithms extension a NULL pointer dereference will
1971 occur. This can be exploited in a DoS attack against the server.
1973 This issue was was reported to OpenSSL by David Ramos of Stanford
1976 [Stephen Henson and Matt Caswell]
1978 *) Multiblock corrupted pointer fix
1980 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1981 feature only applies on 64 bit x86 architecture platforms that support AES
1982 NI instructions. A defect in the implementation of "multiblock" can cause
1983 OpenSSL's internal write buffer to become incorrectly set to NULL when
1984 using non-blocking IO. Typically, when the user application is using a
1985 socket BIO for writing, this will only result in a failed connection.
1986 However if some other BIO is used then it is likely that a segmentation
1987 fault will be triggered, thus enabling a potential DoS attack.
1989 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1993 *) Segmentation fault in DTLSv1_listen fix
1995 The DTLSv1_listen function is intended to be stateless and processes the
1996 initial ClientHello from many peers. It is common for user code to loop
1997 over the call to DTLSv1_listen until a valid ClientHello is received with
1998 an associated cookie. A defect in the implementation of DTLSv1_listen means
1999 that state is preserved in the SSL object from one invocation to the next
2000 that can lead to a segmentation fault. Errors processing the initial
2001 ClientHello can trigger this scenario. An example of such an error could be
2002 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2005 This issue was reported to OpenSSL by Per Allansson.
2009 *) Segmentation fault in ASN1_TYPE_cmp fix
2011 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2012 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2013 certificate signature algorithm consistency this can be used to crash any
2014 certificate verification operation and exploited in a DoS attack. Any
2015 application which performs certificate verification is vulnerable including
2016 OpenSSL clients and servers which enable client authentication.
2020 *) Segmentation fault for invalid PSS parameters fix
2022 The signature verification routines will crash with a NULL pointer
2023 dereference if presented with an ASN.1 signature using the RSA PSS
2024 algorithm and invalid parameters. Since these routines are used to verify
2025 certificate signature algorithms this can be used to crash any
2026 certificate verification operation and exploited in a DoS attack. Any
2027 application which performs certificate verification is vulnerable including
2028 OpenSSL clients and servers which enable client authentication.
2030 This issue was was reported to OpenSSL by Brian Carpenter.
2034 *) ASN.1 structure reuse memory corruption fix
2036 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2037 memory corruption via an invalid write. Such reuse is and has been
2038 strongly discouraged and is believed to be rare.
2040 Applications that parse structures containing CHOICE or ANY DEFINED BY
2041 components may be affected. Certificate parsing (d2i_X509 and related
2042 functions) are however not affected. OpenSSL clients and servers are
2047 *) PKCS7 NULL pointer dereferences fix
2049 The PKCS#7 parsing code does not handle missing outer ContentInfo
2050 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2051 missing content and trigger a NULL pointer dereference on parsing.
2053 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2054 otherwise parse PKCS#7 structures from untrusted sources are
2055 affected. OpenSSL clients and servers are not affected.
2057 This issue was reported to OpenSSL by Michal Zalewski (Google).
2061 *) DoS via reachable assert in SSLv2 servers fix
2063 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2064 servers that both support SSLv2 and enable export cipher suites by sending
2065 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2067 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2068 (OpenSSL development team).
2072 *) Empty CKE with client auth and DHE fix
2074 If client auth is used then a server can seg fault in the event of a DHE
2075 ciphersuite being selected and a zero length ClientKeyExchange message
2076 being sent by the client. This could be exploited in a DoS attack.
2080 *) Handshake with unseeded PRNG fix
2082 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2083 with an unseeded PRNG. The conditions are:
2084 - The client is on a platform where the PRNG has not been seeded
2085 automatically, and the user has not seeded manually
2086 - A protocol specific client method version has been used (i.e. not
2087 SSL_client_methodv23)
2088 - A ciphersuite is used that does not require additional random data from
2089 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2091 If the handshake succeeds then the client random that has been used will
2092 have been generated from a PRNG with insufficient entropy and therefore the
2093 output may be predictable.
2095 For example using the following command with an unseeded openssl will
2096 succeed on an unpatched platform:
2098 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2102 *) Use After Free following d2i_ECPrivatekey error fix
2104 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2105 could cause a use after free condition. This, in turn, could cause a double
2106 free in several private key parsing functions (such as d2i_PrivateKey
2107 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2108 for applications that receive EC private keys from untrusted
2109 sources. This scenario is considered rare.
2111 This issue was discovered by the BoringSSL project and fixed in their
2116 *) X509_to_X509_REQ NULL pointer deref fix
2118 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2119 the certificate key is invalid. This function is rarely used in practice.
2121 This issue was discovered by Brian Carpenter.
2125 *) Removed the export ciphers from the DEFAULT ciphers
2128 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2130 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2131 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2132 So far those who have to target multiple platforms would compromise
2133 and argue that binary targeting say ARMv5 would still execute on
2134 ARMv8. "Universal" build resolves this compromise by providing
2135 near-optimal performance even on newer platforms.
2138 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2139 (other platforms pending).
2140 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2142 *) Add support for the SignedCertificateTimestampList certificate and
2143 OCSP response extensions from RFC6962.
2146 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2147 for corner cases. (Certain input points at infinity could lead to
2148 bogus results, with non-infinity inputs mapped to infinity too.)
2151 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2152 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2153 common cases are optimized and there still is room for further
2154 improvements. Vector Permutation AES for Altivec is also added.
2157 *) Add support for little-endian ppc64 Linux target.
2158 [Marcelo Cerri (IBM)]
2160 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2161 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2162 are optimized and there still is room for further improvements.
2163 Both 32- and 64-bit modes are supported.
2164 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2166 *) Improved ARMv7 NEON support.
2169 *) Support for SPARC Architecture 2011 crypto extensions, first
2170 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2171 SHA256/512, MD5, GHASH and modular exponentiation.
2172 [Andy Polyakov, David Miller]
2174 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2176 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2178 *) Support for new and upcoming Intel processors, including AVX2,
2179 BMI and SHA ISA extensions. This includes additional "stitched"
2180 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2183 This work was sponsored by Intel Corp.
2186 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2187 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2188 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2191 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2192 this fixes a limitation in previous versions of OpenSSL.
2195 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2196 MGF1 digest and OAEP label.
2199 *) Add EVP support for key wrapping algorithms, to avoid problems with
2200 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2201 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2202 algorithms and include tests cases.
2205 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2207 [Douglas E. Engert, Steve Henson]
2209 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2210 difference in days and seconds between two tm or ASN1_TIME structures.
2213 *) Add -rev test option to s_server to just reverse order of characters
2214 received by client and send back to server. Also prints an abbreviated
2215 summary of the connection parameters.
2218 *) New option -brief for s_client and s_server to print out a brief summary
2219 of connection parameters.
2222 *) Add callbacks for arbitrary TLS extensions.
2223 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2225 *) New option -crl_download in several openssl utilities to download CRLs
2226 from CRLDP extension in certificates.
2229 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2232 *) New function X509_CRL_diff to generate a delta CRL from the difference
2233 of two full CRLs. Add support to "crl" utility.
2236 *) New functions to set lookup_crls function and to retrieve
2237 X509_STORE from X509_STORE_CTX.
2240 *) Print out deprecated issuer and subject unique ID fields in
2244 *) Extend OCSP I/O functions so they can be used for simple general purpose
2245 HTTP as well as OCSP. New wrapper function which can be used to download
2246 CRLs using the OCSP API.
2249 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2252 *) SSL_CONF* functions. These provide a common framework for application
2253 configuration using configuration files or command lines.
2256 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2257 message callback and prints the results. Needs compile time option
2258 "enable-ssl-trace". New options to s_client and s_server to enable
2262 *) New ctrl and macro to retrieve supported points extensions.
2263 Print out extension in s_server and s_client.
2266 *) New functions to retrieve certificate signature and signature
2270 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2274 *) New Suite B modes for TLS code. These use and enforce the requirements
2275 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2276 only use Suite B curves. The Suite B modes can be set by using the
2277 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2280 *) New chain verification flags for Suite B levels of security. Check
2281 algorithms are acceptable when flags are set in X509_verify_cert.
2284 *) Make tls1_check_chain return a set of flags indicating checks passed
2285 by a certificate chain. Add additional tests to handle client
2286 certificates: checks for matching certificate type and issuer name
2290 *) If an attempt is made to use a signature algorithm not in the peer
2291 preference list abort the handshake. If client has no suitable
2292 signature algorithms in response to a certificate request do not
2293 use the certificate.
2296 *) If server EC tmp key is not in client preference list abort handshake.
2299 *) Add support for certificate stores in CERT structure. This makes it
2300 possible to have different stores per SSL structure or one store in
2301 the parent SSL_CTX. Include distinct stores for certificate chain
2302 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2303 to build and store a certificate chain in CERT structure: returning
2304 an error if the chain cannot be built: this will allow applications
2305 to test if a chain is correctly configured.
2307 Note: if the CERT based stores are not set then the parent SSL_CTX
2308 store is used to retain compatibility with existing behaviour.
2312 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2313 mask based on the current session, check mask when sending client
2314 hello and checking the requested ciphersuite.
2317 *) New ctrls to retrieve and set certificate types in a certificate
2318 request message. Print out received values in s_client. If certificate
2319 types is not set with custom values set sensible values based on
2320 supported signature algorithms.
2323 *) Support for distinct client and server supported signature algorithms.
2326 *) Add certificate callback. If set this is called whenever a certificate
2327 is required by client or server. An application can decide which
2328 certificate chain to present based on arbitrary criteria: for example
2329 supported signature algorithms. Add very simple example to s_server.
2330 This fixes many of the problems and restrictions of the existing client
2331 certificate callback: for example you can now clear an existing
2332 certificate and specify the whole chain.
2335 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2336 the certificate can be used for (if anything). Set valid_flags field
2337 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2338 to have similar checks in it.
2340 Add new "cert_flags" field to CERT structure and include a "strict mode".
2341 This enforces some TLS certificate requirements (such as only permitting
2342 certificate signature algorithms contained in the supported algorithms
2343 extension) which some implementations ignore: this option should be used
2344 with caution as it could cause interoperability issues.
2347 *) Update and tidy signature algorithm extension processing. Work out
2348 shared signature algorithms based on preferences and peer algorithms
2349 and print them out in s_client and s_server. Abort handshake if no
2350 shared signature algorithms.
2353 *) Add new functions to allow customised supported signature algorithms
2354 for SSL and SSL_CTX structures. Add options to s_client and s_server
2358 *) New function SSL_certs_clear() to delete all references to certificates
2359 from an SSL structure. Before this once a certificate had been added
2360 it couldn't be removed.
2363 *) Integrate hostname, email address and IP address checking with certificate
2364 verification. New verify options supporting checking in openssl utility.
2367 *) Fixes and wildcard matching support to hostname and email checking
2368 functions. Add manual page.
2369 [Florian Weimer (Red Hat Product Security Team)]
2371 *) New functions to check a hostname email or IP address against a
2372 certificate. Add options x509 utility to print results of checks against
2376 *) Fix OCSP checking.
2377 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2379 *) Initial experimental support for explicitly trusted non-root CAs.
2380 OpenSSL still tries to build a complete chain to a root but if an
2381 intermediate CA has a trust setting included that is used. The first
2382 setting is used: whether to trust (e.g., -addtrust option to the x509
2386 *) Add -trusted_first option which attempts to find certificates in the
2387 trusted store even if an untrusted chain is also supplied.
2390 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2391 platform support for Linux and Android.
2394 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2397 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2398 When in FIPS mode the approved implementations are used as normal,
2399 when not in FIPS mode the internal unapproved versions are used instead.
2400 This means that the FIPS capable OpenSSL isn't forced to use the
2401 (often lower performance) FIPS implementations outside FIPS mode.
2404 *) Transparently support X9.42 DH parameters when calling
2405 PEM_read_bio_DHparameters. This means existing applications can handle
2406 the new parameter format automatically.
2409 *) Initial experimental support for X9.42 DH parameter format: mainly
2410 to support use of 'q' parameter for RFC5114 parameters.
2413 *) Add DH parameters from RFC5114 including test data to dhtest.
2416 *) Support for automatic EC temporary key parameter selection. If enabled
2417 the most preferred EC parameters are automatically used instead of
2418 hardcoded fixed parameters. Now a server just has to call:
2419 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2420 support ECDH and use the most appropriate parameters.
2423 *) Enhance and tidy EC curve and point format TLS extension code. Use
2424 static structures instead of allocation if default values are used.
2425 New ctrls to set curves we wish to support and to retrieve shared curves.
2426 Print out shared curves in s_server. New options to s_server and s_client
2427 to set list of supported curves.
2430 *) New ctrls to retrieve supported signature algorithms and
2431 supported curve values as an array of NIDs. Extend openssl utility
2432 to print out received values.
2435 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2436 between NIDs and the more common NIST names such as "P-256". Enhance
2437 ecparam utility and ECC method to recognise the NIST names for curves.
2440 *) Enhance SSL/TLS certificate chain handling to support different
2441 chains for each certificate instead of one chain in the parent SSL_CTX.
2444 *) Support for fixed DH ciphersuite client authentication: where both
2445 server and client use DH certificates with common parameters.
2448 *) Support for fixed DH ciphersuites: those requiring DH server
2452 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2454 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2455 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2456 X509_CINF_get_signature were reverted post internal team review.
2458 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2460 *) Build fixes for the Windows and OpenVMS platforms
2461 [Matt Caswell and Richard Levitte]
2463 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2465 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2466 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2467 dereference. This could lead to a Denial Of Service attack. Thanks to
2468 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2472 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2473 dtls1_buffer_record function under certain conditions. In particular this
2474 could occur if an attacker sent repeated DTLS records with the same
2475 sequence number but for the next epoch. The memory leak could be exploited
2476 by an attacker in a Denial of Service attack through memory exhaustion.
2477 Thanks to Chris Mueller for reporting this issue.
2481 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2482 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2483 method would be set to NULL which could later result in a NULL pointer
2484 dereference. Thanks to Frank Schmirler for reporting this issue.
2488 *) Abort handshake if server key exchange message is omitted for ephemeral
2491 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2492 reporting this issue.
2496 *) Remove non-export ephemeral RSA code on client and server. This code
2497 violated the TLS standard by allowing the use of temporary RSA keys in
2498 non-export ciphersuites and could be used by a server to effectively
2499 downgrade the RSA key length used to a value smaller than the server
2500 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2501 INRIA or reporting this issue.
2505 *) Fixed issue where DH client certificates are accepted without verification.
2506 An OpenSSL server will accept a DH certificate for client authentication
2507 without the certificate verify message. This effectively allows a client to
2508 authenticate without the use of a private key. This only affects servers
2509 which trust a client certificate authority which issues certificates
2510 containing DH keys: these are extremely rare and hardly ever encountered.
2511 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2516 *) Ensure that the session ID context of an SSL is updated when its
2517 SSL_CTX is updated via SSL_set_SSL_CTX.
2519 The session ID context is typically set from the parent SSL_CTX,
2520 and can vary with the CTX.
2523 *) Fix various certificate fingerprint issues.
2525 By using non-DER or invalid encodings outside the signed portion of a
2526 certificate the fingerprint can be changed without breaking the signature.
2527 Although no details of the signed portion of the certificate can be changed
2528 this can cause problems with some applications: e.g. those using the
2529 certificate fingerprint for blacklists.
2531 1. Reject signatures with non zero unused bits.
2533 If the BIT STRING containing the signature has non zero unused bits reject
2534 the signature. All current signature algorithms require zero unused bits.
2536 2. Check certificate algorithm consistency.
2538 Check the AlgorithmIdentifier inside TBS matches the one in the
2539 certificate signature. NB: this will result in signature failure
2540 errors for some broken certificates.
2542 Thanks to Konrad Kraszewski from Google for reporting this issue.
2544 3. Check DSA/ECDSA signatures use DER.
2546 Re-encode DSA/ECDSA signatures and compare with the original received
2547 signature. Return an error if there is a mismatch.
2549 This will reject various cases including garbage after signature
2550 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2551 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2552 (negative or with leading zeroes).
2554 Further analysis was conducted and fixes were developed by Stephen Henson
2555 of the OpenSSL core team.
2560 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2561 results on some platforms, including x86_64. This bug occurs at random
2562 with a very low probability, and is not known to be exploitable in any
2563 way, though its exact impact is difficult to determine. Thanks to Pieter
2564 Wuille (Blockstream) who reported this issue and also suggested an initial
2565 fix. Further analysis was conducted by the OpenSSL development team and
2566 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2567 the OpenSSL core team.
2571 *) Do not resume sessions on the server if the negotiated protocol
2572 version does not match the session's version. Resuming with a different
2573 version, while not strictly forbidden by the RFC, is of questionable
2574 sanity and breaks all known clients.
2575 [David Benjamin, Emilia Käsper]
2577 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2578 early CCS messages during renegotiation. (Note that because
2579 renegotiation is encrypted, this early CCS was not exploitable.)
2582 *) Tighten client-side session ticket handling during renegotiation:
2583 ensure that the client only accepts a session ticket if the server sends
2584 the extension anew in the ServerHello. Previously, a TLS client would
2585 reuse the old extension state and thus accept a session ticket if one was
2586 announced in the initial ServerHello.
2588 Similarly, ensure that the client requires a session ticket if one
2589 was advertised in the ServerHello. Previously, a TLS client would
2590 ignore a missing NewSessionTicket message.
2593 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2595 *) SRTP Memory Leak.
2597 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2598 sends a carefully crafted handshake message, to cause OpenSSL to fail
2599 to free up to 64k of memory causing a memory leak. This could be
2600 exploited in a Denial Of Service attack. This issue affects OpenSSL
2601 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2602 whether SRTP is used or configured. Implementations of OpenSSL that
2603 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2605 The fix was developed by the OpenSSL team.
2609 *) Session Ticket Memory Leak.
2611 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2612 integrity of that ticket is first verified. In the event of a session
2613 ticket integrity check failing, OpenSSL will fail to free memory
2614 causing a memory leak. By sending a large number of invalid session
2615 tickets an attacker could exploit this issue in a Denial Of Service
2620 *) Build option no-ssl3 is incomplete.
2622 When OpenSSL is configured with "no-ssl3" as a build option, servers
2623 could accept and complete a SSL 3.0 handshake, and clients could be
2624 configured to send them.
2626 [Akamai and the OpenSSL team]
2628 *) Add support for TLS_FALLBACK_SCSV.
2629 Client applications doing fallback retries should call
2630 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2632 [Adam Langley, Bodo Moeller]
2634 *) Add additional DigestInfo checks.
2636 Re-encode DigestInto in DER and check against the original when
2637 verifying RSA signature: this will reject any improperly encoded
2638 DigestInfo structures.
2640 Note: this is a precautionary measure and no attacks are currently known.
2644 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2646 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2647 SRP code can be overrun an internal buffer. Add sanity check that
2648 g, A, B < N to SRP code.
2650 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2651 Group for discovering this issue.
2655 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2656 TLS 1.0 instead of higher protocol versions when the ClientHello message
2657 is badly fragmented. This allows a man-in-the-middle attacker to force a
2658 downgrade to TLS 1.0 even if both the server and the client support a
2659 higher protocol version, by modifying the client's TLS records.
2661 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2662 researching this issue.
2666 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2667 to a denial of service attack. A malicious server can crash the client
2668 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2669 ciphersuite and sending carefully crafted handshake messages.
2671 Thanks to Felix Gröbert (Google) for discovering and researching this
2676 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2677 to leak memory. This can be exploited through a Denial of Service attack.
2678 Thanks to Adam Langley for discovering and researching this issue.
2682 *) An attacker can force openssl to consume large amounts of memory whilst
2683 processing DTLS handshake messages. This can be exploited through a
2684 Denial of Service attack.
2685 Thanks to Adam Langley for discovering and researching this issue.
2689 *) An attacker can force an error condition which causes openssl to crash
2690 whilst processing DTLS packets due to memory being freed twice. This
2691 can be exploited through a Denial of Service attack.
2692 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2697 *) If a multithreaded client connects to a malicious server using a resumed
2698 session and the server sends an ec point format extension it could write
2699 up to 255 bytes to freed memory.
2701 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2706 *) A malicious server can crash an OpenSSL client with a null pointer
2707 dereference (read) by specifying an SRP ciphersuite even though it was not
2708 properly negotiated with the client. This can be exploited through a
2709 Denial of Service attack.
2711 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2712 discovering and researching this issue.
2716 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2717 X509_name_oneline, X509_name_print_ex et al. to leak some information
2718 from the stack. Applications may be affected if they echo pretty printing
2719 output to the attacker.
2721 Thanks to Ivan Fratric (Google) for discovering this issue.
2723 [Emilia Käsper, and Steve Henson]
2725 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2726 for corner cases. (Certain input points at infinity could lead to
2727 bogus results, with non-infinity inputs mapped to infinity too.)
2730 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2732 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2733 handshake can force the use of weak keying material in OpenSSL
2734 SSL/TLS clients and servers.
2736 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2737 researching this issue. (CVE-2014-0224)
2738 [KIKUCHI Masashi, Steve Henson]
2740 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2741 OpenSSL DTLS client the code can be made to recurse eventually crashing
2744 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2746 [Imre Rad, Steve Henson]
2748 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2749 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2750 client or server. This is potentially exploitable to run arbitrary
2751 code on a vulnerable client or server.
2753 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2754 [Jüri Aedla, Steve Henson]
2756 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2757 are subject to a denial of service attack.
2759 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2760 this issue. (CVE-2014-3470)
2761 [Felix Gröbert, Ivan Fratric, Steve Henson]
2763 *) Harmonize version and its documentation. -f flag is used to display
2765 [mancha <mancha1@zoho.com>]
2767 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2768 in i2d_ECPrivateKey.
2769 [mancha <mancha1@zoho.com>]
2771 *) Fix some double frees. These are not thought to be exploitable.
2772 [mancha <mancha1@zoho.com>]
2774 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2776 *) A missing bounds check in the handling of the TLS heartbeat extension
2777 can be used to reveal up to 64k of memory to a connected client or
2780 Thanks for Neel Mehta of Google Security for discovering this bug and to
2781 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2782 preparing the fix (CVE-2014-0160)
2783 [Adam Langley, Bodo Moeller]
2785 *) Fix for the attack described in the paper "Recovering OpenSSL
2786 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2787 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2788 http://eprint.iacr.org/2014/140
2790 Thanks to Yuval Yarom and Naomi Benger for discovering this
2791 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2792 [Yuval Yarom and Naomi Benger]
2794 *) TLS pad extension: draft-agl-tls-padding-03
2796 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2797 TLS client Hello record length value would otherwise be > 255 and
2798 less that 512 pad with a dummy extension containing zeroes so it
2799 is at least 512 bytes long.
2801 [Adam Langley, Steve Henson]
2803 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2805 *) Fix for TLS record tampering bug. A carefully crafted invalid
2806 handshake could crash OpenSSL with a NULL pointer exception.
2807 Thanks to Anton Johansson for reporting this issues.
2810 *) Keep original DTLS digest and encryption contexts in retransmission
2811 structures so we can use the previous session parameters if they need
2812 to be resent. (CVE-2013-6450)
2815 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2816 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2817 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2818 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2819 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2820 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2821 [Rob Stradling, Adam Langley]
2823 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2825 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2826 supporting platforms or when small records were transferred.
2827 [Andy Polyakov, Steve Henson]
2829 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2831 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2833 This addresses the flaw in CBC record processing discovered by
2834 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2835 at: http://www.isg.rhul.ac.uk/tls/
2837 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2838 Security Group at Royal Holloway, University of London
2839 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2840 Emilia Käsper for the initial patch.
2842 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2844 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2845 ciphersuites which can be exploited in a denial of service attack.
2846 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2847 and detecting this bug and to Wolfgang Ettlinger
2848 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2852 *) Return an error when checking OCSP signatures when key is NULL.
2853 This fixes a DoS attack. (CVE-2013-0166)
2856 *) Make openssl verify return errors.
2857 [Chris Palmer <palmer@google.com> and Ben Laurie]
2859 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2860 the right response is stapled. Also change SSL_get_certificate()
2861 so it returns the certificate actually sent.
2862 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2863 [Rob Stradling <rob.stradling@comodo.com>]
2865 *) Fix possible deadlock when decoding public keys.
2868 *) Don't use TLS 1.0 record version number in initial client hello
2872 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2874 *) Sanity check record length before skipping explicit IV in TLS
2875 1.2, 1.1 and DTLS to fix DoS attack.
2877 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2878 fuzzing as a service testing platform.
2882 *) Initialise tkeylen properly when encrypting CMS messages.
2883 Thanks to Solar Designer of Openwall for reporting this issue.
2886 *) In FIPS mode don't try to use composite ciphers as they are not
2890 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2892 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2893 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2894 mean any application compiled against OpenSSL 1.0.0 headers setting
2895 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
2896 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2897 0x10000000L Any application which was previously compiled against
2898 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2899 will need to be recompiled as a result. Letting be results in
2900 inability to disable specifically TLS 1.1 and in client context,
2901 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2904 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
2905 disable just protocol X, but all protocols above X *if* there are
2906 protocols *below* X still enabled. In more practical terms it means
2907 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2908 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2909 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2913 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2915 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2916 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2917 in CRYPTO_realloc_clean.
2919 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2920 issue and to Adam Langley <agl@chromium.org> for fixing it.
2922 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2924 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2927 *) Workarounds for some broken servers that "hang" if a client hello
2928 record length exceeds 255 bytes.
2930 1. Do not use record version number > TLS 1.0 in initial client
2931 hello: some (but not all) hanging servers will now work.
2932 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2933 the number of ciphers sent in the client hello. This should be
2934 set to an even number, such as 50, for example by passing:
2935 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2936 Most broken servers should now work.
2937 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2938 TLS 1.2 client support entirely.
2941 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2944 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2946 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2947 STRING form instead of a DigestInfo.
2950 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2951 and the RSA_sign/RSA_verify functions. This was made more apparent when
2952 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2953 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2954 the correct format in RSA_verify so both forms transparently work.
2957 *) Some servers which support TLS 1.0 can choke if we initially indicate
2958 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2959 encrypted premaster secret. As a workaround use the maximum permitted
2960 client version in client hello, this should keep such servers happy
2961 and still work with previous versions of OpenSSL.
2964 *) Add support for TLS/DTLS heartbeats.
2965 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2967 *) Add support for SCTP.
2968 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2970 *) Improved PRNG seeding for VOS.
2971 [Paul Green <Paul.Green@stratus.com>]
2973 *) Extensive assembler packs updates, most notably:
2975 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2976 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2977 - x86_64: bit-sliced AES implementation;
2978 - ARM: NEON support, contemporary platforms optimizations;
2979 - s390x: z196 support;
2980 - *: GHASH and GF(2^m) multiplication implementations;
2984 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2985 (removal of unnecessary code)
2986 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2988 *) Add TLS key material exporter from RFC 5705.
2991 *) Add DTLS-SRTP negotiation from RFC 5764.
2994 *) Add Next Protocol Negotiation,
2995 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2996 disabled with a no-npn flag to config or Configure. Code donated
2998 [Adam Langley <agl@google.com> and Ben Laurie]
3000 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3001 NIST-P256, NIST-P521, with constant-time single point multiplication on
3002 typical inputs. Compiler support for the nonstandard type __uint128_t is
3003 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3004 Code made available under Apache License version 2.0.
3006 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3007 line to include this in your build of OpenSSL, and run "make depend" (or
3008 "make update"). This enables the following EC_METHODs:
3010 EC_GFp_nistp224_method()
3011 EC_GFp_nistp256_method()
3012 EC_GFp_nistp521_method()
3014 EC_GROUP_new_by_curve_name() will automatically use these (while
3015 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3017 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3019 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3020 all platforms. Move ssize_t definition from e_os.h to the public
3021 header file e_os2.h as it now appears in public header file cms.h
3024 *) New -sigopt option to the ca, req and x509 utilities. Additional
3025 signature parameters can be passed using this option and in
3029 *) Add RSA PSS signing function. This will generate and set the
3030 appropriate AlgorithmIdentifiers for PSS based on those in the
3031 corresponding EVP_MD_CTX structure. No application support yet.
3034 *) Support for companion algorithm specific ASN1 signing routines.
3035 New function ASN1_item_sign_ctx() signs a pre-initialised
3036 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3037 the appropriate parameters.
3040 *) Add new algorithm specific ASN1 verification initialisation function
3041 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3042 handling will be the same no matter what EVP_PKEY_METHOD is used.
3043 Add a PSS handler to support verification of PSS signatures: checked
3044 against a number of sample certificates.
3047 *) Add signature printing for PSS. Add PSS OIDs.
3048 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3050 *) Add algorithm specific signature printing. An individual ASN1 method
3051 can now print out signatures instead of the standard hex dump.
3053 More complex signatures (e.g. PSS) can print out more meaningful
3054 information. Include DSA version that prints out the signature
3058 *) Password based recipient info support for CMS library: implementing
3062 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3063 neatly separates the code into cipher and PBE sections and is required
3064 for some algorithms that split PBES2 into separate pieces (such as
3065 password based CMS).
3068 *) Session-handling fixes:
3069 - Fix handling of connections that are resuming with a session ID,
3070 but also support Session Tickets.
3071 - Fix a bug that suppressed issuing of a new ticket if the client
3072 presented a ticket with an expired session.
3073 - Try to set the ticket lifetime hint to something reasonable.
3074 - Make tickets shorter by excluding irrelevant information.
3075 - On the client side, don't ignore renewed tickets.
3076 [Adam Langley, Bodo Moeller (Google)]
3078 *) Fix PSK session representation.
3081 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3083 This work was sponsored by Intel.
3086 *) Add GCM support to TLS library. Some custom code is needed to split
3087 the IV between the fixed (from PRF) and explicit (from TLS record)
3088 portions. This adds all GCM ciphersuites supported by RFC5288 and
3089 RFC5289. Generalise some AES* cipherstrings to include GCM and
3090 add a special AESGCM string for GCM only.
3093 *) Expand range of ctrls for AES GCM. Permit setting invocation
3094 field on decrypt and retrieval of invocation field only on encrypt.
3097 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3098 As required by RFC5289 these ciphersuites cannot be used if for
3099 versions of TLS earlier than 1.2.
3102 *) For FIPS capable OpenSSL interpret a NULL default public key method
3103 as unset and return the appropriate default but do *not* set the default.
3104 This means we can return the appropriate method in applications that
3105 switch between FIPS and non-FIPS modes.
3108 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3109 ENGINE is used then we cannot handle that in the FIPS module so we
3110 keep original code iff non-FIPS operations are allowed.
3113 *) Add -attime option to openssl utilities.
3114 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3116 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3119 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3120 FIPS EC methods unconditionally for now.
3123 *) New build option no-ec2m to disable characteristic 2 code.
3126 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3127 all cases can be covered as some introduce binary incompatibilities.
3130 *) Redirect RSA operations to FIPS module including keygen,
3131 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3134 *) Add similar low level API blocking to ciphers.
3137 *) Low level digest APIs are not approved in FIPS mode: any attempt
3138 to use these will cause a fatal error. Applications that *really* want
3139 to use them can use the private_* version instead.
3142 *) Redirect cipher operations to FIPS module for FIPS builds.
3145 *) Redirect digest operations to FIPS module for FIPS builds.
3148 *) Update build system to add "fips" flag which will link in fipscanister.o
3149 for static and shared library builds embedding a signature if needed.
3152 *) Output TLS supported curves in preference order instead of numerical
3153 order. This is currently hardcoded for the highest order curves first.
3154 This should be configurable so applications can judge speed vs strength.
3157 *) Add TLS v1.2 server support for client authentication.
3160 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3164 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3165 FIPS modules versions.
3168 *) Add TLS v1.2 client side support for client authentication. Keep cache
3169 of handshake records longer as we don't know the hash algorithm to use
3170 until after the certificate request message is received.
3173 *) Initial TLS v1.2 client support. Add a default signature algorithms
3174 extension including all the algorithms we support. Parse new signature
3175 format in client key exchange. Relax some ECC signing restrictions for
3176 TLS v1.2 as indicated in RFC5246.
3179 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3180 to new signature format when needed using client digest preference.
3181 All server ciphersuites should now work correctly in TLS v1.2. No client
3182 support yet and no support for client certificates.
3185 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3186 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3187 ciphersuites. At present only RSA key exchange ciphersuites work with
3188 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3189 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3190 and version checking.
3193 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3194 with this defined it will not be affected by any changes to ssl internal
3195 structures. Add several utility functions to allow openssl application
3196 to work with OPENSSL_NO_SSL_INTERN defined.
3200 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
3202 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3205 *) Permit abbreviated handshakes when renegotiating using the function
3206 SSL_renegotiate_abbreviated().
3207 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3209 *) Add call to ENGINE_register_all_complete() to
3210 ENGINE_load_builtin_engines(), so some implementations get used
3211 automatically instead of needing explicit application support.
3214 *) Add support for TLS key exporter as described in RFC5705.
3215 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3217 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3218 a few changes are required:
3220 Add SSL_OP_NO_TLSv1_1 flag.
3221 Add TLSv1_1 methods.
3222 Update version checking logic to handle version 1.1.
3223 Add explicit IV handling (ported from DTLS code).
3224 Add command line options to s_client/s_server.
3227 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3229 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3230 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3231 content decryption and always return the same error. Note: this attack
3232 needs on average 2^20 messages so it only affects automated senders. The
3233 old behaviour can be re-enabled in the CMS code by setting the
3234 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3235 an MMA defence is not necessary.
3236 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3237 this issue. (CVE-2012-0884)
3240 *) Fix CVE-2011-4619: make sure we really are receiving a
3241 client hello before rejecting multiple SGC restarts. Thanks to
3242 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3245 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3247 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3248 Thanks to Antonio Martin, Enterprise Secure Access Research and
3249 Development, Cisco Systems, Inc. for discovering this bug and
3250 preparing a fix. (CVE-2012-0050)
3253 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3255 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3256 of the Vaudenay padding oracle attack on CBC mode encryption
3257 which enables an efficient plaintext recovery attack against
3258 the OpenSSL implementation of DTLS. Their attack exploits timing
3259 differences arising during decryption processing. A research
3260 paper describing this attack can be found at:
3261 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3262 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3263 Security Group at Royal Holloway, University of London
3264 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3265 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3266 for preparing the fix. (CVE-2011-4108)
3267 [Robin Seggelmann, Michael Tuexen]
3269 *) Clear bytes used for block padding of SSL 3.0 records.
3271 [Adam Langley (Google)]
3273 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3274 Kadianakis <desnacked@gmail.com> for discovering this issue and
3275 Adam Langley for preparing the fix. (CVE-2011-4619)
3276 [Adam Langley (Google)]
3278 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3279 [Andrey Kulikov <amdeich@gmail.com>]
3281 *) Prevent malformed RFC3779 data triggering an assertion failure.
3282 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3283 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3284 [Rob Austein <sra@hactrn.net>]
3286 *) Improved PRNG seeding for VOS.
3287 [Paul Green <Paul.Green@stratus.com>]
3289 *) Fix ssl_ciph.c set-up race.
3290 [Adam Langley (Google)]
3292 *) Fix spurious failures in ecdsatest.c.
3293 [Emilia Käsper (Google)]
3295 *) Fix the BIO_f_buffer() implementation (which was mixing different
3296 interpretations of the '..._len' fields).
3297 [Adam Langley (Google)]
3299 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3300 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3301 threads won't reuse the same blinding coefficients.
3303 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3304 lock to call BN_BLINDING_invert_ex, and avoids one use of
3305 BN_BLINDING_update for each BN_BLINDING structure (previously,
3306 the last update always remained unused).
3307 [Emilia Käsper (Google)]
3309 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3310 [Bob Buckholz (Google)]
3312 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3314 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3315 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3316 [Kaspar Brand <ossl@velox.ch>]
3318 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3319 for multi-threaded use of ECDH. (CVE-2011-3210)
3320 [Adam Langley (Google)]
3322 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3325 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3326 signature public key algorithm by using OID xref utilities instead.
3327 Before this you could only use some ECC ciphersuites with SHA1 only.
3330 *) Add protection against ECDSA timing attacks as mentioned in the paper
3331 by Billy Bob Brumley and Nicola Tuveri, see:
3333 http://eprint.iacr.org/2011/232.pdf
3335 [Billy Bob Brumley and Nicola Tuveri]
3337 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3339 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3340 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3342 *) Fix bug in string printing code: if *any* escaping is enabled we must
3343 escape the escape character (backslash) or the resulting string is
3347 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3349 *) Disable code workaround for ancient and obsolete Netscape browsers
3350 and servers: an attacker can use it in a ciphersuite downgrade attack.
3351 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3354 *) Fixed J-PAKE implementation error, originally discovered by
3355 Sebastien Martini, further info and confirmation from Stefan
3356 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3359 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3361 *) Fix extension code to avoid race conditions which can result in a buffer
3362 overrun vulnerability: resumed sessions must not be modified as they can
3363 be shared by multiple threads. CVE-2010-3864
3366 *) Fix WIN32 build system to correctly link an ENGINE directory into
3370 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3372 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3374 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3376 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3378 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3379 context. The operation can be customised via the ctrl mechanism in
3380 case ENGINEs want to include additional functionality.
3383 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3386 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3387 output hashes compatible with older versions of OpenSSL.
3388 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3390 *) Fix compression algorithm handling: if resuming a session use the
3391 compression algorithm of the resumed session instead of determining
3392 it from client hello again. Don't allow server to change algorithm.
3395 *) Add load_crls() function to apps tidying load_certs() too. Add option
3396 to verify utility to allow additional CRLs to be included.
3399 *) Update OCSP request code to permit adding custom headers to the request:
3400 some responders need this.
3403 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3405 [Julia Lawall <julia@diku.dk>]
3407 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3408 needlessly dereferenced structures, used obsolete functions and
3409 didn't handle all updated verify codes correctly.
3412 *) Disable MD2 in the default configuration.
3415 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3416 indicate the initial BIO being pushed or popped. This makes it possible
3417 to determine whether the BIO is the one explicitly called or as a result
3418 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3419 it handles reference counts correctly and doesn't zero out the I/O bio
3420 when it is not being explicitly popped. WARNING: applications which
3421 included workarounds for the old buggy behaviour will need to be modified
3422 or they could free up already freed BIOs.
3425 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3426 renaming to all platforms (within the 0.9.8 branch, this was
3427 done conditionally on Netware platforms to avoid a name clash).
3428 [Guenter <lists@gknw.net>]
3430 *) Add ECDHE and PSK support to DTLS.
3431 [Michael Tuexen <tuexen@fh-muenster.de>]
3433 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3437 *) Add "missing" function EVP_MD_flags() (without this the only way to
3438 retrieve a digest flags is by accessing the structure directly. Update
3439 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3440 or cipher is registered as in the "from" argument. Print out all
3441 registered digests in the dgst usage message instead of manually
3442 attempting to work them out.
3445 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3446 this allows the use of compression and extensions. Change default cipher
3447 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3448 by default unless an application cipher string requests it.
3451 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3452 key ids to find matching certificates and keys but some PKCS#12 files
3453 don't follow the (somewhat unwritten) rules and this strategy fails.
3454 Now just gather all certificates together and the first private key
3455 then look for the first certificate that matches the key.
3458 *) Support use of registered digest and cipher names for dgst and cipher
3459 commands instead of having to add each one as a special case. So now
3466 openssl dgst -sha256 foo
3468 and this works for ENGINE based algorithms too.
3472 *) Update Gost ENGINE to support parameter files.
3473 [Victor B. Wagner <vitus@cryptocom.ru>]
3475 *) Support GeneralizedTime in ca utility.
3476 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3478 *) Enhance the hash format used for certificate directory links. The new
3479 form uses the canonical encoding (meaning equivalent names will work
3480 even if they aren't identical) and uses SHA1 instead of MD5. This form
3481 is incompatible with the older format and as a result c_rehash should
3482 be used to rebuild symbolic links.
3485 *) Make PKCS#8 the default write format for private keys, replacing the
3486 traditional format. This form is standardised, more secure and doesn't
3487 include an implicit MD5 dependency.
3490 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3491 committed to OpenSSL should pass this lot as a minimum.
3494 *) Add session ticket override functionality for use by EAP-FAST.
3495 [Jouni Malinen <j@w1.fi>]
3497 *) Modify HMAC functions to return a value. Since these can be implemented
3498 in an ENGINE errors can occur.
3501 *) Type-checked OBJ_bsearch_ex.
3504 *) Type-checked OBJ_bsearch. Also some constification necessitated
3505 by type-checking. Still to come: TXT_DB, bsearch(?),
3506 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3510 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3511 seconds to a tm structure directly, instead of going through OS
3512 specific date routines. This avoids any issues with OS routines such
3513 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3514 and X509_time_adj_ex() to cover the extended range. The existing
3515 X509_time_adj() is still usable and will no longer have any date issues.
3518 *) Delta CRL support. New use deltas option which will attempt to locate
3519 and search any appropriate delta CRLs available.
3521 This work was sponsored by Google.
3524 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3525 code and add additional score elements. Validate alternate CRL paths
3526 as part of the CRL checking and indicate a new error "CRL path validation
3527 error" in this case. Applications wanting additional details can use
3528 the verify callback and check the new "parent" field. If this is not
3529 NULL CRL path validation is taking place. Existing applications won't
3530 see this because it requires extended CRL support which is off by
3533 This work was sponsored by Google.
3536 *) Support for freshest CRL extension.
3538 This work was sponsored by Google.
3541 *) Initial indirect CRL support. Currently only supported in the CRLs
3542 passed directly and not via lookup. Process certificate issuer
3543 CRL entry extension and lookup CRL entries by bother issuer name
3544 and serial number. Check and process CRL issuer entry in IDP extension.
3546 This work was sponsored by Google.
3549 *) Add support for distinct certificate and CRL paths. The CRL issuer
3550 certificate is validated separately in this case. Only enabled if
3551 an extended CRL support flag is set: this flag will enable additional
3552 CRL functionality in future.
3554 This work was sponsored by Google.
3557 *) Add support for policy mappings extension.
3559 This work was sponsored by Google.
3562 *) Fixes to pathlength constraint, self issued certificate handling,
3563 policy processing to align with RFC3280 and PKITS tests.
3565 This work was sponsored by Google.
3568 *) Support for name constraints certificate extension. DN, email, DNS
3569 and URI types are currently supported.
3571 This work was sponsored by Google.
3574 *) To cater for systems that provide a pointer-based thread ID rather
3575 than numeric, deprecate the current numeric thread ID mechanism and
3576 replace it with a structure and associated callback type. This
3577 mechanism allows a numeric "hash" to be extracted from a thread ID in
3578 either case, and on platforms where pointers are larger than 'long',
3579 mixing is done to help ensure the numeric 'hash' is usable even if it
3580 can't be guaranteed unique. The default mechanism is to use "&errno"
3581 as a pointer-based thread ID to distinguish between threads.
3583 Applications that want to provide their own thread IDs should now use
3584 CRYPTO_THREADID_set_callback() to register a callback that will call
3585 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3587 Note that ERR_remove_state() is now deprecated, because it is tied
3588 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3589 to free the current thread's error state should be replaced by
3590 ERR_remove_thread_state(NULL).
3592 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3593 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3594 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3595 application was previously providing a numeric thread callback that
3596 was inappropriate for distinguishing threads, then uniqueness might
3597 have been obtained with &errno that happened immediately in the
3598 intermediate development versions of OpenSSL; this is no longer the
3599 case, the numeric thread callback will now override the automatic use
3601 [Geoff Thorpe, with help from Bodo Moeller]
3603 *) Initial support for different CRL issuing certificates. This covers a
3604 simple case where the self issued certificates in the chain exist and
3605 the real CRL issuer is higher in the existing chain.
3607 This work was sponsored by Google.
3610 *) Removed effectively defunct crypto/store from the build.
3613 *) Revamp of STACK to provide stronger type-checking. Still to come:
3614 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3615 ASN1_STRING, CONF_VALUE.
3618 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3619 RAM on SSL connections. This option can save about 34k per idle SSL.
3622 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3623 STACK, TXT_DB, bsearch, qsort.
3626 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3627 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3628 support for data, signedData, compressedData, digestedData and
3629 encryptedData, envelopedData types included. Scripts to check against
3630 RFC4134 examples draft and interop and consistency checks of many
3631 content types and variants.
3634 *) Add options to enc utility to support use of zlib compression BIO.
3637 *) Extend mk1mf to support importing of options and assembly language
3638 files from Configure script, currently only included in VC-WIN32.
3639 The assembly language rules can now optionally generate the source
3640 files from the associated perl scripts.
3643 *) Implement remaining functionality needed to support GOST ciphersuites.
3644 Interop testing has been performed using CryptoPro implementations.
3645 [Victor B. Wagner <vitus@cryptocom.ru>]
3647 *) s390x assembler pack.
3650 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3654 *) Implement Opaque PRF Input TLS extension as specified in
3655 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3656 official specification yet and no extension type assignment by
3657 IANA exists, this extension (for now) will have to be explicitly
3658 enabled when building OpenSSL by providing the extension number
3659 to use. For example, specify an option
3661 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3663 to the "config" or "Configure" script to enable the extension,
3664 assuming extension number 0x9527 (which is a completely arbitrary
3665 and unofficial assignment based on the MD5 hash of the Internet
3666 Draft). Note that by doing so, you potentially lose
3667 interoperability with other TLS implementations since these might
3668 be using the same extension number for other purposes.
3670 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3671 opaque PRF input value to use in the handshake. This will create
3672 an internal copy of the length-'len' string at 'src', and will
3673 return non-zero for success.
3675 To get more control and flexibility, provide a callback function
3678 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3679 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3683 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3686 Callback function 'cb' will be called in handshakes, and is
3687 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3688 Argument 'arg' is for application purposes (the value as given to
3689 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3690 be provided to the callback function). The callback function
3691 has to return non-zero to report success: usually 1 to use opaque
3692 PRF input just if possible, or 2 to enforce use of the opaque PRF
3693 input. In the latter case, the library will abort the handshake
3694 if opaque PRF input is not successfully negotiated.
3696 Arguments 'peerinput' and 'len' given to the callback function
3697 will always be NULL and 0 in the case of a client. A server will
3698 see the client's opaque PRF input through these variables if
3699 available (NULL and 0 otherwise). Note that if the server
3700 provides an opaque PRF input, the length must be the same as the
3701 length of the client's opaque PRF input.
3703 Note that the callback function will only be called when creating
3704 a new session (session resumption can resume whatever was
3705 previously negotiated), and will not be called in SSL 2.0
3706 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3707 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3708 for applications that need to enforce opaque PRF input.
3712 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3715 [Victor B. Wagner <vitus@cryptocom.ru>]
3717 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3718 RFC4507bis. The encrypted ticket format is an encrypted encoded
3719 SSL_SESSION structure, that way new session features are automatically
3722 If a client application caches session in an SSL_SESSION structure
3723 support is transparent because tickets are now stored in the encoded
3726 The SSL_CTX structure automatically generates keys for ticket
3727 protection in servers so again support should be possible
3728 with no application modification.
3730 If a client or server wishes to disable RFC4507 support then the option
3731 SSL_OP_NO_TICKET can be set.
3733 Add a TLS extension debugging callback to allow the contents of any client
3734 or server extensions to be examined.
3736 This work was sponsored by Google.
3739 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3740 OpenSSL should now compile cleanly on gcc 4.2
3741 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3743 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3744 support including streaming MAC support: this is required for GOST
3745 ciphersuite support.
3746 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3748 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3749 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3750 to output in BER and PEM format.
3753 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3754 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3755 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3756 ENGINE support for HMAC keys which are unextractable. New -mac and
3757 -macopt options to dgst utility.
3760 *) New option -sigopt to dgst utility. Update dgst to use
3761 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3762 alternative signing parameters such as X9.31 or PSS in the dgst
3766 *) Change ssl_cipher_apply_rule(), the internal function that does
3767 the work each time a ciphersuite string requests enabling
3768 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3769 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3770 the order of disabled ciphersuites such that those ciphersuites
3771 that most recently went from enabled to disabled not only stay
3772 in order with respect to each other, but also have higher priority
3773 than other disabled ciphersuites the next time ciphersuites are
3776 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3777 the same ciphersuites as with "HIGH" alone, but in a specific
3778 order where the PSK ciphersuites come first (since they are the
3779 most recently disabled ciphersuites when "HIGH" is parsed).
3781 Also, change ssl_create_cipher_list() (using this new
3782 functionality) such that between otherwise identical
3783 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
3787 *) Change ssl_create_cipher_list() so that it automatically
3788 arranges the ciphersuites in reasonable order before starting
3789 to process the rule string. Thus, the definition for "DEFAULT"
3790 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3791 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3792 This makes it much easier to arrive at a reasonable default order
3793 in applications for which anonymous ciphers are OK (meaning
3794 that you can't actually use DEFAULT).
3795 [Bodo Moeller; suggested by Victor Duchovni]
3797 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3798 processing) into multiple integers instead of setting
3799 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3800 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3801 (These masks as well as the individual bit definitions are hidden
3802 away into the non-exported interface ssl/ssl_locl.h, so this
3803 change to the definition of the SSL_CIPHER structure shouldn't
3804 affect applications.) This give us more bits for each of these
3805 categories, so there is no longer a need to coagulate AES128 and
3806 AES256 into a single algorithm bit, and to coagulate Camellia128
3807 and Camellia256 into a single algorithm bit, which has led to all
3810 Thus, among other things, the kludge introduced in 0.9.7m and
3811 0.9.8e for masking out AES256 independently of AES128 or masking
3812 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3814 With the change, we also introduce new ciphersuite aliases that
3815 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3819 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3820 Use the leftmost N bytes of the signature input if the input is
3821 larger than the prime q (with N being the size in bytes of q).
3824 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3825 it yet and it is largely untested.
3828 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3831 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3832 some compilers (gcc 4.2 and later) reject their use. Safestack is
3833 reimplemented. Update ASN1 to avoid use of legacy functions.
3836 *) Win32/64 targets are linked with Winsock2.
3839 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3840 to external functions. This can be used to increase CRL handling
3841 efficiency especially when CRLs are very large by (for example) storing
3842 the CRL revoked certificates in a database.
3845 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3846 new CRLs added to a directory can be used. New command line option
3847 -verify_return_error to s_client and s_server. This causes real errors
3848 to be returned by the verify callback instead of carrying on no matter
3849 what. This reflects the way a "real world" verify callback would behave.
3852 *) GOST engine, supporting several GOST algorithms and public key formats.
3853 Kindly donated by Cryptocom.
3856 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3857 partitioned by DP are handled but no indirect CRL or reason partitioning
3858 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3859 selected via a scoring technique which handles IDP and AKID in CRLs.
3862 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3863 will ultimately be used for all verify operations: this will remove the
3864 X509_STORE dependency on certificate verification and allow alternative
3865 lookup methods. X509_STORE based implementations of these two callbacks.
3868 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3869 Modify get_crl() to find a valid (unexpired) CRL if possible.
3872 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3873 this would be called X509_CRL_cmp() but that name is already used by
3874 a function that just compares CRL issuer names. Cache several CRL
3875 extensions in X509_CRL structure and cache CRLDP in X509.
3878 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3879 this maps equivalent X509_NAME structures into a consistent structure.
3880 Name comparison can then be performed rapidly using memcmp().
3883 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3887 *) Allow digests to supply their own micalg string for S/MIME type using
3888 the ctrl EVP_MD_CTRL_MICALG.
3891 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3892 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3893 ctrl. It can then customise the structure before and/or after signing
3897 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3898 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3899 to free up any added signature OIDs.
3902 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3903 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3904 digest and cipher tables. New options added to openssl utility:
3905 list-message-digest-algorithms and list-cipher-algorithms.
3908 *) Change the array representation of binary polynomials: the list
3909 of degrees of non-zero coefficients is now terminated with -1.
3910 Previously it was terminated with 0, which was also part of the
3911 value; thus, the array representation was not applicable to
3912 polynomials where t^0 has coefficient zero. This change makes
3913 the array representation useful in a more general context.
3916 *) Various modifications and fixes to SSL/TLS cipher string
3917 handling. For ECC, the code now distinguishes between fixed ECDH
3918 with RSA certificates on the one hand and with ECDSA certificates
3919 on the other hand, since these are separate ciphersuites. The
3920 unused code for Fortezza ciphersuites has been removed.
3922 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3923 (not "ECDHE"). For consistency with the code for DH
3924 certificates, use of ECDH certificates is now considered ECDH
3925 authentication, not RSA or ECDSA authentication (the latter is
3926 merely the CA's signing algorithm and not actively used in the
3929 The temporary ciphersuite alias "ECCdraft" is no longer
3930 available, and ECC ciphersuites are no longer excluded from "ALL"
3931 and "DEFAULT". The following aliases now exist for RFC 4492
3932 ciphersuites, most of these by analogy with the DH case:
3934 kECDHr - ECDH cert, signed with RSA
3935 kECDHe - ECDH cert, signed with ECDSA
3936 kECDH - ECDH cert (signed with either RSA or ECDSA)
3937 kEECDH - ephemeral ECDH
3938 ECDH - ECDH cert or ephemeral ECDH
3944 AECDH - anonymous ECDH
3945 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3949 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3950 Use correct micalg parameters depending on digest(s) in signed message.
3953 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3954 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3957 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3958 an engine to register a method. Add ENGINE lookups for methods and
3959 functional reference processing.
3962 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
3963 EVP_{Sign,Verify}* which allow an application to customise the signature
3967 *) New -resign option to smime utility. This adds one or more signers
3968 to an existing PKCS#7 signedData structure. Also -md option to use an
3969 alternative message digest algorithm for signing.
3972 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3973 create PKCS7 structures containing multiple signers. Update smime
3974 application to support multiple signers.
3977 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3981 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3982 Reorganize PBE internals to lookup from a static table using NIDs,
3983 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3984 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3985 PRF which will be automatically used with PBES2.
3988 *) Replace the algorithm specific calls to generate keys in "req" with the
3992 *) Update PKCS#7 enveloped data routines to use new API. This is now
3993 supported by any public key method supporting the encrypt operation. A
3994 ctrl is added to allow the public key algorithm to examine or modify
3995 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3999 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4000 a default digest type to use. In most cases this will be SHA1 but some
4001 algorithms (such as GOST) need to specify an alternative digest. The
4002 return value indicates how strong the preference is 1 means optional and
4003 2 is mandatory (that is it is the only supported type). Modify
4004 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4005 use the default md. Update openssl utilities to use the default digest
4006 type for signing if it is not explicitly indicated.
4009 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4010 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4011 signing method from the key type. This effectively removes the link
4012 between digests and public key types.
4015 *) Add an OID cross reference table and utility functions. Its purpose is to
4016 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4017 rsaEncryption. This will allow some of the algorithm specific hackery
4018 needed to use the correct OID to be removed.
4021 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4022 structures for PKCS7_sign(). They are now set up by the relevant public
4026 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4029 *) Add support for key derivation (agreement) in the API, DH method and
4033 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4034 public and private key formats. As a side effect these add additional
4035 command line functionality not previously available: DSA signatures can be
4036 generated and verified using pkeyutl and DH key support and generation in
4041 [Oliver Tappe <zooey@hirschkaefer.de>]
4043 *) New make target "install_html_docs" installs HTML renditions of the
4045 [Oliver Tappe <zooey@hirschkaefer.de>]
4047 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4048 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4049 support key and parameter generation and add initial key generation
4050 functionality for RSA.
4053 *) Add functions for main EVP_PKEY_method operations. The undocumented
4054 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4055 EVP_PKEY_{encrypt,decrypt}_old.
4058 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4059 key API, doesn't do much yet.
4062 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4063 public key algorithms. New option to openssl utility:
4064 "list-public-key-algorithms" to print out info.
4067 *) Implement the Supported Elliptic Curves Extension for
4068 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4071 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4072 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4075 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4076 utilities such as rsa, dsa, dsaparam etc except they process any key
4080 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4081 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4082 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4086 *) Initial support for pluggable public key ASN1.
4087 De-spaghettify the public key ASN1 handling. Move public and private
4088 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4089 algorithm specific handling to a single module within the relevant
4090 algorithm directory. Add functions to allow (near) opaque processing
4091 of public and private key structures.
4094 *) Implement the Supported Point Formats Extension for
4095 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4098 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4099 for the psk identity [hint] and the psk callback functions to the
4100 SSL_SESSION, SSL and SSL_CTX structure.
4103 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4107 SSL_CTX_use_psk_identity_hint
4108 SSL_get_psk_identity_hint
4109 SSL_get_psk_identity
4110 SSL_use_psk_identity_hint
4112 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4114 *) Add RFC 3161 compliant time stamp request creation, response generation
4115 and response verification functionality.
4116 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4118 *) Add initial support for TLS extensions, specifically for the server_name
4119 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4120 have new members for a host name. The SSL data structure has an
4121 additional member SSL_CTX *initial_ctx so that new sessions can be
4122 stored in that context to allow for session resumption, even after the
4123 SSL has been switched to a new SSL_CTX in reaction to a client's
4124 server_name extension.
4126 New functions (subject to change):
4128 SSL_get_servername()
4129 SSL_get_servername_type()
4132 New CTRL codes and macros (subject to change):
4134 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4135 - SSL_CTX_set_tlsext_servername_callback()
4136 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4137 - SSL_CTX_set_tlsext_servername_arg()
4138 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4140 openssl s_client has a new '-servername ...' option.
4142 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4143 '-key2 ...', '-servername_fatal' (subject to change). This allows
4144 testing the HostName extension for a specific single host name ('-cert'
4145 and '-key' remain fallbacks for handshakes without HostName
4146 negotiation). If the unrecognized_name alert has to be sent, this by
4147 default is a warning; it becomes fatal with the '-servername_fatal'
4150 [Peter Sylvester, Remy Allais, Christophe Renou]
4152 *) Whirlpool hash implementation is added.
4155 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4156 bn(64,32). Because of instruction set limitations it doesn't have
4157 any negative impact on performance. This was done mostly in order
4158 to make it possible to share assembler modules, such as bn_mul_mont
4159 implementations, between 32- and 64-bit builds without hassle.
4162 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4163 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4167 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4168 dedicated Montgomery multiplication procedure, is introduced.
4169 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4170 "64-bit" performance on certain 32-bit targets.
4173 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4174 in SSL structures. New SSL ctrl to set maximum send fragment size.
4175 Save memory by setting the I/O buffer sizes dynamically instead of
4176 using the maximum available value.
4179 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4180 in addition to the text details.
4183 *) Very, very preliminary EXPERIMENTAL support for printing of general
4184 ASN1 structures. This currently produces rather ugly output and doesn't
4185 handle several customised structures at all.
4188 *) Integrated support for PVK file format and some related formats such
4189 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4190 these in the 'rsa' and 'dsa' utilities.
4193 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4196 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4197 place for the (very old) "NETSCAPE" format certificates which are now
4198 handled using new ASN1 code equivalents.
4201 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4202 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4203 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4206 *) Modify CRL distribution points extension code to print out previously
4207 unsupported fields. Enhance extension setting code to allow setting of
4211 *) Add print and set support for Issuing Distribution Point CRL extension.
4214 *) Change 'Configure' script to enable Camellia by default.
4217 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4219 *) When rejecting SSL/TLS records due to an incorrect version number, never
4220 update s->server with a new major version number. As of
4221 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4222 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4223 the previous behavior could result in a read attempt at NULL when
4224 receiving specific incorrect SSL/TLS records once record payload
4225 protection is active. (CVE-2010-0740)
4226 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4228 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4229 could be crashed if the relevant tables were not present (e.g. chrooted).
4230 [Tomas Hoger <thoger@redhat.com>]
4232 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4234 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4235 [Martin Olsson, Neel Mehta]
4237 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4238 accommodate for stack sorting, always a write lock!).
4241 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4242 excessive delays in the RAND_poll(): over a minute. As a workaround
4243 include a time check in the inner Heap32Next loop too.
4246 *) The code that handled flushing of data in SSL/TLS originally used the
4247 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4248 the problem outlined in PR#1949. The fix suggested there however can
4249 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4250 of Apache). So instead simplify the code to flush unconditionally.
4251 This should be fine since flushing with no data to flush is a no op.
4254 *) Handle TLS versions 2.0 and later properly and correctly use the
4255 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4256 off ancient servers have a habit of sticking around for a while...
4259 *) Modify compression code so it frees up structures without using the
4260 ex_data callbacks. This works around a problem where some applications
4261 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4262 restarting) then use compression (e.g. SSL with compression) later.
4263 This results in significant per-connection memory leaks and
4264 has caused some security issues including CVE-2008-1678 and
4268 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4269 change when encrypting or decrypting.
4272 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4273 connect and renegotiate with servers which do not support RI.
4274 Until RI is more widely deployed this option is enabled by default.
4277 *) Add "missing" ssl ctrls to clear options and mode.
4280 *) If client attempts to renegotiate and doesn't support RI respond with
4281 a no_renegotiation alert as required by RFC5746. Some renegotiating
4282 TLS clients will continue a connection gracefully when they receive
4283 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4284 waiting for a server hello which it will never receive. Now we treat a
4285 received no_renegotiation alert as a fatal error. This is because
4286 applications requesting a renegotiation might well expect it to succeed
4287 and would have no code in place to handle the server denying it so the
4288 only safe thing to do is to terminate the connection.
4291 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4292 peer supports secure renegotiation and 0 otherwise. Print out peer
4293 renegotiation support in s_client/s_server.
4296 *) Replace the highly broken and deprecated SPKAC certification method with
4297 the updated NID creation version. This should correctly handle UTF8.
4300 *) Implement RFC5746. Re-enable renegotiation but require the extension
4301 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4302 turns out to be a bad idea. It has been replaced by
4303 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4304 SSL_CTX_set_options(). This is really not recommended unless you
4305 know what you are doing.
4306 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4308 *) Fixes to stateless session resumption handling. Use initial_ctx when
4309 issuing and attempting to decrypt tickets in case it has changed during
4310 servername handling. Use a non-zero length session ID when attempting
4311 stateless session resumption: this makes it possible to determine if
4312 a resumption has occurred immediately after receiving server hello
4313 (several places in OpenSSL subtly assume this) instead of later in
4317 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4318 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4319 fixes for a few places where the return code is not checked
4321 [Julia Lawall <julia@diku.dk>]
4323 *) Add --strict-warnings option to Configure script to include devteam
4324 warnings in other configurations.
4327 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4328 makes it possible to install openssl libraries in locations which
4329 have names other than "lib", for example "/usr/lib64" which some
4331 [Steve Henson, based on patch from Jeremy Utley]
4333 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4334 X690 8.9.12 and can produce some misleading textual output of OIDs.
4335 [Steve Henson, reported by Dan Kaminsky]
4337 *) Delete MD2 from algorithm tables. This follows the recommendation in
4338 several standards that it is not used in new applications due to
4339 several cryptographic weaknesses. For binary compatibility reasons
4340 the MD2 API is still compiled in by default.
4343 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4347 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4348 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4350 [Guenter <lists@gknw.net>]
4352 *) Fix the server certificate chain building code to use X509_verify_cert(),
4353 it used to have an ad-hoc builder which was unable to cope with anything
4354 other than a simple chain.
4355 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4357 *) Don't check self signed certificate signatures in X509_verify_cert()
4358 by default (a flag can override this): it just wastes time without
4359 adding any security. As a useful side effect self signed root CAs
4360 with non-FIPS digests are now usable in FIPS mode.
4363 *) In dtls1_process_out_of_seq_message() the check if the current message
4364 is already buffered was missing. For every new message was memory
4365 allocated, allowing an attacker to perform an denial of service attack
4366 with sending out of seq handshake messages until there is no memory
4367 left. Additionally every future message was buffered, even if the
4368 sequence number made no sense and would be part of another handshake.
4369 So only messages with sequence numbers less than 10 in advance will be
4370 buffered. (CVE-2009-1378)
4371 [Robin Seggelmann, discovered by Daniel Mentz]
4373 *) Records are buffered if they arrive with a future epoch to be
4374 processed after finishing the corresponding handshake. There is
4375 currently no limitation to this buffer allowing an attacker to perform
4376 a DOS attack with sending records with future epochs until there is no
4377 memory left. This patch adds the pqueue_size() function to determine
4378 the size of a buffer and limits the record buffer to 100 entries.
4380 [Robin Seggelmann, discovered by Daniel Mentz]
4382 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4383 parent structure is freed. (CVE-2009-1379)
4386 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4387 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4390 [Ilya O. <vrghost@gmail.com>]
4392 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4394 *) Disable renegotiation completely - this fixes a severe security
4395 problem (CVE-2009-3555) at the cost of breaking all
4396 renegotiation. Renegotiation can be re-enabled by setting
4397 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4398 run-time. This is really not recommended unless you know what
4402 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4404 *) Don't set val to NULL when freeing up structures, it is freed up by
4405 underlying code. If sizeof(void *) > sizeof(long) this can result in
4406 zeroing past the valid field. (CVE-2009-0789)
4407 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4409 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4410 checked correctly. This would allow some invalid signed attributes to
4411 appear to verify correctly. (CVE-2009-0591)
4412 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4414 *) Reject UniversalString and BMPString types with invalid lengths. This
4415 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4416 a legal length. (CVE-2009-0590)
4419 *) Set S/MIME signing as the default purpose rather than setting it
4420 unconditionally. This allows applications to override it at the store
4424 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4425 to handle some structures.
4428 *) Improve efficiency of mem_gets: don't search whole buffer each time
4430 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4432 *) New -hex option for openssl rand.
4435 *) Print out UTF8String and NumericString when parsing ASN1.
4438 *) Support NumericString type for name components.
4441 *) Allow CC in the environment to override the automatically chosen
4442 compiler. Note that nothing is done to ensure flags work with the
4446 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4448 *) Properly check EVP_VerifyFinal() and similar return values
4450 [Ben Laurie, Bodo Moeller, Google Security Team]
4452 *) Enable TLS extensions by default.
4455 *) Allow the CHIL engine to be loaded, whether the application is
4456 multithreaded or not. (This does not release the developer from the
4457 obligation to set up the dynamic locking callbacks.)
4458 [Sander Temme <sander@temme.net>]
4460 *) Use correct exit code if there is an error in dgst command.
4461 [Steve Henson; problem pointed out by Roland Dirlewanger]
4463 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4464 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4467 *) Add experimental JPAKE support, including demo authentication in
4468 s_client and s_server.
4471 *) Set the comparison function in v3_addr_canonize().
4472 [Rob Austein <sra@hactrn.net>]
4474 *) Add support for XMPP STARTTLS in s_client.
4475 [Philip Paeps <philip@freebsd.org>]
4477 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4478 to ensure that even with this option, only ciphersuites in the
4479 server's preference list will be accepted. (Note that the option
4480 applies only when resuming a session, so the earlier behavior was
4481 just about the algorithm choice for symmetric cryptography.)
4484 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4486 *) Fix NULL pointer dereference if a DTLS server received
4487 ChangeCipherSpec as first record (CVE-2009-1386).
4490 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4491 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4494 *) The fix in 0.9.8c that supposedly got rid of unsafe
4495 double-checked locking was incomplete for RSA blinding,
4496 addressing just one layer of what turns out to have been
4497 doubly unsafe triple-checked locking.
4499 So now fix this for real by retiring the MONT_HELPER macro
4500 in crypto/rsa/rsa_eay.c.
4502 [Bodo Moeller; problem pointed out by Marius Schilder]
4504 *) Various precautionary measures:
4506 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4508 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4509 (NB: This would require knowledge of the secret session ticket key
4510 to exploit, in which case you'd be SOL either way.)
4512 - Change bn_nist.c so that it will properly handle input BIGNUMs
4513 outside the expected range.
4515 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4518 [Neel Mehta, Bodo Moeller]
4520 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4521 the load fails. Useful for distros.
4522 [Ben Laurie and the FreeBSD team]
4524 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4527 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4530 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4532 This work was sponsored by Logica.
4535 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4536 keystores. Support for SSL/TLS client authentication too.
4537 Not compiled unless enable-capieng specified to Configure.
4539 This work was sponsored by Logica.
4542 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4543 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4544 attribute creation routines such as certificate requests and PKCS#12
4548 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4550 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4551 handshake which could lead to a client crash as found using the
4552 Codenomicon TLS test suite (CVE-2008-1672)
4553 [Steve Henson, Mark Cox]
4555 *) Fix double free in TLS server name extensions which could lead to
4556 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4559 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4561 Clear the error queue to ensure that error entries left from
4562 older function calls do not interfere with the correct operation.
4563 [Lutz Jaenicke, Erik de Castro Lopo]
4565 *) Remove root CA certificates of commercial CAs:
4567 The OpenSSL project does not recommend any specific CA and does not
4568 have any policy with respect to including or excluding any CA.
4569 Therefore it does not make any sense to ship an arbitrary selection
4570 of root CA certificates with the OpenSSL software.
4573 *) RSA OAEP patches to fix two separate invalid memory reads.
4574 The first one involves inputs when 'lzero' is greater than
4575 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4576 before the beginning of from). The second one involves inputs where
4577 the 'db' section contains nothing but zeroes (there is a one-byte
4578 invalid read after the end of 'db').
4579 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4581 *) Partial backport from 0.9.9-dev:
4583 Introduce bn_mul_mont (dedicated Montgomery multiplication
4584 procedure) as a candidate for BIGNUM assembler implementation.
4585 While 0.9.9-dev uses assembler for various architectures, only
4586 x86_64 is available by default here in the 0.9.8 branch, and
4587 32-bit x86 is available through a compile-time setting.
4589 To try the 32-bit x86 assembler implementation, use Configure
4590 option "enable-montasm" (which exists only for this backport).
4592 As "enable-montasm" for 32-bit x86 disclaims code stability
4593 anyway, in this constellation we activate additional code
4594 backported from 0.9.9-dev for further performance improvements,
4595 namely BN_from_montgomery_word. (To enable this otherwise,
4596 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4598 [Andy Polyakov (backport partially by Bodo Moeller)]
4600 *) Add TLS session ticket callback. This allows an application to set
4601 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4602 values. This is useful for key rollover for example where several key
4603 sets may exist with different names.
4606 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4607 This was broken until now in 0.9.8 releases, such that the only way
4608 a registered ENGINE could be used (assuming it initialises
4609 successfully on the host) was to explicitly set it as the default
4610 for the relevant algorithms. This is in contradiction with 0.9.7
4611 behaviour and the documentation. With this fix, when an ENGINE is
4612 registered into a given algorithm's table of implementations, the
4613 'uptodate' flag is reset so that auto-discovery will be used next
4614 time a new context for that algorithm attempts to select an
4616 [Ian Lister (tweaked by Geoff Thorpe)]
4618 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4619 implementation in the following ways:
4621 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4624 Lack of BER streaming support means one pass streaming processing is
4625 only supported if data is detached: setting the streaming flag is
4626 ignored for embedded content.
4628 CMS support is disabled by default and must be explicitly enabled
4629 with the enable-cms configuration option.
4632 *) Update the GMP engine glue to do direct copies between BIGNUM and
4633 mpz_t when openssl and GMP use the same limb size. Otherwise the
4634 existing "conversion via a text string export" trick is still used.
4635 [Paul Sheer <paulsheer@gmail.com>]
4637 *) Zlib compression BIO. This is a filter BIO which compressed and
4638 uncompresses any data passed through it.
4641 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4642 RFC3394 compatible AES key wrapping.
4645 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4646 sets string data without copying. X509_ALGOR_set0() and
4647 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4648 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4649 from an X509_ATTRIBUTE structure optionally checking it occurs only
4650 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4654 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4655 to get the expected BN_FLG_CONSTTIME behavior.
4656 [Bodo Moeller (Google)]
4660 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4661 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4662 - added some more tests to do_tests.pl
4663 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4664 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4665 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4666 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4667 - various changes to netware.pl to enable gcc-cross builds on Win32
4669 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4670 - various changes to fix missing prototype warnings
4671 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4672 - added AES, WHIRLPOOL and CPUID assembler code to build files
4673 - added missing AES assembler make rules to mk1mf.pl
4674 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4675 [Guenter Knauf <eflash@gmx.net>]
4677 *) Implement certificate status request TLS extension defined in RFC3546.
4678 A client can set the appropriate parameters and receive the encoded
4679 OCSP response via a callback. A server can query the supplied parameters
4680 and set the encoded OCSP response in the callback. Add simplified examples
4681 to s_client and s_server.
4684 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4686 *) Fix various bugs:
4687 + Binary incompatibility of ssl_ctx_st structure
4688 + DTLS interoperation with non-compliant servers
4689 + Don't call get_session_cb() without proposed session
4690 + Fix ia64 assembler code
4691 [Andy Polyakov, Steve Henson]
4693 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4695 *) DTLS Handshake overhaul. There were longstanding issues with
4696 OpenSSL DTLS implementation, which were making it impossible for
4697 RFC 4347 compliant client to communicate with OpenSSL server.
4698 Unfortunately just fixing these incompatibilities would "cut off"
4699 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4700 server keeps tolerating non RFC compliant syntax. The opposite is
4701 not true, 0.9.8f client can not communicate with earlier server.
4702 This update even addresses CVE-2007-4995.
4705 *) Changes to avoid need for function casts in OpenSSL: some compilers
4706 (gcc 4.2 and later) reject their use.
4707 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4710 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4711 RFC4507bis. The encrypted ticket format is an encrypted encoded
4712 SSL_SESSION structure, that way new session features are automatically
4715 If a client application caches session in an SSL_SESSION structure
4716 support is transparent because tickets are now stored in the encoded
4719 The SSL_CTX structure automatically generates keys for ticket
4720 protection in servers so again support should be possible
4721 with no application modification.
4723 If a client or server wishes to disable RFC4507 support then the option
4724 SSL_OP_NO_TICKET can be set.
4726 Add a TLS extension debugging callback to allow the contents of any client
4727 or server extensions to be examined.
4729 This work was sponsored by Google.
4732 *) Add initial support for TLS extensions, specifically for the server_name
4733 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4734 have new members for a host name. The SSL data structure has an
4735 additional member SSL_CTX *initial_ctx so that new sessions can be
4736 stored in that context to allow for session resumption, even after the
4737 SSL has been switched to a new SSL_CTX in reaction to a client's
4738 server_name extension.
4740 New functions (subject to change):
4742 SSL_get_servername()
4743 SSL_get_servername_type()
4746 New CTRL codes and macros (subject to change):
4748 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4749 - SSL_CTX_set_tlsext_servername_callback()
4750 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4751 - SSL_CTX_set_tlsext_servername_arg()
4752 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4754 openssl s_client has a new '-servername ...' option.
4756 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4757 '-key2 ...', '-servername_fatal' (subject to change). This allows
4758 testing the HostName extension for a specific single host name ('-cert'
4759 and '-key' remain fallbacks for handshakes without HostName
4760 negotiation). If the unrecognized_name alert has to be sent, this by
4761 default is a warning; it becomes fatal with the '-servername_fatal'
4764 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4766 *) Add AES and SSE2 assembly language support to VC++ build.
4769 *) Mitigate attack on final subtraction in Montgomery reduction.
4772 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4773 (which previously caused an internal error).
4776 *) Squeeze another 10% out of IGE mode when in != out.
4779 *) AES IGE mode speedup.
4780 [Dean Gaudet (Google)]
4782 *) Add the Korean symmetric 128-bit cipher SEED (see
4783 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4784 add SEED ciphersuites from RFC 4162:
4786 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4787 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4788 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4789 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4791 To minimize changes between patchlevels in the OpenSSL 0.9.8
4792 series, SEED remains excluded from compilation unless OpenSSL
4793 is configured with 'enable-seed'.
4794 [KISA, Bodo Moeller]
4796 *) Mitigate branch prediction attacks, which can be practical if a
4797 single processor is shared, allowing a spy process to extract
4798 information. For detailed background information, see
4799 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4800 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4801 and Necessary Software Countermeasures"). The core of the change
4802 are new versions BN_div_no_branch() and
4803 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4804 respectively, which are slower, but avoid the security-relevant
4805 conditional branches. These are automatically called by BN_div()
4806 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4807 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4808 remove a conditional branch.
4810 BN_FLG_CONSTTIME is the new name for the previous
4811 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4812 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4813 in the exponent causes BN_mod_exp_mont() to use the alternative
4814 implementation in BN_mod_exp_mont_consttime().) The old name
4815 remains as a deprecated alias.
4817 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4818 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4819 constant-time implementations for more than just exponentiation.
4820 Here too the old name is kept as a deprecated alias.
4822 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4823 the BN_BLINDING structure gets an independent copy of the
4824 modulus. This means that the previous "BIGNUM *m" argument to
4825 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4826 essentially becomes "const BIGNUM *m", although we can't actually
4827 change this in the header file before 0.9.9. It allows
4828 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4829 enable BN_FLG_CONSTTIME.
4831 [Matthew D Wood (Intel Corp)]
4833 *) In the SSL/TLS server implementation, be strict about session ID
4834 context matching (which matters if an application uses a single
4835 external cache for different purposes). Previously,
4836 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4837 set. This did ensure strict client verification, but meant that,
4838 with applications using a single external cache for quite
4839 different requirements, clients could circumvent ciphersuite
4840 restrictions for a given session ID context by starting a session
4841 in a different context.
4844 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4845 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4846 authentication-only ciphersuites.
4849 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4850 not complete and could lead to a possible single byte overflow
4851 (CVE-2007-5135) [Ben Laurie]
4853 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4855 *) Since AES128 and AES256 (and similarly Camellia128 and
4856 Camellia256) share a single mask bit in the logic of
4857 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4858 kludge to work properly if AES128 is available and AES256 isn't
4859 (or if Camellia128 is available and Camellia256 isn't).
4862 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4863 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4864 When a point or a seed is encoded in a BIT STRING, we need to
4865 prevent the removal of trailing zero bits to get the proper DER
4866 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4867 of a NamedBitList, for which trailing 0 bits need to be removed.)
4870 *) Have SSL/TLS server implementation tolerate "mismatched" record
4871 protocol version while receiving ClientHello even if the
4872 ClientHello is fragmented. (The server can't insist on the
4873 particular protocol version it has chosen before the ServerHello
4874 message has informed the client about his choice.)
4877 *) Add RFC 3779 support.
4878 [Rob Austein for ARIN, Ben Laurie]
4880 *) Load error codes if they are not already present instead of using a
4881 static variable. This allows them to be cleanly unloaded and reloaded.
4882 Improve header file function name parsing.
4885 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4886 or CAPABILITY handshake as required by RFCs.
4889 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4891 *) Introduce limits to prevent malicious keys being able to
4892 cause a denial of service. (CVE-2006-2940)
4893 [Steve Henson, Bodo Moeller]
4895 *) Fix ASN.1 parsing of certain invalid structures that can result
4896 in a denial of service. (CVE-2006-2937) [Steve Henson]
4898 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4899 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4901 *) Fix SSL client code which could crash if connecting to a
4902 malicious SSLv2 server. (CVE-2006-4343)
4903 [Tavis Ormandy and Will Drewry, Google Security Team]
4905 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4906 match only those. Before that, "AES256-SHA" would be interpreted
4907 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4908 the same strength classification in 0.9.7h) as we currently only
4909 have a single AES bit in the ciphersuite description bitmap.
4910 That change, however, also applied to ciphersuite strings such as
4911 "RC4-MD5" that intentionally matched multiple ciphersuites --
4912 namely, SSL 2.0 ciphersuites in addition to the more common ones
4913 from SSL 3.0/TLS 1.0.
4915 So we change the selection algorithm again: Naming an explicit
4916 ciphersuite selects this one ciphersuite, and any other similar
4917 ciphersuite (same bitmap) from *other* protocol versions.
4918 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4919 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4921 Since SSL 2.0 does not have any ciphersuites for which the
4922 128/256 bit distinction would be relevant, this works for now.
4923 The proper fix will be to use different bits for AES128 and
4924 AES256, which would have avoided the problems from the beginning;
4925 however, bits are scarce, so we can only do this in a new release
4926 (not just a patchlevel) when we can change the SSL_CIPHER
4927 definition to split the single 'unsigned long mask' bitmap into
4928 multiple values to extend the available space.
4932 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4934 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4935 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4937 *) Add AES IGE and biIGE modes.
4940 *) Change the Unix randomness entropy gathering to use poll() when
4941 possible instead of select(), since the latter has some
4942 undesirable limitations.
4943 [Darryl Miles via Richard Levitte and Bodo Moeller]
4945 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4946 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4947 cannot be implicitly activated as part of, e.g., the "AES" alias.
4948 However, please upgrade to OpenSSL 0.9.9[-dev] for
4949 non-experimental use of the ECC ciphersuites to get TLS extension
4950 support, which is required for curve and point format negotiation
4951 to avoid potential handshake problems.
4954 *) Disable rogue ciphersuites:
4956 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4957 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4958 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4960 The latter two were purportedly from
4961 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4964 Also deactivate the remaining ciphersuites from
4965 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4966 unofficial, and the ID has long expired.
4969 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4970 dual-core machines) and other potential thread-safety issues.
4973 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4974 versions), which is now available for royalty-free use
4975 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4976 Also, add Camellia TLS ciphersuites from RFC 4132.
4978 To minimize changes between patchlevels in the OpenSSL 0.9.8
4979 series, Camellia remains excluded from compilation unless OpenSSL
4980 is configured with 'enable-camellia'.
4983 *) Disable the padding bug check when compression is in use. The padding
4984 bug check assumes the first packet is of even length, this is not
4985 necessarily true if compression is enabled and can result in false
4986 positives causing handshake failure. The actual bug test is ancient
4987 code so it is hoped that implementations will either have fixed it by
4988 now or any which still have the bug do not support compression.
4991 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4993 *) When applying a cipher rule check to see if string match is an explicit
4994 cipher suite and only match that one cipher suite if it is.
4997 *) Link in manifests for VC++ if needed.
4998 [Austin Ziegler <halostatue@gmail.com>]
5000 *) Update support for ECC-based TLS ciphersuites according to
5001 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5002 TLS extensions, which are supported starting with the 0.9.9
5003 branch, not in the OpenSSL 0.9.8 branch).
5006 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5007 opaque EVP_CIPHER_CTX handling.
5010 *) Fixes and enhancements to zlib compression code. We now only use
5011 "zlib1.dll" and use the default __cdecl calling convention on Win32
5012 to conform with the standards mentioned here:
5013 http://www.zlib.net/DLL_FAQ.txt
5014 Static zlib linking now works on Windows and the new --with-zlib-include
5015 --with-zlib-lib options to Configure can be used to supply the location
5016 of the headers and library. Gracefully handle case where zlib library
5020 *) Several fixes and enhancements to the OID generation code. The old code
5021 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5022 handle numbers larger than ULONG_MAX, truncated printing and had a
5023 non standard OBJ_obj2txt() behaviour.
5026 *) Add support for building of engines under engine/ as shared libraries
5027 under VC++ build system.
5030 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5031 Hopefully, we will not see any false combination of paths any more.
5034 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5036 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5037 (part of SSL_OP_ALL). This option used to disable the
5038 countermeasure against man-in-the-middle protocol-version
5039 rollback in the SSL 2.0 server implementation, which is a bad
5040 idea. (CVE-2005-2969)
5042 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5043 for Information Security, National Institute of Advanced Industrial
5044 Science and Technology [AIST], Japan)]
5046 *) Add two function to clear and return the verify parameter flags.
5049 *) Keep cipherlists sorted in the source instead of sorting them at
5050 runtime, thus removing the need for a lock.
5053 *) Avoid some small subgroup attacks in Diffie-Hellman.
5054 [Nick Mathewson and Ben Laurie]
5056 *) Add functions for well-known primes.
5059 *) Extended Windows CE support.
5060 [Satoshi Nakamura and Andy Polyakov]
5062 *) Initialize SSL_METHOD structures at compile time instead of during
5063 runtime, thus removing the need for a lock.
5066 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5067 attempting to decrypt each encrypted key in turn. Add support to
5071 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5073 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5076 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5079 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5080 key into the same file any more.
5083 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5086 *) Add -utf8 command line and config file option to 'ca'.
5087 [Stefan <stf@udoma.org]
5089 *) Removed the macro des_crypt(), as it seems to conflict with some
5090 libraries. Use DES_crypt().
5093 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5094 involves renaming the source and generated shared-libs for
5095 both. The engines will accept the corrected or legacy ids
5096 ('ncipher' and '4758_cca' respectively) when binding. NB,
5097 this only applies when building 'shared'.
5098 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5100 *) Add attribute functions to EVP_PKEY structure. Modify
5101 PKCS12_create() to recognize a CSP name attribute and
5102 use it. Make -CSP option work again in pkcs12 utility.
5105 *) Add new functionality to the bn blinding code:
5106 - automatic re-creation of the BN_BLINDING parameters after
5107 a fixed number of uses (currently 32)
5108 - add new function for parameter creation
5109 - introduce flags to control the update behaviour of the
5110 BN_BLINDING parameters
5111 - hide BN_BLINDING structure
5112 Add a second BN_BLINDING slot to the RSA structure to improve
5113 performance when a single RSA object is shared among several
5117 *) Add support for DTLS.
5118 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5120 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5121 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5124 *) Remove buggy and incomplete DH cert support from
5125 ssl/ssl_rsa.c and ssl/s3_both.c
5128 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5129 the apps/openssl applications.
5132 *) Compile clean with "-Wall -Wmissing-prototypes
5133 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5134 DEBUG_SAFESTACK must also be set.
5137 *) Change ./Configure so that certain algorithms can be disabled by default.
5138 The new counterpiece to "no-xxx" is "enable-xxx".
5140 The patented RC5 and MDC2 algorithms will now be disabled unless
5141 "enable-rc5" and "enable-mdc2", respectively, are specified.
5143 (IDEA remains enabled despite being patented. This is because IDEA
5144 is frequently required for interoperability, and there is no license
5145 fee for non-commercial use. As before, "no-idea" can be used to
5146 avoid this algorithm.)
5150 *) Add processing of proxy certificates (see RFC 3820). This work was
5151 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5152 EGEE (Enabling Grids for E-science in Europe).
5155 *) RC4 performance overhaul on modern architectures/implementations, such
5156 as Intel P4, IA-64 and AMD64.
5159 *) New utility extract-section.pl. This can be used specify an alternative
5160 section number in a pod file instead of having to treat each file as
5161 a separate case in Makefile. This can be done by adding two lines to the
5164 =for comment openssl_section:XXX
5166 The blank line is mandatory.
5170 *) New arguments -certform, -keyform and -pass for s_client and s_server
5171 to allow alternative format key and certificate files and passphrase
5175 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5176 update associated structures and add various utility functions.
5178 Add new policy related verify parameters, include policy checking in
5179 standard verify code. Enhance 'smime' application with extra parameters
5180 to support policy checking and print out.
5183 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5184 Nehemiah processors. These extensions support AES encryption in hardware
5185 as well as RNG (though RNG support is currently disabled).
5186 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5188 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5191 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5192 [Andy Polyakov and a number of other people]
5194 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5195 implementation contributed by IBM.
5196 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5198 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5199 exponent rather than 'unsigned long'. There is a corresponding change to
5200 the new 'rsa_keygen' element of the RSA_METHOD structure.
5201 [Jelte Jansen, Geoff Thorpe]
5203 *) Functionality for creating the initial serial number file is now
5204 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5206 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5207 number file to 1, which is bound to cause problems. To avoid
5208 the problems while respecting compatibility between different 0.9.7
5209 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5210 CA.pl for serial number initialization. With the new release 0.9.8,
5211 we can fix the problem directly in the 'ca' utility.)
5214 *) Reduced header interdependencies by declaring more opaque objects in
5215 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5216 give fewer recursive includes, which could break lazy source code - so
5217 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5218 developers should define this symbol when building and using openssl to
5219 ensure they track the recommended behaviour, interfaces, [etc], but
5220 backwards-compatible behaviour prevails when this isn't defined.
5223 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5226 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5227 This will generate a random key of the appropriate length based on the
5228 cipher context. The EVP_CIPHER can provide its own random key generation
5229 routine to support keys of a specific form. This is used in the des and
5230 3des routines to generate a key of the correct parity. Update S/MIME
5231 code to use new functions and hence generate correct parity DES keys.
5232 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5233 valid (weak or incorrect parity).
5236 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5237 as looking them up. This is useful when the verified structure may contain
5238 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5239 present unless the new PKCS7_NO_CRL flag is asserted.
5242 *) Extend ASN1 oid configuration module. It now additionally accepts the
5245 shortName = some long name, 1.2.3.4
5248 *) Reimplemented the BN_CTX implementation. There is now no more static
5249 limitation on the number of variables it can handle nor the depth of the
5250 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5251 information can now expand as required, and rather than having a single
5252 static array of bignums, BN_CTX now uses a linked-list of such arrays
5253 allowing it to expand on demand whilst maintaining the usefulness of
5254 BN_CTX's "bundling".
5257 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5258 to allow all RSA operations to function using a single BN_CTX.
5261 *) Preliminary support for certificate policy evaluation and checking. This
5262 is initially intended to pass the tests outlined in "Conformance Testing
5263 of Relying Party Client Certificate Path Processing Logic" v1.07.
5266 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5267 remained unused and not that useful. A variety of other little bignum
5268 tweaks and fixes have also been made continuing on from the audit (see
5272 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5273 associated ASN1, EVP and SSL functions and old ASN1 macros.
5276 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5277 and this should never fail. So the return value from the use of
5278 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5279 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5282 *) BN_CTX_get() should return zero-valued bignums, providing the same
5283 initialised value as BN_new().
5284 [Geoff Thorpe, suggested by Ulf Möller]
5286 *) Support for inhibitAnyPolicy certificate extension.
5289 *) An audit of the BIGNUM code is underway, for which debugging code is
5290 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5291 is considered valid when processing BIGNUMs, and causes execution to
5292 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5293 further steps are taken to deliberately pollute unused data in BIGNUM
5294 structures to try and expose faulty code further on. For now, openssl will
5295 (in its default mode of operation) continue to tolerate the inconsistent
5296 forms that it has tolerated in the past, but authors and packagers should
5297 consider trying openssl and their own applications when compiled with
5298 these debugging symbols defined. It will help highlight potential bugs in
5299 their own code, and will improve the test coverage for OpenSSL itself. At
5300 some point, these tighter rules will become openssl's default to improve
5301 maintainability, though the assert()s and other overheads will remain only
5302 in debugging configurations. See bn.h for more details.
5303 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5305 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5306 that can only be obtained through BN_CTX_new() (which implicitly
5307 initialises it). The presence of this function only made it possible
5308 to overwrite an existing structure (and cause memory leaks).
5311 *) Because of the callback-based approach for implementing LHASH as a
5312 template type, lh_insert() adds opaque objects to hash-tables and
5313 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5314 to clean up those corresponding objects before destroying the hash table
5315 (and losing the object pointers). So some over-zealous constifications in
5316 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5317 objects as "const" and the lh_doall[_arg] callback wrappers are not
5318 prototyped to have "const" restrictions on the object pointers they are
5319 given (and so aren't required to cast them away any more).
5322 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5323 (speed) prefers to use its own implementation. The two implementations
5324 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5325 its object type properly exposed (MS_TM) instead of casting to/from "char
5326 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5327 aren't necessarily the greatest nomenclatures - but this is what was used
5328 internally to the implementation so I've used that for now.
5331 *) Ensure that deprecated functions do not get compiled when
5332 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5333 the self-tests were still using deprecated key-generation functions so
5334 these have been updated also.
5337 *) Reorganise PKCS#7 code to separate the digest location functionality
5338 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5339 New function PKCS7_set_digest() to set the digest type for PKCS#7
5340 digestedData type. Add additional code to correctly generate the
5341 digestedData type and add support for this type in PKCS7 initialization
5345 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5346 structure of type "other".
5349 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5350 sure the loop does correctly stop and breaking ("division by zero")
5351 modulus operations are not performed. The (pre-generated) prime
5352 table crypto/bn/bn_prime.h was already correct, but it could not be
5353 re-generated on some platforms because of the "division by zero"
5354 situation in the script.
5355 [Ralf S. Engelschall]
5357 *) Update support for ECC-based TLS ciphersuites according to
5358 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5359 SHA-1 now is only used for "small" curves (where the
5360 representation of a field element takes up to 24 bytes); for
5361 larger curves, the field element resulting from ECDH is directly
5362 used as premaster secret.
5363 [Douglas Stebila (Sun Microsystems Laboratories)]
5365 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5366 curve secp160r1 to the tests.
5367 [Douglas Stebila (Sun Microsystems Laboratories)]
5369 *) Add the possibility to load symbols globally with DSO.
5370 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5372 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5373 control of the error stack.
5376 *) Add support for STORE in ENGINE.
5379 *) Add the STORE type. The intention is to provide a common interface
5380 to certificate and key stores, be they simple file-based stores, or
5381 HSM-type store, or LDAP stores, or...
5382 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5385 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5386 pass a list of arguments to any function as well as provide a way
5387 for a function to pass data back to the caller.
5390 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5391 works like BUF_strdup() but can be used to duplicate a portion of
5392 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5396 *) Add the function sk_find_ex() which works like sk_find(), but will
5397 return an index to an element even if an exact match couldn't be
5398 found. The index is guaranteed to point at the element where the
5399 searched-for key would be inserted to preserve sorting order.
5402 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5403 takes an extra flags argument for optional functionality. Currently,
5404 the following flags are defined:
5406 OBJ_BSEARCH_VALUE_ON_NOMATCH
5407 This one gets OBJ_bsearch_ex() to return a pointer to the first
5408 element where the comparing function returns a negative or zero
5411 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5412 This one gets OBJ_bsearch_ex() to return a pointer to the first
5413 element where the comparing function returns zero. This is useful
5414 if there are more than one element where the comparing function
5418 *) Make it possible to create self-signed certificates with 'openssl ca'
5419 in such a way that the self-signed certificate becomes part of the
5420 CA database and uses the same mechanisms for serial number generation
5421 as all other certificate signing. The new flag '-selfsign' enables
5422 this functionality. Adapt CA.sh and CA.pl.in.
5425 *) Add functionality to check the public key of a certificate request
5426 against a given private. This is useful to check that a certificate
5427 request can be signed by that key (self-signing).
5430 *) Make it possible to have multiple active certificates with the same
5431 subject in the CA index file. This is done only if the keyword
5432 'unique_subject' is set to 'no' in the main CA section (default
5433 if 'CA_default') of the configuration file. The value is saved
5434 with the database itself in a separate index attribute file,
5435 named like the index file with '.attr' appended to the name.
5438 *) Generate multi-valued AVAs using '+' notation in config files for
5442 *) Support for nameConstraints certificate extension.
5445 *) Support for policyConstraints certificate extension.
5448 *) Support for policyMappings certificate extension.
5451 *) Make sure the default DSA_METHOD implementation only uses its
5452 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5453 and change its own handlers to be NULL so as to remove unnecessary
5454 indirection. This lets alternative implementations fallback to the
5455 default implementation more easily.
5458 *) Support for directoryName in GeneralName related extensions
5462 *) Make it possible to link applications using Makefile.shared.
5463 Make that possible even when linking against static libraries!
5466 *) Support for single pass processing for S/MIME signing. This now
5467 means that S/MIME signing can be done from a pipe, in addition
5468 cleartext signing (multipart/signed type) is effectively streaming
5469 and the signed data does not need to be all held in memory.
5471 This is done with a new flag PKCS7_STREAM. When this flag is set
5472 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5473 is done after the data is output (and digests calculated) in
5474 SMIME_write_PKCS7().
5477 *) Add full support for -rpath/-R, both in shared libraries and
5478 applications, at least on the platforms where it's known how
5482 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5483 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5484 will now compute a table of multiples of the generator that
5485 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5486 faster (notably in the case of a single point multiplication,
5487 scalar * generator).
5488 [Nils Larsch, Bodo Moeller]
5490 *) IPv6 support for certificate extensions. The various extensions
5491 which use the IP:a.b.c.d can now take IPv6 addresses using the
5492 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5496 *) Added an ENGINE that implements RSA by performing private key
5497 exponentiations with the GMP library. The conversions to and from
5498 GMP's mpz_t format aren't optimised nor are any montgomery forms
5499 cached, and on x86 it appears OpenSSL's own performance has caught up.
5500 However there are likely to be other architectures where GMP could
5501 provide a boost. This ENGINE is not built in by default, but it can be
5502 specified at Configure time and should be accompanied by the necessary
5503 linker additions, eg;
5504 ./config -DOPENSSL_USE_GMP -lgmp
5507 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5508 testing availability of engines with "-t" - the old behaviour is
5509 produced by increasing the feature's verbosity with "-tt".
5512 *) ECDSA routines: under certain error conditions uninitialized BN objects
5513 could be freed. Solution: make sure initialization is performed early
5514 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5518 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5519 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5520 software implementations. For DSA and DH, parameter generation can
5521 also be overridden by providing the appropriate method callbacks.
5524 *) Change the "progress" mechanism used in key-generation and
5525 primality testing to functions that take a new BN_GENCB pointer in
5526 place of callback/argument pairs. The new API functions have "_ex"
5527 postfixes and the older functions are reimplemented as wrappers for
5528 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5529 declarations of the old functions to help (graceful) attempts to
5530 migrate to the new functions. Also, the new key-generation API
5531 functions operate on a caller-supplied key-structure and return
5532 success/failure rather than returning a key or NULL - this is to
5533 help make "keygen" another member function of RSA_METHOD etc.
5535 Example for using the new callback interface:
5537 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5541 BN_GENCB_set(&my_cb, my_callback, my_arg);
5543 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5544 /* For the meaning of a, b in calls to my_callback(), see the
5545 * documentation of the function that calls the callback.
5546 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5547 * my_callback should return 1 if it wants BN_is_prime_ex()
5548 * to continue, or 0 to stop.
5553 *) Change the ZLIB compression method to be stateful, and make it
5554 available to TLS with the number defined in
5555 draft-ietf-tls-compression-04.txt.
5558 *) Add the ASN.1 structures and functions for CertificatePair, which
5559 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5561 CertificatePair ::= SEQUENCE {
5562 forward [0] Certificate OPTIONAL,
5563 reverse [1] Certificate OPTIONAL,
5564 -- at least one of the pair shall be present -- }
5566 Also implement the PEM functions to read and write certificate
5567 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5569 This needed to be defined, mostly for the sake of the LDAP
5570 attribute crossCertificatePair, but may prove useful elsewhere as
5574 *) Make it possible to inhibit symlinking of shared libraries in
5575 Makefile.shared, for Cygwin's sake.
5578 *) Extend the BIGNUM API by creating a function
5579 void BN_set_negative(BIGNUM *a, int neg);
5580 and a macro that behave like
5581 int BN_is_negative(const BIGNUM *a);
5583 to avoid the need to access 'a->neg' directly in applications.
5586 *) Implement fast modular reduction for pseudo-Mersenne primes
5587 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5588 EC_GROUP_new_curve_GFp() will now automatically use this
5590 [Nils Larsch <nla@trustcenter.de>]
5592 *) Add new lock type (CRYPTO_LOCK_BN).
5595 *) Change the ENGINE framework to automatically load engines
5596 dynamically from specific directories unless they could be
5597 found to already be built in or loaded. Move all the
5598 current engines except for the cryptodev one to a new
5600 The engines in engines/ are built as shared libraries if
5601 the "shared" options was given to ./Configure or ./config.
5602 Otherwise, they are inserted in libcrypto.a.
5603 /usr/local/ssl/engines is the default directory for dynamic
5604 engines, but that can be overridden at configure time through
5605 the usual use of --prefix and/or --openssldir, and at run
5606 time with the environment variable OPENSSL_ENGINES.
5607 [Geoff Thorpe and Richard Levitte]
5609 *) Add Makefile.shared, a helper makefile to build shared
5610 libraries. Adapt Makefile.org.
5613 *) Add version info to Win32 DLLs.
5614 [Peter 'Luna' Runestig" <peter@runestig.com>]
5616 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5617 can be added using this API to created arbitrary PKCS#12
5618 files while avoiding the low level API.
5620 New options to PKCS12_create(), key or cert can be NULL and
5621 will then be omitted from the output file. The encryption
5622 algorithm NIDs can be set to -1 for no encryption, the mac
5623 iteration count can be set to 0 to omit the mac.
5625 Enhance pkcs12 utility by making the -nokeys and -nocerts
5626 options work when creating a PKCS#12 file. New option -nomac
5627 to omit the mac, NONE can be set for an encryption algorithm.
5628 New code is modified to use the enhanced PKCS12_create()
5629 instead of the low level API.
5632 *) Extend ASN1 encoder to support indefinite length constructed
5633 encoding. This can output sequences tags and octet strings in
5634 this form. Modify pk7_asn1.c to support indefinite length
5635 encoding. This is experimental and needs additional code to
5636 be useful, such as an ASN1 bio and some enhanced streaming
5639 Extend template encode functionality so that tagging is passed
5640 down to the template encoder.
5643 *) Let 'openssl req' fail if an argument to '-newkey' is not
5644 recognized instead of using RSA as a default.
5647 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5648 As these are not official, they are not included in "ALL";
5649 the "ECCdraft" ciphersuite group alias can be used to select them.
5650 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5652 *) Add ECDH engine support.
5653 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5655 *) Add ECDH in new directory crypto/ecdh/.
5656 [Douglas Stebila (Sun Microsystems Laboratories)]
5658 *) Let BN_rand_range() abort with an error after 100 iterations
5659 without success (which indicates a broken PRNG).
5662 *) Change BN_mod_sqrt() so that it verifies that the input value
5663 is really the square of the return value. (Previously,
5664 BN_mod_sqrt would show GIGO behaviour.)
5667 *) Add named elliptic curves over binary fields from X9.62, SECG,
5668 and WAP/WTLS; add OIDs that were still missing.
5670 [Sheueling Chang Shantz and Douglas Stebila
5671 (Sun Microsystems Laboratories)]
5673 *) Extend the EC library for elliptic curves over binary fields
5674 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5677 EC_GF2m_simple_method
5681 EC_GROUP_new_curve_GF2m
5682 EC_GROUP_set_curve_GF2m
5683 EC_GROUP_get_curve_GF2m
5684 EC_POINT_set_affine_coordinates_GF2m
5685 EC_POINT_get_affine_coordinates_GF2m
5686 EC_POINT_set_compressed_coordinates_GF2m
5688 Point compression for binary fields is disabled by default for
5689 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5692 As binary polynomials are represented as BIGNUMs, various members
5693 of the EC_GROUP and EC_POINT data structures can be shared
5694 between the implementations for prime fields and binary fields;
5695 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5696 are essentially identical to their ..._GFp counterparts.
5697 (For simplicity, the '..._GFp' prefix has been dropped from
5698 various internal method names.)
5700 An internal 'field_div' method (similar to 'field_mul' and
5701 'field_sqr') has been added; this is used only for binary fields.
5703 [Sheueling Chang Shantz and Douglas Stebila
5704 (Sun Microsystems Laboratories)]
5706 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5707 through methods ('mul', 'precompute_mult').
5709 The generic implementations (now internally called 'ec_wNAF_mul'
5710 and 'ec_wNAF_precomputed_mult') remain the default if these
5711 methods are undefined.
5713 [Sheueling Chang Shantz and Douglas Stebila
5714 (Sun Microsystems Laboratories)]
5716 *) New function EC_GROUP_get_degree, which is defined through
5717 EC_METHOD. For curves over prime fields, this returns the bit
5718 length of the modulus.
5720 [Sheueling Chang Shantz and Douglas Stebila
5721 (Sun Microsystems Laboratories)]
5723 *) New functions EC_GROUP_dup, EC_POINT_dup.
5724 (These simply call ..._new and ..._copy).
5726 [Sheueling Chang Shantz and Douglas Stebila
5727 (Sun Microsystems Laboratories)]
5729 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5730 Polynomials are represented as BIGNUMs (where the sign bit is not
5731 used) in the following functions [macros]:
5734 BN_GF2m_sub [= BN_GF2m_add]
5735 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5736 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5737 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5739 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5740 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5741 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5742 BN_GF2m_cmp [= BN_ucmp]
5744 (Note that only the 'mod' functions are actually for fields GF(2^m).
5745 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5747 For some functions, an the irreducible polynomial defining a
5748 field can be given as an 'unsigned int[]' with strictly
5749 decreasing elements giving the indices of those bits that are set;
5750 i.e., p[] represents the polynomial
5751 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5753 p[0] > p[1] > ... > p[k] = 0.
5754 This applies to the following functions:
5759 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5760 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5762 BN_GF2m_mod_sqrt_arr
5763 BN_GF2m_mod_solve_quad_arr
5767 Conversion can be performed by the following functions:
5772 bntest.c has additional tests for binary polynomial arithmetic.
5774 Two implementations for BN_GF2m_mod_div() are available.
5775 The default algorithm simply uses BN_GF2m_mod_inv() and
5776 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5777 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5778 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5780 [Sheueling Chang Shantz and Douglas Stebila
5781 (Sun Microsystems Laboratories)]
5783 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5784 functionality is disabled at compile-time.
5785 [Douglas Stebila <douglas.stebila@sun.com>]
5787 *) Change default behaviour of 'openssl asn1parse' so that more
5788 information is visible when viewing, e.g., a certificate:
5790 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5791 mode the content of non-printable OCTET STRINGs is output in a
5792 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5793 avoid the appearance of a printable string.
5794 [Nils Larsch <nla@trustcenter.de>]
5796 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5798 EC_GROUP_set_asn1_flag()
5799 EC_GROUP_get_asn1_flag()
5800 EC_GROUP_set_point_conversion_form()
5801 EC_GROUP_get_point_conversion_form()
5802 These control ASN1 encoding details:
5803 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5804 has been set to OPENSSL_EC_NAMED_CURVE.
5805 - Points are encoded in uncompressed form by default; options for
5806 asn1_for are as for point2oct, namely
5807 POINT_CONVERSION_COMPRESSED
5808 POINT_CONVERSION_UNCOMPRESSED
5809 POINT_CONVERSION_HYBRID
5811 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5814 EC_GROUP_get0_seed()
5815 EC_GROUP_get_seed_len()
5816 This is used only for ASN1 purposes (so far).
5817 [Nils Larsch <nla@trustcenter.de>]
5819 *) Add 'field_type' member to EC_METHOD, which holds the NID
5820 of the appropriate field type OID. The new function
5821 EC_METHOD_get_field_type() returns this value.
5822 [Nils Larsch <nla@trustcenter.de>]
5827 EC_POINT_point2hex()
5828 EC_POINT_hex2point()
5829 providing useful interfaces to EC_POINT_point2oct() and
5830 EC_POINT_oct2point().
5831 [Nils Larsch <nla@trustcenter.de>]
5833 *) Change internals of the EC library so that the functions
5834 EC_GROUP_set_generator()
5835 EC_GROUP_get_generator()
5836 EC_GROUP_get_order()
5837 EC_GROUP_get_cofactor()
5838 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5839 to methods, which would lead to unnecessary code duplication when
5840 adding different types of curves.
5841 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5843 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5844 arithmetic, and such that modified wNAFs are generated
5845 (which avoid length expansion in many cases).
5848 *) Add a function EC_GROUP_check_discriminant() (defined via
5849 EC_METHOD) that verifies that the curve discriminant is non-zero.
5851 Add a function EC_GROUP_check() that makes some sanity tests
5852 on a EC_GROUP, its generator and order. This includes
5853 EC_GROUP_check_discriminant().
5854 [Nils Larsch <nla@trustcenter.de>]
5856 *) Add ECDSA in new directory crypto/ecdsa/.
5858 Add applications 'openssl ecparam' and 'openssl ecdsa'
5859 (these are based on 'openssl dsaparam' and 'openssl dsa').
5861 ECDSA support is also included in various other files across the
5862 library. Most notably,
5863 - 'openssl req' now has a '-newkey ecdsa:file' option;
5864 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5865 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5866 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5867 them suitable for ECDSA where domain parameters must be
5868 extracted before the specific public key;
5869 - ECDSA engine support has been added.
5870 [Nils Larsch <nla@trustcenter.de>]
5872 *) Include some named elliptic curves, and add OIDs from X9.62,
5873 SECG, and WAP/WTLS. Each curve can be obtained from the new
5875 EC_GROUP_new_by_curve_name(),
5876 and the list of available named curves can be obtained with
5877 EC_get_builtin_curves().
5878 Also add a 'curve_name' member to EC_GROUP objects, which can be
5880 EC_GROUP_set_curve_name()
5881 EC_GROUP_get_curve_name()
5882 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5884 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5885 was actually never needed) and in BN_mul(). The removal in BN_mul()
5886 required a small change in bn_mul_part_recursive() and the addition
5887 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5888 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5889 bn_sub_words() and bn_add_words() except they take arrays with
5893 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5895 *) Cleanse PEM buffers before freeing them since they may contain
5897 [Benjamin Bennett <ben@psc.edu>]
5899 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5900 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5901 authentication-only ciphersuites.
5904 *) Since AES128 and AES256 share a single mask bit in the logic of
5905 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5906 kludge to work properly if AES128 is available and AES256 isn't.
5909 *) Expand security boundary to match 1.1.1 module.
5912 *) Remove redundant features: hash file source, editing of test vectors
5913 modify fipsld to use external fips_premain.c signature.
5916 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5917 run algorithm test programs.
5920 *) Make algorithm test programs more tolerant of whitespace.
5923 *) Have SSL/TLS server implementation tolerate "mismatched" record
5924 protocol version while receiving ClientHello even if the
5925 ClientHello is fragmented. (The server can't insist on the
5926 particular protocol version it has chosen before the ServerHello
5927 message has informed the client about his choice.)
5930 *) Load error codes if they are not already present instead of using a
5931 static variable. This allows them to be cleanly unloaded and reloaded.
5934 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5936 *) Introduce limits to prevent malicious keys being able to
5937 cause a denial of service. (CVE-2006-2940)
5938 [Steve Henson, Bodo Moeller]
5940 *) Fix ASN.1 parsing of certain invalid structures that can result
5941 in a denial of service. (CVE-2006-2937) [Steve Henson]
5943 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5944 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5946 *) Fix SSL client code which could crash if connecting to a
5947 malicious SSLv2 server. (CVE-2006-4343)
5948 [Tavis Ormandy and Will Drewry, Google Security Team]
5950 *) Change ciphersuite string processing so that an explicit
5951 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5952 will no longer include "AES128-SHA"), and any other similar
5953 ciphersuite (same bitmap) from *other* protocol versions (so that
5954 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5955 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5956 changes from 0.9.8b and 0.9.8d.
5959 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5961 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5962 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5964 *) Change the Unix randomness entropy gathering to use poll() when
5965 possible instead of select(), since the latter has some
5966 undesirable limitations.
5967 [Darryl Miles via Richard Levitte and Bodo Moeller]
5969 *) Disable rogue ciphersuites:
5971 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5972 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5973 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5975 The latter two were purportedly from
5976 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5979 Also deactivate the remaining ciphersuites from
5980 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5981 unofficial, and the ID has long expired.
5984 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5985 dual-core machines) and other potential thread-safety issues.
5988 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5990 *) Adapt fipsld and the build system to link against the validated FIPS
5991 module in FIPS mode.
5994 *) Fixes for VC++ 2005 build under Windows.
5997 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5998 from a Windows bash shell such as MSYS. It is autodetected from the
5999 "config" script when run from a VC++ environment. Modify standard VC++
6000 build to use fipscanister.o from the GNU make build.
6003 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6005 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6006 The value now differs depending on if you build for FIPS or not.
6007 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6008 safely run with a non-FIPSed libcrypto, as it may crash because of
6009 the difference induced by this change.
6012 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6014 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6015 (part of SSL_OP_ALL). This option used to disable the
6016 countermeasure against man-in-the-middle protocol-version
6017 rollback in the SSL 2.0 server implementation, which is a bad
6018 idea. (CVE-2005-2969)
6020 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6021 for Information Security, National Institute of Advanced Industrial
6022 Science and Technology [AIST], Japan)]
6024 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6025 mainly for FIPS compliance and not fully integrated at this stage.
6028 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6029 the exponentiation using a fixed-length exponent. (Otherwise,
6030 the information leaked through timing could expose the secret key
6031 after many signatures; cf. Bleichenbacher's attack on DSA with
6035 *) Make a new fixed-window mod_exp implementation the default for
6036 RSA, DSA, and DH private-key operations so that the sequence of
6037 squares and multiplies and the memory access pattern are
6038 independent of the particular secret key. This will mitigate
6039 cache-timing and potential related attacks.
6041 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6042 and this is automatically used by BN_mod_exp_mont() if the new flag
6043 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6044 will use this BN flag for private exponents unless the flag
6045 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6046 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6048 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6050 *) Change the client implementation for SSLv23_method() and
6051 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6052 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6053 (Previously, the SSL 2.0 backwards compatible Client Hello
6054 message format would be used even with SSL_OP_NO_SSLv2.)
6057 *) Add support for smime-type MIME parameter in S/MIME messages which some
6061 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6062 a threadsafe manner. Modify rsa code to use new function and add calls
6063 to dsa and dh code (which had race conditions before).
6066 *) Include the fixed error library code in the C error file definitions
6067 instead of fixing them up at runtime. This keeps the error code
6068 structures constant.
6071 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6073 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6076 *) Fixes for newer kerberos headers. NB: the casts are needed because
6077 the 'length' field is signed on one version and unsigned on another
6078 with no (?) obvious way to tell the difference, without these VC++
6079 complains. Also the "definition" of FAR (blank) is no longer included
6080 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6081 some needed definitions.
6084 *) Undo Cygwin change.
6087 *) Added support for proxy certificates according to RFC 3820.
6088 Because they may be a security thread to unaware applications,
6089 they must be explicitly allowed in run-time. See
6090 docs/HOWTO/proxy_certificates.txt for further information.
6093 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6095 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6096 server and client random values. Previously
6097 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6098 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6100 This change has negligible security impact because:
6102 1. Server and client random values still have 24 bytes of pseudo random
6105 2. Server and client random values are sent in the clear in the initial
6108 3. The master secret is derived using the premaster secret (48 bytes in
6109 size for static RSA ciphersuites) as well as client server and random
6112 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6115 [Stephen Henson, reported by UK NISCC]
6117 *) Use Windows randomness collection on Cygwin.
6120 *) Fix hang in EGD/PRNGD query when communication socket is closed
6121 prematurely by EGD/PRNGD.
6122 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6124 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6127 *) Back-port of selected performance improvements from development
6128 branch, as well as improved support for PowerPC platforms.
6131 *) Add lots of checks for memory allocation failure, error codes to indicate
6132 failure and freeing up memory if a failure occurs.
6133 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6135 *) Add new -passin argument to dgst.
6138 *) Perform some character comparisons of different types in X509_NAME_cmp:
6139 this is needed for some certificates that re-encode DNs into UTF8Strings
6140 (in violation of RFC3280) and can't or won't issue name rollover
6144 *) Make an explicit check during certificate validation to see that
6145 the CA setting in each certificate on the chain is correct. As a
6146 side effect always do the following basic checks on extensions,
6147 not just when there's an associated purpose to the check:
6149 - if there is an unhandled critical extension (unless the user
6150 has chosen to ignore this fault)
6151 - if the path length has been exceeded (if one is set at all)
6152 - that certain extensions fit the associated purpose (if one has
6156 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6158 *) Avoid a race condition when CRLs are checked in a multi threaded
6159 environment. This would happen due to the reordering of the revoked
6160 entries during signature checking and serial number lookup. Now the
6161 encoding is cached and the serial number sort performed under a lock.
6162 Add new STACK function sk_is_sorted().
6165 *) Add Delta CRL to the extension code.
6168 *) Various fixes to s3_pkt.c so alerts are sent properly.
6169 [David Holmes <d.holmes@f5.com>]
6171 *) Reduce the chances of duplicate issuer name and serial numbers (in
6172 violation of RFC3280) using the OpenSSL certificate creation utilities.
6173 This is done by creating a random 64 bit value for the initial serial
6174 number when a serial number file is created or when a self signed
6175 certificate is created using 'openssl req -x509'. The initial serial
6176 number file is created using 'openssl x509 -next_serial' in CA.pl
6177 rather than being initialized to 1.
6180 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6182 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6183 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6184 [Joe Orton, Steve Henson]
6186 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6188 [Joe Orton, Steve Henson]
6190 *) Make it possible to have multiple active certificates with the same
6191 subject in the CA index file. This is done only if the keyword
6192 'unique_subject' is set to 'no' in the main CA section (default
6193 if 'CA_default') of the configuration file. The value is saved
6194 with the database itself in a separate index attribute file,
6195 named like the index file with '.attr' appended to the name.
6198 *) X509 verify fixes. Disable broken certificate workarounds when
6199 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6200 keyUsage extension present. Don't accept CRLs with unhandled critical
6201 extensions: since verify currently doesn't process CRL extensions this
6202 rejects a CRL with *any* critical extensions. Add new verify error codes
6206 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6207 A clarification of RFC2560 will require the use of OCTET STRINGs and
6208 some implementations cannot handle the current raw format. Since OpenSSL
6209 copies and compares OCSP nonces as opaque blobs without any attempt at
6210 parsing them this should not create any compatibility issues.
6213 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6214 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6215 this HMAC (and other) operations are several times slower than OpenSSL
6219 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6220 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6222 *) Use the correct content when signing type "other".
6225 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6227 *) Fix various bugs revealed by running the NISCC test suite:
6229 Stop out of bounds reads in the ASN1 code when presented with
6230 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6232 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6234 If verify callback ignores invalid public key errors don't try to check
6235 certificate signature with the NULL public key.
6239 *) New -ignore_err option in ocsp application to stop the server
6240 exiting on the first error in a request.
6243 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6244 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6248 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6249 extra data after the compression methods not only for TLS 1.0
6250 but also for SSL 3.0 (as required by the specification).
6251 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6253 *) Change X509_certificate_type() to mark the key as exported/exportable
6254 when it's 512 *bits* long, not 512 bytes.
6257 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6258 blocks during encryption.
6261 *) Various fixes to base64 BIO and non blocking I/O. On write
6262 flushes were not handled properly if the BIO retried. On read
6263 data was not being buffered properly and had various logic bugs.
6264 This also affects blocking I/O when the data being decoded is a
6268 *) Various S/MIME bugfixes and compatibility changes:
6269 output correct application/pkcs7 MIME type if
6270 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6271 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6272 of files as .eml work). Correctly handle very long lines in MIME
6276 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6278 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6279 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6280 a protocol version number mismatch like a decryption error
6281 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6284 *) Turn on RSA blinding by default in the default implementation
6285 to avoid a timing attack. Applications that don't want it can call
6286 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6287 They would be ill-advised to do so in most cases.
6288 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6290 *) Change RSA blinding code so that it works when the PRNG is not
6291 seeded (in this case, the secret RSA exponent is abused as
6292 an unpredictable seed -- if it is not unpredictable, there
6293 is no point in blinding anyway). Make RSA blinding thread-safe
6294 by remembering the creator's thread ID in rsa->blinding and
6295 having all other threads use local one-time blinding factors
6296 (this requires more computation than sharing rsa->blinding, but
6297 avoids excessive locking; and if an RSA object is not shared
6298 between threads, blinding will still be very fast).
6301 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6302 ENGINE as defaults for all supported algorithms irrespective of
6303 the 'flags' parameter. 'flags' is now honoured, so applications
6304 should make sure they are passing it correctly.
6307 *) Target "mingw" now allows native Windows code to be generated in
6308 the Cygwin environment as well as with the MinGW compiler.
6311 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6313 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6314 via timing by performing a MAC computation even if incorrect
6315 block cipher padding has been found. This is a countermeasure
6316 against active attacks where the attacker has to distinguish
6317 between bad padding and a MAC verification error. (CVE-2003-0078)
6319 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6320 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6321 Martin Vuagnoux (EPFL, Ilion)]
6323 *) Make the no-err option work as intended. The intention with no-err
6324 is not to have the whole error stack handling routines removed from
6325 libcrypto, it's only intended to remove all the function name and
6326 reason texts, thereby removing some of the footprint that may not
6327 be interesting if those errors aren't displayed anyway.
6329 NOTE: it's still possible for any application or module to have it's
6330 own set of error texts inserted. The routines are there, just not
6331 used by default when no-err is given.
6334 *) Add support for FreeBSD on IA64.
6335 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6337 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6338 Kerberos function mit_des_cbc_cksum(). Before this change,
6339 the value returned by DES_cbc_cksum() was like the one from
6340 mit_des_cbc_cksum(), except the bytes were swapped.
6341 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6343 *) Allow an application to disable the automatic SSL chain building.
6344 Before this a rather primitive chain build was always performed in
6345 ssl3_output_cert_chain(): an application had no way to send the
6346 correct chain if the automatic operation produced an incorrect result.
6348 Now the chain builder is disabled if either:
6350 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6352 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6354 The reasoning behind this is that an application would not want the
6355 auto chain building to take place if extra chain certificates are
6356 present and it might also want a means of sending no additional
6357 certificates (for example the chain has two certificates and the
6361 *) Add the possibility to build without the ENGINE framework.
6362 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6364 *) Under Win32 gmtime() can return NULL: check return value in
6365 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6368 *) DSA routines: under certain error conditions uninitialized BN objects
6369 could be freed. Solution: make sure initialization is performed early
6370 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6371 Nils Larsch <nla@trustcenter.de> via PR#459)
6374 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6375 checked on reconnect on the client side, therefore session resumption
6376 could still fail with a "ssl session id is different" error. This
6377 behaviour is masked when SSL_OP_ALL is used due to
6378 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6379 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6380 followup to PR #377.
6383 *) IA-32 assembler support enhancements: unified ELF targets, support
6384 for SCO/Caldera platforms, fix for Cygwin shared build.
6387 *) Add support for FreeBSD on sparc64. As a consequence, support for
6388 FreeBSD on non-x86 processors is separate from x86 processors on
6389 the config script, much like the NetBSD support.
6390 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6392 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6394 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6397 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6398 code (06) was taken as the first octet of the session ID and the last
6399 octet was ignored consequently. As a result SSLv2 client side session
6400 caching could not have worked due to the session ID mismatch between
6402 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6406 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6407 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6411 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6412 seems that in spite of existing for more than a year, many application
6413 author have done nothing to provide the necessary callbacks, which
6414 means that this particular engine will not work properly anywhere.
6415 This is a very unfortunate situation which forces us, in the name
6416 of usability, to give the hw_ncipher.c a static lock, which is part
6418 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6419 appear in 0.9.8 or later. We EXPECT application authors to have
6420 dealt properly with this when 0.9.8 is released (unless we actually
6421 make such changes in the libcrypto locking code that changes will
6422 have to be made anyway).
6425 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6426 octets have been read, EOF or an error occurs. Without this change
6427 some truncated ASN1 structures will not produce an error.
6430 *) Disable Heimdal support, since it hasn't been fully implemented.
6431 Still give the possibility to force the use of Heimdal, but with
6432 warnings and a request that patches get sent to openssl-dev.
6435 *) Add the VC-CE target, introduce the WINCE sysname, and add
6436 INSTALL.WCE and appropriate conditionals to make it build.
6437 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6439 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6440 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6441 edit numbers of the version.
6442 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6444 *) Introduce safe string copy and catenation functions
6445 (BUF_strlcpy() and BUF_strlcat()).
6446 [Ben Laurie (CHATS) and Richard Levitte]
6448 *) Avoid using fixed-size buffers for one-line DNs.
6449 [Ben Laurie (CHATS)]
6451 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6452 resizing buffers containing secrets, and use where appropriate.
6453 [Ben Laurie (CHATS)]
6455 *) Avoid using fixed size buffers for configuration file location.
6456 [Ben Laurie (CHATS)]
6458 *) Avoid filename truncation for various CA files.
6459 [Ben Laurie (CHATS)]
6461 *) Use sizeof in preference to magic numbers.
6462 [Ben Laurie (CHATS)]
6464 *) Avoid filename truncation in cert requests.
6465 [Ben Laurie (CHATS)]
6467 *) Add assertions to check for (supposedly impossible) buffer
6469 [Ben Laurie (CHATS)]
6471 *) Don't cache truncated DNS entries in the local cache (this could
6472 potentially lead to a spoofing attack).
6473 [Ben Laurie (CHATS)]
6475 *) Fix various buffers to be large enough for hex/decimal
6476 representations in a platform independent manner.
6477 [Ben Laurie (CHATS)]
6479 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6480 resizing buffers containing secrets, and use where appropriate.
6481 [Ben Laurie (CHATS)]
6483 *) Add BIO_indent() to avoid much slightly worrying code to do
6485 [Ben Laurie (CHATS)]
6487 *) Convert sprintf()/BIO_puts() to BIO_printf().
6488 [Ben Laurie (CHATS)]
6490 *) buffer_gets() could terminate with the buffer only half
6492 [Ben Laurie (CHATS)]
6494 *) Add assertions to prevent user-supplied crypto functions from
6495 overflowing internal buffers by having large block sizes, etc.
6496 [Ben Laurie (CHATS)]
6498 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6500 [Ben Laurie (CHATS)]
6502 *) Eliminate unused copy of key in RC4.
6503 [Ben Laurie (CHATS)]
6505 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6506 [Ben Laurie (CHATS)]
6508 *) Fix off-by-one error in EGD path.
6509 [Ben Laurie (CHATS)]
6511 *) If RANDFILE path is too long, ignore instead of truncating.
6512 [Ben Laurie (CHATS)]
6514 *) Eliminate unused and incorrectly sized X.509 structure
6516 [Ben Laurie (CHATS)]
6518 *) Eliminate unused and dangerous function knumber().
6519 [Ben Laurie (CHATS)]
6521 *) Eliminate unused and dangerous structure, KSSL_ERR.
6522 [Ben Laurie (CHATS)]
6524 *) Protect against overlong session ID context length in an encoded
6525 session object. Since these are local, this does not appear to be
6527 [Ben Laurie (CHATS)]
6529 *) Change from security patch (see 0.9.6e below) that did not affect
6530 the 0.9.6 release series:
6532 Remote buffer overflow in SSL3 protocol - an attacker could
6533 supply an oversized master key in Kerberos-enabled versions.
6535 [Ben Laurie (CHATS)]
6537 *) Change the SSL kerb5 codes to match RFC 2712.
6540 *) Make -nameopt work fully for req and add -reqopt switch.
6541 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6543 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6544 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6546 *) Make sure tests can be performed even if the corresponding algorithms
6547 have been removed entirely. This was also the last step to make
6548 OpenSSL compilable with DJGPP under all reasonable conditions.
6549 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6551 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6552 to allow version independent disabling of normally unselected ciphers,
6553 which may be activated as a side-effect of selecting a single cipher.
6555 (E.g., cipher list string "RSA" enables ciphersuites that are left
6556 out of "ALL" because they do not provide symmetric encryption.
6557 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6558 [Lutz Jaenicke, Bodo Moeller]
6560 *) Add appropriate support for separate platform-dependent build
6561 directories. The recommended way to make a platform-dependent
6562 build directory is the following (tested on Linux), maybe with
6565 # Place yourself outside of the OpenSSL source tree. In
6566 # this example, the environment variable OPENSSL_SOURCE
6567 # is assumed to contain the absolute OpenSSL source directory.
6568 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6569 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6570 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6571 mkdir -p `dirname $F`
6572 ln -s $OPENSSL_SOURCE/$F $F
6575 To be absolutely sure not to disturb the source tree, a "make clean"
6576 is a good thing. If it isn't successful, don't worry about it,
6577 it probably means the source directory is very clean.
6580 *) Make sure any ENGINE control commands make local copies of string
6581 pointers passed to them whenever necessary. Otherwise it is possible
6582 the caller may have overwritten (or deallocated) the original string
6583 data when a later ENGINE operation tries to use the stored values.
6584 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6586 *) Improve diagnostics in file reading and command-line digests.
6587 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6589 *) Add AES modes CFB and OFB to the object database. Correct an
6590 error in AES-CFB decryption.
6593 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6594 allows existing EVP_CIPHER_CTX structures to be reused after
6595 calling EVP_*Final(). This behaviour is used by encryption
6596 BIOs and some applications. This has the side effect that
6597 applications must explicitly clean up cipher contexts with
6598 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6601 *) Check the values of dna and dnb in bn_mul_recursive before calling
6602 bn_mul_comba (a non zero value means the a or b arrays do not contain
6603 n2 elements) and fallback to bn_mul_normal if either is not zero.
6606 *) Fix escaping of non-ASCII characters when using the -subj option
6607 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6610 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6611 form for "surname", serialNumber has no short form.
6612 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6613 therefore remove "mail" short name for "internet 7".
6614 The OID for unique identifiers in X509 certificates is
6615 x500UniqueIdentifier, not uniqueIdentifier.
6616 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6619 *) Add an "init" command to the ENGINE config module and auto initialize
6620 ENGINEs. Without any "init" command the ENGINE will be initialized
6621 after all ctrl commands have been executed on it. If init=1 the
6622 ENGINE is initialized at that point (ctrls before that point are run
6623 on the uninitialized ENGINE and after on the initialized one). If
6624 init=0 then the ENGINE will not be initialized at all.
6627 *) Fix the 'app_verify_callback' interface so that the user-defined
6628 argument is actually passed to the callback: In the
6629 SSL_CTX_set_cert_verify_callback() prototype, the callback
6630 declaration has been changed from
6633 int (*cb)(X509_STORE_CTX *,void *);
6634 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6635 i=s->ctx->app_verify_callback(&ctx)
6636 has been changed into
6637 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6639 To update applications using SSL_CTX_set_cert_verify_callback(),
6640 a dummy argument can be added to their callback functions.
6641 [D. K. Smetters <smetters@parc.xerox.com>]
6643 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6644 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6646 *) Add and OPENSSL_LOAD_CONF define which will cause
6647 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6648 This allows older applications to transparently support certain
6649 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6650 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6651 load the config file and OPENSSL_add_all_algorithms_conf() which will
6652 always load it have also been added.
6655 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6656 Adjust NIDs and EVP layer.
6657 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6659 *) Config modules support in openssl utility.
6661 Most commands now load modules from the config file,
6662 though in a few (such as version) this isn't done
6663 because it couldn't be used for anything.
6665 In the case of ca and req the config file used is
6666 the same as the utility itself: that is the -config
6667 command line option can be used to specify an
6671 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6672 use "openssl_conf" if filename is NULL use default openssl config file.
6675 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6676 config section name. Add a new flag to tolerate a missing config file
6677 and move code to CONF_modules_load_file().
6680 *) Support for crypto accelerator cards from Accelerated Encryption
6681 Processing, www.aep.ie. (Use engine 'aep')
6682 The support was copied from 0.9.6c [engine] and adapted/corrected
6683 to work with the new engine framework.
6684 [AEP Inc. and Richard Levitte]
6686 *) Support for SureWare crypto accelerator cards from Baltimore
6687 Technologies. (Use engine 'sureware')
6688 The support was copied from 0.9.6c [engine] and adapted
6689 to work with the new engine framework.
6692 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6693 make the newer ENGINE framework commands for the CHIL engine work.
6694 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6696 *) Make it possible to produce shared libraries on ReliantUNIX.
6697 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6699 *) Add the configuration target debug-linux-ppro.
6700 Make 'openssl rsa' use the general key loading routines
6701 implemented in apps.c, and make those routines able to
6702 handle the key format FORMAT_NETSCAPE and the variant
6704 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6706 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6707 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6709 *) Add -keyform to rsautl, and document -engine.
6710 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6712 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6713 BIO_R_NO_SUCH_FILE error code rather than the generic
6714 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6717 *) Add new functions
6719 ERR_peek_last_error_line
6720 ERR_peek_last_error_line_data.
6721 These are similar to
6724 ERR_peek_error_line_data,
6725 but report on the latest error recorded rather than the first one
6726 still in the error queue.
6727 [Ben Laurie, Bodo Moeller]
6729 *) default_algorithms option in ENGINE config module. This allows things
6731 default_algorithms = ALL
6732 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6735 *) Preliminary ENGINE config module.
6738 *) New experimental application configuration code.
6741 *) Change the AES code to follow the same name structure as all other
6742 symmetric ciphers, and behave the same way. Move everything to
6743 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6744 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6746 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6747 [Ben Laurie and Theo de Raadt]
6749 *) Add option to output public keys in req command.
6750 [Massimiliano Pala madwolf@openca.org]
6752 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6753 (up to about 10% better than before for P-192 and P-224).
6756 *) New functions/macros
6758 SSL_CTX_set_msg_callback(ctx, cb)
6759 SSL_CTX_set_msg_callback_arg(ctx, arg)
6760 SSL_set_msg_callback(ssl, cb)
6761 SSL_set_msg_callback_arg(ssl, arg)
6763 to request calling a callback function
6765 void cb(int write_p, int version, int content_type,
6766 const void *buf, size_t len, SSL *ssl, void *arg)
6768 whenever a protocol message has been completely received
6769 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6770 protocol version according to which the SSL library interprets
6771 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6772 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6773 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6774 specification (change_cipher_spec(20), alert(21), handshake(22)).
6775 'buf' and 'len' point to the actual message, 'ssl' to the
6776 SSL object, and 'arg' is the application-defined value set by
6777 SSL[_CTX]_set_msg_callback_arg().
6779 'openssl s_client' and 'openssl s_server' have new '-msg' options
6780 to enable a callback that displays all protocol messages.
6783 *) Change the shared library support so shared libraries are built as
6784 soon as the corresponding static library is finished, and thereby get
6785 openssl and the test programs linked against the shared library.
6786 This still only happens when the keyword "shard" has been given to
6787 the configuration scripts.
6789 NOTE: shared library support is still an experimental thing, and
6790 backward binary compatibility is still not guaranteed.
6791 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6793 *) Add support for Subject Information Access extension.
6794 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6796 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6797 additional bytes when new memory had to be allocated, not just
6798 when reusing an existing buffer.
6801 *) New command line and configuration option 'utf8' for the req command.
6802 This allows field values to be specified as UTF8 strings.
6805 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6806 runs for the former and machine-readable output for the latter.
6809 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6810 of the e-mail address in the DN (i.e., it will go into a certificate
6811 extension only). The new configuration file option 'email_in_dn = no'
6812 has the same effect.
6813 [Massimiliano Pala madwolf@openca.org]
6815 *) Change all functions with names starting with des_ to be starting
6816 with DES_ instead. Add wrappers that are compatible with libdes,
6817 but are named _ossl_old_des_*. Finally, add macros that map the
6818 des_* symbols to the corresponding _ossl_old_des_* if libdes
6819 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6820 desired, the des_* symbols will be mapped to DES_*, with one
6823 Since we provide two compatibility mappings, the user needs to
6824 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6825 compatibility is desired. The default (i.e., when that macro
6826 isn't defined) is OpenSSL 0.9.6c compatibility.
6828 There are also macros that enable and disable the support of old
6829 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6830 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6831 are defined, the default will apply: to support the old des routines.
6833 In either case, one must include openssl/des.h to get the correct
6834 definitions. Do not try to just include openssl/des_old.h, that
6837 NOTE: This is a major break of an old API into a new one. Software
6838 authors are encouraged to switch to the DES_ style functions. Some
6839 time in the future, des_old.h and the libdes compatibility functions
6840 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6841 default), and then completely removed.
6844 *) Test for certificates which contain unsupported critical extensions.
6845 If such a certificate is found during a verify operation it is
6846 rejected by default: this behaviour can be overridden by either
6847 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6848 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6849 X509_supported_extension() has also been added which returns 1 if a
6850 particular extension is supported.
6853 *) Modify the behaviour of EVP cipher functions in similar way to digests
6854 to retain compatibility with existing code.
6857 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6858 compatibility with existing code. In particular the 'ctx' parameter does
6859 not have to be to be initialized before the call to EVP_DigestInit() and
6860 it is tidied up after a call to EVP_DigestFinal(). New function
6861 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6862 EVP_MD_CTX_copy() changed to not require the destination to be
6863 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6864 requires the destination to be valid.
6866 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6867 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6870 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6871 so that complete 'Handshake' protocol structures are kept in memory
6872 instead of overwriting 'msg_type' and 'length' with 'body' data.
6875 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6876 [Massimo Santin via Richard Levitte]
6878 *) Major restructuring to the underlying ENGINE code. This includes
6879 reduction of linker bloat, separation of pure "ENGINE" manipulation
6880 (initialisation, etc) from functionality dealing with implementations
6881 of specific crypto interfaces. This change also introduces integrated
6882 support for symmetric ciphers and digest implementations - so ENGINEs
6883 can now accelerate these by providing EVP_CIPHER and EVP_MD
6884 implementations of their own. This is detailed in crypto/engine/README
6885 as it couldn't be adequately described here. However, there are a few
6886 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6887 were changed in the original introduction of ENGINE code have now
6888 reverted back - the hooking from this code to ENGINE is now a good
6889 deal more passive and at run-time, operations deal directly with
6890 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6891 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6892 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6893 they were not being used by the framework as there is no concept of a
6894 BIGNUM_METHOD and they could not be generalised to the new
6895 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6896 ENGINE_cpy() has been removed as it cannot be consistently defined in
6900 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6903 *) Change mkdef.pl to sort symbols that get the same entry number,
6904 and make sure the automatically generated functions ERR_load_*
6905 become part of libeay.num as well.
6908 *) New function SSL_renegotiate_pending(). This returns true once
6909 renegotiation has been requested (either SSL_renegotiate() call
6910 or HelloRequest/ClientHello received from the peer) and becomes
6911 false once a handshake has been completed.
6912 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6913 sends a HelloRequest, but does not ensure that a handshake takes
6914 place. SSL_renegotiate_pending() is useful for checking if the
6915 client has followed the request.)
6918 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6919 By default, clients may request session resumption even during
6920 renegotiation (if session ID contexts permit); with this option,
6921 session resumption is possible only in the first handshake.
6923 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6924 more bits available for options that should not be part of
6925 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6928 *) Add some demos for certificate and certificate request creation.
6931 *) Make maximum certificate chain size accepted from the peer application
6932 settable (SSL*_get/set_max_cert_list()), as proposed by
6933 "Douglas E. Engert" <deengert@anl.gov>.
6936 *) Add support for shared libraries for Unixware-7
6937 (Boyd Lynn Gerber <gerberb@zenez.com>).
6940 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6941 be done prior to destruction. Use this to unload error strings from
6942 ENGINEs that load their own error strings. NB: This adds two new API
6943 functions to "get" and "set" this destroy handler in an ENGINE.
6946 *) Alter all existing ENGINE implementations (except "openssl" and
6947 "openbsd") to dynamically instantiate their own error strings. This
6948 makes them more flexible to be built both as statically-linked ENGINEs
6949 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6950 Also, add stub code to each that makes building them as self-contained
6951 shared-libraries easier (see README.ENGINE).
6954 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6955 implementations into applications that are completely implemented in
6956 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6957 commands that can be used to configure what shared-library to load and
6958 to control aspects of the way it is handled. Also, made an update to
6959 the README.ENGINE file that brings its information up-to-date and
6960 provides some information and instructions on the "dynamic" ENGINE
6961 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6964 *) Make it possible to unload ranges of ERR strings with a new
6965 "ERR_unload_strings" function.
6968 *) Add a copy() function to EVP_MD.
6971 *) Make EVP_MD routines take a context pointer instead of just the
6972 md_data void pointer.
6975 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6976 that the digest can only process a single chunk of data
6977 (typically because it is provided by a piece of
6978 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6979 is only going to provide a single chunk of data, and hence the
6980 framework needn't accumulate the data for oneshot drivers.
6983 *) As with "ERR", make it possible to replace the underlying "ex_data"
6984 functions. This change also alters the storage and management of global
6985 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6986 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6987 index counters. The API functions that use this state have been changed
6988 to take a "class_index" rather than pointers to the class's local STACK
6989 and counter, and there is now an API function to dynamically create new
6990 classes. This centralisation allows us to (a) plug a lot of the
6991 thread-safety problems that existed, and (b) makes it possible to clean
6992 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6993 such data would previously have always leaked in application code and
6994 workarounds were in place to make the memory debugging turn a blind eye
6995 to it. Application code that doesn't use this new function will still
6996 leak as before, but their memory debugging output will announce it now
6997 rather than letting it slide.
6999 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7000 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7001 has a return value to indicate success or failure.
7004 *) Make it possible to replace the underlying "ERR" functions such that the
7005 global state (2 LHASH tables and 2 locks) is only used by the "default"
7006 implementation. This change also adds two functions to "get" and "set"
7007 the implementation prior to it being automatically set the first time
7008 any other ERR function takes place. Ie. an application can call "get",
7009 pass the return value to a module it has just loaded, and that module
7010 can call its own "set" function using that value. This means the
7011 module's "ERR" operations will use (and modify) the error state in the
7012 application and not in its own statically linked copy of OpenSSL code.
7015 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7016 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7017 the operation, and provides a more encapsulated way for external code
7018 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7019 to use these functions rather than manually incrementing the counts.
7021 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7024 *) Add EVP test program.
7027 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7030 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7031 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7032 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7033 These allow a CRL to be built without having to access X509_CRL fields
7034 directly. Modify 'ca' application to use new functions.
7037 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7038 bug workarounds. Rollback attack detection is a security feature.
7039 The problem will only arise on OpenSSL servers when TLSv1 is not
7040 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7041 Software authors not wanting to support TLSv1 will have special reasons
7042 for their choice and can explicitly enable this option.
7043 [Bodo Moeller, Lutz Jaenicke]
7045 *) Rationalise EVP so it can be extended: don't include a union of
7046 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7047 (similar to those existing for EVP_CIPHER_CTX).
7052 EVP_MD_CTX_init(&md); /* new function call */
7053 EVP_DigestInit(&md, EVP_sha1());
7054 EVP_DigestUpdate(&md, in, len);
7055 EVP_DigestFinal(&md, out, NULL);
7056 EVP_MD_CTX_cleanup(&md); /* new function call */
7060 *) Make DES key schedule conform to the usual scheme, as well as
7061 correcting its structure. This means that calls to DES functions
7062 now have to pass a pointer to a des_key_schedule instead of a
7063 plain des_key_schedule (which was actually always a pointer
7066 des_key_schedule ks;
7068 des_set_key_checked(..., &ks);
7069 des_ncbc_encrypt(..., &ks, ...);
7071 (Note that a later change renames 'des_...' into 'DES_...'.)
7074 *) Initial reduction of linker bloat: the use of some functions, such as
7075 PEM causes large amounts of unused functions to be linked in due to
7076 poor organisation. For example pem_all.c contains every PEM function
7077 which has a knock on effect of linking in large amounts of (unused)
7078 ASN1 code. Grouping together similar functions and splitting unrelated
7079 functions prevents this.
7082 *) Cleanup of EVP macros.
7085 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7086 correct _ecb suffix.
7089 *) Add initial OCSP responder support to ocsp application. The
7090 revocation information is handled using the text based index
7091 use by the ca application. The responder can either handle
7092 requests generated internally, supplied in files (for example
7093 via a CGI script) or using an internal minimal server.
7096 *) Add configuration choices to get zlib compression for TLS.
7099 *) Changes to Kerberos SSL for RFC 2712 compliance:
7100 1. Implemented real KerberosWrapper, instead of just using
7101 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7102 2. Implemented optional authenticator field of KerberosWrapper.
7104 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7105 and authenticator structs; see crypto/krb5/.
7107 Generalized Kerberos calls to support multiple Kerberos libraries.
7108 [Vern Staats <staatsvr@asc.hpc.mil>,
7109 Jeffrey Altman <jaltman@columbia.edu>
7110 via Richard Levitte]
7112 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7113 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7114 values for each of the key sizes rather than having just
7115 parameters (and 'speed' generating keys each time).
7118 *) Speed up EVP routines.
7121 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7122 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7123 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7124 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7126 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7127 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7128 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7131 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7133 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7136 *) Added the OS2-EMX target.
7137 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7139 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7140 to support NCONF routines in extension code. New function CONF_set_nconf()
7141 to allow functions which take an NCONF to also handle the old LHASH
7142 structure: this means that the old CONF compatible routines can be
7143 retained (in particular wrt extensions) without having to duplicate the
7144 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7147 *) Enhance the general user interface with mechanisms for inner control
7148 and with possibilities to have yes/no kind of prompts.
7151 *) Change all calls to low level digest routines in the library and
7152 applications to use EVP. Add missing calls to HMAC_cleanup() and
7153 don't assume HMAC_CTX can be copied using memcpy().
7154 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7156 *) Add the possibility to control engines through control names but with
7157 arbitrary arguments instead of just a string.
7158 Change the key loaders to take a UI_METHOD instead of a callback
7159 function pointer. NOTE: this breaks binary compatibility with earlier
7160 versions of OpenSSL [engine].
7161 Adapt the nCipher code for these new conditions and add a card insertion
7165 *) Enhance the general user interface with mechanisms to better support
7166 dialog box interfaces, application-defined prompts, the possibility
7167 to use defaults (for example default passwords from somewhere else)
7168 and interrupts/cancellations.
7171 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7172 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7175 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7176 tidy up some unnecessarily weird code in 'sk_new()').
7177 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7179 *) Change the key loading routines for ENGINEs to use the same kind
7180 callback (pem_password_cb) as all other routines that need this
7184 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7185 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7186 than this minimum value is recommended.
7189 *) New random seeder for OpenVMS, using the system process statistics
7190 that are easily reachable.
7193 *) Windows apparently can't transparently handle global
7194 variables defined in DLLs. Initialisations such as:
7196 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7198 won't compile. This is used by the any applications that need to
7199 declare their own ASN1 modules. This was fixed by adding the option
7200 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7201 needed for static libraries under Win32.
7204 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7205 setting of purpose and trust fields. New X509_STORE trust and
7206 purpose functions and tidy up setting in other SSL functions.
7209 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7210 structure. These are inherited by X509_STORE_CTX when it is
7211 initialised. This allows various defaults to be set in the
7212 X509_STORE structure (such as flags for CRL checking and custom
7213 purpose or trust settings) for functions which only use X509_STORE_CTX
7214 internally such as S/MIME.
7216 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7217 trust settings if they are not set in X509_STORE. This allows X509_STORE
7218 purposes and trust (in S/MIME for example) to override any set by default.
7220 Add command line options for CRL checking to smime, s_client and s_server
7224 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7225 are set then the CRL is looked up in the X509_STORE structure and
7226 its validity and signature checked, then if the certificate is found
7227 in the CRL the verify fails with a revoked error.
7229 Various new CRL related callbacks added to X509_STORE_CTX structure.
7231 Command line options added to 'verify' application to support this.
7233 This needs some additional work, such as being able to handle multiple
7234 CRLs with different times, extension based lookup (rather than just
7235 by subject name) and ultimately more complete V2 CRL extension
7239 *) Add a general user interface API (crypto/ui/). This is designed
7240 to replace things like des_read_password and friends (backward
7241 compatibility functions using this new API are provided).
7242 The purpose is to remove prompting functions from the DES code
7243 section as well as provide for prompting through dialog boxes in
7244 a window system and the like.
7247 *) Add "ex_data" support to ENGINE so implementations can add state at a
7248 per-structure level rather than having to store it globally.
7251 *) Make it possible for ENGINE structures to be copied when retrieved by
7252 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7253 This causes the "original" ENGINE structure to act like a template,
7254 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7255 operational state can be localised to each ENGINE structure, despite the
7256 fact they all share the same "methods". New ENGINE structures returned in
7257 this case have no functional references and the return value is the single
7258 structural reference. This matches the single structural reference returned
7259 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7263 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7264 needs to match any other type at all we need to manually clear the
7268 *) Changes to the "openssl engine" utility to include;
7269 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7270 about an ENGINE's available control commands.
7271 - executing control commands from command line arguments using the
7272 '-pre' and '-post' switches. '-post' is only used if '-t' is
7273 specified and the ENGINE is successfully initialised. The syntax for
7274 the individual commands are colon-separated, for example;
7275 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7278 *) New dynamic control command support for ENGINEs. ENGINEs can now
7279 declare their own commands (numbers), names (strings), descriptions,
7280 and input types for run-time discovery by calling applications. A
7281 subset of these commands are implicitly classed as "executable"
7282 depending on their input type, and only these can be invoked through
7283 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7284 can be based on user input, config files, etc). The distinction is
7285 that "executable" commands cannot return anything other than a boolean
7286 result and can only support numeric or string input, whereas some
7287 discoverable commands may only be for direct use through
7288 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7289 pointers, or other custom uses. The "executable" commands are to
7290 support parameterisations of ENGINE behaviour that can be
7291 unambiguously defined by ENGINEs and used consistently across any
7292 OpenSSL-based application. Commands have been added to all the
7293 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7294 control over shared-library paths without source code alterations.
7297 *) Changed all ENGINE implementations to dynamically allocate their
7298 ENGINEs rather than declaring them statically. Apart from this being
7299 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7300 this also allows the implementations to compile without using the
7301 internal engine_int.h header.
7304 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7305 'const' value. Any code that should be able to modify a RAND_METHOD
7306 should already have non-const pointers to it (ie. they should only
7307 modify their own ones).
7310 *) Made a variety of little tweaks to the ENGINE code.
7311 - "atalla" and "ubsec" string definitions were moved from header files
7312 to C code. "nuron" string definitions were placed in variables
7313 rather than hard-coded - allowing parameterisation of these values
7314 later on via ctrl() commands.
7315 - Removed unused "#if 0"'d code.
7316 - Fixed engine list iteration code so it uses ENGINE_free() to release
7317 structural references.
7318 - Constified the RAND_METHOD element of ENGINE structures.
7319 - Constified various get/set functions as appropriate and added
7320 missing functions (including a catch-all ENGINE_cpy that duplicates
7321 all ENGINE values onto a new ENGINE except reference counts/state).
7322 - Removed NULL parameter checks in get/set functions. Setting a method
7323 or function to NULL is a way of cancelling out a previously set
7324 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7325 and doesn't justify the extra error symbols and code.
7326 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7327 flags from engine_int.h to engine.h.
7328 - Changed prototypes for ENGINE handler functions (init(), finish(),
7329 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7332 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7333 to the algorithm using long division. The binary algorithm can be
7334 used only if the modulus is odd. On 32-bit systems, it is faster
7335 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7336 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7337 up to 450 bits. In 64-bit environments, the binary algorithm
7338 appears to be advantageous for much longer moduli; here we use it
7339 for moduli up to 2048 bits.
7342 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7343 could not support the combine flag in choice fields.
7346 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7347 extensions from a certificate request to the certificate.
7350 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7351 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7352 file: this allows the display of the certificate about to be
7353 signed to be customised, to allow certain fields to be included
7354 or excluded and extension details. The old system didn't display
7355 multicharacter strings properly, omitted fields not in the policy
7356 and couldn't display additional details such as extensions.
7359 *) Function EC_POINTs_mul for multiple scalar multiplication
7360 of an arbitrary number of elliptic curve points
7361 \sum scalars[i]*points[i],
7362 optionally including the generator defined for the EC_GROUP:
7363 scalar*generator + \sum scalars[i]*points[i].
7365 EC_POINT_mul is a simple wrapper function for the typical case
7366 that the point list has just one item (besides the optional
7370 *) First EC_METHODs for curves over GF(p):
7372 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7373 operations and provides various method functions that can also
7374 operate with faster implementations of modular arithmetic.
7376 EC_GFp_mont_method() reuses most functions that are part of
7377 EC_GFp_simple_method, but uses Montgomery arithmetic.
7379 [Bodo Moeller; point addition and point doubling
7380 implementation directly derived from source code provided by
7381 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7383 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7384 crypto/ec/ec_lib.c):
7386 Curves are EC_GROUP objects (with an optional group generator)
7387 based on EC_METHODs that are built into the library.
7389 Points are EC_POINT objects based on EC_GROUP objects.
7391 Most of the framework would be able to handle curves over arbitrary
7392 finite fields, but as there are no obvious types for fields other
7393 than GF(p), some functions are limited to that for now.
7396 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7397 that the file contains a complete HTTP response.
7400 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7401 change the def and num file printf format specifier from "%-40sXXX"
7402 to "%-39s XXX". The latter will always guarantee a space after the
7403 field while the former will cause them to run together if the field
7404 is 40 of more characters long.
7407 *) Constify the cipher and digest 'method' functions and structures
7408 and modify related functions to take constant EVP_MD and EVP_CIPHER
7412 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7413 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7416 *) Modify EVP_Digest*() routines so they now return values. Although the
7417 internal software routines can never fail additional hardware versions
7421 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7423 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7424 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7427 ERR_R_NESTED_ASN1_ERROR
7429 ERR_R_MISSING_ASN1_EOS
7430 were 4 .. 9, conflicting with
7431 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7433 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7434 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7436 Add new error code 'ERR_R_INTERNAL_ERROR'.
7439 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7443 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7444 sets the subject name for a new request or supersedes the
7445 subject name in a given request. Formats that can be parsed are
7446 'CN=Some Name, OU=myOU, C=IT'
7448 'CN=Some Name/OU=myOU/C=IT'.
7450 Add options '-batch' and '-verbose' to 'openssl req'.
7451 [Massimiliano Pala <madwolf@hackmasters.net>]
7453 *) Introduce the possibility to access global variables through
7454 functions on platform were that's the best way to handle exporting
7455 global variables in shared libraries. To enable this functionality,
7456 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7457 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7458 is normally done by Configure or something similar).
7460 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7461 in the source file (foo.c) like this:
7463 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7464 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7466 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7467 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7469 OPENSSL_DECLARE_GLOBAL(int,foo);
7470 #define foo OPENSSL_GLOBAL_REF(foo)
7471 OPENSSL_DECLARE_GLOBAL(double,bar);
7472 #define bar OPENSSL_GLOBAL_REF(bar)
7474 The #defines are very important, and therefore so is including the
7475 header file everywhere where the defined globals are used.
7477 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7478 of ASN.1 items, but that structure is a bit different.
7480 The largest change is in util/mkdef.pl which has been enhanced with
7481 better and easier to understand logic to choose which symbols should
7482 go into the Windows .def files as well as a number of fixes and code
7483 cleanup (among others, algorithm keywords are now sorted
7484 lexicographically to avoid constant rewrites).
7487 *) In BN_div() keep a copy of the sign of 'num' before writing the
7488 result to 'rm' because if rm==num the value will be overwritten
7489 and produce the wrong result if 'num' is negative: this caused
7490 problems with BN_mod() and BN_nnmod().
7493 *) Function OCSP_request_verify(). This checks the signature on an
7494 OCSP request and verifies the signer certificate. The signer
7495 certificate is just checked for a generic purpose and OCSP request
7499 *) Add OCSP_check_validity() function to check the validity of OCSP
7500 responses. OCSP responses are prepared in real time and may only
7501 be a few seconds old. Simply checking that the current time lies
7502 between thisUpdate and nextUpdate max reject otherwise valid responses
7503 caused by either OCSP responder or client clock inaccuracy. Instead
7504 we allow thisUpdate and nextUpdate to fall within a certain period of
7505 the current time. The age of the response can also optionally be
7506 checked. Two new options -validity_period and -status_age added to
7510 *) If signature or public key algorithm is unrecognized print out its
7511 OID rather that just UNKNOWN.
7514 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7515 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7516 ID to be generated from the issuer certificate alone which can then be
7517 passed to OCSP_id_issuer_cmp().
7520 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7521 ASN1 modules to export functions returning ASN1_ITEM pointers
7522 instead of the ASN1_ITEM structures themselves. This adds several
7523 new macros which allow the underlying ASN1 function/structure to
7524 be accessed transparently. As a result code should not use ASN1_ITEM
7525 references directly (such as &X509_it) but instead use the relevant
7526 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7527 use of the new ASN1 code on platforms where exporting structures
7528 is problematical (for example in shared libraries) but exporting
7529 functions returning pointers to structures is not.
7532 *) Add support for overriding the generation of SSL/TLS session IDs.
7533 These callbacks can be registered either in an SSL_CTX or per SSL.
7534 The purpose of this is to allow applications to control, if they wish,
7535 the arbitrary values chosen for use as session IDs, particularly as it
7536 can be useful for session caching in multiple-server environments. A
7537 command-line switch for testing this (and any client code that wishes
7538 to use such a feature) has been added to "s_server".
7539 [Geoff Thorpe, Lutz Jaenicke]
7541 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7542 of the form '#if defined(...) || defined(...) || ...' and
7543 '#if !defined(...) && !defined(...) && ...'. This also avoids
7544 the growing number of special cases it was previously handling.
7547 *) Make all configuration macros available for application by making
7548 sure they are available in opensslconf.h, by giving them names starting
7549 with "OPENSSL_" to avoid conflicts with other packages and by making
7550 sure e_os2.h will cover all platform-specific cases together with
7552 Additionally, it is now possible to define configuration/platform-
7553 specific names (called "system identities"). In the C code, these
7554 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7555 macro with the name beginning with "OPENSSL_SYS_", which is determined
7556 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7560 *) New option -set_serial to 'req' and 'x509' this allows the serial
7561 number to use to be specified on the command line. Previously self
7562 signed certificates were hard coded with serial number 0 and the
7563 CA options of 'x509' had to use a serial number in a file which was
7567 *) New options to 'ca' utility to support V2 CRL entry extensions.
7568 Currently CRL reason, invalidity date and hold instruction are
7569 supported. Add new CRL extensions to V3 code and some new objects.
7572 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7573 disable standard block padding (aka PKCS#5 padding) in the EVP
7574 API, which was previously mandatory. This means that the data is
7575 not padded in any way and so the total length much be a multiple
7576 of the block size, otherwise an error occurs.
7579 *) Initial (incomplete) OCSP SSL support.
7582 *) New function OCSP_parse_url(). This splits up a URL into its host,
7583 port and path components: primarily to parse OCSP URLs. New -url
7584 option to ocsp utility.
7587 *) New nonce behavior. The return value of OCSP_check_nonce() now
7588 reflects the various checks performed. Applications can decide
7589 whether to tolerate certain situations such as an absent nonce
7590 in a response when one was present in a request: the ocsp application
7591 just prints out a warning. New function OCSP_add1_basic_nonce()
7592 this is to allow responders to include a nonce in a response even if
7593 the request is nonce-less.
7596 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7597 skipped when using openssl x509 multiple times on a single input file,
7598 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7601 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7602 set string type: to handle setting ASN1_TIME structures. Fix ca
7603 utility to correctly initialize revocation date of CRLs.
7606 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7607 the clients preferred ciphersuites and rather use its own preferences.
7608 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7609 Internet Explorer by ensuring unchanged hash method during stepup.
7610 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7613 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7614 to aes and add a new 'exist' option to print out symbols that don't
7618 *) Additional options to ocsp utility to allow flags to be set and
7619 additional certificates supplied.
7622 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7623 OCSP client a number of certificate to only verify the response
7627 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7628 handle the new API. Currently only ECB, CBC modes supported. Add new
7631 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7632 Encryption Standard (AES) Ciphersuites for Transport Layer
7633 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7634 not enabled by default and were not part of the "ALL" ciphersuite
7635 alias because they were not yet official; they could be
7636 explicitly requested by specifying the "AESdraft" ciphersuite
7637 group alias. In the final release of OpenSSL 0.9.7, the group
7638 alias is called "AES" and is part of "ALL".)
7639 [Ben Laurie, Steve Henson, Bodo Moeller]
7641 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7642 request to response.
7645 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7646 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7647 extract information from a certificate request. OCSP_response_create()
7648 creates a response and optionally adds a basic response structure.
7649 OCSP_basic_add1_status() adds a complete single response to a basic
7650 response and returns the OCSP_SINGLERESP structure just added (to allow
7651 extensions to be included for example). OCSP_basic_add1_cert() adds a
7652 certificate to a basic response and OCSP_basic_sign() signs a basic
7653 response with various flags. New helper functions ASN1_TIME_check()
7654 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7655 (converts ASN1_TIME to GeneralizedTime).
7658 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7659 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7660 structure from a certificate. X509_pubkey_digest() digests the public_key
7661 contents: this is used in various key identifiers.
7664 *) Make sk_sort() tolerate a NULL argument.
7665 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7667 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7668 passed by the function are trusted implicitly. If any of them signed the
7669 response then it is assumed to be valid and is not verified.
7672 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7673 to data. This was previously part of the PKCS7 ASN1 code. This
7674 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7675 [Steve Henson, reported by Kenneth R. Robinette
7676 <support@securenetterm.com>]
7678 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7679 routines: without these tracing memory leaks is very painful.
7680 Fix leaks in PKCS12 and PKCS7 routines.
7683 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7684 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7685 effectively meant GeneralizedTime would never be used. Now it
7686 is initialised to -1 but X509_time_adj() now has to check the value
7687 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7688 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7689 [Steve Henson, reported by Kenneth R. Robinette
7690 <support@securenetterm.com>]
7692 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7693 result in a zero length in the ASN1_INTEGER structure which was
7694 not consistent with the structure when d2i_ASN1_INTEGER() was used
7695 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7696 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7697 where it did not print out a minus for negative ASN1_INTEGER.
7700 *) Add summary printout to ocsp utility. The various functions which
7701 convert status values to strings have been renamed to:
7702 OCSP_response_status_str(), OCSP_cert_status_str() and
7703 OCSP_crl_reason_str() and are no longer static. New options
7704 to verify nonce values and to disable verification. OCSP response
7705 printout format cleaned up.
7708 *) Add additional OCSP certificate checks. These are those specified
7709 in RFC2560. This consists of two separate checks: the CA of the
7710 certificate being checked must either be the OCSP signer certificate
7711 or the issuer of the OCSP signer certificate. In the latter case the
7712 OCSP signer certificate must contain the OCSP signing extended key
7713 usage. This check is performed by attempting to match the OCSP
7714 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7715 in the OCSP_CERTID structures of the response.
7718 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7719 and related routines. This uses the standard OpenSSL certificate
7720 verify routines to perform initial checks (just CA validity) and
7721 to obtain the certificate chain. Then additional checks will be
7722 performed on the chain. Currently the root CA is checked to see
7723 if it is explicitly trusted for OCSP signing. This is used to set
7724 a root CA as a global signing root: that is any certificate that
7725 chains to that CA is an acceptable OCSP signing certificate.
7728 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7729 extensions from a separate configuration file.
7730 As when reading extensions from the main configuration file,
7731 the '-extensions ...' option may be used for specifying the
7733 [Massimiliano Pala <madwolf@comune.modena.it>]
7735 *) New OCSP utility. Allows OCSP requests to be generated or
7736 read. The request can be sent to a responder and the output
7737 parsed, outputed or printed in text form. Not complete yet:
7738 still needs to check the OCSP response validity.
7741 *) New subcommands for 'openssl ca':
7742 'openssl ca -status <serial>' prints the status of the cert with
7743 the given serial number (according to the index file).
7744 'openssl ca -updatedb' updates the expiry status of certificates
7746 [Massimiliano Pala <madwolf@comune.modena.it>]
7748 *) New '-newreq-nodes' command option to CA.pl. This is like
7749 '-newreq', but calls 'openssl req' with the '-nodes' option
7750 so that the resulting key is not encrypted.
7751 [Damien Miller <djm@mindrot.org>]
7753 *) New configuration for the GNU Hurd.
7754 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7756 *) Initial code to implement OCSP basic response verify. This
7757 is currently incomplete. Currently just finds the signer's
7758 certificate and verifies the signature on the response.
7761 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7762 value of OPENSSLDIR. This is available via the new '-d' option
7763 to 'openssl version', and is also included in 'openssl version -a'.
7766 *) Allowing defining memory allocation callbacks that will be given
7767 file name and line number information in additional arguments
7768 (a const char* and an int). The basic functionality remains, as
7769 well as the original possibility to just replace malloc(),
7770 realloc() and free() by functions that do not know about these
7771 additional arguments. To register and find out the current
7772 settings for extended allocation functions, the following
7773 functions are provided:
7775 CRYPTO_set_mem_ex_functions
7776 CRYPTO_set_locked_mem_ex_functions
7777 CRYPTO_get_mem_ex_functions
7778 CRYPTO_get_locked_mem_ex_functions
7780 These work the same way as CRYPTO_set_mem_functions and friends.
7781 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7782 extended allocation function is enabled.
7783 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7784 a conventional allocation function is enabled.
7785 [Richard Levitte, Bodo Moeller]
7787 *) Finish off removing the remaining LHASH function pointer casts.
7788 There should no longer be any prototype-casting required when using
7789 the LHASH abstraction, and any casts that remain are "bugs". See
7790 the callback types and macros at the head of lhash.h for details
7791 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7794 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7795 If /dev/[u]random devices are not available or do not return enough
7796 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7798 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7799 /etc/entropy will be queried once each in this sequence, querying stops
7800 when enough entropy was collected without querying more sockets.
7803 *) Change the Unix RAND_poll() variant to be able to poll several
7804 random devices, as specified by DEVRANDOM, until a sufficient amount
7805 of data has been collected. We spend at most 10 ms on each file
7806 (select timeout) and read in non-blocking mode. DEVRANDOM now
7807 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7808 (previously it was just the string "/dev/urandom"), so on typical
7809 platforms the 10 ms delay will never occur.
7810 Also separate out the Unix variant to its own file, rand_unix.c.
7811 For VMS, there's a currently-empty rand_vms.c.
7814 *) Move OCSP client related routines to ocsp_cl.c. These
7815 provide utility functions which an application needing
7816 to issue a request to an OCSP responder and analyse the
7817 response will typically need: as opposed to those which an
7818 OCSP responder itself would need which will be added later.
7820 OCSP_request_sign() signs an OCSP request with an API similar
7821 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7822 response. OCSP_response_get1_basic() extracts basic response
7823 from response. OCSP_resp_find_status(): finds and extracts status
7824 information from an OCSP_CERTID structure (which will be created
7825 when the request structure is built). These are built from lower
7826 level functions which work on OCSP_SINGLERESP structures but
7827 won't normally be used unless the application wishes to examine
7828 extensions in the OCSP response for example.
7830 Replace nonce routines with a pair of functions.
7831 OCSP_request_add1_nonce() adds a nonce value and optionally
7832 generates a random value. OCSP_check_nonce() checks the
7833 validity of the nonce in an OCSP response.
7836 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7837 This doesn't copy the supplied OCSP_CERTID and avoids the
7838 need to free up the newly created id. Change return type
7839 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7840 This can then be used to add extensions to the request.
7841 Deleted OCSP_request_new(), since most of its functionality
7842 is now in OCSP_REQUEST_new() (and the case insensitive name
7843 clash) apart from the ability to set the request name which
7844 will be added elsewhere.
7847 *) Update OCSP API. Remove obsolete extensions argument from
7848 various functions. Extensions are now handled using the new
7849 OCSP extension code. New simple OCSP HTTP function which
7850 can be used to send requests and parse the response.
7853 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7854 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7855 uses the special reorder version of SET OF to sort the attributes
7856 and reorder them to match the encoded order. This resolves a long
7857 standing problem: a verify on a PKCS7 structure just after signing
7858 it used to fail because the attribute order did not match the
7859 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7860 it uses the received order. This is necessary to tolerate some broken
7861 software that does not order SET OF. This is handled by encoding
7862 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7863 to produce the required SET OF.
7866 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7867 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7868 files to get correct declarations of the ASN.1 item variables.
7871 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7872 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7873 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7874 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7875 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7876 ASN1_ITEM and no wrapper functions.
7879 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7880 replace the old function pointer based I/O routines. Change most of
7881 the *_d2i_bio() and *_d2i_fp() functions to use these.
7884 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7885 lines, recognize more "algorithms" that can be deselected, and make
7886 it complain about algorithm deselection that isn't recognised.
7889 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7890 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7891 to use new functions. Add NO_ASN1_OLD which can be set to remove
7892 some old style ASN1 functions: this can be used to determine if old
7893 code will still work when these eventually go away.
7896 *) New extension functions for OCSP structures, these follow the
7897 same conventions as certificates and CRLs.
7900 *) New function X509V3_add1_i2d(). This automatically encodes and
7901 adds an extension. Its behaviour can be customised with various
7902 flags to append, replace or delete. Various wrappers added for
7903 certificates and CRLs.
7906 *) Fix to avoid calling the underlying ASN1 print routine when
7907 an extension cannot be parsed. Correct a typo in the
7908 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7911 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7912 entries for variables.
7915 *) Add functionality to apps/openssl.c for detecting locking
7916 problems: As the program is single-threaded, all we have
7917 to do is register a locking callback using an array for
7918 storing which locks are currently held by the program.
7921 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7922 SSL_get_ex_data_X509_STORE_idx(), which is used in
7923 ssl_verify_cert_chain() and thus can be called at any time
7924 during TLS/SSL handshakes so that thread-safety is essential.
7925 Unfortunately, the ex_data design is not at all suited
7926 for multi-threaded use, so it probably should be abolished.
7929 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7930 [Broadcom, tweaked and integrated by Geoff Thorpe]
7932 *) Move common extension printing code to new function
7933 X509V3_print_extensions(). Reorganise OCSP print routines and
7934 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7937 *) New function X509_signature_print() to remove duplication in some
7941 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7942 set (this was treated exactly the same as SET OF previously). This
7943 is used to reorder the STACK representing the structure to match the
7944 encoding. This will be used to get round a problem where a PKCS7
7945 structure which was signed could not be verified because the STACK
7946 order did not reflect the encoded order.
7949 *) Reimplement the OCSP ASN1 module using the new code.
7952 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7953 for its ASN1 operations. The old style function pointers still exist
7954 for now but they will eventually go away.
7957 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7958 completely replaces the old ASN1 functionality with a table driven
7959 encoder and decoder which interprets an ASN1_ITEM structure describing
7960 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7961 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7962 has also been converted to the new form.
7965 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7966 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7967 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7968 for negative moduli.
7971 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7972 of not touching the result's sign bit.
7975 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7979 *) Changed the LHASH code to use prototypes for callbacks, and created
7980 macros to declare and implement thin (optionally static) functions
7981 that provide type-safety and avoid function pointer casting for the
7982 type-specific callbacks.
7985 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7987 [Veers Staats <staatsvr@asc.hpc.mil>,
7988 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7990 *) Reformat the FAQ so the different questions and answers can be divided
7991 in sections depending on the subject.
7994 *) Have the zlib compression code load ZLIB.DLL dynamically under
7998 *) New function BN_mod_sqrt for computing square roots modulo a prime
7999 (using the probabilistic Tonelli-Shanks algorithm unless
8000 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8001 be handled deterministically).
8002 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8004 *) Make BN_mod_inverse faster by explicitly handling small quotients
8005 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8006 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8009 *) New function BN_kronecker.
8012 *) Fix BN_gcd so that it works on negative inputs; the result is
8013 positive unless both parameters are zero.
8014 Previously something reasonably close to an infinite loop was
8015 possible because numbers could be growing instead of shrinking
8016 in the implementation of Euclid's algorithm.
8019 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8020 sign of the number in question.
8022 Fix BN_is_word(a,w) to work correctly for w == 0.
8024 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8025 because its test if the absolute value of 'a' equals 'w'.
8026 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8027 it exists mostly for use in the implementations of BN_is_zero(),
8028 BN_is_one(), and BN_is_word().
8031 *) New function BN_swap.
8034 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8035 the exponentiation functions are more likely to produce reasonable
8036 results on negative inputs.
8039 *) Change BN_mod_mul so that the result is always non-negative.
8040 Previously, it could be negative if one of the factors was negative;
8041 I don't think anyone really wanted that behaviour.
8044 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8045 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8046 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8047 and add new functions:
8056 BN_mod_lshift1_quick
8060 These functions always generate non-negative results.
8062 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8063 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8065 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8066 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8067 be reduced modulo m.
8068 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8071 The following entry accidentally appeared in the CHANGES file
8072 distributed with OpenSSL 0.9.7. The modifications described in
8073 it do *not* apply to OpenSSL 0.9.7.
8075 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8076 was actually never needed) and in BN_mul(). The removal in BN_mul()
8077 required a small change in bn_mul_part_recursive() and the addition
8078 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8079 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8080 bn_sub_words() and bn_add_words() except they take arrays with
8085 *) In 'openssl passwd', verify passwords read from the terminal
8086 unless the '-salt' option is used (which usually means that
8087 verification would just waste user's time since the resulting
8088 hash is going to be compared with some given password hash)
8089 or the new '-noverify' option is used.
8091 This is an incompatible change, but it does not affect
8092 non-interactive use of 'openssl passwd' (passwords on the command
8093 line, '-stdin' option, '-in ...' option) and thus should not
8097 *) Remove all references to RSAref, since there's no more need for it.
8100 *) Make DSO load along a path given through an environment variable
8101 (SHLIB_PATH) with shl_load().
8104 *) Constify the ENGINE code as a result of BIGNUM constification.
8105 Also constify the RSA code and most things related to it. In a
8106 few places, most notable in the depth of the ASN.1 code, ugly
8107 casts back to non-const were required (to be solved at a later
8111 *) Make it so the openssl application has all engines loaded by default.
8114 *) Constify the BIGNUM routines a little more.
8117 *) Add the following functions:
8119 ENGINE_load_cswift()
8121 ENGINE_load_atalla()
8123 ENGINE_load_builtin_engines()
8125 That way, an application can itself choose if external engines that
8126 are built-in in OpenSSL shall ever be used or not. The benefit is
8127 that applications won't have to be linked with libdl or other dso
8128 libraries unless it's really needed.
8130 Changed 'openssl engine' to load all engines on demand.
8131 Changed the engine header files to avoid the duplication of some
8132 declarations (they differed!).
8135 *) 'openssl engine' can now list capabilities.
8138 *) Better error reporting in 'openssl engine'.
8141 *) Never call load_dh_param(NULL) in s_server.
8144 *) Add engine application. It can currently list engines by name and
8145 identity, and test if they are actually available.
8148 *) Improve RPM specification file by forcing symbolic linking and making
8149 sure the installed documentation is also owned by root.root.
8150 [Damien Miller <djm@mindrot.org>]
8152 *) Give the OpenSSL applications more possibilities to make use of
8153 keys (public as well as private) handled by engines.
8156 *) Add OCSP code that comes from CertCo.
8159 *) Add VMS support for the Rijndael code.
8162 *) Added untested support for Nuron crypto accelerator.
8165 *) Add support for external cryptographic devices. This code was
8166 previously distributed separately as the "engine" branch.
8167 [Geoff Thorpe, Richard Levitte]
8169 *) Rework the filename-translation in the DSO code. It is now possible to
8170 have far greater control over how a "name" is turned into a filename
8171 depending on the operating environment and any oddities about the
8172 different shared library filenames on each system.
8175 *) Support threads on FreeBSD-elf in Configure.
8178 *) Fix for SHA1 assembly problem with MASM: it produces
8179 warnings about corrupt line number information when assembling
8180 with debugging information. This is caused by the overlapping
8182 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8185 NCONF_get_number() has no error checking at all. As a replacement,
8186 NCONF_get_number_e() is defined (_e for "error checking") and is
8187 promoted strongly. The old NCONF_get_number is kept around for
8188 binary backward compatibility.
8189 Make it possible for methods to load from something other than a BIO,
8190 by providing a function pointer that is given a name instead of a BIO.
8191 For example, this could be used to load configuration data from an
8195 *) Fix for non blocking accept BIOs. Added new I/O special reason
8196 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8197 with non blocking I/O was not possible because no retry code was
8198 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8202 *) Added the beginnings of Rijndael support.
8205 *) Fix for bug in DirectoryString mask setting. Add support for
8206 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8207 to allow certificate printing to more controllable, additional
8208 'certopt' option to 'x509' to allow new printing options to be
8212 *) Clean old EAY MD5 hack from e_os.h.
8215 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8217 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8218 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8219 [Joe Orton, Steve Henson]
8221 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8223 *) Fix additional bug revealed by the NISCC test suite:
8225 Stop bug triggering large recursion when presented with
8226 certain ASN.1 tags (CVE-2003-0851)
8229 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8231 *) Fix various bugs revealed by running the NISCC test suite:
8233 Stop out of bounds reads in the ASN1 code when presented with
8234 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8236 If verify callback ignores invalid public key errors don't try to check
8237 certificate signature with the NULL public key.
8241 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8242 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8246 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8247 extra data after the compression methods not only for TLS 1.0
8248 but also for SSL 3.0 (as required by the specification).
8249 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8251 *) Change X509_certificate_type() to mark the key as exported/exportable
8252 when it's 512 *bits* long, not 512 bytes.
8255 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8257 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8258 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8259 a protocol version number mismatch like a decryption error
8260 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8263 *) Turn on RSA blinding by default in the default implementation
8264 to avoid a timing attack. Applications that don't want it can call
8265 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8266 They would be ill-advised to do so in most cases.
8267 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8269 *) Change RSA blinding code so that it works when the PRNG is not
8270 seeded (in this case, the secret RSA exponent is abused as
8271 an unpredictable seed -- if it is not unpredictable, there
8272 is no point in blinding anyway). Make RSA blinding thread-safe
8273 by remembering the creator's thread ID in rsa->blinding and
8274 having all other threads use local one-time blinding factors
8275 (this requires more computation than sharing rsa->blinding, but
8276 avoids excessive locking; and if an RSA object is not shared
8277 between threads, blinding will still be very fast).
8280 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8282 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8283 via timing by performing a MAC computation even if incorrect
8284 block cipher padding has been found. This is a countermeasure
8285 against active attacks where the attacker has to distinguish
8286 between bad padding and a MAC verification error. (CVE-2003-0078)
8288 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8289 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8290 Martin Vuagnoux (EPFL, Ilion)]
8292 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8294 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8295 memory from it's contents. This is done with a counter that will
8296 place alternating values in each byte. This can be used to solve
8297 two issues: 1) the removal of calls to memset() by highly optimizing
8298 compilers, and 2) cleansing with other values than 0, since those can
8299 be read through on certain media, for example a swap space on disk.
8302 *) Bugfix: client side session caching did not work with external caching,
8303 because the session->cipher setting was not restored when reloading
8304 from the external cache. This problem was masked, when
8305 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8306 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8309 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8310 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8311 [Zeev Lieber <zeev-l@yahoo.com>]
8313 *) Undo an undocumented change introduced in 0.9.6e which caused
8314 repeated calls to OpenSSL_add_all_ciphers() and
8315 OpenSSL_add_all_digests() to be ignored, even after calling
8319 *) Change the default configuration reader to deal with last line not
8320 being properly terminated.
8323 *) Change X509_NAME_cmp() so it applies the special rules on handling
8324 DN values that are of type PrintableString, as well as RDNs of type
8325 emailAddress where the value has the type ia5String.
8326 [stefank@valicert.com via Richard Levitte]
8328 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8329 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8330 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8331 the bitwise-OR of the two for use by the majority of applications
8332 wanting this behaviour, and update the docs. The documented
8333 behaviour and actual behaviour were inconsistent and had been
8334 changing anyway, so this is more a bug-fix than a behavioural
8336 [Geoff Thorpe, diagnosed by Nadav Har'El]
8338 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8339 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8342 *) Fix initialization code race conditions in
8343 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8344 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8345 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8346 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8347 ssl2_get_cipher_by_char(),
8348 ssl3_get_cipher_by_char().
8349 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8351 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8352 the cached sessions are flushed, as the remove_cb() might use ex_data
8353 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8354 (see [openssl.org #212]).
8355 [Geoff Thorpe, Lutz Jaenicke]
8357 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8358 length, instead of the encoding length to d2i_ASN1_OBJECT.
8361 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8363 *) [In 0.9.6g-engine release:]
8364 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8365 [Lynn Gazis <lgazis@rainbow.com>]
8367 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8369 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8370 and get fix the header length calculation.
8371 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8372 Alon Kantor <alonk@checkpoint.com> (and others),
8375 *) Use proper error handling instead of 'assertions' in buffer
8376 overflow checks added in 0.9.6e. This prevents DoS (the
8377 assertions could call abort()).
8378 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8380 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8382 *) Add various sanity checks to asn1_get_length() to reject
8383 the ASN1 length bytes if they exceed sizeof(long), will appear
8384 negative or the content length exceeds the length of the
8386 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8388 *) Fix cipher selection routines: ciphers without encryption had no flags
8389 for the cipher strength set and where therefore not handled correctly
8390 by the selection routines (PR #130).
8393 *) Fix EVP_dsa_sha macro.
8397 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8398 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8399 that was added in OpenSSL 0.9.6d.
8401 As the countermeasure turned out to be incompatible with some
8402 broken SSL implementations, the new option is part of SSL_OP_ALL.
8403 SSL_OP_ALL is usually employed when compatibility with weird SSL
8404 implementations is desired (e.g. '-bugs' option to 's_client' and
8405 's_server'), so the new option is automatically set in many
8409 *) Changes in security patch:
8411 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8412 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8413 Air Force Materiel Command, USAF, under agreement number
8416 *) Add various sanity checks to asn1_get_length() to reject
8417 the ASN1 length bytes if they exceed sizeof(long), will appear
8418 negative or the content length exceeds the length of the
8419 supplied buffer. (CVE-2002-0659)
8420 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8422 *) Assertions for various potential buffer overflows, not known to
8424 [Ben Laurie (CHATS)]
8426 *) Various temporary buffers to hold ASCII versions of integers were
8427 too small for 64 bit platforms. (CVE-2002-0655)
8428 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8430 *) Remote buffer overflow in SSL3 protocol - an attacker could
8431 supply an oversized session ID to a client. (CVE-2002-0656)
8432 [Ben Laurie (CHATS)]
8434 *) Remote buffer overflow in SSL2 protocol - an attacker could
8435 supply an oversized client master key. (CVE-2002-0656)
8436 [Ben Laurie (CHATS)]
8438 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8440 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8441 encoded as NULL) with id-dsa-with-sha1.
8442 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8444 *) Check various X509_...() return values in apps/req.c.
8445 [Nils Larsch <nla@trustcenter.de>]
8447 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8448 an end-of-file condition would erroneously be flagged, when the CRLF
8449 was just at the end of a processed block. The bug was discovered when
8450 processing data through a buffering memory BIO handing the data to a
8451 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8452 <ptsekov@syntrex.com> and Nedelcho Stanev.
8455 *) Implement a countermeasure against a vulnerability recently found
8456 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8457 before application data chunks to avoid the use of known IVs
8458 with data potentially chosen by the attacker.
8461 *) Fix length checks in ssl3_get_client_hello().
8464 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8465 to prevent ssl3_read_internal() from incorrectly assuming that
8466 ssl3_read_bytes() found application data while handshake
8467 processing was enabled when in fact s->s3->in_read_app_data was
8468 merely automatically cleared during the initial handshake.
8469 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8471 *) Fix object definitions for Private and Enterprise: they were not
8472 recognized in their shortname (=lowercase) representation. Extend
8473 obj_dat.pl to issue an error when using undefined keywords instead
8474 of silently ignoring the problem (Svenning Sorensen
8475 <sss@sss.dnsalias.net>).
8478 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8479 generators, i.e. generators other than 2 and 5. (Previously, the
8480 code did not properly initialise the 'add' and 'rem' values to
8481 BN_generate_prime().)
8483 In the new general case, we do not insist that 'generator' is
8484 actually a primitive root: This requirement is rather pointless;
8485 a generator of the order-q subgroup is just as good, if not
8489 *) Map new X509 verification errors to alerts. Discovered and submitted by
8490 Tom Wu <tom@arcot.com>.
8493 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8494 returning non-zero before the data has been completely received
8495 when using non-blocking I/O.
8496 [Bodo Moeller; problem pointed out by John Hughes]
8498 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8499 [Ben Laurie, Lutz Jaenicke]
8501 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8502 Yoram Zahavi <YoramZ@gilian.com>).
8505 *) Add information about CygWin 1.3 and on, and preserve proper
8506 configuration for the versions before that.
8507 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8509 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8510 check whether we deal with a copy of a session and do not delete from
8511 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8512 <izhar@checkpoint.com>.
8515 *) Do not store session data into the internal session cache, if it
8516 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8517 flag is set). Proposed by Aslam <aslam@funk.com>.
8520 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8524 *) [In 0.9.6d-engine release:]
8525 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8526 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8528 *) Add the configuration target linux-s390x.
8529 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8531 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8532 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8533 variable as an indication that a ClientHello message has been
8534 received. As the flag value will be lost between multiple
8535 invocations of ssl3_accept when using non-blocking I/O, the
8536 function may not be aware that a handshake has actually taken
8537 place, thus preventing a new session from being added to the
8540 To avoid this problem, we now set s->new_session to 2 instead of
8541 using a local variable.
8542 [Lutz Jaenicke, Bodo Moeller]
8544 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8545 if the SSL_R_LENGTH_MISMATCH error is detected.
8546 [Geoff Thorpe, Bodo Moeller]
8548 *) New 'shared_ldflag' column in Configure platform table.
8551 *) Fix EVP_CIPHER_mode macro.
8552 ["Dan S. Camper" <dan@bti.net>]
8554 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8555 type, we must throw them away by setting rr->length to 0.
8556 [D P Chang <dpc@qualys.com>]
8558 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8560 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8561 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8562 worked incorrectly for those cases where range = 10..._2 and
8563 3*range is two bits longer than range.)
8566 *) Only add signing time to PKCS7 structures if it is not already
8570 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8571 OBJ_ld_ce should be OBJ_id_ce.
8572 Also some ip-pda OIDs in crypto/objects/objects.txt were
8573 incorrect (cf. RFC 3039).
8574 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8576 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8577 returns early because it has nothing to do.
8578 [Andy Schneider <andy.schneider@bjss.co.uk>]
8580 *) [In 0.9.6c-engine release:]
8581 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8582 [Andy Schneider <andy.schneider@bjss.co.uk>]
8584 *) [In 0.9.6c-engine release:]
8585 Add support for Cryptographic Appliance's keyserver technology.
8586 (Use engine 'keyclient')
8587 [Cryptographic Appliances and Geoff Thorpe]
8589 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8590 is called via tools/c89.sh because arguments have to be
8591 rearranged (all '-L' options must appear before the first object
8593 [Richard Shapiro <rshapiro@abinitio.com>]
8595 *) [In 0.9.6c-engine release:]
8596 Add support for Broadcom crypto accelerator cards, backported
8598 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8600 *) [In 0.9.6c-engine release:]
8601 Add support for SureWare crypto accelerator cards from
8602 Baltimore Technologies. (Use engine 'sureware')
8603 [Baltimore Technologies and Mark Cox]
8605 *) [In 0.9.6c-engine release:]
8606 Add support for crypto accelerator cards from Accelerated
8607 Encryption Processing, www.aep.ie. (Use engine 'aep')
8608 [AEP Inc. and Mark Cox]
8610 *) Add a configuration entry for gcc on UnixWare.
8611 [Gary Benson <gbenson@redhat.com>]
8613 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8614 messages are stored in a single piece (fixed-length part and
8615 variable-length part combined) and fix various bugs found on the way.
8618 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8619 instead. BIO_gethostbyname() does not know what timeouts are
8620 appropriate, so entries would stay in cache even when they have
8622 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8624 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8625 faced with a pathologically small ClientHello fragment that does
8626 not contain client_version: Instead of aborting with an error,
8627 simply choose the highest available protocol version (i.e.,
8628 TLS 1.0 unless it is disabled). In practice, ClientHello
8629 messages are never sent like this, but this change gives us
8630 strictly correct behaviour at least for TLS.
8633 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8634 never resets s->method to s->ctx->method when called from within
8635 one of the SSL handshake functions.
8636 [Bodo Moeller; problem pointed out by Niko Baric]
8638 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8639 (sent using the client's version number) if client_version is
8640 smaller than the protocol version in use. Also change
8641 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8642 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8643 the client will at least see that alert.
8646 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8650 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8651 client receives HelloRequest while in a handshake.
8652 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8654 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8655 should end in 'break', not 'goto end' which circumvents various
8656 cleanups done in state SSL_ST_OK. But session related stuff
8657 must be disabled for SSL_ST_OK in the case that we just sent a
8660 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8661 before just sending a HelloRequest.
8662 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8664 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8665 reveal whether illegal block cipher padding was found or a MAC
8666 verification error occurred. (Neither SSLerr() codes nor alerts
8667 are directly visible to potential attackers, but the information
8668 may leak via logfiles.)
8670 Similar changes are not required for the SSL 2.0 implementation
8671 because the number of padding bytes is sent in clear for SSL 2.0,
8672 and the extra bytes are just ignored. However ssl/s2_pkt.c
8673 failed to verify that the purported number of padding bytes is in
8677 *) Add OpenUNIX-8 support including shared libraries
8678 (Boyd Lynn Gerber <gerberb@zenez.com>).
8681 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8682 'wristwatch attack' using huge encoding parameters (cf.
8683 James H. Manger's CRYPTO 2001 paper). Note that the
8684 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8685 encoding parameters and hence was not vulnerable.
8688 *) BN_sqr() bug fix.
8689 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8691 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8692 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8693 followed by modular reduction.
8694 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8696 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8697 equivalent based on BN_pseudo_rand() instead of BN_rand().
8700 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8701 This function was broken, as the check for a new client hello message
8702 to handle SGC did not allow these large messages.
8703 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8706 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8709 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8710 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8713 *) Rework the configuration and shared library support for Tru64 Unix.
8714 The configuration part makes use of modern compiler features and
8715 still retains old compiler behavior for those that run older versions
8716 of the OS. The shared library support part includes a variant that
8717 uses the RPATH feature, and is available through the special
8718 configuration target "alpha-cc-rpath", which will never be selected
8720 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8722 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8723 with the same message size as in ssl3_get_certificate_request().
8724 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8725 messages might inadvertently be reject as too long.
8726 [Petr Lampa <lampa@fee.vutbr.cz>]
8728 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8731 *) Modified SSL library such that the verify_callback that has been set
8732 specificly for an SSL object with SSL_set_verify() is actually being
8733 used. Before the change, a verify_callback set with this function was
8734 ignored and the verify_callback() set in the SSL_CTX at the time of
8735 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8736 to allow the necessary settings.
8739 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8740 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8741 done automatically (in contradiction to the requirements of the C
8742 standard). This made problems when used from OpenSSH.
8745 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8746 dh->length and always used
8748 BN_rand_range(priv_key, dh->p).
8750 BN_rand_range() is not necessary for Diffie-Hellman, and this
8751 specific range makes Diffie-Hellman unnecessarily inefficient if
8752 dh->length (recommended exponent length) is much smaller than the
8753 length of dh->p. We could use BN_rand_range() if the order of
8754 the subgroup was stored in the DH structure, but we only have
8759 BN_rand(priv_key, l, ...)
8761 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8767 RSA_eay_public_encrypt
8768 RSA_eay_private_decrypt
8769 RSA_eay_private_encrypt (signing)
8770 RSA_eay_public_decrypt (signature verification)
8772 (default implementations for RSA_public_encrypt,
8773 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8774 always reject numbers >= n.
8777 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8778 to synchronize access to 'locking_thread'. This is necessary on
8779 systems where access to 'locking_thread' (an 'unsigned long'
8780 variable) is not atomic.
8783 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8784 *before* setting the 'crypto_lock_rand' flag. The previous code had
8785 a race condition if 0 is a valid thread ID.
8786 [Travis Vitek <vitek@roguewave.com>]
8788 *) Add support for shared libraries under Irix.
8789 [Albert Chin-A-Young <china@thewrittenword.com>]
8791 *) Add configuration option to build on Linux on both big-endian and
8793 [Ralf Baechle <ralf@uni-koblenz.de>]
8795 *) Add the possibility to create shared libraries on HP-UX.
8798 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8800 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8801 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8802 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8803 PRNG state recovery was possible based on the output of
8804 one PRNG request appropriately sized to gain knowledge on
8805 'md' followed by enough consecutive 1-byte PRNG requests
8806 to traverse all of 'state'.
8808 1. When updating 'md_local' (the current thread's copy of 'md')
8809 during PRNG output generation, hash all of the previous
8810 'md_local' value, not just the half used for PRNG output.
8812 2. Make the number of bytes from 'state' included into the hash
8813 independent from the number of PRNG bytes requested.
8815 The first measure alone would be sufficient to avoid
8816 Markku-Juhani's attack. (Actually it had never occurred
8817 to me that the half of 'md_local' used for chaining was the
8818 half from which PRNG output bytes were taken -- I had always
8819 assumed that the secret half would be used.) The second
8820 measure makes sure that additional data from 'state' is never
8821 mixed into 'md_local' in small portions; this heuristically
8822 further strengthens the PRNG.
8825 *) Fix crypto/bn/asm/mips3.s.
8828 *) When only the key is given to "enc", the IV is undefined. Print out
8829 an error message in this case.
8832 *) Handle special case when X509_NAME is empty in X509 printing routines.
8835 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8836 positive and less than q.
8839 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8840 used: it isn't thread safe and the add_lock_callback should handle
8842 [Paul Rose <Paul.Rose@bridge.com>]
8844 *) Verify that incoming data obeys the block size in
8845 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8849 [Ulf Möller, Bodo Möller]
8851 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8852 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8853 when fixing the server behaviour for backwards-compatible 'client
8854 hello' messages. (Note that the attack is impractical against
8855 SSL 3.0 and TLS 1.0 anyway because length and version checking
8856 means that the probability of guessing a valid ciphertext is
8857 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8860 Before 0.9.5, the countermeasure (hide the error by generating a
8861 random 'decryption result') did not work properly because
8862 ERR_clear_error() was missing, meaning that SSL_get_error() would
8863 detect the supposedly ignored error.
8865 Both problems are now fixed.
8868 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8869 (previously it was 1024).
8872 *) Fix for compatibility mode trust settings: ignore trust settings
8873 unless some valid trust or reject settings are present.
8876 *) Fix for blowfish EVP: its a variable length cipher.
8879 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8880 parameters in DSA public key structures and return an error in the
8881 DSA routines if parameters are absent.
8884 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8885 in the current directory if neither $RANDFILE nor $HOME was set.
8886 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8887 caused some confusion to Windows users who haven't defined $HOME.
8888 Thus RAND_file_name() is changed again: e_os.h can define a
8889 DEFAULT_HOME, which will be used if $HOME is not set.
8890 For Windows, we use "C:"; on other platforms, we still require
8891 environment variables.
8893 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8894 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8895 having multiple threads call RAND_poll() concurrently.
8898 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8899 combination of a flag and a thread ID variable.
8900 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8901 flag), *other* threads can enter ssleay_add_bytes without obeying
8902 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8903 that they do not hold after the first thread unsets add_do_not_lock).
8906 *) Change bctest again: '-x' expressions are not available in all
8910 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8912 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8913 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8915 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8916 the default extension for executables, if any. Also, make the perl
8917 scripts that use symlink() to test if it really exists and use "cp"
8918 if it doesn't. All this made OpenSSL compilable and installable in
8922 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8923 If SEQUENCE is length is indefinite just set c->slen to the total
8924 amount of data available.
8925 [Steve Henson, reported by shige@FreeBSD.org]
8926 [This change does not apply to 0.9.7.]
8928 *) Change bctest to avoid here-documents inside command substitution
8929 (workaround for FreeBSD /bin/sh bug).
8930 For compatibility with Ultrix, avoid shell functions (introduced
8931 in the bctest version that searches along $PATH).
8934 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8935 with des_encrypt() defined on some operating systems, like Solaris
8939 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8940 On the Importance of Eliminating Errors in Cryptographic
8941 Computations, J. Cryptology 14 (2001) 2, 101-119,
8942 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8945 *) MIPS assembler BIGNUM division bug fix.
8948 *) Disabled incorrect Alpha assembler code.
8951 *) Fix PKCS#7 decode routines so they correctly update the length
8952 after reading an EOC for the EXPLICIT tag.
8954 [This change does not apply to 0.9.7.]
8956 *) Fix bug in PKCS#12 key generation routines. This was triggered
8957 if a 3DES key was generated with a 0 initial byte. Include
8958 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8959 (but broken) behaviour.
8962 *) Enhance bctest to search for a working bc along $PATH and print
8964 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8966 *) Fix memory leaks in err.c: free err_data string if necessary;
8967 don't write to the wrong index in ERR_set_error_data.
8970 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8974 *) Replace rdtsc with _emit statements for VC++ version 5.
8975 [Jeremy Cooper <jeremy@baymoo.org>]
8977 *) Make it possible to reuse SSLv2 sessions.
8980 *) In copy_email() check for >= 0 as a return value for
8981 X509_NAME_get_index_by_NID() since 0 is a valid index.
8982 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8984 *) Avoid coredump with unsupported or invalid public keys by checking if
8985 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8986 PKCS7_verify() fails with non detached data.
8989 *) Don't use getenv in library functions when run as setuid/setgid.
8990 New function OPENSSL_issetugid().
8993 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8994 due to incorrect handling of multi-threading:
8996 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8998 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9000 3. Count how many times MemCheck_off() has been called so that
9001 nested use can be treated correctly. This also avoids
9002 inband-signalling in the previous code (which relied on the
9003 assumption that thread ID 0 is impossible).
9006 *) Add "-rand" option also to s_client and s_server.
9009 *) Fix CPU detection on Irix 6.x.
9010 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9011 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9013 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9016 [This change does not apply to 0.9.7.]
9018 *) Use the cached encoding of an X509_NAME structure rather than
9019 copying it. This is apparently the reason for the libsafe "errors"
9020 but the code is actually correct.
9023 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9024 Bleichenbacher's DSA attack.
9025 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9026 to be set and top=0 forces the highest bit to be set; top=-1 is new
9027 and leaves the highest bit random.
9028 [Ulf Moeller, Bodo Moeller]
9030 *) In the NCONF_...-based implementations for CONF_... queries
9031 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9032 a temporary CONF structure with the data component set to NULL
9033 (which gives segmentation faults in lh_retrieve).
9034 Instead, use NULL for the CONF pointer in CONF_get_string and
9035 CONF_get_number (which may use environment variables) and directly
9036 return NULL from CONF_get_section.
9039 *) Fix potential buffer overrun for EBCDIC.
9042 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9043 keyUsage if basicConstraints absent for a CA.
9046 *) Make SMIME_write_PKCS7() write mail header values with a format that
9047 is more generally accepted (no spaces before the semicolon), since
9048 some programs can't parse those values properly otherwise. Also make
9049 sure BIO's that break lines after each write do not create invalid
9053 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9054 macros previously used would not encode an empty SEQUENCE OF
9055 and break the signature.
9057 [This change does not apply to 0.9.7.]
9059 *) Zero the premaster secret after deriving the master secret in
9063 *) Add some EVP_add_digest_alias registrations (as found in
9064 OpenSSL_add_all_digests()) to SSL_library_init()
9065 aka OpenSSL_add_ssl_algorithms(). This provides improved
9066 compatibility with peers using X.509 certificates
9067 with unconventional AlgorithmIdentifier OIDs.
9070 *) Fix for Irix with NO_ASM.
9071 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9073 *) ./config script fixes.
9074 [Ulf Moeller, Richard Levitte]
9076 *) Fix 'openssl passwd -1'.
9079 *) Change PKCS12_key_gen_asc() so it can cope with non null
9080 terminated strings whose length is passed in the passlen
9081 parameter, for example from PEM callbacks. This was done
9082 by adding an extra length parameter to asc2uni().
9083 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9085 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9086 call failed, free the DSA structure.
9089 *) Fix to uni2asc() to cope with zero length Unicode strings.
9090 These are present in some PKCS#12 files.
9093 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9094 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9095 when writing a 32767 byte record.
9096 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9098 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9099 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9101 (RSA objects have a reference count access to which is protected
9102 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9103 so they are meant to be shared between threads.)
9104 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9105 "Reddie, Steven" <Steven.Reddie@ca.com>]
9107 *) Fix a deadlock in CRYPTO_mem_leaks().
9110 *) Use better test patterns in bntest.
9113 *) rand_win.c fix for Borland C.
9116 *) BN_rshift bugfix for n == 0.
9119 *) Add a 'bctest' script that checks for some known 'bc' bugs
9120 so that 'make test' does not abort just because 'bc' is broken.
9123 *) Store verify_result within SSL_SESSION also for client side to
9124 avoid potential security hole. (Re-used sessions on the client side
9125 always resulted in verify_result==X509_V_OK, not using the original
9126 result of the server certificate verification.)
9129 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9130 SSL3_RT_APPLICATION_DATA, return 0.
9131 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9135 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9136 releases, have been re-implemented by renaming the previous
9137 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9138 and ssl3_read_internal, respectively, and adding 'peek' parameters
9139 to them. The new ssl[23]_{read,peek} functions are calls to
9140 ssl[23]_read_internal with the 'peek' flag set appropriately.
9141 A 'peek' parameter has also been added to ssl3_read_bytes, which
9142 does the actual work for ssl3_read_internal.
9145 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9146 the method-specific "init()" handler. Also clean up ex_data after
9147 calling the method-specific "finish()" handler. Previously, this was
9148 happening the other way round.
9151 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9152 The previous value, 12, was not always sufficient for BN_mod_exp().
9155 *) Make sure that shared libraries get the internal name engine with
9156 the full version number and not just 0. This should mark the
9157 shared libraries as not backward compatible. Of course, this should
9158 be changed again when we can guarantee backward binary compatibility.
9161 *) Fix typo in get_cert_by_subject() in by_dir.c
9162 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9164 *) Rework the system to generate shared libraries:
9166 - Make note of the expected extension for the shared libraries and
9167 if there is a need for symbolic links from for example libcrypto.so.0
9168 to libcrypto.so.0.9.7. There is extended info in Configure for
9171 - Make as few rebuilds of the shared libraries as possible.
9173 - Still avoid linking the OpenSSL programs with the shared libraries.
9175 - When installing, install the shared libraries separately from the
9179 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9181 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9182 and not in SSL_clear because the latter is also used by the
9183 accept/connect functions; previously, the settings made by
9184 SSL_set_read_ahead would be lost during the handshake.
9185 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9187 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9188 Previously, it would create entries for disabled algorithms no
9192 *) Added several new manual pages for SSL_* function.
9195 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9197 *) In ssl23_get_client_hello, generate an error message when faced
9198 with an initial SSL 3.0/TLS record that is too small to contain the
9199 first two bytes of the ClientHello message, i.e. client_version.
9200 (Note that this is a pathologic case that probably has never happened
9201 in real life.) The previous approach was to use the version number
9202 from the record header as a substitute; but our protocol choice
9203 should not depend on that one because it is not authenticated
9204 by the Finished messages.
9207 *) More robust randomness gathering functions for Windows.
9208 [Jeffrey Altman <jaltman@columbia.edu>]
9210 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9211 not set then we don't setup the error code for issuer check errors
9212 to avoid possibly overwriting other errors which the callback does
9213 handle. If an application does set the flag then we assume it knows
9214 what it is doing and can handle the new informational codes
9218 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9219 a general "ANY" type, as such it should be able to decode anything
9220 including tagged types. However it didn't check the class so it would
9221 wrongly interpret tagged types in the same way as their universal
9222 counterpart and unknown types were just rejected. Changed so that the
9223 tagged and unknown types are handled in the same way as a SEQUENCE:
9224 that is the encoding is stored intact. There is also a new type
9225 "V_ASN1_OTHER" which is used when the class is not universal, in this
9226 case we have no idea what the actual type is so we just lump them all
9230 *) On VMS, stdout may very well lead to a file that is written to
9231 in a record-oriented fashion. That means that every write() will
9232 write a separate record, which will be read separately by the
9233 programs trying to read from it. This can be very confusing.
9235 The solution is to put a BIO filter in the way that will buffer
9236 text until a linefeed is reached, and then write everything a
9237 line at a time, so every record written will be an actual line,
9238 not chunks of lines and not (usually doesn't happen, but I've
9239 seen it once) several lines in one record. BIO_f_linebuffer() is
9242 Currently, it's a VMS-only method, because that's where it has
9243 been tested well enough.
9246 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9247 it can return incorrect results.
9248 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9249 but it was in 0.9.6-beta[12].)
9252 *) Disable the check for content being present when verifying detached
9253 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9254 include zero length content when signing messages.
9257 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9258 BIO_ctrl (for BIO pairs).
9261 *) Add DSO method for VMS.
9264 *) Bug fix: Montgomery multiplication could produce results with the
9268 *) Add RPM specification openssl.spec and modify it to build three
9269 packages. The default package contains applications, application
9270 documentation and run-time libraries. The devel package contains
9271 include files, static libraries and function documentation. The
9272 doc package contains the contents of the doc directory. The original
9273 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9276 *) Add a large number of documentation files for many SSL routines.
9277 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9279 *) Add a configuration entry for Sony News 4.
9280 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9282 *) Don't set the two most significant bits to one when generating a
9283 random number < q in the DSA library.
9286 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9287 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9288 the underlying transport is blocking) if a handshake took place.
9289 (The default behaviour is needed by applications such as s_client
9290 and s_server that use select() to determine when to use SSL_read;
9291 but for applications that know in advance when to expect data, it
9292 just makes things more complicated.)
9295 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9299 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9300 work better on such systems.
9301 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9303 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9304 Update PKCS12_parse() so it copies the friendlyName and the
9305 keyid to the certificates aux info.
9308 *) Fix bug in PKCS7_verify() which caused an infinite loop
9309 if there was more than one signature.
9310 [Sven Uszpelkat <su@celocom.de>]
9312 *) Major change in util/mkdef.pl to include extra information
9313 about each symbol, as well as presenting variables as well
9314 as functions. This change means that there's n more need
9315 to rebuild the .num files when some algorithms are excluded.
9318 *) Allow the verify time to be set by an application,
9319 rather than always using the current time.
9322 *) Phase 2 verify code reorganisation. The certificate
9323 verify code now looks up an issuer certificate by a
9324 number of criteria: subject name, authority key id
9325 and key usage. It also verifies self signed certificates
9326 by the same criteria. The main comparison function is
9327 X509_check_issued() which performs these checks.
9329 Lot of changes were necessary in order to support this
9330 without completely rewriting the lookup code.
9332 Authority and subject key identifier are now cached.
9334 The LHASH 'certs' is X509_STORE has now been replaced
9335 by a STACK_OF(X509_OBJECT). This is mainly because an
9336 LHASH can't store or retrieve multiple objects with
9337 the same hash value.
9339 As a result various functions (which were all internal
9340 use only) have changed to handle the new X509_STORE
9341 structure. This will break anything that messed round
9342 with X509_STORE internally.
9344 The functions X509_STORE_add_cert() now checks for an
9345 exact match, rather than just subject name.
9347 The X509_STORE API doesn't directly support the retrieval
9348 of multiple certificates matching a given criteria, however
9349 this can be worked round by performing a lookup first
9350 (which will fill the cache with candidate certificates)
9351 and then examining the cache for matches. This is probably
9352 the best we can do without throwing out X509_LOOKUP
9353 entirely (maybe later...).
9355 The X509_VERIFY_CTX structure has been enhanced considerably.
9357 All certificate lookup operations now go via a get_issuer()
9358 callback. Although this currently uses an X509_STORE it
9359 can be replaced by custom lookups. This is a simple way
9360 to bypass the X509_STORE hackery necessary to make this
9361 work and makes it possible to use more efficient techniques
9362 in future. A very simple version which uses a simple
9363 STACK for its trusted certificate store is also provided
9364 using X509_STORE_CTX_trusted_stack().
9366 The verify_cb() and verify() callbacks now have equivalents
9367 in the X509_STORE_CTX structure.
9369 X509_STORE_CTX also has a 'flags' field which can be used
9370 to customise the verify behaviour.
9373 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9374 excludes S/MIME capabilities.
9377 *) When a certificate request is read in keep a copy of the
9378 original encoding of the signed data and use it when outputting
9379 again. Signatures then use the original encoding rather than
9380 a decoded, encoded version which may cause problems if the
9381 request is improperly encoded.
9384 *) For consistency with other BIO_puts implementations, call
9385 buffer_write(b, ...) directly in buffer_puts instead of calling
9388 In BIO_puts, increment b->num_write as in BIO_write.
9389 [Peter.Sylvester@EdelWeb.fr]
9391 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9392 BN_zero, we may not return a BIGNUM with an array consisting of
9396 *) Avoid calling abort() from within the library when problems are
9397 detected, except if preprocessor symbols have been defined
9398 (such as REF_CHECK, BN_DEBUG etc.).
9401 *) New openssl application 'rsautl'. This utility can be
9402 used for low level RSA operations. DER public key
9403 BIO/fp routines also added.
9406 *) New Configure entry and patches for compiling on QNX 4.
9407 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9409 *) A demo state-machine implementation was sponsored by
9410 Nuron (http://www.nuron.com/) and is now available in
9411 demos/state_machine.
9414 *) New options added to the 'dgst' utility for signature
9415 generation and verification.
9418 *) Unrecognized PKCS#7 content types are now handled via a
9419 catch all ASN1_TYPE structure. This allows unsupported
9420 types to be stored as a "blob" and an application can
9421 encode and decode it manually.
9424 *) Fix various signed/unsigned issues to make a_strex.c
9426 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9428 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9429 length if passed a buffer. ASN1_INTEGER_to_BN failed
9430 if passed a NULL BN and its argument was negative.
9431 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9433 *) Modification to PKCS#7 encoding routines to output definite
9434 length encoding. Since currently the whole structures are in
9435 memory there's not real point in using indefinite length
9436 constructed encoding. However if OpenSSL is compiled with
9437 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9440 *) Added BIO_vprintf() and BIO_vsnprintf().
9443 *) Added more prefixes to parse for in the strings written
9444 through a logging bio, to cover all the levels that are available
9445 through syslog. The prefixes are now:
9447 PANIC, EMERG, EMR => LOG_EMERG
9448 ALERT, ALR => LOG_ALERT
9449 CRIT, CRI => LOG_CRIT
9450 ERROR, ERR => LOG_ERR
9451 WARNING, WARN, WAR => LOG_WARNING
9452 NOTICE, NOTE, NOT => LOG_NOTICE
9453 INFO, INF => LOG_INFO
9454 DEBUG, DBG => LOG_DEBUG
9456 and as before, if none of those prefixes are present at the
9457 beginning of the string, LOG_ERR is chosen.
9459 On Win32, the LOG_* levels are mapped according to this:
9461 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9462 LOG_WARNING => EVENTLOG_WARNING_TYPE
9463 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9467 *) Made it possible to reconfigure with just the configuration
9468 argument "reconf" or "reconfigure". The command line arguments
9469 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9470 and are retrieved from there when reconfiguring.
9474 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9476 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9479 *) The obj_dat.pl script was messing up the sorting of object
9480 names. The reason was that it compared the quoted version
9481 of strings as a result "OCSP" > "OCSP Signing" because
9482 " > SPACE. Changed script to store unquoted versions of
9483 names and add quotes on output. It was also omitting some
9484 names from the lookup table if they were given a default
9485 value (that is if SN is missing it is given the same
9486 value as LN and vice versa), these are now added on the
9487 grounds that if an object has a name we should be able to
9488 look it up. Finally added warning output when duplicate
9489 short or long names are found.
9492 *) Changes needed for Tandem NSK.
9493 [Scott Uroff <scott@xypro.com>]
9495 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9496 RSA_padding_check_SSLv23(), special padding was never detected
9497 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9498 version rollback attacks was not effective.
9500 In s23_clnt.c, don't use special rollback-attack detection padding
9501 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9502 client; similarly, in s23_srvr.c, don't do the rollback check if
9503 SSL 2.0 is the only protocol enabled in the server.
9506 *) Make it possible to get hexdumps of unprintable data with 'openssl
9507 asn1parse'. By implication, the functions ASN1_parse_dump() and
9508 BIO_dump_indent() are added.
9511 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9512 these print out strings and name structures based on various
9513 flags including RFC2253 support and proper handling of
9514 multibyte characters. Added options to the 'x509' utility
9515 to allow the various flags to be set.
9518 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9519 Also change the functions X509_cmp_current_time() and
9520 X509_gmtime_adj() work with an ASN1_TIME structure,
9521 this will enable certificates using GeneralizedTime in validity
9522 dates to be checked.
9525 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9526 negative public key encodings) on by default,
9527 NO_NEG_PUBKEY_BUG can be set to disable it.
9530 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9531 content octets. An i2c_ASN1_OBJECT is unnecessary because
9532 the encoding can be trivially obtained from the structure.
9535 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9536 not read locks (CRYPTO_r_[un]lock).
9539 *) A first attempt at creating official support for shared
9540 libraries through configuration. I've kept it so the
9541 default is static libraries only, and the OpenSSL programs
9542 are always statically linked for now, but there are
9543 preparations for dynamic linking in place.
9544 This has been tested on Linux and Tru64.
9547 *) Randomness polling function for Win9x, as described in:
9548 Peter Gutmann, Software Generation of Practically Strong
9552 *) Fix so PRNG is seeded in req if using an already existing
9556 *) New options to smime application. -inform and -outform
9557 allow alternative formats for the S/MIME message including
9558 PEM and DER. The -content option allows the content to be
9559 specified separately. This should allow things like Netscape
9560 form signing output easier to verify.
9563 *) Fix the ASN1 encoding of tags using the 'long form'.
9566 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9567 STRING types. These convert content octets to and from the
9568 underlying type. The actual tag and length octets are
9569 already assumed to have been read in and checked. These
9570 are needed because all other string types have virtually
9571 identical handling apart from the tag. By having versions
9572 of the ASN1 functions that just operate on content octets
9573 IMPLICIT tagging can be handled properly. It also allows
9574 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9575 and ASN1_INTEGER are identical apart from the tag.
9578 *) Change the handling of OID objects as follows:
9580 - New object identifiers are inserted in objects.txt, following
9581 the syntax given in objects.README.
9582 - objects.pl is used to process obj_mac.num and create a new
9584 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9587 This is currently kind of a hack, and the perl code in objects.pl
9588 isn't very elegant, but it works as I intended. The simplest way
9589 to check that it worked correctly is to look in obj_dat.h and
9590 check the array nid_objs and make sure the objects haven't moved
9591 around (this is important!). Additions are OK, as well as
9592 consistent name changes.
9595 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9598 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9599 The given file adds to whatever has already been seeded into the
9600 random pool through the RANDFILE configuration file option or
9601 environment variable, or the default random state file.
9604 *) mkstack.pl now sorts each macro group into lexical order.
9605 Previously the output order depended on the order the files
9606 appeared in the directory, resulting in needless rewriting
9610 *) Patches to make OpenSSL compile under Win32 again. Mostly
9611 work arounds for the VC++ problem that it treats func() as
9612 func(void). Also stripped out the parts of mkdef.pl that
9613 added extra typesafe functions: these no longer exist.
9616 *) Reorganisation of the stack code. The macros are now all
9617 collected in safestack.h . Each macro is defined in terms of
9618 a "stack macro" of the form SKM_<name>(type, a, b). The
9619 DEBUG_SAFESTACK is now handled in terms of function casts,
9620 this has the advantage of retaining type safety without the
9621 use of additional functions. If DEBUG_SAFESTACK is not defined
9622 then the non typesafe macros are used instead. Also modified the
9623 mkstack.pl script to handle the new form. Needs testing to see
9624 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9625 the default if no major problems. Similar behaviour for ASN1_SET_OF
9626 and PKCS12_STACK_OF.
9629 *) When some versions of IIS use the 'NET' form of private key the
9630 key derivation algorithm is different. Normally MD5(password) is
9631 used as a 128 bit RC4 key. In the modified case
9632 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9633 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9634 as the old Netscape_RSA functions except they have an additional
9635 'sgckey' parameter which uses the modified algorithm. Also added
9636 an -sgckey command line option to the rsa utility. Thanks to
9637 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9638 algorithm to openssl-dev.
9641 *) The evp_local.h macros were using 'c.##kname' which resulted in
9642 invalid expansion on some systems (SCO 5.0.5 for example).
9643 Corrected to 'c.kname'.
9644 [Phillip Porch <root@theporch.com>]
9646 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9647 a STACK of email addresses from a certificate or request, these look
9648 in the subject name and the subject alternative name extensions and
9649 omit any duplicate addresses.
9652 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9653 This makes DSA verification about 2 % faster.
9656 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9657 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9658 plus overhead for 1024 bit moduli).
9659 This makes exponentiations about 0.5 % faster for 1024 bit
9660 exponents (as measured by "openssl speed rsa2048").
9663 *) Rename memory handling macros to avoid conflicts with other
9665 Malloc => OPENSSL_malloc
9666 Malloc_locked => OPENSSL_malloc_locked
9667 Realloc => OPENSSL_realloc
9668 Free => OPENSSL_free
9671 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9672 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9675 *) CygWin32 support.
9676 [John Jarvie <jjarvie@newsguy.com>]
9678 *) The type-safe stack code has been rejigged. It is now only compiled
9679 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9680 by default all type-specific stack functions are "#define"d back to
9681 standard stack functions. This results in more streamlined output
9682 but retains the type-safety checking possibilities of the original
9686 *) The STACK code has been cleaned up, and certain type declarations
9687 that didn't make a lot of sense have been brought in line. This has
9688 also involved a cleanup of sorts in safestack.h to more correctly
9689 map type-safe stack functions onto their plain stack counterparts.
9690 This work has also resulted in a variety of "const"ifications of
9691 lots of the code, especially "_cmp" operations which should normally
9692 be prototyped with "const" parameters anyway.
9695 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9696 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9697 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9698 where all of 'md' is used each time the PRNG is used, but 'state'
9699 is used only indexed by a cyclic counter. As entropy may not be
9700 well distributed from the beginning, 'md' is important as a
9701 chaining variable. However, the output function chains only half
9702 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9703 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9704 in all of 'state' being rewritten, with the new values depending
9705 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9708 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9709 the handshake is continued after ssl_verify_cert_chain();
9710 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9711 can lead to 'unexplainable' connection aborts later.
9712 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9714 *) Major EVP API cipher revision.
9715 Add hooks for extra EVP features. This allows various cipher
9716 parameters to be set in the EVP interface. Support added for variable
9717 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9718 setting of RC2 and RC5 parameters.
9720 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9723 Remove lots of duplicated code from the EVP library. For example *every*
9724 cipher init() function handles the 'iv' in the same way according to the
9725 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9726 for CFB and OFB modes they zero ctx->num.
9728 New functionality allows removal of S/MIME code RC2 hack.
9730 Most of the routines have the same form and so can be declared in terms
9733 By shifting this to the top level EVP_CipherInit() it can be removed from
9734 all individual ciphers. If the cipher wants to handle IVs or keys
9735 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9738 Change lots of functions like EVP_EncryptUpdate() to now return a
9739 value: although software versions of the algorithms cannot fail
9740 any installed hardware versions can.
9743 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9744 this option is set, tolerate broken clients that send the negotiated
9745 protocol version number instead of the requested protocol version
9749 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9750 i.e. non-zero for export ciphersuites, zero otherwise.
9751 Previous versions had this flag inverted, inconsistent with
9752 rsa_tmp_cb (..._TMP_RSA_CB).
9753 [Bodo Moeller; problem reported by Amit Chopra]
9755 *) Add missing DSA library text string. Work around for some IIS
9756 key files with invalid SEQUENCE encoding.
9759 *) Add a document (doc/standards.txt) that list all kinds of standards
9760 and so on that are implemented in OpenSSL.
9763 *) Enhance c_rehash script. Old version would mishandle certificates
9764 with the same subject name hash and wouldn't handle CRLs at all.
9765 Added -fingerprint option to crl utility, to support new c_rehash
9769 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9772 *) Fix for SSL server purpose checking. Server checking was
9773 rejecting certificates which had extended key usage present
9774 but no ssl client purpose.
9775 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9777 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9778 is a little unclear about how a blank password is handled.
9779 Since the password in encoded as a BMPString with terminating
9780 double NULL a zero length password would end up as just the
9781 double NULL. However no password at all is different and is
9782 handled differently in the PKCS#12 key generation code. NS
9783 treats a blank password as zero length. MSIE treats it as no
9784 password on export: but it will try both on import. We now do
9785 the same: PKCS12_parse() tries zero length and no password if
9786 the password is set to "" or NULL (NULL is now a valid password:
9787 it wasn't before) as does the pkcs12 application.
9790 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9791 perror when PEM_read_bio_X509_REQ fails, the error message must
9792 be obtained from the error queue.
9795 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9796 it in ERR_remove_state if appropriate, and change ERR_get_state
9797 accordingly to avoid race conditions (this is necessary because
9798 thread_hash is no longer constant once set).
9801 *) Bugfix for linux-elf makefile.one.
9804 *) RSA_get_default_method() will now cause a default
9805 RSA_METHOD to be chosen if one doesn't exist already.
9806 Previously this was only set during a call to RSA_new()
9807 or RSA_new_method(NULL) meaning it was possible for
9808 RSA_get_default_method() to return NULL.
9811 *) Added native name translation to the existing DSO code
9812 that will convert (if the flag to do so is set) filenames
9813 that are sufficiently small and have no path information
9814 into a canonical native form. Eg. "blah" converted to
9815 "libblah.so" or "blah.dll" etc.
9818 *) New function ERR_error_string_n(e, buf, len) which is like
9819 ERR_error_string(e, buf), but writes at most 'len' bytes
9820 including the 0 terminator. For ERR_error_string_n, 'buf'
9822 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9824 *) CONF library reworked to become more general. A new CONF
9825 configuration file reader "class" is implemented as well as a
9826 new functions (NCONF_*, for "New CONF") to handle it. The now
9827 old CONF_* functions are still there, but are reimplemented to
9828 work in terms of the new functions. Also, a set of functions
9829 to handle the internal storage of the configuration data is
9830 provided to make it easier to write new configuration file
9831 reader "classes" (I can definitely see something reading a
9832 configuration file in XML format, for example), called _CONF_*,
9833 or "the configuration storage API"...
9835 The new configuration file reading functions are:
9837 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9838 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9840 NCONF_default, NCONF_WIN32
9842 NCONF_dump_fp, NCONF_dump_bio
9844 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9845 NCONF_new creates a new CONF object. This works in the same way
9846 as other interfaces in OpenSSL, like the BIO interface.
9847 NCONF_dump_* dump the internal storage of the configuration file,
9848 which is useful for debugging. All other functions take the same
9849 arguments as the old CONF_* functions wth the exception of the
9850 first that must be a `CONF *' instead of a `LHASH *'.
9852 To make it easer to use the new classes with the old CONF_* functions,
9853 the function CONF_set_default_method is provided.
9856 *) Add '-tls1' option to 'openssl ciphers', which was already
9857 mentioned in the documentation but had not been implemented.
9858 (This option is not yet really useful because even the additional
9859 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9862 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9863 OpenSSL-based applications) load shared libraries and bind to
9864 them in a portable way.
9865 [Geoff Thorpe, with contributions from Richard Levitte]
9867 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9869 *) Make sure _lrotl and _lrotr are only used with MSVC.
9871 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9872 (the default implementation of RAND_status).
9874 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9875 to '-clrext' (= clear extensions), as intended and documented.
9876 [Bodo Moeller; inconsistency pointed out by Michael Attili
9879 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9880 was larger than the MD block size.
9881 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9883 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9884 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9885 using the passed key: if the passed key was a private key the result
9886 of X509_print(), for example, would be to print out all the private key
9890 *) des_quad_cksum() byte order bug fix.
9891 [Ulf Möller, using the problem description in krb4-0.9.7, where
9892 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9894 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9896 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9898 *) For easily testing in shell scripts whether some command
9899 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9900 returns with exit code 0 iff no command of the given name is available.
9901 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9902 the output goes to stdout and nothing is printed to stderr.
9903 Additional arguments are always ignored.
9905 Since for each cipher there is a command of the same name,
9906 the 'no-cipher' compilation switches can be tested this way.
9908 ('openssl no-XXX' is not able to detect pseudo-commands such
9909 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9912 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9915 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9916 is set; it will be thrown away anyway because each handshake creates
9918 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9919 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9920 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9921 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
9924 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9925 'Q' and 'R' lose their special meanings (quit/renegotiate).
9926 This is part of what -quiet does; unlike -quiet, -ign_eof
9927 does not suppress any output.
9930 *) Add compatibility options to the purpose and trust code. The
9931 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9932 accepts a certificate or CA, this was the previous behaviour,
9933 with all the associated security issues.
9935 X509_TRUST_COMPAT is the old trust behaviour: only and
9936 automatically trust self signed roots in certificate store. A
9937 new trust setting X509_TRUST_DEFAULT is used to specify that
9938 a purpose has no associated trust setting and it should instead
9939 use the value in the default purpose.
9942 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9943 and fix a memory leak.
9946 *) In util/mkerr.pl (which implements 'make errors'), preserve
9947 reason strings from the previous version of the .c file, as
9948 the default to have only downcase letters (and digits) in
9949 automatically generated reasons codes is not always appropriate.
9952 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9953 using strerror. Previously, ERR_reason_error_string() returned
9954 library names as reason strings for SYSerr; but SYSerr is a special
9955 case where small numbers are errno values, not library numbers.
9958 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9959 converts DSA parameters into DH parameters. (When creating parameters,
9960 DSA_generate_parameters is used.)
9963 *) Include 'length' (recommended exponent length) in C code generated
9964 by 'openssl dhparam -C'.
9967 *) The second argument to set_label in perlasm was already being used
9968 so couldn't be used as a "file scope" flag. Moved to third argument
9972 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9973 instead of RAND_bytes for encryption IVs and salts.
9976 *) Include RAND_status() into RAND_METHOD instead of implementing
9977 it only for md_rand.c Otherwise replacing the PRNG by calling
9978 RAND_set_rand_method would be impossible.
9981 *) Don't let DSA_generate_key() enter an infinite loop if the random
9982 number generation fails.
9985 *) New 'rand' application for creating pseudo-random output.
9988 *) Added configuration support for Linux/IA64
9989 [Rolf Haberrecker <rolf@suse.de>]
9991 *) Assembler module support for Mingw32.
9994 *) Shared library support for HPUX (in shlib/).
9995 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9997 *) Shared library support for Solaris gcc.
9998 [Lutz Behnke <behnke@trustcenter.de>]
10000 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10002 *) PKCS7_encrypt() was adding text MIME headers twice because they
10003 were added manually and by SMIME_crlf_copy().
10006 *) In bntest.c don't call BN_rand with zero bits argument.
10007 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10009 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10010 case was implemented. This caused BN_div_recp() to fail occasionally.
10013 *) Add an optional second argument to the set_label() in the perl
10014 assembly language builder. If this argument exists and is set
10015 to 1 it signals that the assembler should use a symbol whose
10016 scope is the entire file, not just the current function. This
10017 is needed with MASM which uses the format label:: for this scope.
10018 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10020 *) Change the ASN1 types so they are typedefs by default. Before
10021 almost all types were #define'd to ASN1_STRING which was causing
10022 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10026 *) Change names of new functions to the new get1/get0 naming
10027 convention: After 'get1', the caller owns a reference count
10028 and has to call ..._free; 'get0' returns a pointer to some
10029 data structure without incrementing reference counters.
10030 (Some of the existing 'get' functions increment a reference
10031 counter, some don't.)
10032 Similarly, 'set1' and 'add1' functions increase reference
10033 counters or duplicate objects.
10036 *) Allow for the possibility of temp RSA key generation failure:
10037 the code used to assume it always worked and crashed on failure.
10040 *) Fix potential buffer overrun problem in BIO_printf().
10041 [Ulf Möller, using public domain code by Patrick Powell; problem
10042 pointed out by David Sacerdote <das33@cornell.edu>]
10044 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10045 RAND_egd() and RAND_status(). In the command line application,
10046 the EGD socket can be specified like a seed file using RANDFILE
10050 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10051 Some CAs (e.g. Verisign) distribute certificates in this form.
10054 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10055 list to exclude them. This means that no special compilation option
10056 is needed to use anonymous DH: it just needs to be included in the
10060 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10061 EVP_MD_type. The old functionality is available in a new macro called
10062 EVP_MD_md(). Change code that uses it and update docs.
10065 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10066 where the 'void *' argument is replaced by a function pointer argument.
10067 Previously 'void *' was abused to point to functions, which works on
10068 many platforms, but is not correct. As these functions are usually
10069 called by macros defined in OpenSSL header files, most source code
10070 should work without changes.
10073 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10074 sections with information on -D... compiler switches used for
10075 compiling the library so that applications can see them. To enable
10076 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10077 must be defined. E.g.,
10078 #define OPENSSL_ALGORITHM_DEFINES
10079 #include <openssl/opensslconf.h>
10080 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10081 [Richard Levitte, Ulf and Bodo Möller]
10083 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10087 *) Change the 'other' type in certificate aux info to a STACK_OF
10088 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10089 the required ASN1 format: arbitrary types determined by an OID.
10092 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10093 argument to 'req'. This is not because the function is newer or
10094 better than others it just uses the work 'NEW' in the certificate
10095 request header lines. Some software needs this.
10098 *) Reorganise password command line arguments: now passwords can be
10099 obtained from various sources. Delete the PEM_cb function and make
10100 it the default behaviour: i.e. if the callback is NULL and the
10101 usrdata argument is not NULL interpret it as a null terminated pass
10102 phrase. If usrdata and the callback are NULL then the pass phrase
10103 is prompted for as usual.
10106 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10107 the support is automatically enabled. The resulting binaries will
10108 autodetect the card and use it if present.
10109 [Ben Laurie and Compaq Inc.]
10111 *) Work around for Netscape hang bug. This sends certificate request
10112 and server done in one record. Since this is perfectly legal in the
10113 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10114 the bugs/SSLv3 entry for more info.
10117 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10120 *) Add -rand argument to smime and pkcs12 applications and read/write
10124 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10127 *) Add command line password options to the remaining applications.
10130 *) Bug fix for BN_div_recp() for numerators with an even number of
10134 *) More tests in bntest.c, and changed test_bn output.
10137 *) ./config recognizes MacOS X now.
10140 *) Bug fix for BN_div() when the first words of num and divisor are
10141 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10144 *) Add support for various broken PKCS#8 formats, and command line
10145 options to produce them.
10148 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10149 get temporary BIGNUMs from a BN_CTX.
10152 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10156 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10157 include a #define from the old name to the new. The original intent
10158 was that statically linked binaries could for example just call
10159 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10160 link with digests. This never worked because SSLeay_add_all_digests()
10161 and SSLeay_add_all_ciphers() were in the same source file so calling
10162 one would link with the other. They are now in separate source files.
10165 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10168 *) Use a less unusual form of the Miller-Rabin primality test (it used
10169 a binary algorithm for exponentiation integrated into the Miller-Rabin
10170 loop, our standard modexp algorithms are faster).
10173 *) Support for the EBCDIC character set completed.
10174 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10176 *) Source code cleanups: use const where appropriate, eliminate casts,
10177 use void * instead of char * in lhash.
10180 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10181 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10182 this the server could overwrite ephemeral keys that the client
10186 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10187 using 50 iterations of the Rabin-Miller test.
10189 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10190 iterations of the Rabin-Miller test as required by the appendix
10191 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10192 As BN_is_prime_fasttest includes trial division, DSA parameter
10193 generation becomes much faster.
10195 This implies a change for the callback functions in DSA_is_prime
10196 and DSA_generate_parameters: The callback function is called once
10197 for each positive witness in the Rabin-Miller test, not just
10198 occasionally in the inner loop; and the parameters to the
10199 callback function now provide an iteration count for the outer
10200 loop rather than for the current invocation of the inner loop.
10201 DSA_generate_parameters additionally can call the callback
10202 function with an 'iteration count' of -1, meaning that a
10203 candidate has passed the trial division test (when q is generated
10204 from an application-provided seed, trial division is skipped).
10207 *) New function BN_is_prime_fasttest that optionally does trial
10208 division before starting the Rabin-Miller test and has
10209 an additional BN_CTX * argument (whereas BN_is_prime always
10210 has to allocate at least one BN_CTX).
10211 'callback(1, -1, cb_arg)' is called when a number has passed the
10212 trial division stage.
10215 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10219 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10222 *) New function BN_pseudo_rand().
10225 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10226 bignum version of BN_from_montgomery() with the working code from
10227 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10231 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10232 made it impossible to use the same SSL_SESSION data structure in
10233 SSL2 clients in multiple threads.
10236 *) The return value of RAND_load_file() no longer counts bytes obtained
10237 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10238 to seed the PRNG (previously an explicit byte count was required).
10239 [Ulf Möller, Bodo Möller]
10241 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10242 used (char *) instead of (void *) and had casts all over the place.
10245 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10248 *) Retain source code compatibility for BN_prime_checks macro:
10249 BN_is_prime(..., BN_prime_checks, ...) now uses
10250 BN_prime_checks_for_size to determine the appropriate number of
10251 Rabin-Miller iterations.
10254 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10255 DH_CHECK_P_NOT_SAFE_PRIME.
10256 (Check if this is true? OpenPGP calls them "strong".)
10259 *) Merge the functionality of "dh" and "gendh" programs into a new program
10260 "dhparam". The old programs are retained for now but will handle DH keys
10261 (instead of parameters) in future.
10264 *) Make the ciphers, s_server and s_client programs check the return values
10265 when a new cipher list is set.
10268 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10269 ciphers. Before when the 56bit ciphers were enabled the sorting was
10272 The syntax for the cipher sorting has been extended to support sorting by
10273 cipher-strength (using the strength_bits hard coded in the tables).
10274 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10276 Fix a bug in the cipher-command parser: when supplying a cipher command
10277 string with an "undefined" symbol (neither command nor alphanumeric
10278 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10279 an error is flagged.
10281 Due to the strength-sorting extension, the code of the
10282 ssl_create_cipher_list() function was completely rearranged. I hope that
10283 the readability was also increased :-)
10284 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10286 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10287 for the first serial number and places 2 in the serial number file. This
10288 avoids problems when the root CA is created with serial number zero and
10289 the first user certificate has the same issuer name and serial number
10293 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10294 the new code. Add documentation for this stuff.
10297 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10298 X509_*() to X509at_*() on the grounds that they don't handle X509
10299 structures and behave in an analogous way to the X509v3 functions:
10300 they shouldn't be called directly but wrapper functions should be used
10303 So we also now have some wrapper functions that call the X509at functions
10304 when passed certificate requests. (TO DO: similar things can be done with
10305 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10306 things. Some of these need some d2i or i2d and print functionality
10307 because they handle more complex structures.)
10310 *) Add missing #ifndefs that caused missing symbols when building libssl
10311 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10312 NO_RSA in ssl/s2*.c.
10313 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10315 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10316 has a return value which indicates the quality of the random data
10317 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10318 error queue. New function RAND_pseudo_bytes() generates output that is
10319 guaranteed to be unique but not unpredictable. RAND_add is like
10320 RAND_seed, but takes an extra argument for an entropy estimate
10321 (RAND_seed always assumes full entropy).
10324 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10325 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10326 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10327 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10328 false-positive rate of at most 2^-80 for random input.
10331 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10334 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10335 in the 0.9.5 release), this returns the chain
10336 from an X509_CTX structure with a dup of the stack and all
10337 the X509 reference counts upped: so the stack will exist
10338 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10341 Also make SSL_SESSION_print() print out the verify return
10345 *) Add manpage for the pkcs12 command. Also change the default
10346 behaviour so MAC iteration counts are used unless the new
10347 -nomaciter option is used. This improves file security and
10348 only older versions of MSIE (4.0 for example) need it.
10351 *) Honor the no-xxx Configure options when creating .DEF files.
10354 *) Add PKCS#10 attributes to field table: challengePassword,
10355 unstructuredName and unstructuredAddress. These are taken from
10356 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10357 international characters are used.
10359 More changes to X509_ATTRIBUTE code: allow the setting of types
10360 based on strings. Remove the 'loc' parameter when adding
10361 attributes because these will be a SET OF encoding which is sorted
10365 *) Initial changes to the 'req' utility to allow request generation
10366 automation. This will allow an application to just generate a template
10367 file containing all the field values and have req construct the
10370 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10371 used all over the place including certificate requests and PKCS#7
10372 structures. They are currently handled manually where necessary with
10373 some primitive wrappers for PKCS#7. The new functions behave in a
10374 manner analogous to the X509 extension functions: they allow
10375 attributes to be looked up by NID and added.
10377 Later something similar to the X509V3 code would be desirable to
10378 automatically handle the encoding, decoding and printing of the
10379 more complex types. The string types like challengePassword can
10380 be handled by the string table functions.
10382 Also modified the multi byte string table handling. Now there is
10383 a 'global mask' which masks out certain types. The table itself
10384 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10385 is useful when for example there is only one permissible type
10386 (as in countryName) and using the mask might result in no valid
10390 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10391 SSL_get_peer_finished to allow applications to obtain the latest
10392 Finished messages sent to the peer or expected from the peer,
10393 respectively. (SSL_get_peer_finished is usually the Finished message
10394 actually received from the peer, otherwise the protocol will be aborted.)
10396 As the Finished message are message digests of the complete handshake
10397 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10398 be used for external authentication procedures when the authentication
10399 provided by SSL/TLS is not desired or is not enough.
10402 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10403 the host supports BWX extension and if Compaq C is present on the
10404 $PATH. Just exploiting of the BWX extension results in 20-30%
10405 performance kick for some algorithms, e.g. DES and RC4 to mention
10406 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10410 *) Add support for MS "fast SGC". This is arguably a violation of the
10411 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10412 weak crypto and after checking the certificate is SGC a second one
10413 with strong crypto. MS SGC stops the first handshake after receiving
10414 the server certificate message and sends a second client hello. Since
10415 a server will typically do all the time consuming operations before
10416 expecting any further messages from the client (server key exchange
10417 is the most expensive) there is little difference between the two.
10419 To get OpenSSL to support MS SGC we have to permit a second client
10420 hello message after we have sent server done. In addition we have to
10421 reset the MAC if we do get this second client hello.
10424 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10425 if a DER encoded private key is RSA or DSA traditional format. Changed
10426 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10427 format DER encoded private key. Newer code should use PKCS#8 format which
10428 has the key type encoded in the ASN1 structure. Added DER private key
10429 support to pkcs8 application.
10432 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10433 ciphersuites has been selected (as required by the SSL 3/TLS 1
10434 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10435 is set, we interpret this as a request to violate the specification
10436 (the worst that can happen is a handshake failure, and 'correct'
10437 behaviour would result in a handshake failure anyway).
10440 *) In SSL_CTX_add_session, take into account that there might be multiple
10441 SSL_SESSION structures with the same session ID (e.g. when two threads
10442 concurrently obtain them from an external cache).
10443 The internal cache can handle only one SSL_SESSION with a given ID,
10444 so if there's a conflict, we now throw out the old one to achieve
10448 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10449 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10450 some routines that use cipher OIDs: some ciphers do not have OIDs
10451 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10455 *) Simplify the trust setting structure and code. Now we just have
10456 two sequences of OIDs for trusted and rejected settings. These will
10457 typically have values the same as the extended key usage extension
10458 and any application specific purposes.
10460 The trust checking code now has a default behaviour: it will just
10461 check for an object with the same NID as the passed id. Functions can
10462 be provided to override either the default behaviour or the behaviour
10463 for a given id. SSL client, server and email already have functions
10464 in place for compatibility: they check the NID and also return "trusted"
10465 if the certificate is self signed.
10468 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10469 traditional format into an EVP_PKEY structure.
10472 *) Add a password callback function PEM_cb() which either prompts for
10473 a password if usr_data is NULL or otherwise assumes it is a null
10474 terminated password. Allow passwords to be passed on command line
10475 environment or config files in a few more utilities.
10478 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10479 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10480 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10481 Update documentation.
10484 *) Support for ASN1 "NULL" type. This could be handled before by using
10485 ASN1_TYPE but there wasn't any function that would try to read a NULL
10486 and produce an error if it couldn't. For compatibility we also have
10487 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10488 don't allocate anything because they don't need to.
10491 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10493 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10495 *) Rebuild of the memory allocation routines used by OpenSSL code and
10496 possibly others as well. The purpose is to make an interface that
10497 provide hooks so anyone can build a separate set of allocation and
10498 deallocation routines to be used by OpenSSL, for example memory
10499 pool implementations, or something else, which was previously hard
10500 since Malloc(), Realloc() and Free() were defined as macros having
10501 the values malloc, realloc and free, respectively (except for Win32
10502 compilations). The same is provided for memory debugging code.
10503 OpenSSL already comes with functionality to find memory leaks, but
10504 this gives people a chance to debug other memory problems.
10506 With these changes, a new set of functions and macros have appeared:
10508 CRYPTO_set_mem_debug_functions() [F]
10509 CRYPTO_get_mem_debug_functions() [F]
10510 CRYPTO_dbg_set_options() [F]
10511 CRYPTO_dbg_get_options() [F]
10512 CRYPTO_malloc_debug_init() [M]
10514 The memory debug functions are NULL by default, unless the library
10515 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10516 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10517 gives the standard debugging functions that come with OpenSSL) or
10518 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10519 provided by the library user) must be used. When the standard
10520 debugging functions are used, CRYPTO_dbg_set_options can be used to
10521 request additional information:
10522 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10523 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10525 Also, things like CRYPTO_set_mem_functions will always give the
10526 expected result (the new set of functions is used for allocation
10527 and deallocation) at all times, regardless of platform and compiler
10530 To finish it up, some functions that were never use in any other
10531 way than through macros have a new API and new semantic:
10533 CRYPTO_dbg_malloc()
10534 CRYPTO_dbg_realloc()
10537 All macros of value have retained their old syntax.
10538 [Richard Levitte and Bodo Moeller]
10540 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10541 ordering of SMIMECapabilities wasn't in "strength order" and there
10542 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10546 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10547 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10548 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10550 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10551 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10552 functionality to handle multipart/signed properly) and a utility
10553 called 'smime' to call all this stuff. This is based on code I
10554 originally wrote for Celo who have kindly allowed it to be
10555 included in OpenSSL.
10558 *) Add variants des_set_key_checked and des_set_key_unchecked of
10559 des_set_key (aka des_key_sched). Global variable des_check_key
10560 decides which of these is called by des_set_key; this way
10561 des_check_key behaves as it always did, but applications and
10562 the library itself, which was buggy for des_check_key == 1,
10563 have a cleaner way to pick the version they need.
10566 *) New function PKCS12_newpass() which changes the password of a
10570 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10571 dynamic mix. In both cases the ids can be used as an index into the
10572 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10573 functions so they accept a list of the field values and the
10574 application doesn't need to directly manipulate the X509_TRUST
10578 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10582 *) Modify the way the V3 extension code looks up extensions. This now
10583 works in a similar way to the object code: we have some "standard"
10584 extensions in a static table which is searched with OBJ_bsearch()
10585 and the application can add dynamic ones if needed. The file
10586 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10587 updated whenever a new extension is added to the core code and kept
10588 in ext_nid order. There is a simple program 'tabtest.c' which checks
10589 this. New extensions are not added too often so this file can readily
10590 be maintained manually.
10592 There are two big advantages in doing things this way. The extensions
10593 can be looked up immediately and no longer need to be "added" using
10594 X509V3_add_standard_extensions(): this function now does nothing.
10595 [Side note: I get *lots* of email saying the extension code doesn't
10596 work because people forget to call this function]
10597 Also no dynamic allocation is done unless new extensions are added:
10598 so if we don't add custom extensions there is no need to call
10599 X509V3_EXT_cleanup().
10602 *) Modify enc utility's salting as follows: make salting the default. Add a
10603 magic header, so unsalted files fail gracefully instead of just decrypting
10604 to garbage. This is because not salting is a big security hole, so people
10605 should be discouraged from doing it.
10608 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10609 digest to be passed on the command line but it only used this
10610 parameter when signing a certificate. Modified so all relevant
10611 operations are affected by the digest parameter including the
10612 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10613 DSA key was used because it didn't fix the digest.
10616 *) Initial certificate chain verify code. Currently tests the untrusted
10617 certificates for consistency with the verify purpose (which is set
10618 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10620 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10621 this is because it will reject chains with invalid extensions whereas
10622 every previous version of OpenSSL and SSLeay made no checks at all.
10624 Trust code: checks the root CA for the relevant trust settings. Trust
10625 settings have an initial value consistent with the verify purpose: e.g.
10626 if the verify purpose is for SSL client use it expects the CA to be
10627 trusted for SSL client use. However the default value can be changed to
10628 permit custom trust settings: one example of this would be to only trust
10629 certificates from a specific "secure" set of CAs.
10631 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10632 which should be used for version portability: especially since the
10633 verify structure is likely to change more often now.
10635 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10636 to set them. If not set then assume SSL clients will verify SSL servers
10639 Two new options to the verify program: -untrusted allows a set of
10640 untrusted certificates to be passed in and -purpose which sets the
10641 intended purpose of the certificate. If a purpose is set then the
10642 new chain verify code is used to check extension consistency.
10645 *) Support for the authority information access extension.
10648 *) Modify RSA and DSA PEM read routines to transparently handle
10649 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10650 public keys in a format compatible with certificate
10651 SubjectPublicKeyInfo structures. Unfortunately there were already
10652 functions called *_PublicKey_* which used various odd formats so
10653 these are retained for compatibility: however the DSA variants were
10654 never in a public release so they have been deleted. Changed dsa/rsa
10655 utilities to handle the new format: note no releases ever handled public
10656 keys so we should be OK.
10658 The primary motivation for this change is to avoid the same fiasco
10659 that dogs private keys: there are several incompatible private key
10660 formats some of which are standard and some OpenSSL specific and
10661 require various evil hacks to allow partial transparent handling and
10662 even then it doesn't work with DER formats. Given the option anything
10663 other than PKCS#8 should be dumped: but the other formats have to
10664 stay in the name of compatibility.
10666 With public keys and the benefit of hindsight one standard format
10667 is used which works with EVP_PKEY, RSA or DSA structures: though
10668 it clearly returns an error if you try to read the wrong kind of key.
10670 Added a -pubkey option to the 'x509' utility to output the public key.
10671 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10672 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10673 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10674 that do the same as the EVP_PKEY_assign_*() except they up the
10675 reference count of the added key (they don't "swallow" the
10679 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10680 CRLs would fail if the file contained no certificates or no CRLs:
10681 added a new function to read in both types and return the number
10682 read: this means that if none are read it will be an error. The
10683 DER versions of the certificate and CRL reader would always fail
10684 because it isn't possible to mix certificates and CRLs in DER format
10685 without choking one or the other routine. Changed this to just read
10686 a certificate: this is the best we can do. Also modified the code
10687 in apps/verify.c to take notice of return codes: it was previously
10688 attempting to read in certificates from NULL pointers and ignoring
10689 any errors: this is one reason why the cert and CRL reader seemed
10690 to work. It doesn't check return codes from the default certificate
10691 routines: these may well fail if the certificates aren't installed.
10694 *) Code to support otherName option in GeneralName.
10697 *) First update to verify code. Change the verify utility
10698 so it warns if it is passed a self signed certificate:
10699 for consistency with the normal behaviour. X509_verify
10700 has been modified to it will now verify a self signed
10701 certificate if *exactly* the same certificate appears
10702 in the store: it was previously impossible to trust a
10703 single self signed certificate. This means that:
10704 openssl verify ss.pem
10705 now gives a warning about a self signed certificate but
10706 openssl verify -CAfile ss.pem ss.pem
10710 *) For servers, store verify_result in SSL_SESSION data structure
10711 (and add it to external session representation).
10712 This is needed when client certificate verifications fails,
10713 but an application-provided verification callback (set by
10714 SSL_CTX_set_cert_verify_callback) allows accepting the session
10715 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10716 but returns 1): When the session is reused, we have to set
10717 ssl->verify_result to the appropriate error code to avoid
10719 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10721 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10722 case in PKCS7_dataInit() where the signed PKCS7 structure
10723 didn't contain any existing data because it was being created.
10724 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10726 *) Add a salt to the key derivation routines in enc.c. This
10727 forms the first 8 bytes of the encrypted file. Also add a
10728 -S option to allow a salt to be input on the command line.
10731 *) New function X509_cmp(). Oddly enough there wasn't a function
10732 to compare two certificates. We do this by working out the SHA1
10733 hash and comparing that. X509_cmp() will be needed by the trust
10737 *) SSL_get1_session() is like SSL_get_session(), but increments
10738 the reference count in the SSL_SESSION returned.
10739 [Geoff Thorpe <geoff@eu.c2.net>]
10741 *) Fix for 'req': it was adding a null to request attributes.
10742 Also change the X509_LOOKUP and X509_INFO code to handle
10743 certificate auxiliary information.
10746 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10750 *) Add the possibility to add extra information to the memory leak
10751 detecting output, to form tracebacks, showing from where each
10752 allocation was originated: CRYPTO_push_info("constant string") adds
10753 the string plus current file name and line number to a per-thread
10754 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10755 is like calling CYRPTO_pop_info() until the stack is empty.
10756 Also updated memory leak detection code to be multi-thread-safe.
10759 *) Add options -text and -noout to pkcs7 utility and delete the
10760 encryption options which never did anything. Update docs.
10763 *) Add options to some of the utilities to allow the pass phrase
10764 to be included on either the command line (not recommended on
10765 OSes like Unix) or read from the environment. Update the
10766 manpages and fix a few bugs.
10769 *) Add a few manpages for some of the openssl commands.
10772 *) Fix the -revoke option in ca. It was freeing up memory twice,
10773 leaking and not finding already revoked certificates.
10776 *) Extensive changes to support certificate auxiliary information.
10777 This involves the use of X509_CERT_AUX structure and X509_AUX
10778 functions. An X509_AUX function such as PEM_read_X509_AUX()
10779 can still read in a certificate file in the usual way but it
10780 will also read in any additional "auxiliary information". By
10781 doing things this way a fair degree of compatibility can be
10782 retained: existing certificates can have this information added
10783 using the new 'x509' options.
10785 Current auxiliary information includes an "alias" and some trust
10786 settings. The trust settings will ultimately be used in enhanced
10787 certificate chain verification routines: currently a certificate
10788 can only be trusted if it is self signed and then it is trusted
10792 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10793 The problem was that one of the replacement routines had not been working
10794 since SSLeay releases. For now the offending routine has been replaced
10795 with non-optimised assembler. Even so, this now gives around 95%
10796 performance improvement for 1024 bit RSA signs.
10799 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10800 handling. Most clients have the effective key size in bits equal to
10801 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10802 A few however don't do this and instead use the size of the decrypted key
10803 to determine the RC2 key length and the AlgorithmIdentifier to determine
10804 the effective key length. In this case the effective key length can still
10805 be 40 bits but the key length can be 168 bits for example. This is fixed
10806 by manually forcing an RC2 key into the EVP_PKEY structure because the
10807 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10808 the key length and effective key length are equal.
10811 *) Add a bunch of functions that should simplify the creation of
10812 X509_NAME structures. Now you should be able to do:
10813 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10814 and have it automatically work out the correct field type and fill in
10815 the structures. The more adventurous can try:
10816 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10817 and it will (hopefully) work out the correct multibyte encoding.
10820 *) Change the 'req' utility to use the new field handling and multibyte
10821 copy routines. Before the DN field creation was handled in an ad hoc
10822 way in req, ca, and x509 which was rather broken and didn't support
10823 BMPStrings or UTF8Strings. Since some software doesn't implement
10824 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10825 using the dirstring_type option. See the new comment in the default
10826 openssl.cnf for more info.
10829 *) Make crypto/rand/md_rand.c more robust:
10830 - Assure unique random numbers after fork().
10831 - Make sure that concurrent threads access the global counter and
10832 md serializably so that we never lose entropy in them
10833 or use exactly the same state in multiple threads.
10834 Access to the large state is not always serializable because
10835 the additional locking could be a performance killer, and
10836 md should be large enough anyway.
10839 *) New file apps/app_rand.c with commonly needed functionality
10840 for handling the random seed file.
10842 Use the random seed file in some applications that previously did not:
10844 dsaparam -genkey (which also ignored its '-rand' option),
10847 x509 (when signing).
10848 Except on systems with /dev/urandom, it is crucial to have a random
10849 seed file at least for key creation, DSA signing, and for DH exchanges;
10850 for RSA signatures we could do without one.
10852 gendh and gendsa (unlike genrsa) used to read only the first byte
10853 of each file listed in the '-rand' option. The function as previously
10854 found in genrsa is now in app_rand.c and is used by all programs
10855 that support '-rand'.
10858 *) In RAND_write_file, use mode 0600 for creating files;
10859 don't just chmod when it may be too late.
10862 *) Report an error from X509_STORE_load_locations
10863 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10866 *) New function ASN1_mbstring_copy() this copies a string in either
10867 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10868 into an ASN1_STRING type. A mask of permissible types is passed
10869 and it chooses the "minimal" type to use or an error if not type
10873 *) Add function equivalents to the various macros in asn1.h. The old
10874 macros are retained with an M_ prefix. Code inside the library can
10875 use the M_ macros. External code (including the openssl utility)
10876 should *NOT* in order to be "shared library friendly".
10879 *) Add various functions that can check a certificate's extensions
10880 to see if it usable for various purposes such as SSL client,
10881 server or S/MIME and CAs of these types. This is currently
10882 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10883 verification. Also added a -purpose flag to x509 utility to
10884 print out all the purposes.
10887 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10891 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10892 for, obtain and decode and extension and obtain its critical flag.
10893 This allows all the necessary extension code to be handled in a
10894 single function call.
10897 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10898 platforms. See crypto/rc4/rc4_enc.c for further details.
10901 *) New -noout option to asn1parse. This causes no output to be produced
10902 its main use is when combined with -strparse and -out to extract data
10903 from a file (which may not be in ASN.1 format).
10906 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10907 when producing the local key id.
10908 [Richard Levitte <levitte@stacken.kth.se>]
10910 *) New option -dhparam in s_server. This allows a DH parameter file to be
10911 stated explicitly. If it is not stated then it tries the first server
10912 certificate file. The previous behaviour hard coded the filename
10916 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10917 a public key to be input or output. For example:
10918 openssl rsa -in key.pem -pubout -out pubkey.pem
10919 Also added necessary DSA public key functions to handle this.
10922 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10923 in the message. This was handled by allowing
10924 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10925 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10927 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10928 to the end of the strings whereas this didn't. This would cause problems
10929 if strings read with d2i_ASN1_bytes() were later modified.
10930 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10932 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10933 data and it contains EOF it will end up returning an error. This is
10934 caused by input 46 bytes long. The cause is due to the way base64
10935 BIOs find the start of base64 encoded data. They do this by trying a
10936 trial decode on each line until they find one that works. When they
10937 do a flag is set and it starts again knowing it can pass all the
10938 data directly through the decoder. Unfortunately it doesn't reset
10939 the context it uses. This means that if EOF is reached an attempt
10940 is made to pass two EOFs through the context and this causes the
10941 resulting error. This can also cause other problems as well. As is
10942 usual with these problems it takes *ages* to find and the fix is
10943 trivial: move one line.
10944 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10946 *) Ugly workaround to get s_client and s_server working under Windows. The
10947 old code wouldn't work because it needed to select() on sockets and the
10948 tty (for keypresses and to see if data could be written). Win32 only
10949 supports select() on sockets so we select() with a 1s timeout on the
10950 sockets and then see if any characters are waiting to be read, if none
10951 are present then we retry, we also assume we can always write data to
10952 the tty. This isn't nice because the code then blocks until we've
10953 received a complete line of data and it is effectively polling the
10954 keyboard at 1s intervals: however it's quite a bit better than not
10955 working at all :-) A dedicated Windows application might handle this
10956 with an event loop for example.
10959 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10960 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10961 will be called when RSA_sign() and RSA_verify() are used. This is useful
10962 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10963 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10964 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10965 This necessitated the support of an extra signature type NID_md5_sha1
10966 for SSL signatures and modifications to the SSL library to use it instead
10967 of calling RSA_public_decrypt() and RSA_private_encrypt().
10970 *) Add new -verify -CAfile and -CApath options to the crl program, these
10971 will lookup a CRL issuers certificate and verify the signature in a
10972 similar way to the verify program. Tidy up the crl program so it
10973 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10974 less strict. It will now permit CRL extensions even if it is not
10975 a V2 CRL: this will allow it to tolerate some broken CRLs.
10978 *) Initialize all non-automatic variables each time one of the openssl
10979 sub-programs is started (this is necessary as they may be started
10980 multiple times from the "OpenSSL>" prompt).
10981 [Lennart Bang, Bodo Moeller]
10983 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10984 removing all other RSA functionality (this is what NO_RSA does). This
10985 is so (for example) those in the US can disable those operations covered
10986 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10990 *) Non-copying interface to BIO pairs.
10991 (still largely untested)
10994 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10995 ASCII string. This was handled independently in various places before.
10998 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10999 UTF8 strings a character at a time.
11002 *) Use client_version from client hello to select the protocol
11003 (s23_srvr.c) and for RSA client key exchange verification
11004 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11007 *) Add various utility functions to handle SPKACs, these were previously
11008 handled by poking round in the structure internals. Added new function
11009 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11010 print, verify and generate SPKACs. Based on an original idea from
11011 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11014 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11017 *) Allow the config file extension section to be overwritten on the
11018 command line. Based on an original idea from Massimiliano Pala
11019 <madwolf@comune.modena.it>. The new option is called -extensions
11020 and can be applied to ca, req and x509. Also -reqexts to override
11021 the request extensions in req and -crlexts to override the crl extensions
11025 *) Add new feature to the SPKAC handling in ca. Now you can include
11026 the same field multiple times by preceding it by "XXXX." for example:
11029 this is the same syntax as used in the req config file.
11032 *) Allow certificate extensions to be added to certificate requests. These
11033 are specified in a 'req_extensions' option of the req section of the
11034 config file. They can be printed out with the -text option to req but
11035 are otherwise ignored at present.
11038 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11039 data read consists of only the final block it would not decrypted because
11040 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11041 A misplaced 'break' also meant the decrypted final block might not be
11042 copied until the next read.
11045 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11046 a few extra parameters to the DH structure: these will be useful if
11047 for example we want the value of 'q' or implement X9.42 DH.
11050 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11051 provides hooks that allow the default DSA functions or functions on a
11052 "per key" basis to be replaced. This allows hardware acceleration and
11053 hardware key storage to be handled without major modification to the
11054 library. Also added low level modexp hooks and CRYPTO_EX structure and
11055 associated functions.
11058 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11059 as "read only": it can't be written to and the buffer it points to will
11060 not be freed. Reading from a read only BIO is much more efficient than
11061 a normal memory BIO. This was added because there are several times when
11062 an area of memory needs to be read from a BIO. The previous method was
11063 to create a memory BIO and write the data to it, this results in two
11064 copies of the data and an O(n^2) reading algorithm. There is a new
11065 function BIO_new_mem_buf() which creates a read only memory BIO from
11066 an area of memory. Also modified the PKCS#7 routines to use read only
11070 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11071 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11072 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11073 but a retry condition occurred while trying to read the rest.
11076 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11077 NID_pkcs7_encrypted by default: this was wrong since this should almost
11078 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11079 the encrypted data type: this is a more sensible place to put it and it
11080 allows the PKCS#12 code to be tidied up that duplicated this
11084 *) Changed obj_dat.pl script so it takes its input and output files on
11085 the command line. This should avoid shell escape redirection problems
11089 *) Initial support for certificate extension requests, these are included
11090 in things like Xenroll certificate requests. Included functions to allow
11091 extensions to be obtained and added.
11094 *) -crlf option to s_client and s_server for sending newlines as
11095 CRLF (as required by many protocols).
11098 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11100 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11101 [Ralf S. Engelschall]
11103 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11104 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11106 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11110 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11111 DH parameters/keys (q is lost during that conversion, but the resulting
11112 DH parameters contain its length).
11114 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11115 much faster than DH_generate_parameters (which creates parameters
11116 where p = 2*q + 1), and also the smaller q makes DH computations
11117 much more efficient (160-bit exponentiation instead of 1024-bit
11118 exponentiation); so this provides a convenient way to support DHE
11119 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11120 utter importance to use
11121 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11123 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11124 when such DH parameters are used, because otherwise small subgroup
11125 attacks may become possible!
11128 *) Avoid memory leak in i2d_DHparams.
11131 *) Allow the -k option to be used more than once in the enc program:
11132 this allows the same encrypted message to be read by multiple recipients.
11135 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11136 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11137 it will always use the numerical form of the OID, even if it has a short
11141 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11142 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11143 otherwise bn_mod_exp was called. In the case of hardware keys for example
11144 no private key components need be present and it might store extra data
11145 in the RSA structure, which cannot be accessed from bn_mod_exp.
11146 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11147 private key operations.
11150 *) Added support for SPARC Linux.
11153 *) pem_password_cb function type incompatibly changed from
11154 typedef int pem_password_cb(char *buf, int size, int rwflag);
11156 ....(char *buf, int size, int rwflag, void *userdata);
11157 so that applications can pass data to their callbacks:
11158 The PEM[_ASN1]_{read,write}... functions and macros now take an
11159 additional void * argument, which is just handed through whenever
11160 the password callback is called.
11161 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11163 New function SSL_CTX_set_default_passwd_cb_userdata.
11165 Compatibility note: As many C implementations push function arguments
11166 onto the stack in reverse order, the new library version is likely to
11167 interoperate with programs that have been compiled with the old
11168 pem_password_cb definition (PEM_whatever takes some data that
11169 happens to be on the stack as its last argument, and the callback
11170 just ignores this garbage); but there is no guarantee whatsoever that
11173 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11174 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11175 problems not only on Windows, but also on some Unix platforms.
11176 To avoid problematic command lines, these definitions are now in an
11177 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11178 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11181 *) MIPS III/IV assembler module is reimplemented.
11184 *) More DES library cleanups: remove references to srand/rand and
11185 delete an unused file.
11188 *) Add support for the free Netwide assembler (NASM) under Win32,
11189 since not many people have MASM (ml) and it can be hard to obtain.
11190 This is currently experimental but it seems to work OK and pass all
11191 the tests. Check out INSTALL.W32 for info.
11194 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11195 without temporary keys kept an extra copy of the server key,
11196 and connections with temporary keys did not free everything in case
11200 *) New function RSA_check_key and new openssl rsa option -check
11201 for verifying the consistency of RSA keys.
11202 [Ulf Moeller, Bodo Moeller]
11204 *) Various changes to make Win32 compile work:
11205 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11206 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11207 comparison" warnings.
11208 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11211 *) Add a debugging option to PKCS#5 v2 key generation function: when
11212 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11213 derived keys are printed to stderr.
11216 *) Copy the flags in ASN1_STRING_dup().
11217 [Roman E. Pavlov <pre@mo.msk.ru>]
11219 *) The x509 application mishandled signing requests containing DSA
11220 keys when the signing key was also DSA and the parameters didn't match.
11222 It was supposed to omit the parameters when they matched the signing key:
11223 the verifying software was then supposed to automatically use the CA's
11224 parameters if they were absent from the end user certificate.
11226 Omitting parameters is no longer recommended. The test was also
11227 the wrong way round! This was probably due to unusual behaviour in
11228 EVP_cmp_parameters() which returns 1 if the parameters match.
11229 This meant that parameters were omitted when they *didn't* match and
11230 the certificate was useless. Certificates signed with 'ca' didn't have
11232 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11234 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11235 The interface is as follows:
11236 Applications can use
11237 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11238 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11239 "off" is now the default.
11240 The library internally uses
11241 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11242 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11243 to disable memory-checking temporarily.
11245 Some inconsistent states that previously were possible (and were
11246 even the default) are now avoided.
11248 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11249 with each memory chunk allocated; this is occasionally more helpful
11250 than just having a counter.
11252 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11254 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11258 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11259 which largely parallels "options", but is for changing API behaviour,
11260 whereas "options" are about protocol behaviour.
11261 Initial "mode" flags are:
11263 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11264 a single record has been written.
11265 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11266 retries use the same buffer location.
11267 (But all of the contents must be
11271 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11274 *) Fix problems with no-hmac etc.
11275 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11277 *) New functions RSA_get_default_method(), RSA_set_method() and
11278 RSA_get_method(). These allows replacement of RSA_METHODs without having
11279 to mess around with the internals of an RSA structure.
11282 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11283 Also really enable memory leak checks in openssl.c and in some
11285 [Chad C. Mulligan, Bodo Moeller]
11287 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11288 up the length of negative integers. This has now been simplified to just
11289 store the length when it is first determined and use it later, rather
11290 than trying to keep track of where data is copied and updating it to
11292 [Steve Henson, reported by Brien Wheeler
11293 <bwheeler@authentica-security.com>]
11295 *) Add a new function PKCS7_signatureVerify. This allows the verification
11296 of a PKCS#7 signature but with the signing certificate passed to the
11297 function itself. This contrasts with PKCS7_dataVerify which assumes the
11298 certificate is present in the PKCS#7 structure. This isn't always the
11299 case: certificates can be omitted from a PKCS#7 structure and be
11300 distributed by "out of band" means (such as a certificate database).
11303 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11304 function prototypes in pem.h, also change util/mkdef.pl to add the
11305 necessary function names.
11308 *) mk1mf.pl (used by Windows builds) did not properly read the
11309 options set by Configure in the top level Makefile, and Configure
11310 was not even able to write more than one option correctly.
11311 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11314 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11315 file to be loaded from a BIO or FILE pointer. The BIO version will
11316 for example allow memory BIOs to contain config info.
11319 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11320 Whoever hopes to achieve shared-library compatibility across versions
11321 must use this, not the compile-time macro.
11322 (Exercise 0.9.4: Which is the minimum library version required by
11324 Note: All this applies only to multi-threaded programs, others don't
11328 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11329 through a BIO pair triggered the default case, i.e.
11330 SSLerr(...,SSL_R_UNKNOWN_STATE).
11333 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11334 can use the SSL library even if none of the specific BIOs is
11338 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11339 for the encoded length.
11340 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11342 *) Add initial documentation of the X509V3 functions.
11345 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11346 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11347 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11348 secure PKCS#8 private key format with a high iteration count.
11351 *) Fix determination of Perl interpreter: A perl or perl5
11352 _directory_ in $PATH was also accepted as the interpreter.
11353 [Ralf S. Engelschall]
11355 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11356 wrong with it but it was very old and did things like calling
11357 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11358 unusual formatting.
11361 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11362 to use the new extension code.
11365 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11366 with macros. This should make it easier to change their form, add extra
11367 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11371 *) Add to configuration table a new entry that can specify an alternative
11372 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11373 according to Mark Crispin <MRC@Panda.COM>.
11377 *) DES CBC did not update the IV. Weird.
11380 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11381 Changing the behaviour of the former might break existing programs --
11382 where IV updating is needed, des_ncbc_encrypt can be used.
11385 *) When bntest is run from "make test" it drives bc to check its
11386 calculations, as well as internally checking them. If an internal check
11387 fails, it needs to cause bc to give a non-zero result or make test carries
11388 on without noticing the failure. Fixed.
11391 *) DES library cleanups.
11394 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11395 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11396 ciphers. NOTE: although the key derivation function has been verified
11397 against some published test vectors it has not been extensively tested
11398 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11402 *) Instead of "mkdir -p", which is not fully portable, use new
11403 Perl script "util/mkdir-p.pl".
11406 *) Rewrite the way password based encryption (PBE) is handled. It used to
11407 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11408 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11409 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11410 the 'parameter' field of the AlgorithmIdentifier is passed to the
11411 underlying key generation function so it must do its own ASN1 parsing.
11412 This has also changed the EVP_PBE_CipherInit() function which now has a
11413 'parameter' argument instead of literal salt and iteration count values
11414 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11417 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11418 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11419 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11420 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11421 value was just used as a "magic string" and not used directly its
11422 value doesn't matter.
11425 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11429 *) "linux-sparc64" configuration (ultrapenguin).
11430 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11431 "linux-sparc" configuration.
11432 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11434 *) config now generates no-xxx options for missing ciphers.
11437 *) Support the EBCDIC character set (work in progress).
11438 File ebcdic.c not yet included because it has a different license.
11439 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11441 *) Support BS2000/OSD-POSIX.
11442 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11444 *) Make callbacks for key generation use void * instead of char *.
11447 *) Make S/MIME samples compile (not yet tested).
11450 *) Additional typesafe stacks.
11453 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11457 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11459 *) New configuration variant "sco5-gcc".
11461 *) Updated some demos.
11462 [Sean O Riordain, Wade Scholine]
11464 *) Add missing BIO_free at exit of pkcs12 application.
11467 *) Fix memory leak in conf.c.
11470 *) Updates for Win32 to assembler version of MD5.
11473 *) Set #! path to perl in apps/der_chop to where we found it
11474 instead of using a fixed path.
11477 *) SHA library changes for irix64-mips4-cc.
11480 *) Improvements for VMS support.
11484 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11486 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11487 This also avoids the problems with SC4.2 and unpatched SC5.
11488 [Andy Polyakov <appro@fy.chalmers.se>]
11490 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11491 These are required because of the typesafe stack would otherwise break
11492 existing code. If old code used a structure member which used to be STACK
11493 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11494 sk_num or sk_value it would produce an error because the num, data members
11495 are not present in STACK_OF. Now it just produces a warning. sk_set
11496 replaces the old method of assigning a value to sk_value
11497 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11498 that does this will no longer work (and should use sk_set instead) but
11499 this could be regarded as a "questionable" behaviour anyway.
11502 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11503 correctly handle encrypted S/MIME data.
11506 *) Change type of various DES function arguments from des_cblock
11507 (which means, in function argument declarations, pointer to char)
11508 to des_cblock * (meaning pointer to array with 8 char elements),
11509 which allows the compiler to do more typechecking; it was like
11510 that back in SSLeay, but with lots of ugly casts.
11512 Introduce new type const_des_cblock.
11515 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11516 problems: find RecipientInfo structure that matches recipient certificate
11517 and initialise the ASN1 structures properly based on passed cipher.
11520 *) Belatedly make the BN tests actually check the results.
11523 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11524 to and from BNs: it was completely broken. New compilation option
11525 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11526 key elements as negative integers.
11529 *) Reorganize and speed up MD5.
11530 [Andy Polyakov <appro@fy.chalmers.se>]
11533 [Richard Levitte <richard@levitte.org>]
11535 *) New option -out to asn1parse to allow the parsed structure to be
11536 output to a file. This is most useful when combined with the -strparse
11537 option to examine the output of things like OCTET STRINGS.
11540 *) Make SSL library a little more fool-proof by not requiring any longer
11541 that SSL_set_{accept,connect}_state be called before
11542 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11543 in many applications because usually everything *appeared* to work as
11544 intended anyway -- now it really works as intended).
11547 *) Move openssl.cnf out of lib/.
11550 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11551 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11552 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11553 [Ralf S. Engelschall]
11555 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11556 handle PKCS#7 enveloped data properly.
11557 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11559 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11560 copying pointers. The cert_st handling is changed by this in
11561 various ways (and thus what used to be known as ctx->default_cert
11562 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11563 any longer when s->cert does not give us what we need).
11564 ssl_cert_instantiate becomes obsolete by this change.
11565 As soon as we've got the new code right (possibly it already is?),
11566 we have solved a couple of bugs of the earlier code where s->cert
11567 was used as if it could not have been shared with other SSL structures.
11569 Note that using the SSL API in certain dirty ways now will result
11570 in different behaviour than observed with earlier library versions:
11571 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11572 does not influence s as it used to.
11574 In order to clean up things more thoroughly, inside SSL_SESSION
11575 we don't use CERT any longer, but a new structure SESS_CERT
11576 that holds per-session data (if available); currently, this is
11577 the peer's certificate chain and, for clients, the server's certificate
11578 and temporary key. CERT holds only those values that can have
11579 meaningful defaults in an SSL_CTX.
11582 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11583 from the internal representation. Various PKCS#7 fixes: remove some
11584 evil casts and set the enc_dig_alg field properly based on the signing
11588 *) Allow PKCS#12 password to be set from the command line or the
11589 environment. Let 'ca' get its config file name from the environment
11590 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11594 *) Allow certificate policies extension to use an IA5STRING for the
11595 organization field. This is contrary to the PKIX definition but
11596 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11600 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11601 without disallowing inline assembler and the like for non-pedantic builds.
11604 *) Support Borland C++ builder.
11605 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11607 *) Support Mingw32.
11610 *) SHA-1 cleanups and performance enhancements.
11611 [Andy Polyakov <appro@fy.chalmers.se>]
11613 *) Sparc v8plus assembler for the bignum library.
11614 [Andy Polyakov <appro@fy.chalmers.se>]
11616 *) Accept any -xxx and +xxx compiler options in Configure.
11619 *) Update HPUX configuration.
11622 *) Add missing sk_<type>_unshift() function to safestack.h
11623 [Ralf S. Engelschall]
11625 *) New function SSL_CTX_use_certificate_chain_file that sets the
11626 "extra_cert"s in addition to the certificate. (This makes sense
11627 only for "PEM" format files, as chains as a whole are not
11631 *) Support verify_depth from the SSL API.
11632 x509_vfy.c had what can be considered an off-by-one-error:
11633 Its depth (which was not part of the external interface)
11634 was actually counting the number of certificates in a chain;
11635 now it really counts the depth.
11638 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11639 instead of X509err, which often resulted in confusing error
11640 messages since the error codes are not globally unique
11641 (e.g. an alleged error in ssl3_accept when a certificate
11642 didn't match the private key).
11644 *) New function SSL_CTX_set_session_id_context that allows to set a default
11645 value (so that you don't need SSL_set_session_id_context for each
11646 connection using the SSL_CTX).
11649 *) OAEP decoding bug fix.
11652 *) Support INSTALL_PREFIX for package builders, as proposed by
11656 *) New Configure options "threads" and "no-threads". For systems
11657 where the proper compiler options are known (currently Solaris
11658 and Linux), "threads" is the default.
11661 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11664 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11665 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11666 such as /usr/local/bin.
11669 *) "make linux-shared" to build shared libraries.
11670 [Niels Poppe <niels@netbox.org>]
11672 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11675 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11676 extension adding in x509 utility.
11679 *) Remove NOPROTO sections and error code comments.
11682 *) Partial rewrite of the DEF file generator to now parse the ANSI
11686 *) New Configure options --prefix=DIR and --openssldir=DIR.
11689 *) Complete rewrite of the error code script(s). It is all now handled
11690 by one script at the top level which handles error code gathering,
11691 header rewriting and C source file generation. It should be much better
11692 than the old method: it now uses a modified version of Ulf's parser to
11693 read the ANSI prototypes in all header files (thus the old K&R definitions
11694 aren't needed for error creation any more) and do a better job of
11695 translating function codes into names. The old 'ASN1 error code imbedded
11696 in a comment' is no longer necessary and it doesn't use .err files which
11697 have now been deleted. Also the error code call doesn't have to appear all
11698 on one line (which resulted in some large lines...).
11701 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11704 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11705 0 (which usually indicates a closed connection), but continue reading.
11708 *) Fix some race conditions.
11711 *) Add support for CRL distribution points extension. Add Certificate
11712 Policies and CRL distribution points documentation.
11715 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11718 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11719 8 of keying material. Merlin has also confirmed interop with this fix
11720 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11721 [Merlin Hughes <merlin@baltimore.ie>]
11723 *) Fix lots of warnings.
11724 [Richard Levitte <levitte@stacken.kth.se>]
11726 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11727 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11728 [Richard Levitte <levitte@stacken.kth.se>]
11730 *) Fix problems with sizeof(long) == 8.
11731 [Andy Polyakov <appro@fy.chalmers.se>]
11733 *) Change functions to ANSI C.
11736 *) Fix typos in error codes.
11737 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11739 *) Remove defunct assembler files from Configure.
11742 *) SPARC v8 assembler BIGNUM implementation.
11743 [Andy Polyakov <appro@fy.chalmers.se>]
11745 *) Support for Certificate Policies extension: both print and set.
11746 Various additions to support the r2i method this uses.
11749 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11750 return a const string when you are expecting an allocated buffer.
11753 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11754 types DirectoryString and DisplayText.
11757 *) Add code to allow r2i extensions to access the configuration database,
11758 add an LHASH database driver and add several ctx helper functions.
11761 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11762 fail when they extended the size of a BIGNUM.
11765 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11766 support typesafe stack.
11769 *) Fix typo in SSL_[gs]et_options().
11770 [Nils Frostberg <nils@medcom.se>]
11772 *) Delete various functions and files that belonged to the (now obsolete)
11773 old X509V3 handling code.
11776 *) New Configure option "rsaref".
11779 *) Don't auto-generate pem.h.
11782 *) Introduce type-safe ASN.1 SETs.
11785 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11786 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11788 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11789 that links with OpenSSL (well at least cause lots of warnings), but fear
11790 not: the conversion is trivial, and it eliminates loads of evil casts. A
11791 few STACKed things have been converted already. Feel free to convert more.
11792 In the fullness of time, I'll do away with the STACK type altogether.
11795 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11796 specified in <certfile> by updating the entry in the index.txt file.
11797 This way one no longer has to edit the index.txt file manually for
11798 revoking a certificate. The -revoke option does the gory details now.
11799 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11801 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11802 `-text' option at all and this way the `-noout -text' combination was
11803 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11804 [Ralf S. Engelschall]
11806 *) Make sure a corresponding plain text error message exists for the
11807 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11808 verify callback function determined that a certificate was revoked.
11809 [Ralf S. Engelschall]
11811 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11812 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11813 all available ciphers including rc5, which was forgotten until now.
11814 In order to let the testing shell script know which algorithms
11815 are available, a new (up to now undocumented) command
11816 "openssl list-cipher-commands" is used.
11819 *) Bugfix: s_client occasionally would sleep in select() when
11820 it should have checked SSL_pending() first.
11823 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11824 the raw DSA values prior to ASN.1 encoding.
11827 *) Tweaks to Configure
11828 [Niels Poppe <niels@netbox.org>]
11830 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11834 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11837 *) New config option to avoid instructions that are illegal on the 80386.
11838 The default code is faster, but requires at least a 486.
11841 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11842 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11843 same as SSL2_VERSION anyway.
11846 *) New "-showcerts" option for s_client.
11849 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11850 application. Various cleanups and fixes.
11853 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11854 modify error routines to work internally. Add error codes and PBE init
11855 to library startup routines.
11858 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11859 packing functions to asn1 and evp. Changed function names and error
11860 codes along the way.
11863 *) PKCS12 integration: and so it begins... First of several patches to
11864 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11865 objects to objects.h
11868 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11869 and display support for Thawte strong extranet extension.
11872 *) Add LinuxPPC support.
11873 [Jeff Dubrule <igor@pobox.org>]
11875 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11876 bn_div_words in alpha.s.
11877 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11879 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11880 OAEP isn't supported when OpenSSL is built with RSAref.
11881 [Ulf Moeller <ulf@fitug.de>]
11883 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11884 so they no longer are missing under -DNOPROTO.
11885 [Soren S. Jorvang <soren@t.dk>]
11888 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11890 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11891 doesn't work when the session is reused. Coming soon!
11894 *) Fix a security hole, that allows sessions to be reused in the wrong
11895 context thus bypassing client cert protection! All software that uses
11896 client certs and session caches in multiple contexts NEEDS PATCHING to
11897 allow session reuse! A fuller solution is in the works.
11898 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11900 *) Some more source tree cleanups (removed obsolete files
11901 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11902 permission on "config" script to be executable) and a fix for the INSTALL
11904 [Ulf Moeller <ulf@fitug.de>]
11906 *) Remove some legacy and erroneous uses of malloc, free instead of
11908 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11910 *) Make rsa_oaep_test return non-zero on error.
11911 [Ulf Moeller <ulf@fitug.de>]
11913 *) Add support for native Solaris shared libraries. Configure
11914 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11915 if someone would make that last step automatic.
11916 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11918 *) ctx_size was not built with the right compiler during "make links". Fixed.
11921 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11922 except NULL ciphers". This means the default cipher list will no longer
11923 enable NULL ciphers. They need to be specifically enabled e.g. with
11924 the string "DEFAULT:eNULL".
11927 *) Fix to RSA private encryption routines: if p < q then it would
11928 occasionally produce an invalid result. This will only happen with
11929 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11932 *) Be less restrictive and allow also `perl util/perlpath.pl
11933 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11934 because this way one can also use an interpreter named `perl5' (which is
11935 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11936 installed as `perl').
11937 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11939 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11940 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11942 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11943 advapi32.lib to Win32 build and change the pem test comparison
11944 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11945 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11946 and crypto/des/ede_cbcm_enc.c.
11949 *) DES quad checksum was broken on big-endian architectures. Fixed.
11952 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11953 Win32 test batch file so it (might) work again. The Win32 test batch file
11954 is horrible: I feel ill....
11957 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11958 in e_os.h. Audit of header files to check ANSI and non ANSI
11959 sections: 10 functions were absent from non ANSI section and not exported
11960 from Windows DLLs. Fixed up libeay.num for new functions.
11963 *) Make `openssl version' output lines consistent.
11964 [Ralf S. Engelschall]
11966 *) Fix Win32 symbol export lists for BIO functions: Added
11967 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11968 to ms/libeay{16,32}.def.
11969 [Ralf S. Engelschall]
11971 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11972 fine under Unix and passes some trivial tests I've now added. But the
11973 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11974 added to make sure no one expects that this stuff really works in the
11975 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11976 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11978 [Ralf S. Engelschall]
11980 *) Fix the generation of two part addresses in perl.
11981 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11983 *) Add config entry for Linux on MIPS.
11984 [John Tobey <jtobey@channel1.com>]
11986 *) Make links whenever Configure is run, unless we are on Windoze.
11989 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11990 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11994 *) Add a useful kludge to allow package maintainers to specify compiler and
11995 other platforms details on the command line without having to patch the
11996 Configure script everytime: One now can use ``perl Configure
11997 <id>:<details>'', i.e. platform ids are allowed to have details appended
11998 to them (separated by colons). This is treated as there would be a static
11999 pre-configured entry in Configure's %table under key <id> with value
12000 <details> and ``perl Configure <id>'' is called. So, when you want to
12001 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12002 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12003 now, which overrides the FreeBSD-elf entry on-the-fly.
12004 [Ralf S. Engelschall]
12006 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12009 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12010 on the `perl Configure ...' command line. This way one can compile
12011 OpenSSL libraries with Position Independent Code (PIC) which is needed
12012 for linking it into DSOs.
12013 [Ralf S. Engelschall]
12015 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12019 *) Cleaned up the LICENSE document: The official contact for any license
12020 questions now is the OpenSSL core team under openssl-core@openssl.org.
12021 And add a paragraph about the dual-license situation to make sure people
12022 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12023 to the OpenSSL toolkit.
12024 [Ralf S. Engelschall]
12026 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12027 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12028 Additionally cleaned up the `make links' target: Remove unnecessary
12029 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12030 to speed processing and no longer clutter the display with confusing
12031 stuff. Instead only the actually done links are displayed.
12032 [Ralf S. Engelschall]
12034 *) Permit null encryption ciphersuites, used for authentication only. It used
12035 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12036 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12040 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12041 signed attributes when verifying signatures (this would break them),
12042 the detached data encoding was wrong and public keys obtained using
12043 X509_get_pubkey() weren't freed.
12046 *) Add text documentation for the BUFFER functions. Also added a work around
12047 to a Win95 console bug. This was triggered by the password read stuff: the
12048 last character typed gets carried over to the next fread(). If you were
12049 generating a new cert request using 'req' for example then the last
12050 character of the passphrase would be CR which would then enter the first
12054 *) Added the new `Includes OpenSSL Cryptography Software' button as
12055 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12056 button and can be used by applications based on OpenSSL to show the
12057 relationship to the OpenSSL project.
12058 [Ralf S. Engelschall]
12060 *) Remove confusing variables in function signatures in files
12061 ssl/ssl_lib.c and ssl/ssl.h.
12062 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12064 *) Don't install bss_file.c under PREFIX/include/
12065 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12067 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12068 functions that return function pointers and has support for NT specific
12069 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12070 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12071 unsigned to signed types: this was killing the Win32 compile.
12074 *) Add new certificate file to stack functions,
12075 SSL_add_dir_cert_subjects_to_stack() and
12076 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12077 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12078 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12079 This means that Apache-SSL and similar packages don't have to mess around
12080 to add as many CAs as they want to the preferred list.
12083 *) Experiment with doxygen documentation. Currently only partially applied to
12085 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12086 openssl.doxy as the configuration file.
12089 *) Get rid of remaining C++-style comments which strict C compilers hate.
12090 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12092 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12093 compiled in by default: it has problems with large keys.
12096 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12097 DH private keys and/or callback functions which directly correspond to
12098 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12099 is needed for applications which have to configure certificates on a
12100 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12102 For the RSA certificate situation is makes no difference, but
12103 for the DSA certificate situation this fixes the "no shared cipher"
12104 problem where the OpenSSL cipher selection procedure failed because the
12105 temporary keys were not overtaken from the context and the API provided
12106 no way to reconfigure them.
12107 The new functions now let applications reconfigure the stuff and they
12108 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12109 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12110 non-public-API function ssl_cert_instantiate() is used as a helper
12111 function and also to reduce code redundancy inside ssl_rsa.c.
12112 [Ralf S. Engelschall]
12114 *) Move s_server -dcert and -dkey options out of the undocumented feature
12115 area because they are useful for the DSA situation and should be
12116 recognized by the users.
12117 [Ralf S. Engelschall]
12119 *) Fix the cipher decision scheme for export ciphers: the export bits are
12120 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12121 SSL_EXP_MASK. So, the original variable has to be used instead of the
12122 already masked variable.
12123 [Richard Levitte <levitte@stacken.kth.se>]
12125 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12126 [Richard Levitte <levitte@stacken.kth.se>]
12128 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12129 from `int' to `unsigned int' because it's a length and initialized by
12130 EVP_DigestFinal() which expects an `unsigned int *'.
12131 [Richard Levitte <levitte@stacken.kth.se>]
12133 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12134 script. Instead use the usual Shell->Perl transition trick.
12135 [Ralf S. Engelschall]
12137 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12138 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12139 -noout -modulus' as it's already the case for `openssl rsa -noout
12140 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12141 currently the public key is printed (a decision which was already done by
12142 `openssl dsa -modulus' in the past) which serves a similar purpose.
12143 Additionally the NO_RSA no longer completely removes the whole -modulus
12144 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12146 [Ralf S. Engelschall]
12148 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12149 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12150 [Arne Ansper <arne@ats.cyber.ee>]
12152 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12153 to be added. Now both 'req' and 'ca' can use new objects defined in the
12157 *) Add cool BIO that does syslog (or event log on NT).
12158 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12160 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12161 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12162 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12163 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12166 *) Add preliminary config info for new extension code.
12169 *) Make RSA_NO_PADDING really use no padding.
12170 [Ulf Moeller <ulf@fitug.de>]
12172 *) Generate errors when private/public key check is done.
12175 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12176 for some CRL extensions and new objects added.
12179 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12180 key usage extension and fuller support for authority key id.
12183 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12184 padding method for RSA, which is recommended for new applications in PKCS
12185 #1 v2.0 (RFC 2437, October 1998).
12186 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12187 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12188 against Bleichbacher's attack on RSA.
12189 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12192 *) Updates to the new SSL compression code
12193 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12195 *) Fix so that the version number in the master secret, when passed
12196 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12197 (because the server will not accept higher), that the version number
12198 is 0x03,0x01, not 0x03,0x00
12199 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12201 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12202 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12203 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12206 *) Support for RAW extensions where an arbitrary extension can be
12207 created by including its DER encoding. See apps/openssl.cnf for
12211 *) Make sure latest Perl versions don't interpret some generated C array
12212 code as Perl array code in the crypto/err/err_genc.pl script.
12213 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12215 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12216 not many people have the assembler. Various Win32 compilation fixes and
12217 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12218 build instructions.
12221 *) Modify configure script 'Configure' to automatically create crypto/date.h
12222 file under Win32 and also build pem.h from pem.org. New script
12223 util/mkfiles.pl to create the MINFO file on environments that can't do a
12224 'make files': perl util/mkfiles.pl >MINFO should work.
12227 *) Major rework of DES function declarations, in the pursuit of correctness
12228 and purity. As a result, many evil casts evaporated, and some weirdness,
12229 too. You may find this causes warnings in your code. Zapping your evil
12230 casts will probably fix them. Mostly.
12233 *) Fix for a typo in asn1.h. Bug fix to object creation script
12234 obj_dat.pl. It considered a zero in an object definition to mean
12235 "end of object": none of the objects in objects.h have any zeros
12236 so it wasn't spotted.
12237 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12239 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12240 Masking (CBCM). In the absence of test vectors, the best I have been able
12241 to do is check that the decrypt undoes the encrypt, so far. Send me test
12242 vectors if you have them.
12245 *) Correct calculation of key length for export ciphers (too much space was
12246 allocated for null ciphers). This has not been tested!
12249 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12250 message is now correct (it understands "crypto" and "ssl" on its
12251 command line). There is also now an "update" option. This will update
12252 the util/ssleay.num and util/libeay.num files with any new functions.
12254 perl util/mkdef.pl crypto ssl update
12255 it will update them.
12258 *) Overhauled the Perl interface (perl/*):
12259 - ported BN stuff to OpenSSL's different BN library
12260 - made the perl/ source tree CVS-aware
12261 - renamed the package from SSLeay to OpenSSL (the files still contain
12262 their history because I've copied them in the repository)
12263 - removed obsolete files (the test scripts will be replaced
12264 by better Test::Harness variants in the future)
12265 [Ralf S. Engelschall]
12267 *) First cut for a very conservative source tree cleanup:
12268 1. merge various obsolete readme texts into doc/ssleay.txt
12269 where we collect the old documents and readme texts.
12270 2. remove the first part of files where I'm already sure that we no
12271 longer need them because of three reasons: either they are just temporary
12272 files which were left by Eric or they are preserved original files where
12273 I've verified that the diff is also available in the CVS via "cvs diff
12274 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12275 the crypto/md/ stuff).
12276 [Ralf S. Engelschall]
12278 *) More extension code. Incomplete support for subject and issuer alt
12279 name, issuer and authority key id. Change the i2v function parameters
12280 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12281 what that's for :-) Fix to ASN1 macro which messed up
12282 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12285 *) Preliminary support for ENUMERATED type. This is largely copied from the
12289 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12290 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12292 *) Make sure `make rehash' target really finds the `openssl' program.
12293 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12295 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12296 like to hear about it if this slows down other processors.
12299 *) Add CygWin32 platform information to Configure script.
12300 [Alan Batie <batie@aahz.jf.intel.com>]
12302 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12303 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12305 *) New program nseq to manipulate netscape certificate sequences
12308 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12312 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12313 but the BN code had some problems that would cause failures when
12314 doing certificate verification and some other functions.
12315 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12317 *) Add ASN1 and PEM code to support netscape certificate sequences.
12320 *) Add ASN1 and PEM code to support netscape certificate sequences.
12323 *) Add several PKIX and private extended key usage OIDs.
12326 *) Modify the 'ca' program to handle the new extension code. Modify
12327 openssl.cnf for new extension format, add comments.
12330 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12331 and add a sample to openssl.cnf so req -x509 now adds appropriate
12335 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12336 error code, add initial support to X509_print() and x509 application.
12339 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12340 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12341 stuff is currently isolated and isn't even compiled yet.
12344 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12345 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12346 Removed the versions check from X509 routines when loading extensions:
12347 this allows certain broken certificates that don't set the version
12348 properly to be processed.
12351 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12352 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12353 can still be regenerated with "make depend".
12356 *) Spelling mistake in C version of CAST-128.
12357 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12359 *) Changes to the error generation code. The perl script err-code.pl
12360 now reads in the old error codes and retains the old numbers, only
12361 adding new ones if necessary. It also only changes the .err files if new
12362 codes are added. The makefiles have been modified to only insert errors
12363 when needed (to avoid needlessly modifying header files). This is done
12364 by only inserting errors if the .err file is newer than the auto generated
12365 C file. To rebuild all the error codes from scratch (the old behaviour)
12366 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12367 or delete all the .err files.
12370 *) CAST-128 was incorrectly implemented for short keys. The C version has
12371 been fixed, but is untested. The assembler versions are also fixed, but
12372 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12373 to regenerate it if needed.
12374 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12375 Hagino <itojun@kame.net>]
12377 *) File was opened incorrectly in randfile.c.
12378 [Ulf Möller <ulf@fitug.de>]
12380 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12381 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12382 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12383 al: it's just almost always a UTCTime. Note this patch adds new error
12384 codes so do a "make errors" if there are problems.
12387 *) Correct Linux 1 recognition in config.
12388 [Ulf Möller <ulf@fitug.de>]
12390 *) Remove pointless MD5 hash when using DSA keys in ca.
12391 [Anonymous <nobody@replay.com>]
12393 *) Generate an error if given an empty string as a cert directory. Also
12394 generate an error if handed NULL (previously returned 0 to indicate an
12395 error, but didn't set one).
12396 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12398 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12401 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12402 parameters. This was causing a warning which killed off the Win32 compile.
12405 *) Remove C++ style comments from crypto/bn/bn_local.h.
12406 [Neil Costigan <neil.costigan@celocom.com>]
12408 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12409 based on a text string, looking up short and long names and finally
12410 "dot" format. The "dot" format stuff didn't work. Added new function
12411 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12412 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12413 OID is not part of the table.
12416 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12417 X509_LOOKUP_by_alias().
12420 *) Sort openssl functions by name.
12423 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12424 encryption from sample DSA keys (in case anyone is interested the password
12428 *) Make _all_ *_free functions accept a NULL pointer.
12429 [Frans Heymans <fheymans@isaserver.be>]
12431 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12433 [Anonymous <nobody@replay.com>]
12435 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12436 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12438 *) Don't blow it for numeric -newkey arguments to apps/req.
12439 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12441 *) Temp key "for export" tests were wrong in s3_srvr.c.
12442 [Anonymous <nobody@replay.com>]
12444 *) Add prototype for temp key callback functions
12445 SSL_CTX_set_tmp_{rsa,dh}_callback().
12448 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12449 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12452 *) X509_name_add_entry() freed the wrong thing after an error.
12453 [Arne Ansper <arne@ats.cyber.ee>]
12455 *) rsa_eay.c would attempt to free a NULL context.
12456 [Arne Ansper <arne@ats.cyber.ee>]
12458 *) BIO_s_socket() had a broken should_retry() on Windoze.
12459 [Arne Ansper <arne@ats.cyber.ee>]
12461 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12462 [Arne Ansper <arne@ats.cyber.ee>]
12464 *) Make sure the already existing X509_STORE->depth variable is initialized
12465 in X509_STORE_new(), but document the fact that this variable is still
12466 unused in the certificate verification process.
12467 [Ralf S. Engelschall]
12469 *) Fix the various library and apps files to free up pkeys obtained from
12470 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12473 *) Fix reference counting in X509_PUBKEY_get(). This makes
12474 demos/maurice/example2.c work, amongst others, probably.
12475 [Steve Henson and Ben Laurie]
12477 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12478 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12479 are no longer created. This way we have a single and consistent command
12480 line interface `openssl <command>', similar to `cvs <command>'.
12481 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12483 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12484 BIT STRING wrapper always have zero unused bits.
12487 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12490 *) Make the top-level INSTALL documentation easier to understand.
12493 *) Makefiles updated to exit if an error occurs in a sub-directory
12494 make (including if user presses ^C) [Paul Sutton]
12496 *) Make Montgomery context stuff explicit in RSA data structure.
12499 *) Fix build order of pem and err to allow for generated pem.h.
12502 *) Fix renumbering bug in X509_NAME_delete_entry().
12505 *) Enhanced the err-ins.pl script so it makes the error library number
12506 global and can add a library name. This is needed for external ASN1 and
12507 other error libraries.
12510 *) Fixed sk_insert which never worked properly.
12513 *) Fix ASN1 macros so they can handle indefinite length constructed
12514 EXPLICIT tags. Some non standard certificates use these: they can now
12518 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12519 into a single doc/ssleay.txt bundle. This way the information is still
12520 preserved but no longer messes up this directory. Now it's new room for
12521 the new set of documentation files.
12522 [Ralf S. Engelschall]
12524 *) SETs were incorrectly DER encoded. This was a major pain, because they
12525 shared code with SEQUENCEs, which aren't coded the same. This means that
12526 almost everything to do with SETs or SEQUENCEs has either changed name or
12527 number of arguments.
12528 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12530 *) Fix test data to work with the above.
12533 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12534 was already fixed by Eric for 0.9.1 it seems.
12535 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12537 *) Autodetect FreeBSD3.
12540 *) Fix various bugs in Configure. This affects the following platforms:
12544 unixware-2.0-pentium
12548 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12549 before they are needed.
12552 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12556 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12558 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12559 changed SSLeay to OpenSSL in version strings.
12560 [Ralf S. Engelschall]
12562 *) Some fixups to the top-level documents.
12565 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12566 because the symlink to include/ was missing.
12567 [Ralf S. Engelschall]
12569 *) Incorporated the popular no-RSA/DSA-only patches
12570 which allow to compile a RSA-free SSLeay.
12571 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12573 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12574 when "ssleay" is still not found.
12575 [Ralf S. Engelschall]
12577 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12578 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12580 *) Updated the README file.
12581 [Ralf S. Engelschall]
12583 *) Added various .cvsignore files in the CVS repository subdirs
12584 to make a "cvs update" really silent.
12585 [Ralf S. Engelschall]
12587 *) Recompiled the error-definition header files and added
12588 missing symbols to the Win32 linker tables.
12589 [Ralf S. Engelschall]
12591 *) Cleaned up the top-level documents;
12592 o new files: CHANGES and LICENSE
12593 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12594 o merged COPYRIGHT into LICENSE
12595 o removed obsolete TODO file
12596 o renamed MICROSOFT to INSTALL.W32
12597 [Ralf S. Engelschall]
12599 *) Removed dummy files from the 0.9.1b source tree:
12600 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12601 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12602 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12603 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12604 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12605 [Ralf S. Engelschall]
12607 *) Added various platform portability fixes.
12610 *) The Genesis of the OpenSSL rpject:
12611 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12612 Young and Tim J. Hudson created while they were working for C2Net until
12614 [The OpenSSL Project]
12617 Changes between 0.9.0b and 0.9.1b [not released]
12619 *) Updated a few CA certificates under certs/
12622 *) Changed some BIGNUM api stuff.
12625 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12626 DGUX x86, Linux Alpha, etc.
12629 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12630 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12634 *) Add -strparse option to asn1pars program which parses nested
12636 [Dr Stephen Henson <shenson@bigfoot.com>]
12638 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12641 *) DSA fix for "ca" program.
12644 *) Added "-genkey" option to "dsaparam" program.
12647 *) Added RIPE MD160 (rmd160) message digest.
12650 *) Added -a (all) option to "ssleay version" command.
12653 *) Added PLATFORM define which is the id given to Configure.
12656 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12659 *) Extended the ASN.1 parser routines.
12662 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12665 *) Added a BN_CTX to the BN library.
12668 *) Fixed the weak key values in DES library
12671 *) Changed API in EVP library for cipher aliases.
12674 *) Added support for RC2/64bit cipher.
12677 *) Converted the lhash library to the crypto/mem.c functions.
12680 *) Added more recognized ASN.1 object ids.
12683 *) Added more RSA padding checks for SSL/TLS.
12686 *) Added BIO proxy/filter functionality.
12689 *) Added extra_certs to SSL_CTX which can be used
12690 send extra CA certificates to the client in the CA cert chain sending
12691 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12694 *) Now Fortezza is denied in the authentication phase because
12695 this is key exchange mechanism is not supported by SSLeay at all.
12698 *) Additional PKCS1 checks.
12701 *) Support the string "TLSv1" for all TLS v1 ciphers.
12704 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12705 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12708 *) Fixed a few memory leaks.
12711 *) Fixed various code and comment typos.
12714 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12715 bytes sent in the client random.
12716 [Edward Bishop <ebishop@spyglass.com>]