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 *) Add SM3 implemented according to GB/T 32905-2016
13 [ Jack Lloyd <jack.lloyd@ribose.com>,
14 Ronald Tse <ronald.tse@ribose.com>,
15 Erick Borsboom <erick.borsboom@ribose.com> ]
17 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
18 as documented in RFC6066.
19 Based on a patch from Tomasz Moń
20 [Filipe Raimundo da Silva]
22 *) Add SM4 implemented according to GB/T 32907-2016.
23 [ Jack Lloyd <jack.lloyd@ribose.com>,
24 Ronald Tse <ronald.tse@ribose.com>,
25 Erick Borsboom <erick.borsboom@ribose.com> ]
27 *) Reimplement -newreq-nodes and ERR_error_string_n; the
28 original author does not agree with the license change.
31 *) Add ARIA AEAD TLS support.
34 *) Some macro definitions to support VS6 have been removed. Visual
35 Studio 6 has not worked since 1.1.0
38 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
39 without clearing the errors.
42 *) Add "atfork" functions. If building on a system that without
43 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
44 requirements. The RAND facility now uses/requires this.
50 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
51 not possible to disable entirely. However, it's still possible to
52 disable the console reading UI method, UI_OpenSSL() (use UI_null()
55 To disable, configure with 'no-ui-console'. 'no-ui' is still
56 possible to use as an alias. Check at compile time with the
57 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
58 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
61 *) Add a STORE module, which implements a uniform and URI based reader of
62 stores that can contain keys, certificates, CRLs and numerous other
63 objects. The main API is loosely based on a few stdio functions,
64 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
65 OSSL_STORE_error and OSSL_STORE_close.
66 The implementation uses backends called "loaders" to implement arbitrary
67 URI schemes. There is one built in "loader" for the 'file' scheme.
70 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
71 then adjusted to work on FreeBSD 8.4 as well.
72 Enable by configuring with 'enable-devcryptoeng'. This is done by default
73 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
76 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
77 util/mkerr.pl, which is adapted to allow those prefixes, leading to
78 error code calls like this:
80 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
82 With this change, we claim the namespaces OSSL and OPENSSL in a manner
83 that can be encoded in C. For the foreseeable future, this will only
85 [Richard Levitte and Tim Hudson]
87 *) Removed BSD cryptodev engine.
90 *) Add a build target 'build_all_generated', to build all generated files
91 and only that. This can be used to prepare everything that requires
92 things like perl for a system that lacks perl and then move everything
93 to that system and do the rest of the build there.
96 *) In the UI interface, make it possible to duplicate the user data. This
97 can be used by engines that need to retain the data for a longer time
98 than just the call where this user data is passed.
101 *) Ignore the '-named_curve auto' value for compatibility of applications
103 [Tomas Mraz <tmraz@fedoraproject.org>]
105 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
106 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
107 alerts across multiple records (some of which could be empty). In practice
108 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
109 prohibts this altogether and other libraries (BoringSSL, NSS) do not
110 support this at all. Supporting it adds significant complexity to the
111 record layer, and its removal is unlikely to cause inter-operability
115 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
116 with Z. These are meant to replace LONG and ZLONG and to be size safe.
117 The use of LONG and ZLONG is discouraged and scheduled for deprecation
121 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
122 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
123 [Richard Levitte, Andy Polyakov]
125 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
129 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
130 platform rather than 'mingw'.
133 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
134 success if they are asked to add an object which already exists
135 in the store. This change cascades to other functions which load
136 certificates and CRLs.
139 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
140 facilitate stack unwinding even from assembly subroutines.
143 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
144 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
147 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
148 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
149 which is the minimum version we support.
152 *) Certificate time validation (X509_cmp_time) enforces stricter
153 compliance with RFC 5280. Fractional seconds and timezone offsets
154 are no longer allowed.
157 *) Add support for ARIA
160 *) s_client will now send the Server Name Indication (SNI) extension by
161 default unless the new "-noservername" option is used. The server name is
162 based on the host provided to the "-connect" option unless overridden by
166 *) Add support for SipHash
169 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
170 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
171 prevent issues where no progress is being made and the peer continually
172 sends unrecognised record types, using up resources processing them.
175 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
176 using the algorithm defined in
177 https://www.akkadia.org/drepper/SHA-crypt.txt
180 *) Heartbeat support has been removed; the ABI is changed for now.
181 [Richard Levitte, Rich Salz]
183 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
186 *) The RSA "null" method, which was partially supported to avoid patent
187 issues, has been replaced to always returns NULL.
191 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
193 *) bn_sqrx8x_internal carry bug on x86_64
195 There is a carry propagating bug in the x86_64 Montgomery squaring
196 procedure. No EC algorithms are affected. Analysis suggests that attacks
197 against RSA and DSA as a result of this defect would be very difficult to
198 perform and are not believed likely. Attacks against DH are considered just
199 feasible (although very difficult) because most of the work necessary to
200 deduce information about a private key may be performed offline. The amount
201 of resources required for such an attack would be very significant and
202 likely only accessible to a limited number of attackers. An attacker would
203 additionally need online access to an unpatched system using the target
204 private key in a scenario with persistent DH parameters and a private
205 key that is shared between multiple clients.
207 This only affects processors that support the BMI1, BMI2 and ADX extensions
208 like Intel Broadwell (5th generation) and later or AMD Ryzen.
210 This issue was reported to OpenSSL by the OSS-Fuzz project.
214 *) Malformed X.509 IPAddressFamily could cause OOB read
216 If an X.509 certificate has a malformed IPAddressFamily extension,
217 OpenSSL could do a one-byte buffer overread. The most likely result
218 would be an erroneous display of the certificate in text format.
220 This issue was reported to OpenSSL by the OSS-Fuzz project.
224 Changes between 1.1.0e and 1.1.0f [25 May 2017]
226 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
227 platform rather than 'mingw'.
230 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
231 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
232 which is the minimum version we support.
235 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
237 *) Encrypt-Then-Mac renegotiation crash
239 During a renegotiation handshake if the Encrypt-Then-Mac extension is
240 negotiated where it was not in the original handshake (or vice-versa) then
241 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
242 and servers are affected.
244 This issue was reported to OpenSSL by Joe Orton (Red Hat).
248 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
250 *) Truncated packet could crash via OOB read
252 If one side of an SSL/TLS path is running on a 32-bit host and a specific
253 cipher is being used, then a truncated packet can cause that host to
254 perform an out-of-bounds read, usually resulting in a crash.
256 This issue was reported to OpenSSL by Robert Święcki of Google.
260 *) Bad (EC)DHE parameters cause a client crash
262 If a malicious server supplies bad parameters for a DHE or ECDHE key
263 exchange then this can result in the client attempting to dereference a
264 NULL pointer leading to a client crash. This could be exploited in a Denial
267 This issue was reported to OpenSSL by Guido Vranken.
271 *) BN_mod_exp may produce incorrect results on x86_64
273 There is a carry propagating bug in the x86_64 Montgomery squaring
274 procedure. No EC algorithms are affected. Analysis suggests that attacks
275 against RSA and DSA as a result of this defect would be very difficult to
276 perform and are not believed likely. Attacks against DH are considered just
277 feasible (although very difficult) because most of the work necessary to
278 deduce information about a private key may be performed offline. The amount
279 of resources required for such an attack would be very significant and
280 likely only accessible to a limited number of attackers. An attacker would
281 additionally need online access to an unpatched system using the target
282 private key in a scenario with persistent DH parameters and a private
283 key that is shared between multiple clients. For example this can occur by
284 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
285 similar to CVE-2015-3193 but must be treated as a separate problem.
287 This issue was reported to OpenSSL by the OSS-Fuzz project.
291 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
293 *) ChaCha20/Poly1305 heap-buffer-overflow
295 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
296 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
297 crash. This issue is not considered to be exploitable beyond a DoS.
299 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
303 *) CMS Null dereference
305 Applications parsing invalid CMS structures can crash with a NULL pointer
306 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
307 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
308 structure callback if an attempt is made to free certain invalid encodings.
309 Only CHOICE structures using a callback which do not handle NULL value are
312 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
316 *) Montgomery multiplication may produce incorrect results
318 There is a carry propagating bug in the Broadwell-specific Montgomery
319 multiplication procedure that handles input lengths divisible by, but
320 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
321 and DH private keys are impossible. This is because the subroutine in
322 question is not used in operations with the private key itself and an input
323 of the attacker's direct choice. Otherwise the bug can manifest itself as
324 transient authentication and key negotiation failures or reproducible
325 erroneous outcome of public-key operations with specially crafted input.
326 Among EC algorithms only Brainpool P-512 curves are affected and one
327 presumably can attack ECDH key negotiation. Impact was not analyzed in
328 detail, because pre-requisites for attack are considered unlikely. Namely
329 multiple clients have to choose the curve in question and the server has to
330 share the private key among them, neither of which is default behaviour.
331 Even then only clients that chose the curve will be affected.
333 This issue was publicly reported as transient failures and was not
334 initially recognized as a security issue. Thanks to Richard Morgan for
335 providing reproducible case.
339 *) Removed automatic addition of RPATH in shared libraries and executables,
340 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
343 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
345 *) Fix Use After Free for large message sizes
347 The patch applied to address CVE-2016-6307 resulted in an issue where if a
348 message larger than approx 16k is received then the underlying buffer to
349 store the incoming message is reallocated and moved. Unfortunately a
350 dangling pointer to the old location is left which results in an attempt to
351 write to the previously freed location. This is likely to result in a
352 crash, however it could potentially lead to execution of arbitrary code.
354 This issue only affects OpenSSL 1.1.0a.
356 This issue was reported to OpenSSL by Robert Święcki.
360 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
362 *) OCSP Status Request extension unbounded memory growth
364 A malicious client can send an excessively large OCSP Status Request
365 extension. If that client continually requests renegotiation, sending a
366 large OCSP Status Request extension each time, then there will be unbounded
367 memory growth on the server. This will eventually lead to a Denial Of
368 Service attack through memory exhaustion. Servers with a default
369 configuration are vulnerable even if they do not support OCSP. Builds using
370 the "no-ocsp" build time option are not affected.
372 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
376 *) SSL_peek() hang on empty record
378 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
379 sends an empty record. This could be exploited by a malicious peer in a
380 Denial Of Service attack.
382 This issue was reported to OpenSSL by Alex Gaynor.
386 *) Excessive allocation of memory in tls_get_message_header() and
387 dtls1_preprocess_fragment()
389 A (D)TLS message includes 3 bytes for its length in the header for the
390 message. This would allow for messages up to 16Mb in length. Messages of
391 this length are excessive and OpenSSL includes a check to ensure that a
392 peer is sending reasonably sized messages in order to avoid too much memory
393 being consumed to service a connection. A flaw in the logic of version
394 1.1.0 means that memory for the message is allocated too early, prior to
395 the excessive message length check. Due to way memory is allocated in
396 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
397 to service a connection. This could lead to a Denial of Service through
398 memory exhaustion. However, the excessive message length check still takes
399 place, and this would cause the connection to immediately fail. Assuming
400 that the application calls SSL_free() on the failed connection in a timely
401 manner then the 21Mb of allocated memory will then be immediately freed
402 again. Therefore the excessive memory allocation will be transitory in
403 nature. This then means that there is only a security impact if:
405 1) The application does not call SSL_free() in a timely manner in the event
406 that the connection fails
408 2) The application is working in a constrained environment where there is
409 very little free memory
411 3) The attacker initiates multiple connection attempts such that there are
412 multiple connections in a state where memory has been allocated for the
413 connection; SSL_free() has not yet been called; and there is insufficient
414 memory to service the multiple requests.
416 Except in the instance of (1) above any Denial Of Service is likely to be
417 transitory because as soon as the connection fails the memory is
418 subsequently freed again in the SSL_free() call. However there is an
419 increased risk during this period of application crashes due to the lack of
420 memory - which would then mean a more serious Denial of Service.
422 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
423 (CVE-2016-6307 and CVE-2016-6308)
426 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
427 had to be removed. Primary reason is that vendor assembler can't
428 assemble our modules with -KPIC flag. As result it, assembly
429 support, was not even available as option. But its lack means
430 lack of side-channel resistant code, which is incompatible with
431 security by todays standards. Fortunately gcc is readily available
432 prepackaged option, which we firmly point at...
435 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
437 *) Windows command-line tool supports UTF-8 opt-in option for arguments
438 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
439 (to any value) allows Windows user to access PKCS#12 file generated
440 with Windows CryptoAPI and protected with non-ASCII password, as well
441 as files generated under UTF-8 locale on Linux also protected with
445 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
446 have been disabled by default and removed from DEFAULT, just like RC4.
447 See the RC4 item below to re-enable both.
450 *) The method for finding the storage location for the Windows RAND seed file
451 has changed. First we check %RANDFILE%. If that is not set then we check
452 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
453 all else fails we fall back to C:\.
456 *) The EVP_EncryptUpdate() function has had its return type changed from void
457 to int. A return of 0 indicates and error while a return of 1 indicates
461 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
462 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
463 off the constant time implementation for RSA, DSA and DH have been made
464 no-ops and deprecated.
467 *) Windows RAND implementation was simplified to only get entropy by
468 calling CryptGenRandom(). Various other RAND-related tickets
470 [Joseph Wylie Yandle, Rich Salz]
472 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
473 and OPENSSL_LH_, respectively. The old names are available
474 with API compatibility. They new names are now completely documented.
477 *) Unify TYPE_up_ref(obj) methods signature.
478 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
479 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
480 int (instead of void) like all others TYPE_up_ref() methods.
481 So now these methods also check the return value of CRYPTO_atomic_add(),
482 and the validity of object reference counter.
483 [fdasilvayy@gmail.com]
485 *) With Windows Visual Studio builds, the .pdb files are installed
486 alongside the installed libraries and executables. For a static
487 library installation, ossl_static.pdb is the associate compiler
488 generated .pdb file to be used when linking programs.
491 *) Remove openssl.spec. Packaging files belong with the packagers.
494 *) Automatic Darwin/OSX configuration has had a refresh, it will now
495 recognise x86_64 architectures automatically. You can still decide
496 to build for a different bitness with the environment variable
497 KERNEL_BITS (can be 32 or 64), for example:
499 KERNEL_BITS=32 ./config
503 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
504 256 bit AES and HMAC with SHA256.
507 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
510 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
513 *) To enable users to have their own config files and build file templates,
514 Configure looks in the directory indicated by the environment variable
515 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
516 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
517 name and is used as is.
520 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
521 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
522 X509_CERT_FILE_CTX was removed.
525 *) "shared" builds are now the default. To create only static libraries use
526 the "no-shared" Configure option.
529 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
530 All of these option have not worked for some while and are fundamental
534 *) Make various cleanup routines no-ops and mark them as deprecated. Most
535 global cleanup functions are no longer required because they are handled
536 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
537 Explicitly de-initing can cause problems (e.g. where a library that uses
538 OpenSSL de-inits, but an application is still using it). The affected
539 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
540 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
541 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
545 *) --strict-warnings no longer enables runtime debugging options
546 such as REF_DEBUG. Instead, debug options are automatically
547 enabled with '--debug' builds.
548 [Andy Polyakov, Emilia Käsper]
550 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
551 have been moved out of the public header files. New functions for managing
552 these have been added.
555 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
556 objects have been moved out of the public header files. New
557 functions for managing these have been added.
560 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
561 have been moved out of the public header files. New functions for managing
562 these have been added.
565 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
566 moved out of the public header files. New functions for managing these
570 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
573 *) Removed the mk1mf build scripts.
576 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
577 it is always safe to #include a header now.
580 *) Removed the aged BC-32 config and all its supporting scripts
583 *) Removed support for Ultrix, Netware, and OS/2.
586 *) Add support for HKDF.
589 *) Add support for blake2b and blake2s
592 *) Added support for "pipelining". Ciphers that have the
593 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
594 encryptions/decryptions simultaneously. There are currently no built-in
595 ciphers with this property but the expectation is that engines will be able
596 to offer it to significantly improve throughput. Support has been extended
597 into libssl so that multiple records for a single connection can be
598 processed in one go (for >=TLS 1.1).
601 *) Added the AFALG engine. This is an async capable engine which is able to
602 offload work to the Linux kernel. In this initial version it only supports
603 AES128-CBC. The kernel must be version 4.1.0 or greater.
606 *) OpenSSL now uses a new threading API. It is no longer necessary to
607 set locking callbacks to use OpenSSL in a multi-threaded environment. There
608 are two supported threading models: pthreads and windows threads. It is
609 also possible to configure OpenSSL at compile time for "no-threads". The
610 old threading API should no longer be used. The functions have been
611 replaced with "no-op" compatibility macros.
612 [Alessandro Ghedini, Matt Caswell]
614 *) Modify behavior of ALPN to invoke callback after SNI/servername
615 callback, such that updates to the SSL_CTX affect ALPN.
618 *) Add SSL_CIPHER queries for authentication and key-exchange.
621 *) Changes to the DEFAULT cipherlist:
622 - Prefer (EC)DHE handshakes over plain RSA.
623 - Prefer AEAD ciphers over legacy ciphers.
624 - Prefer ECDSA over RSA when both certificates are available.
625 - Prefer TLSv1.2 ciphers/PRF.
626 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
630 *) Change the ECC default curve list to be this, in order: x25519,
631 secp256r1, secp521r1, secp384r1.
634 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
635 disabled by default. They can be re-enabled using the
636 enable-weak-ssl-ciphers option to Configure.
639 *) If the server has ALPN configured, but supports no protocols that the
640 client advertises, send a fatal "no_application_protocol" alert.
641 This behaviour is SHALL in RFC 7301, though it isn't universally
642 implemented by other servers.
645 *) Add X25519 support.
646 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
647 for public and private key encoding using the format documented in
648 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
649 key generation and key derivation.
651 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
655 *) Deprecate SRP_VBASE_get_by_user.
656 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
657 In order to fix an unavoidable memory leak (CVE-2016-0798),
658 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
659 seed, even if the seed is configured.
661 Users should use SRP_VBASE_get1_by_user instead. Note that in
662 SRP_VBASE_get1_by_user, caller must free the returned value. Note
663 also that even though configuring the SRP seed attempts to hide
664 invalid usernames by continuing the handshake with fake
665 credentials, this behaviour is not constant time and no strong
666 guarantees are made that the handshake is indistinguishable from
667 that of a valid user.
670 *) Configuration change; it's now possible to build dynamic engines
671 without having to build shared libraries and vice versa. This
672 only applies to the engines in engines/, those in crypto/engine/
673 will always be built into libcrypto (i.e. "static").
675 Building dynamic engines is enabled by default; to disable, use
676 the configuration option "disable-dynamic-engine".
678 The only requirements for building dynamic engines are the
679 presence of the DSO module and building with position independent
680 code, so they will also automatically be disabled if configuring
681 with "disable-dso" or "disable-pic".
683 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
684 are also taken away from openssl/opensslconf.h, as they are
688 *) Configuration change; if there is a known flag to compile
689 position independent code, it will always be applied on the
690 libcrypto and libssl object files, and never on the application
691 object files. This means other libraries that use routines from
692 libcrypto / libssl can be made into shared libraries regardless
693 of how OpenSSL was configured.
695 If this isn't desirable, the configuration options "disable-pic"
696 or "no-pic" can be used to disable the use of PIC. This will
697 also disable building shared libraries and dynamic engines.
700 *) Removed JPAKE code. It was experimental and has no wide use.
703 *) The INSTALL_PREFIX Makefile variable has been renamed to
704 DESTDIR. That makes for less confusion on what this variable
705 is for. Also, the configuration option --install_prefix is
709 *) Heartbeat for TLS has been removed and is disabled by default
710 for DTLS; configure with enable-heartbeats. Code that uses the
711 old #define's might need to be updated.
712 [Emilia Käsper, Rich Salz]
714 *) Rename REF_CHECK to REF_DEBUG.
717 *) New "unified" build system
719 The "unified" build system is aimed to be a common system for all
720 platforms we support. With it comes new support for VMS.
722 This system builds supports building in a different directory tree
723 than the source tree. It produces one Makefile (for unix family
724 or lookalikes), or one descrip.mms (for VMS).
726 The source of information to make the Makefile / descrip.mms is
727 small files called 'build.info', holding the necessary
728 information for each directory with source to compile, and a
729 template in Configurations, like unix-Makefile.tmpl or
732 With this change, the library names were also renamed on Windows
733 and on VMS. They now have names that are closer to the standard
734 on Unix, and include the major version number, and in certain
735 cases, the architecture they are built for. See "Notes on shared
736 libraries" in INSTALL.
738 We rely heavily on the perl module Text::Template.
741 *) Added support for auto-initialisation and de-initialisation of the library.
742 OpenSSL no longer requires explicit init or deinit routines to be called,
743 except in certain circumstances. See the OPENSSL_init_crypto() and
744 OPENSSL_init_ssl() man pages for further information.
747 *) The arguments to the DTLSv1_listen function have changed. Specifically the
748 "peer" argument is now expected to be a BIO_ADDR object.
750 *) Rewrite of BIO networking library. The BIO library lacked consistent
751 support of IPv6, and adding it required some more extensive
752 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
753 which hold all types of addresses and chains of address information.
754 It also introduces a new API, with functions like BIO_socket,
755 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
756 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
757 have been adapted accordingly.
760 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
764 *) CRIME protection: disable compression by default, even if OpenSSL is
765 compiled with zlib enabled. Applications can still enable compression
766 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
767 using the SSL_CONF library to configure compression.
770 *) The signature of the session callback configured with
771 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
772 was explicitly marked as 'const unsigned char*' instead of
776 *) Always DPURIFY. Remove the use of uninitialized memory in the
777 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
780 *) Removed many obsolete configuration items, including
781 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
782 MD2_CHAR, MD2_INT, MD2_LONG
784 IDEA_SHORT, IDEA_LONG
785 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
786 [Rich Salz, with advice from Andy Polyakov]
788 *) Many BN internals have been moved to an internal header file.
789 [Rich Salz with help from Andy Polyakov]
791 *) Configuration and writing out the results from it has changed.
792 Files such as Makefile include/openssl/opensslconf.h and are now
793 produced through general templates, such as Makefile.in and
794 crypto/opensslconf.h.in and some help from the perl module
797 Also, the center of configuration information is no longer
798 Makefile. Instead, Configure produces a perl module in
799 configdata.pm which holds most of the config data (in the hash
800 table %config), the target data that comes from the target
801 configuration in one of the Configurations/*.conf files (in
805 *) To clarify their intended purposes, the Configure options
806 --prefix and --openssldir change their semantics, and become more
807 straightforward and less interdependent.
809 --prefix shall be used exclusively to give the location INSTALLTOP
810 where programs, scripts, libraries, include files and manuals are
811 going to be installed. The default is now /usr/local.
813 --openssldir shall be used exclusively to give the default
814 location OPENSSLDIR where certificates, private keys, CRLs are
815 managed. This is also where the default openssl.cnf gets
817 If the directory given with this option is a relative path, the
818 values of both the --prefix value and the --openssldir value will
819 be combined to become OPENSSLDIR.
820 The default for --openssldir is INSTALLTOP/ssl.
822 Anyone who uses --openssldir to specify where OpenSSL is to be
823 installed MUST change to use --prefix instead.
826 *) The GOST engine was out of date and therefore it has been removed. An up
827 to date GOST engine is now being maintained in an external repository.
828 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
829 support for GOST ciphersuites (these are only activated if a GOST engine
833 *) EGD is no longer supported by default; use enable-egd when
835 [Ben Kaduk and Rich Salz]
837 *) The distribution now has Makefile.in files, which are used to
838 create Makefile's when Configure is run. *Configure must be run
839 before trying to build now.*
842 *) The return value for SSL_CIPHER_description() for error conditions
846 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
848 Obtaining and performing DNSSEC validation of TLSA records is
849 the application's responsibility. The application provides
850 the TLSA records of its choice to OpenSSL, and these are then
851 used to authenticate the peer.
853 The TLSA records need not even come from DNS. They can, for
854 example, be used to implement local end-entity certificate or
855 trust-anchor "pinning", where the "pin" data takes the form
856 of TLSA records, which can augment or replace verification
857 based on the usual WebPKI public certification authorities.
860 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
861 continues to support deprecated interfaces in default builds.
862 However, applications are strongly advised to compile their
863 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
864 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
865 or the 1.1.0 releases.
867 In environments in which all applications have been ported to
868 not use any deprecated interfaces OpenSSL's Configure script
869 should be used with the --api=1.1.0 option to entirely remove
870 support for the deprecated features from the library and
871 unconditionally disable them in the installed headers.
872 Essentially the same effect can be achieved with the "no-deprecated"
873 argument to Configure, except that this will always restrict
874 the build to just the latest API, rather than a fixed API
877 As applications are ported to future revisions of the API,
878 they should update their compile-time OPENSSL_API_COMPAT define
879 accordingly, but in most cases should be able to continue to
880 compile with later releases.
882 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
883 0x10000000L and 0x00908000L, respectively. However those
884 versions did not support the OPENSSL_API_COMPAT feature, and
885 so applications are not typically tested for explicit support
886 of just the undeprecated features of either release.
889 *) Add support for setting the minimum and maximum supported protocol.
890 It can bet set via the SSL_set_min_proto_version() and
891 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
892 MaxProtcol. It's recommended to use the new APIs to disable
893 protocols instead of disabling individual protocols using
894 SSL_set_options() or SSL_CONF's Protocol. This change also
895 removes support for disabling TLS 1.2 in the OpenSSL TLS
896 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
899 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
902 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
903 and integrates ECDSA and ECDH functionality into EC. Implementations can
904 now redirect key generation and no longer need to convert to or from
907 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
908 include the ec.h header file instead.
911 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
912 ciphers who are no longer supported and drops support the ephemeral RSA key
913 exchange. The LOW ciphers currently doesn't have any ciphers in it.
916 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
917 opaque. For HMAC_CTX, the following constructors and destructors
920 HMAC_CTX *HMAC_CTX_new(void);
921 void HMAC_CTX_free(HMAC_CTX *ctx);
923 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
924 destroy such methods has been added. See EVP_MD_meth_new(3) and
925 EVP_CIPHER_meth_new(3) for documentation.
928 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
929 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
930 EVP_MD_CTX_reset() should be called instead to reinitialise
931 an already created structure.
932 2) For consistency with the majority of our object creators and
933 destructors, EVP_MD_CTX_(create|destroy) were renamed to
934 EVP_MD_CTX_(new|free). The old names are retained as macros
935 for deprecated builds.
938 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
939 cryptographic operations to be performed asynchronously as long as an
940 asynchronous capable engine is used. See the ASYNC_start_job() man page for
941 further details. Libssl has also had this capability integrated with the
942 introduction of the new mode SSL_MODE_ASYNC and associated error
943 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
944 pages. This work was developed in partnership with Intel Corp.
947 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
948 always enabled now. If you want to disable the support you should
949 exclude it using the list of supported ciphers. This also means that the
950 "-no_ecdhe" option has been removed from s_server.
953 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
954 SSL_{CTX_}set1_curves() which can set a list.
957 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
958 curve you want to support using SSL_{CTX_}set1_curves().
961 *) State machine rewrite. The state machine code has been significantly
962 refactored in order to remove much duplication of code and solve issues
963 with the old code (see ssl/statem/README for further details). This change
964 does have some associated API changes. Notably the SSL_state() function
965 has been removed and replaced by SSL_get_state which now returns an
966 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
967 altogether. The previous handshake states defined in ssl.h and ssl3.h have
971 *) All instances of the string "ssleay" in the public API were replaced
972 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
973 Some error codes related to internal RSA_eay API's were renamed.
976 *) The demo files in crypto/threads were moved to demo/threads.
979 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
981 [Matt Caswell, Rich Salz]
983 *) New ASN.1 embed macro.
985 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
986 structure is not allocated: it is part of the parent. That is instead of
994 This reduces memory fragmentation and make it impossible to accidentally
995 set a mandatory field to NULL.
997 This currently only works for some fields specifically a SEQUENCE, CHOICE,
998 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
999 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1003 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1006 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1007 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1008 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1009 DES and RC4 ciphersuites.
1012 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1013 This changes the decoding behaviour for some invalid messages,
1014 though the change is mostly in the more lenient direction, and
1015 legacy behaviour is preserved as much as possible.
1018 *) Fix no-stdio build.
1019 [ David Woodhouse <David.Woodhouse@intel.com> and also
1020 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1022 *) New testing framework
1023 The testing framework has been largely rewritten and is now using
1024 perl and the perl modules Test::Harness and an extended variant of
1025 Test::More called OpenSSL::Test to do its work. All test scripts in
1026 test/ have been rewritten into test recipes, and all direct calls to
1027 executables in test/Makefile have become individual recipes using the
1028 simplified testing OpenSSL::Test::Simple.
1030 For documentation on our testing modules, do:
1032 perldoc test/testlib/OpenSSL/Test/Simple.pm
1033 perldoc test/testlib/OpenSSL/Test.pm
1037 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1038 are used; the latter aborts on memory leaks (usually checked on exit).
1039 Some undocumented "set malloc, etc., hooks" functions were removed
1040 and others were changed. All are now documented.
1043 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1045 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1047 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1048 from RFC4279, RFC4785, RFC5487, RFC5489.
1050 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1051 original RSA_PSK patch.
1054 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1055 era flag was never set throughout the codebase (only read). Also removed
1056 SSL3_FLAGS_POP_BUFFER which was only used if
1057 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1060 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1061 to be "oneline" instead of "compat".
1064 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1065 not aware of clients that still exhibit this bug, and the workaround
1066 hasn't been working properly for a while.
1069 *) The return type of BIO_number_read() and BIO_number_written() as well as
1070 the corresponding num_read and num_write members in the BIO structure has
1071 changed from unsigned long to uint64_t. On platforms where an unsigned
1072 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1076 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1077 OpenSSL without support for them. It also means that maintaining
1078 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1079 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1082 *) Removed support for the two export grade static DH ciphersuites
1083 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1084 were newly added (along with a number of other static DH ciphersuites) to
1085 1.0.2. However the two export ones have *never* worked since they were
1086 introduced. It seems strange in any case to be adding new export
1087 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1090 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1091 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1092 and turned into macros which simply call the new preferred function names
1093 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1094 should use the new names instead. Also as part of this change the ssl23.h
1095 header file has been removed.
1098 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1099 code and the associated standard is no longer considered fit-for-purpose.
1102 *) RT2547 was closed. When generating a private key, try to make the
1103 output file readable only by the owner. This behavior change might
1104 be noticeable when interacting with other software.
1106 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1110 *) Added HTTP GET support to the ocsp command.
1113 *) Changed default digest for the dgst and enc commands from MD5 to
1117 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1120 *) Added support for TLS extended master secret from
1121 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1122 initial patch which was a great help during development.
1125 *) All libssl internal structures have been removed from the public header
1126 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1127 now redundant). Users should not attempt to access internal structures
1128 directly. Instead they should use the provided API functions.
1131 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1132 Access to deprecated functions can be re-enabled by running config with
1133 "enable-deprecated". In addition applications wishing to use deprecated
1134 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1135 will, by default, disable some transitive includes that previously existed
1136 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1139 *) Added support for OCB mode. OpenSSL has been granted a patent license
1140 compatible with the OpenSSL license for use of OCB. Details are available
1141 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1142 for OCB can be removed by calling config with no-ocb.
1145 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1146 compatible client hello.
1149 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1150 done while fixing the error code for the key-too-small case.
1151 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1153 *) CA.sh has been removed; use CA.pl instead.
1156 *) Removed old DES API.
1159 *) Remove various unsupported platforms:
1165 Sinix/ReliantUNIX RM400
1170 16-bit platforms such as WIN16
1173 *) Clean up OPENSSL_NO_xxx #define's
1174 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1175 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1176 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1177 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1178 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1179 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1180 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1181 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1182 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1183 Remove MS_STATIC; it's a relic from platforms <32 bits.
1186 *) Cleaned up dead code
1187 Remove all but one '#ifdef undef' which is to be looked at.
1190 *) Clean up calling of xxx_free routines.
1191 Just like free(), fix most of the xxx_free routines to accept
1192 NULL. Remove the non-null checks from callers. Save much code.
1195 *) Add secure heap for storage of private keys (when possible).
1196 Add BIO_s_secmem(), CBIGNUM, etc.
1197 Contributed by Akamai Technologies under our Corporate CLA.
1200 *) Experimental support for a new, fast, unbiased prime candidate generator,
1201 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1202 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1204 *) New output format NSS in the sess_id command line tool. This allows
1205 exporting the session id and the master key in NSS keylog format.
1206 [Martin Kaiser <martin@kaiser.cx>]
1208 *) Harmonize version and its documentation. -f flag is used to display
1210 [mancha <mancha1@zoho.com>]
1212 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1213 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1214 [mancha <mancha1@zoho.com>]
1216 *) Fix some double frees. These are not thought to be exploitable.
1217 [mancha <mancha1@zoho.com>]
1219 *) A missing bounds check in the handling of the TLS heartbeat extension
1220 can be used to reveal up to 64k of memory to a connected client or
1223 Thanks for Neel Mehta of Google Security for discovering this bug and to
1224 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1225 preparing the fix (CVE-2014-0160)
1226 [Adam Langley, Bodo Moeller]
1228 *) Fix for the attack described in the paper "Recovering OpenSSL
1229 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1230 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1231 http://eprint.iacr.org/2014/140
1233 Thanks to Yuval Yarom and Naomi Benger for discovering this
1234 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1235 [Yuval Yarom and Naomi Benger]
1237 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1238 this fixes a limitation in previous versions of OpenSSL.
1241 *) Experimental encrypt-then-mac support.
1243 Experimental support for encrypt then mac from
1244 draft-gutmann-tls-encrypt-then-mac-02.txt
1246 To enable it set the appropriate extension number (0x42 for the test
1247 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1249 For non-compliant peers (i.e. just about everything) this should have no
1252 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1256 *) Add EVP support for key wrapping algorithms, to avoid problems with
1257 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1258 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1259 algorithms and include tests cases.
1262 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1266 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1267 MGF1 digest and OAEP label.
1270 *) Make openssl verify return errors.
1271 [Chris Palmer <palmer@google.com> and Ben Laurie]
1273 *) New function ASN1_TIME_diff to calculate the difference between two
1274 ASN1_TIME structures or one structure and the current time.
1277 *) Update fips_test_suite to support multiple command line options. New
1278 test to induce all self test errors in sequence and check expected
1282 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1283 sign or verify all in one operation.
1286 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1287 test programs and fips_test_suite. Includes functionality to parse
1288 the minimal script output of fipsalgest.pl directly.
1291 *) Add authorisation parameter to FIPS_module_mode_set().
1294 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1297 *) Use separate DRBG fields for internal and external flags. New function
1298 FIPS_drbg_health_check() to perform on demand health checking. Add
1299 generation tests to fips_test_suite with reduced health check interval to
1300 demonstrate periodic health checking. Add "nodh" option to
1301 fips_test_suite to skip very slow DH test.
1304 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1308 *) More extensive health check for DRBG checking many more failure modes.
1309 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1310 combination: call this in fips_test_suite.
1313 *) Add support for canonical generation of DSA parameter 'g'. See
1316 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1317 POST to handle HMAC cases.
1320 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1321 to return numerical and string versions of the FIPS module number.
1324 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1325 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1326 outside the validated module in the FIPS capable OpenSSL.
1329 *) Minor change to DRBG entropy callback semantics. In some cases
1330 there is no multiple of the block length between min_len and
1331 max_len. Allow the callback to return more than max_len bytes
1332 of entropy but discard any extra: it is the callback's responsibility
1333 to ensure that the extra data discarded does not impact the
1334 requested amount of entropy.
1337 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1338 information in FIPS186-3, SP800-57 and SP800-131A.
1341 *) CCM support via EVP. Interface is very similar to GCM case except we
1342 must supply all data in one chunk (i.e. no update, final) and the
1343 message length must be supplied if AAD is used. Add algorithm test
1347 *) Initial version of POST overhaul. Add POST callback to allow the status
1348 of POST to be monitored and/or failures induced. Modify fips_test_suite
1349 to use callback. Always run all selftests even if one fails.
1352 *) XTS support including algorithm test driver in the fips_gcmtest program.
1353 Note: this does increase the maximum key length from 32 to 64 bytes but
1354 there should be no binary compatibility issues as existing applications
1355 will never use XTS mode.
1358 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1359 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1360 performs algorithm blocking for unapproved PRNG types. Also do not
1361 set PRNG type in FIPS_mode_set(): leave this to the application.
1362 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1363 the standard OpenSSL PRNG: set additional data to a date time vector.
1366 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1367 This shouldn't present any incompatibility problems because applications
1368 shouldn't be using these directly and any that are will need to rethink
1369 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1372 *) Extensive self tests and health checking required by SP800-90 DRBG.
1373 Remove strength parameter from FIPS_drbg_instantiate and always
1374 instantiate at maximum supported strength.
1377 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1380 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1383 *) New function DH_compute_key_padded() to compute a DH key and pad with
1384 leading zeroes if needed: this complies with SP800-56A et al.
1387 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1388 anything, incomplete, subject to change and largely untested at present.
1391 *) Modify fipscanisteronly build option to only build the necessary object
1392 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1395 *) Add experimental option FIPSSYMS to give all symbols in
1396 fipscanister.o and FIPS or fips prefix. This will avoid
1397 conflicts with future versions of OpenSSL. Add perl script
1398 util/fipsas.pl to preprocess assembly language source files
1399 and rename any affected symbols.
1402 *) Add selftest checks and algorithm block of non-fips algorithms in
1403 FIPS mode. Remove DES2 from selftests.
1406 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1407 return internal method without any ENGINE dependencies. Add new
1408 tiny fips sign and verify functions.
1411 *) New build option no-ec2m to disable characteristic 2 code.
1414 *) New build option "fipscanisteronly". This only builds fipscanister.o
1415 and (currently) associated fips utilities. Uses the file Makefile.fips
1416 instead of Makefile.org as the prototype.
1419 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1420 Update fips_gcmtest to use IV generator.
1423 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1424 setting output buffer to NULL. The *Final function must be
1425 called although it will not retrieve any additional data. The tag
1426 can be set or retrieved with a ctrl. The IV length is by default 12
1427 bytes (96 bits) but can be set to an alternative value. If the IV
1428 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1432 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1433 underlying do_cipher function handles all cipher semantics itself
1434 including padding and finalisation. This is useful if (for example)
1435 an ENGINE cipher handles block padding itself. The behaviour of
1436 do_cipher is subtly changed if this flag is set: the return value
1437 is the number of characters written to the output buffer (zero is
1438 no longer an error code) or a negative error code. Also if the
1439 input buffer is NULL and length 0 finalisation should be performed.
1442 *) If a candidate issuer certificate is already part of the constructed
1443 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1446 *) Improve forward-security support: add functions
1448 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1449 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1451 for use by SSL/TLS servers; the callback function will be called whenever a
1452 new session is created, and gets to decide whether the session may be
1453 cached to make it resumable (return 0) or not (return 1). (As by the
1454 SSL/TLS protocol specifications, the session_id sent by the server will be
1455 empty to indicate that the session is not resumable; also, the server will
1456 not generate RFC 4507 (RFC 5077) session tickets.)
1458 A simple reasonable callback implementation is to return is_forward_secure.
1459 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1460 by the SSL/TLS server library, indicating whether it can provide forward
1462 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1464 *) New -verify_name option in command line utilities to set verification
1468 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1469 Add CMAC pkey methods.
1472 *) Experimental renegotiation in s_server -www mode. If the client
1473 browses /reneg connection is renegotiated. If /renegcert it is
1474 renegotiated requesting a certificate.
1477 *) Add an "external" session cache for debugging purposes to s_server. This
1478 should help trace issues which normally are only apparent in deployed
1479 multi-process servers.
1482 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1483 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1484 BIO_set_cipher() and some obscure PEM functions were changed so they
1485 can now return an error. The RAND changes required a change to the
1486 RAND_METHOD structure.
1489 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1490 a gcc attribute to warn if the result of a function is ignored. This
1491 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1492 whose return value is often ignored.
1495 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1496 These allow SCTs (signed certificate timestamps) to be requested and
1497 validated when establishing a connection.
1498 [Rob Percival <robpercival@google.com>]
1500 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1502 *) Prevent padding oracle in AES-NI CBC MAC check
1504 A MITM attacker can use a padding oracle attack to decrypt traffic
1505 when the connection uses an AES CBC cipher and the server support
1508 This issue was introduced as part of the fix for Lucky 13 padding
1509 attack (CVE-2013-0169). The padding check was rewritten to be in
1510 constant time by making sure that always the same bytes are read and
1511 compared against either the MAC or padding bytes. But it no longer
1512 checked that there was enough data to have both the MAC and padding
1515 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1519 *) Fix EVP_EncodeUpdate overflow
1521 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1522 Base64 encoding of binary data. If an attacker is able to supply very large
1523 amounts of input data then a length check can overflow resulting in a heap
1526 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1527 the PEM_write_bio* family of functions. These are mainly used within the
1528 OpenSSL command line applications, so any application which processes data
1529 from an untrusted source and outputs it as a PEM file should be considered
1530 vulnerable to this issue. User applications that call these APIs directly
1531 with large amounts of untrusted data may also be vulnerable.
1533 This issue was reported by Guido Vranken.
1537 *) Fix EVP_EncryptUpdate overflow
1539 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1540 is able to supply very large amounts of input data after a previous call to
1541 EVP_EncryptUpdate() with a partial block then a length check can overflow
1542 resulting in a heap corruption. Following an analysis of all OpenSSL
1543 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1544 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1545 the first called function after an EVP_EncryptInit(), and therefore that
1546 specific call must be safe. The second form is where the length passed to
1547 EVP_EncryptUpdate() can be seen from the code to be some small value and
1548 therefore there is no possibility of an overflow. Since all instances are
1549 one of these two forms, it is believed that there can be no overflows in
1550 internal code due to this problem. It should be noted that
1551 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1552 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1553 of these calls have also been analysed too and it is believed there are no
1554 instances in internal usage where an overflow could occur.
1556 This issue was reported by Guido Vranken.
1560 *) Prevent ASN.1 BIO excessive memory allocation
1562 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1563 a short invalid encoding can cause allocation of large amounts of memory
1564 potentially consuming excessive resources or exhausting memory.
1566 Any application parsing untrusted data through d2i BIO functions is
1567 affected. The memory based functions such as d2i_X509() are *not* affected.
1568 Since the memory based functions are used by the TLS library, TLS
1569 applications are not affected.
1571 This issue was reported by Brian Carpenter.
1577 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1578 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1579 in arbitrary stack data being returned in the buffer.
1581 This issue was reported by Guido Vranken.
1585 *) Modify behavior of ALPN to invoke callback after SNI/servername
1586 callback, such that updates to the SSL_CTX affect ALPN.
1589 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1593 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1594 methods are enabled and ssl2 is disabled the methods return NULL.
1597 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1599 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1600 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1601 provide any "EXPORT" or "LOW" strength ciphers.
1604 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1605 is by default disabled at build-time. Builds that are not configured with
1606 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1607 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1608 will need to explicitly call either of:
1610 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1612 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1614 as appropriate. Even if either of those is used, or the application
1615 explicitly uses the version-specific SSLv2_method() or its client and
1616 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1617 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1618 ciphers, and SSLv2 56-bit DES are no longer available.
1622 *) Fix a double-free in DSA code
1624 A double free bug was discovered when OpenSSL parses malformed DSA private
1625 keys and could lead to a DoS attack or memory corruption for applications
1626 that receive DSA private keys from untrusted sources. This scenario is
1629 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1634 *) Disable SRP fake user seed to address a server memory leak.
1636 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1638 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1639 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1640 was changed to ignore the "fake user" SRP seed, even if the seed
1643 Users should use SRP_VBASE_get1_by_user instead. Note that in
1644 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1645 also that even though configuring the SRP seed attempts to hide
1646 invalid usernames by continuing the handshake with fake
1647 credentials, this behaviour is not constant time and no strong
1648 guarantees are made that the handshake is indistinguishable from
1649 that of a valid user.
1653 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1655 In the BN_hex2bn function the number of hex digits is calculated using an
1656 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1657 large values of |i| this can result in |bn_expand| not allocating any
1658 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1659 field as NULL leading to a subsequent NULL ptr deref. For very large values
1660 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1661 In this case memory is allocated to the internal BIGNUM data field, but it
1662 is insufficiently sized leading to heap corruption. A similar issue exists
1663 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1664 is ever called by user applications with very large untrusted hex/dec data.
1665 This is anticipated to be a rare occurrence.
1667 All OpenSSL internal usage of these functions use data that is not expected
1668 to be untrusted, e.g. config file data or application command line
1669 arguments. If user developed applications generate config file data based
1670 on untrusted data then it is possible that this could also lead to security
1671 consequences. This is also anticipated to be rare.
1673 This issue was reported to OpenSSL by Guido Vranken.
1677 *) Fix memory issues in BIO_*printf functions
1679 The internal |fmtstr| function used in processing a "%s" format string in
1680 the BIO_*printf functions could overflow while calculating the length of a
1681 string and cause an OOB read when printing very long strings.
1683 Additionally the internal |doapr_outch| function can attempt to write to an
1684 OOB memory location (at an offset from the NULL pointer) in the event of a
1685 memory allocation failure. In 1.0.2 and below this could be caused where
1686 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1687 could be in processing a very long "%s" format string. Memory leaks can
1690 The first issue may mask the second issue dependent on compiler behaviour.
1691 These problems could enable attacks where large amounts of untrusted data
1692 is passed to the BIO_*printf functions. If applications use these functions
1693 in this way then they could be vulnerable. OpenSSL itself uses these
1694 functions when printing out human-readable dumps of ASN.1 data. Therefore
1695 applications that print this data could be vulnerable if the data is from
1696 untrusted sources. OpenSSL command line applications could also be
1697 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1698 as command line arguments.
1700 Libssl is not considered directly vulnerable. Additionally certificates etc
1701 received via remote connections via libssl are also unlikely to be able to
1702 trigger these issues because of message size limits enforced within libssl.
1704 This issue was reported to OpenSSL Guido Vranken.
1708 *) Side channel attack on modular exponentiation
1710 A side-channel attack was found which makes use of cache-bank conflicts on
1711 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1712 of RSA keys. The ability to exploit this issue is limited as it relies on
1713 an attacker who has control of code in a thread running on the same
1714 hyper-threaded core as the victim thread which is performing decryptions.
1716 This issue was reported to OpenSSL by Yuval Yarom, The University of
1717 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1718 Nadia Heninger, University of Pennsylvania with more information at
1719 http://cachebleed.info.
1723 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1724 if no keysize is specified with default_bits. This fixes an
1725 omission in an earlier change that changed all RSA/DSA key generation
1726 apps to use 2048 bits by default.
1729 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1730 *) DH small subgroups
1732 Historically OpenSSL only ever generated DH parameters based on "safe"
1733 primes. More recently (in version 1.0.2) support was provided for
1734 generating X9.42 style parameter files such as those required for RFC 5114
1735 support. The primes used in such files may not be "safe". Where an
1736 application is using DH configured with parameters based on primes that are
1737 not "safe" then an attacker could use this fact to find a peer's private
1738 DH exponent. This attack requires that the attacker complete multiple
1739 handshakes in which the peer uses the same private DH exponent. For example
1740 this could be used to discover a TLS server's private DH exponent if it's
1741 reusing the private DH exponent or it's using a static DH ciphersuite.
1743 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1744 TLS. It is not on by default. If the option is not set then the server
1745 reuses the same private DH exponent for the life of the server process and
1746 would be vulnerable to this attack. It is believed that many popular
1747 applications do set this option and would therefore not be at risk.
1749 The fix for this issue adds an additional check where a "q" parameter is
1750 available (as is the case in X9.42 based parameters). This detects the
1751 only known attack, and is the only possible defense for static DH
1752 ciphersuites. This could have some performance impact.
1754 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1755 default and cannot be disabled. This could have some performance impact.
1757 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1761 *) SSLv2 doesn't block disabled ciphers
1763 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1764 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1765 been disabled, provided that the SSLv2 protocol was not also disabled via
1768 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1769 and Sebastian Schinzel.
1773 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1775 *) BN_mod_exp may produce incorrect results on x86_64
1777 There is a carry propagating bug in the x86_64 Montgomery squaring
1778 procedure. No EC algorithms are affected. Analysis suggests that attacks
1779 against RSA and DSA as a result of this defect would be very difficult to
1780 perform and are not believed likely. Attacks against DH are considered just
1781 feasible (although very difficult) because most of the work necessary to
1782 deduce information about a private key may be performed offline. The amount
1783 of resources required for such an attack would be very significant and
1784 likely only accessible to a limited number of attackers. An attacker would
1785 additionally need online access to an unpatched system using the target
1786 private key in a scenario with persistent DH parameters and a private
1787 key that is shared between multiple clients. For example this can occur by
1788 default in OpenSSL DHE based SSL/TLS ciphersuites.
1790 This issue was reported to OpenSSL by Hanno Böck.
1794 *) Certificate verify crash with missing PSS parameter
1796 The signature verification routines will crash with a NULL pointer
1797 dereference if presented with an ASN.1 signature using the RSA PSS
1798 algorithm and absent mask generation function parameter. Since these
1799 routines are used to verify certificate signature algorithms this can be
1800 used to crash any certificate verification operation and exploited in a
1801 DoS attack. Any application which performs certificate verification is
1802 vulnerable including OpenSSL clients and servers which enable client
1805 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1809 *) X509_ATTRIBUTE memory leak
1811 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1812 memory. This structure is used by the PKCS#7 and CMS routines so any
1813 application which reads PKCS#7 or CMS data from untrusted sources is
1814 affected. SSL/TLS is not affected.
1816 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1821 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1822 This changes the decoding behaviour for some invalid messages,
1823 though the change is mostly in the more lenient direction, and
1824 legacy behaviour is preserved as much as possible.
1827 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1829 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1831 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1833 *) Alternate chains certificate forgery
1835 During certificate verification, OpenSSL will attempt to find an
1836 alternative certificate chain if the first attempt to build such a chain
1837 fails. An error in the implementation of this logic can mean that an
1838 attacker could cause certain checks on untrusted certificates to be
1839 bypassed, such as the CA flag, enabling them to use a valid leaf
1840 certificate to act as a CA and "issue" an invalid certificate.
1842 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1846 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1848 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1849 incompatibility in the handling of HMAC. The previous ABI has now been
1853 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1855 *) Malformed ECParameters causes infinite loop
1857 When processing an ECParameters structure OpenSSL enters an infinite loop
1858 if the curve specified is over a specially malformed binary polynomial
1861 This can be used to perform denial of service against any
1862 system which processes public keys, certificate requests or
1863 certificates. This includes TLS clients and TLS servers with
1864 client authentication enabled.
1866 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1870 *) Exploitable out-of-bounds read in X509_cmp_time
1872 X509_cmp_time does not properly check the length of the ASN1_TIME
1873 string and can read a few bytes out of bounds. In addition,
1874 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1877 An attacker can use this to craft malformed certificates and CRLs of
1878 various sizes and potentially cause a segmentation fault, resulting in
1879 a DoS on applications that verify certificates or CRLs. TLS clients
1880 that verify CRLs are affected. TLS clients and servers with client
1881 authentication enabled may be affected if they use custom verification
1884 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1885 independently by Hanno Böck.
1889 *) PKCS7 crash with missing EnvelopedContent
1891 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1892 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1893 with missing content and trigger a NULL pointer dereference on parsing.
1895 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1896 structures from untrusted sources are affected. OpenSSL clients and
1897 servers are not affected.
1899 This issue was reported to OpenSSL by Michal Zalewski (Google).
1903 *) CMS verify infinite loop with unknown hash function
1905 When verifying a signedData message the CMS code can enter an infinite loop
1906 if presented with an unknown hash function OID. This can be used to perform
1907 denial of service against any system which verifies signedData messages using
1909 This issue was reported to OpenSSL by Johannes Bauer.
1913 *) Race condition handling NewSessionTicket
1915 If a NewSessionTicket is received by a multi-threaded client when attempting to
1916 reuse a previous ticket then a race condition can occur potentially leading to
1917 a double free of the ticket data.
1921 *) Only support 256-bit or stronger elliptic curves with the
1922 'ecdh_auto' setting (server) or by default (client). Of supported
1923 curves, prefer P-256 (both).
1926 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1928 *) ClientHello sigalgs DoS fix
1930 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1931 invalid signature algorithms extension a NULL pointer dereference will
1932 occur. This can be exploited in a DoS attack against the server.
1934 This issue was was reported to OpenSSL by David Ramos of Stanford
1937 [Stephen Henson and Matt Caswell]
1939 *) Multiblock corrupted pointer fix
1941 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1942 feature only applies on 64 bit x86 architecture platforms that support AES
1943 NI instructions. A defect in the implementation of "multiblock" can cause
1944 OpenSSL's internal write buffer to become incorrectly set to NULL when
1945 using non-blocking IO. Typically, when the user application is using a
1946 socket BIO for writing, this will only result in a failed connection.
1947 However if some other BIO is used then it is likely that a segmentation
1948 fault will be triggered, thus enabling a potential DoS attack.
1950 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1954 *) Segmentation fault in DTLSv1_listen fix
1956 The DTLSv1_listen function is intended to be stateless and processes the
1957 initial ClientHello from many peers. It is common for user code to loop
1958 over the call to DTLSv1_listen until a valid ClientHello is received with
1959 an associated cookie. A defect in the implementation of DTLSv1_listen means
1960 that state is preserved in the SSL object from one invocation to the next
1961 that can lead to a segmentation fault. Errors processing the initial
1962 ClientHello can trigger this scenario. An example of such an error could be
1963 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1966 This issue was reported to OpenSSL by Per Allansson.
1970 *) Segmentation fault in ASN1_TYPE_cmp fix
1972 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1973 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1974 certificate signature algorithm consistency this can be used to crash any
1975 certificate verification operation and exploited in a DoS attack. Any
1976 application which performs certificate verification is vulnerable including
1977 OpenSSL clients and servers which enable client authentication.
1981 *) Segmentation fault for invalid PSS parameters fix
1983 The signature verification routines will crash with a NULL pointer
1984 dereference if presented with an ASN.1 signature using the RSA PSS
1985 algorithm and invalid parameters. Since these routines are used to verify
1986 certificate signature algorithms this can be used to crash any
1987 certificate verification operation and exploited in a DoS attack. Any
1988 application which performs certificate verification is vulnerable including
1989 OpenSSL clients and servers which enable client authentication.
1991 This issue was was reported to OpenSSL by Brian Carpenter.
1995 *) ASN.1 structure reuse memory corruption fix
1997 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1998 memory corruption via an invalid write. Such reuse is and has been
1999 strongly discouraged and is believed to be rare.
2001 Applications that parse structures containing CHOICE or ANY DEFINED BY
2002 components may be affected. Certificate parsing (d2i_X509 and related
2003 functions) are however not affected. OpenSSL clients and servers are
2008 *) PKCS7 NULL pointer dereferences fix
2010 The PKCS#7 parsing code does not handle missing outer ContentInfo
2011 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2012 missing content and trigger a NULL pointer dereference on parsing.
2014 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2015 otherwise parse PKCS#7 structures from untrusted sources are
2016 affected. OpenSSL clients and servers are not affected.
2018 This issue was reported to OpenSSL by Michal Zalewski (Google).
2022 *) DoS via reachable assert in SSLv2 servers fix
2024 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2025 servers that both support SSLv2 and enable export cipher suites by sending
2026 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2028 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2029 (OpenSSL development team).
2033 *) Empty CKE with client auth and DHE fix
2035 If client auth is used then a server can seg fault in the event of a DHE
2036 ciphersuite being selected and a zero length ClientKeyExchange message
2037 being sent by the client. This could be exploited in a DoS attack.
2041 *) Handshake with unseeded PRNG fix
2043 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2044 with an unseeded PRNG. The conditions are:
2045 - The client is on a platform where the PRNG has not been seeded
2046 automatically, and the user has not seeded manually
2047 - A protocol specific client method version has been used (i.e. not
2048 SSL_client_methodv23)
2049 - A ciphersuite is used that does not require additional random data from
2050 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2052 If the handshake succeeds then the client random that has been used will
2053 have been generated from a PRNG with insufficient entropy and therefore the
2054 output may be predictable.
2056 For example using the following command with an unseeded openssl will
2057 succeed on an unpatched platform:
2059 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2063 *) Use After Free following d2i_ECPrivatekey error fix
2065 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2066 could cause a use after free condition. This, in turn, could cause a double
2067 free in several private key parsing functions (such as d2i_PrivateKey
2068 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2069 for applications that receive EC private keys from untrusted
2070 sources. This scenario is considered rare.
2072 This issue was discovered by the BoringSSL project and fixed in their
2077 *) X509_to_X509_REQ NULL pointer deref fix
2079 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2080 the certificate key is invalid. This function is rarely used in practice.
2082 This issue was discovered by Brian Carpenter.
2086 *) Removed the export ciphers from the DEFAULT ciphers
2089 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2091 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2092 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2093 So far those who have to target multiple platforms would compromise
2094 and argue that binary targeting say ARMv5 would still execute on
2095 ARMv8. "Universal" build resolves this compromise by providing
2096 near-optimal performance even on newer platforms.
2099 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2100 (other platforms pending).
2101 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2103 *) Add support for the SignedCertificateTimestampList certificate and
2104 OCSP response extensions from RFC6962.
2107 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2108 for corner cases. (Certain input points at infinity could lead to
2109 bogus results, with non-infinity inputs mapped to infinity too.)
2112 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2113 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2114 common cases are optimized and there still is room for further
2115 improvements. Vector Permutation AES for Altivec is also added.
2118 *) Add support for little-endian ppc64 Linux target.
2119 [Marcelo Cerri (IBM)]
2121 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2122 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2123 are optimized and there still is room for further improvements.
2124 Both 32- and 64-bit modes are supported.
2125 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2127 *) Improved ARMv7 NEON support.
2130 *) Support for SPARC Architecture 2011 crypto extensions, first
2131 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2132 SHA256/512, MD5, GHASH and modular exponentiation.
2133 [Andy Polyakov, David Miller]
2135 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2137 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2139 *) Support for new and upcoming Intel processors, including AVX2,
2140 BMI and SHA ISA extensions. This includes additional "stitched"
2141 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2144 This work was sponsored by Intel Corp.
2147 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2148 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2149 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2152 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2153 this fixes a limitation in previous versions of OpenSSL.
2156 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2157 MGF1 digest and OAEP label.
2160 *) Add EVP support for key wrapping algorithms, to avoid problems with
2161 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2162 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2163 algorithms and include tests cases.
2166 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2168 [Douglas E. Engert, Steve Henson]
2170 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2171 difference in days and seconds between two tm or ASN1_TIME structures.
2174 *) Add -rev test option to s_server to just reverse order of characters
2175 received by client and send back to server. Also prints an abbreviated
2176 summary of the connection parameters.
2179 *) New option -brief for s_client and s_server to print out a brief summary
2180 of connection parameters.
2183 *) Add callbacks for arbitrary TLS extensions.
2184 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2186 *) New option -crl_download in several openssl utilities to download CRLs
2187 from CRLDP extension in certificates.
2190 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2193 *) New function X509_CRL_diff to generate a delta CRL from the difference
2194 of two full CRLs. Add support to "crl" utility.
2197 *) New functions to set lookup_crls function and to retrieve
2198 X509_STORE from X509_STORE_CTX.
2201 *) Print out deprecated issuer and subject unique ID fields in
2205 *) Extend OCSP I/O functions so they can be used for simple general purpose
2206 HTTP as well as OCSP. New wrapper function which can be used to download
2207 CRLs using the OCSP API.
2210 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2213 *) SSL_CONF* functions. These provide a common framework for application
2214 configuration using configuration files or command lines.
2217 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2218 message callback and prints the results. Needs compile time option
2219 "enable-ssl-trace". New options to s_client and s_server to enable
2223 *) New ctrl and macro to retrieve supported points extensions.
2224 Print out extension in s_server and s_client.
2227 *) New functions to retrieve certificate signature and signature
2231 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2235 *) New Suite B modes for TLS code. These use and enforce the requirements
2236 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2237 only use Suite B curves. The Suite B modes can be set by using the
2238 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2241 *) New chain verification flags for Suite B levels of security. Check
2242 algorithms are acceptable when flags are set in X509_verify_cert.
2245 *) Make tls1_check_chain return a set of flags indicating checks passed
2246 by a certificate chain. Add additional tests to handle client
2247 certificates: checks for matching certificate type and issuer name
2251 *) If an attempt is made to use a signature algorithm not in the peer
2252 preference list abort the handshake. If client has no suitable
2253 signature algorithms in response to a certificate request do not
2254 use the certificate.
2257 *) If server EC tmp key is not in client preference list abort handshake.
2260 *) Add support for certificate stores in CERT structure. This makes it
2261 possible to have different stores per SSL structure or one store in
2262 the parent SSL_CTX. Include distinct stores for certificate chain
2263 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2264 to build and store a certificate chain in CERT structure: returning
2265 an error if the chain cannot be built: this will allow applications
2266 to test if a chain is correctly configured.
2268 Note: if the CERT based stores are not set then the parent SSL_CTX
2269 store is used to retain compatibility with existing behaviour.
2273 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2274 mask based on the current session, check mask when sending client
2275 hello and checking the requested ciphersuite.
2278 *) New ctrls to retrieve and set certificate types in a certificate
2279 request message. Print out received values in s_client. If certificate
2280 types is not set with custom values set sensible values based on
2281 supported signature algorithms.
2284 *) Support for distinct client and server supported signature algorithms.
2287 *) Add certificate callback. If set this is called whenever a certificate
2288 is required by client or server. An application can decide which
2289 certificate chain to present based on arbitrary criteria: for example
2290 supported signature algorithms. Add very simple example to s_server.
2291 This fixes many of the problems and restrictions of the existing client
2292 certificate callback: for example you can now clear an existing
2293 certificate and specify the whole chain.
2296 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2297 the certificate can be used for (if anything). Set valid_flags field
2298 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2299 to have similar checks in it.
2301 Add new "cert_flags" field to CERT structure and include a "strict mode".
2302 This enforces some TLS certificate requirements (such as only permitting
2303 certificate signature algorithms contained in the supported algorithms
2304 extension) which some implementations ignore: this option should be used
2305 with caution as it could cause interoperability issues.
2308 *) Update and tidy signature algorithm extension processing. Work out
2309 shared signature algorithms based on preferences and peer algorithms
2310 and print them out in s_client and s_server. Abort handshake if no
2311 shared signature algorithms.
2314 *) Add new functions to allow customised supported signature algorithms
2315 for SSL and SSL_CTX structures. Add options to s_client and s_server
2319 *) New function SSL_certs_clear() to delete all references to certificates
2320 from an SSL structure. Before this once a certificate had been added
2321 it couldn't be removed.
2324 *) Integrate hostname, email address and IP address checking with certificate
2325 verification. New verify options supporting checking in openssl utility.
2328 *) Fixes and wildcard matching support to hostname and email checking
2329 functions. Add manual page.
2330 [Florian Weimer (Red Hat Product Security Team)]
2332 *) New functions to check a hostname email or IP address against a
2333 certificate. Add options x509 utility to print results of checks against
2337 *) Fix OCSP checking.
2338 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2340 *) Initial experimental support for explicitly trusted non-root CAs.
2341 OpenSSL still tries to build a complete chain to a root but if an
2342 intermediate CA has a trust setting included that is used. The first
2343 setting is used: whether to trust (e.g., -addtrust option to the x509
2347 *) Add -trusted_first option which attempts to find certificates in the
2348 trusted store even if an untrusted chain is also supplied.
2351 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2352 platform support for Linux and Android.
2355 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2358 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2359 When in FIPS mode the approved implementations are used as normal,
2360 when not in FIPS mode the internal unapproved versions are used instead.
2361 This means that the FIPS capable OpenSSL isn't forced to use the
2362 (often lower performance) FIPS implementations outside FIPS mode.
2365 *) Transparently support X9.42 DH parameters when calling
2366 PEM_read_bio_DHparameters. This means existing applications can handle
2367 the new parameter format automatically.
2370 *) Initial experimental support for X9.42 DH parameter format: mainly
2371 to support use of 'q' parameter for RFC5114 parameters.
2374 *) Add DH parameters from RFC5114 including test data to dhtest.
2377 *) Support for automatic EC temporary key parameter selection. If enabled
2378 the most preferred EC parameters are automatically used instead of
2379 hardcoded fixed parameters. Now a server just has to call:
2380 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2381 support ECDH and use the most appropriate parameters.
2384 *) Enhance and tidy EC curve and point format TLS extension code. Use
2385 static structures instead of allocation if default values are used.
2386 New ctrls to set curves we wish to support and to retrieve shared curves.
2387 Print out shared curves in s_server. New options to s_server and s_client
2388 to set list of supported curves.
2391 *) New ctrls to retrieve supported signature algorithms and
2392 supported curve values as an array of NIDs. Extend openssl utility
2393 to print out received values.
2396 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2397 between NIDs and the more common NIST names such as "P-256". Enhance
2398 ecparam utility and ECC method to recognise the NIST names for curves.
2401 *) Enhance SSL/TLS certificate chain handling to support different
2402 chains for each certificate instead of one chain in the parent SSL_CTX.
2405 *) Support for fixed DH ciphersuite client authentication: where both
2406 server and client use DH certificates with common parameters.
2409 *) Support for fixed DH ciphersuites: those requiring DH server
2413 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2415 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2416 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2417 X509_CINF_get_signature were reverted post internal team review.
2419 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2421 *) Build fixes for the Windows and OpenVMS platforms
2422 [Matt Caswell and Richard Levitte]
2424 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2426 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2427 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2428 dereference. This could lead to a Denial Of Service attack. Thanks to
2429 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2433 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2434 dtls1_buffer_record function under certain conditions. In particular this
2435 could occur if an attacker sent repeated DTLS records with the same
2436 sequence number but for the next epoch. The memory leak could be exploited
2437 by an attacker in a Denial of Service attack through memory exhaustion.
2438 Thanks to Chris Mueller for reporting this issue.
2442 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2443 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2444 method would be set to NULL which could later result in a NULL pointer
2445 dereference. Thanks to Frank Schmirler for reporting this issue.
2449 *) Abort handshake if server key exchange message is omitted for ephemeral
2452 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2453 reporting this issue.
2457 *) Remove non-export ephemeral RSA code on client and server. This code
2458 violated the TLS standard by allowing the use of temporary RSA keys in
2459 non-export ciphersuites and could be used by a server to effectively
2460 downgrade the RSA key length used to a value smaller than the server
2461 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2462 INRIA or reporting this issue.
2466 *) Fixed issue where DH client certificates are accepted without verification.
2467 An OpenSSL server will accept a DH certificate for client authentication
2468 without the certificate verify message. This effectively allows a client to
2469 authenticate without the use of a private key. This only affects servers
2470 which trust a client certificate authority which issues certificates
2471 containing DH keys: these are extremely rare and hardly ever encountered.
2472 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2477 *) Ensure that the session ID context of an SSL is updated when its
2478 SSL_CTX is updated via SSL_set_SSL_CTX.
2480 The session ID context is typically set from the parent SSL_CTX,
2481 and can vary with the CTX.
2484 *) Fix various certificate fingerprint issues.
2486 By using non-DER or invalid encodings outside the signed portion of a
2487 certificate the fingerprint can be changed without breaking the signature.
2488 Although no details of the signed portion of the certificate can be changed
2489 this can cause problems with some applications: e.g. those using the
2490 certificate fingerprint for blacklists.
2492 1. Reject signatures with non zero unused bits.
2494 If the BIT STRING containing the signature has non zero unused bits reject
2495 the signature. All current signature algorithms require zero unused bits.
2497 2. Check certificate algorithm consistency.
2499 Check the AlgorithmIdentifier inside TBS matches the one in the
2500 certificate signature. NB: this will result in signature failure
2501 errors for some broken certificates.
2503 Thanks to Konrad Kraszewski from Google for reporting this issue.
2505 3. Check DSA/ECDSA signatures use DER.
2507 Re-encode DSA/ECDSA signatures and compare with the original received
2508 signature. Return an error if there is a mismatch.
2510 This will reject various cases including garbage after signature
2511 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2512 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2513 (negative or with leading zeroes).
2515 Further analysis was conducted and fixes were developed by Stephen Henson
2516 of the OpenSSL core team.
2521 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2522 results on some platforms, including x86_64. This bug occurs at random
2523 with a very low probability, and is not known to be exploitable in any
2524 way, though its exact impact is difficult to determine. Thanks to Pieter
2525 Wuille (Blockstream) who reported this issue and also suggested an initial
2526 fix. Further analysis was conducted by the OpenSSL development team and
2527 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2528 the OpenSSL core team.
2532 *) Do not resume sessions on the server if the negotiated protocol
2533 version does not match the session's version. Resuming with a different
2534 version, while not strictly forbidden by the RFC, is of questionable
2535 sanity and breaks all known clients.
2536 [David Benjamin, Emilia Käsper]
2538 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2539 early CCS messages during renegotiation. (Note that because
2540 renegotiation is encrypted, this early CCS was not exploitable.)
2543 *) Tighten client-side session ticket handling during renegotiation:
2544 ensure that the client only accepts a session ticket if the server sends
2545 the extension anew in the ServerHello. Previously, a TLS client would
2546 reuse the old extension state and thus accept a session ticket if one was
2547 announced in the initial ServerHello.
2549 Similarly, ensure that the client requires a session ticket if one
2550 was advertised in the ServerHello. Previously, a TLS client would
2551 ignore a missing NewSessionTicket message.
2554 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2556 *) SRTP Memory Leak.
2558 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2559 sends a carefully crafted handshake message, to cause OpenSSL to fail
2560 to free up to 64k of memory causing a memory leak. This could be
2561 exploited in a Denial Of Service attack. This issue affects OpenSSL
2562 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2563 whether SRTP is used or configured. Implementations of OpenSSL that
2564 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2566 The fix was developed by the OpenSSL team.
2570 *) Session Ticket Memory Leak.
2572 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2573 integrity of that ticket is first verified. In the event of a session
2574 ticket integrity check failing, OpenSSL will fail to free memory
2575 causing a memory leak. By sending a large number of invalid session
2576 tickets an attacker could exploit this issue in a Denial Of Service
2581 *) Build option no-ssl3 is incomplete.
2583 When OpenSSL is configured with "no-ssl3" as a build option, servers
2584 could accept and complete a SSL 3.0 handshake, and clients could be
2585 configured to send them.
2587 [Akamai and the OpenSSL team]
2589 *) Add support for TLS_FALLBACK_SCSV.
2590 Client applications doing fallback retries should call
2591 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2593 [Adam Langley, Bodo Moeller]
2595 *) Add additional DigestInfo checks.
2597 Re-encode DigestInto in DER and check against the original when
2598 verifying RSA signature: this will reject any improperly encoded
2599 DigestInfo structures.
2601 Note: this is a precautionary measure and no attacks are currently known.
2605 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2607 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2608 SRP code can be overrun an internal buffer. Add sanity check that
2609 g, A, B < N to SRP code.
2611 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2612 Group for discovering this issue.
2616 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2617 TLS 1.0 instead of higher protocol versions when the ClientHello message
2618 is badly fragmented. This allows a man-in-the-middle attacker to force a
2619 downgrade to TLS 1.0 even if both the server and the client support a
2620 higher protocol version, by modifying the client's TLS records.
2622 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2623 researching this issue.
2627 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2628 to a denial of service attack. A malicious server can crash the client
2629 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2630 ciphersuite and sending carefully crafted handshake messages.
2632 Thanks to Felix Gröbert (Google) for discovering and researching this
2637 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2638 to leak memory. This can be exploited through a Denial of Service attack.
2639 Thanks to Adam Langley for discovering and researching this issue.
2643 *) An attacker can force openssl to consume large amounts of memory whilst
2644 processing DTLS handshake messages. This can be exploited through a
2645 Denial of Service attack.
2646 Thanks to Adam Langley for discovering and researching this issue.
2650 *) An attacker can force an error condition which causes openssl to crash
2651 whilst processing DTLS packets due to memory being freed twice. This
2652 can be exploited through a Denial of Service attack.
2653 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2658 *) If a multithreaded client connects to a malicious server using a resumed
2659 session and the server sends an ec point format extension it could write
2660 up to 255 bytes to freed memory.
2662 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2667 *) A malicious server can crash an OpenSSL client with a null pointer
2668 dereference (read) by specifying an SRP ciphersuite even though it was not
2669 properly negotiated with the client. This can be exploited through a
2670 Denial of Service attack.
2672 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2673 discovering and researching this issue.
2677 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2678 X509_name_oneline, X509_name_print_ex et al. to leak some information
2679 from the stack. Applications may be affected if they echo pretty printing
2680 output to the attacker.
2682 Thanks to Ivan Fratric (Google) for discovering this issue.
2684 [Emilia Käsper, and Steve Henson]
2686 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2687 for corner cases. (Certain input points at infinity could lead to
2688 bogus results, with non-infinity inputs mapped to infinity too.)
2691 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2693 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2694 handshake can force the use of weak keying material in OpenSSL
2695 SSL/TLS clients and servers.
2697 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2698 researching this issue. (CVE-2014-0224)
2699 [KIKUCHI Masashi, Steve Henson]
2701 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2702 OpenSSL DTLS client the code can be made to recurse eventually crashing
2705 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2707 [Imre Rad, Steve Henson]
2709 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2710 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2711 client or server. This is potentially exploitable to run arbitrary
2712 code on a vulnerable client or server.
2714 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2715 [Jüri Aedla, Steve Henson]
2717 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2718 are subject to a denial of service attack.
2720 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2721 this issue. (CVE-2014-3470)
2722 [Felix Gröbert, Ivan Fratric, Steve Henson]
2724 *) Harmonize version and its documentation. -f flag is used to display
2726 [mancha <mancha1@zoho.com>]
2728 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2729 in i2d_ECPrivateKey.
2730 [mancha <mancha1@zoho.com>]
2732 *) Fix some double frees. These are not thought to be exploitable.
2733 [mancha <mancha1@zoho.com>]
2735 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2737 *) A missing bounds check in the handling of the TLS heartbeat extension
2738 can be used to reveal up to 64k of memory to a connected client or
2741 Thanks for Neel Mehta of Google Security for discovering this bug and to
2742 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2743 preparing the fix (CVE-2014-0160)
2744 [Adam Langley, Bodo Moeller]
2746 *) Fix for the attack described in the paper "Recovering OpenSSL
2747 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2748 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2749 http://eprint.iacr.org/2014/140
2751 Thanks to Yuval Yarom and Naomi Benger for discovering this
2752 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2753 [Yuval Yarom and Naomi Benger]
2755 *) TLS pad extension: draft-agl-tls-padding-03
2757 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2758 TLS client Hello record length value would otherwise be > 255 and
2759 less that 512 pad with a dummy extension containing zeroes so it
2760 is at least 512 bytes long.
2762 [Adam Langley, Steve Henson]
2764 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2766 *) Fix for TLS record tampering bug. A carefully crafted invalid
2767 handshake could crash OpenSSL with a NULL pointer exception.
2768 Thanks to Anton Johansson for reporting this issues.
2771 *) Keep original DTLS digest and encryption contexts in retransmission
2772 structures so we can use the previous session parameters if they need
2773 to be resent. (CVE-2013-6450)
2776 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2777 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2778 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2779 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2780 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2781 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2782 [Rob Stradling, Adam Langley]
2784 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2786 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2787 supporting platforms or when small records were transferred.
2788 [Andy Polyakov, Steve Henson]
2790 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2792 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2794 This addresses the flaw in CBC record processing discovered by
2795 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2796 at: http://www.isg.rhul.ac.uk/tls/
2798 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2799 Security Group at Royal Holloway, University of London
2800 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2801 Emilia Käsper for the initial patch.
2803 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2805 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2806 ciphersuites which can be exploited in a denial of service attack.
2807 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2808 and detecting this bug and to Wolfgang Ettlinger
2809 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2813 *) Return an error when checking OCSP signatures when key is NULL.
2814 This fixes a DoS attack. (CVE-2013-0166)
2817 *) Make openssl verify return errors.
2818 [Chris Palmer <palmer@google.com> and Ben Laurie]
2820 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2821 the right response is stapled. Also change SSL_get_certificate()
2822 so it returns the certificate actually sent.
2823 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2824 [Rob Stradling <rob.stradling@comodo.com>]
2826 *) Fix possible deadlock when decoding public keys.
2829 *) Don't use TLS 1.0 record version number in initial client hello
2833 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2835 *) Sanity check record length before skipping explicit IV in TLS
2836 1.2, 1.1 and DTLS to fix DoS attack.
2838 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2839 fuzzing as a service testing platform.
2843 *) Initialise tkeylen properly when encrypting CMS messages.
2844 Thanks to Solar Designer of Openwall for reporting this issue.
2847 *) In FIPS mode don't try to use composite ciphers as they are not
2851 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2853 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2854 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2855 mean any application compiled against OpenSSL 1.0.0 headers setting
2856 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2857 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2858 0x10000000L Any application which was previously compiled against
2859 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2860 will need to be recompiled as a result. Letting be results in
2861 inability to disable specifically TLS 1.1 and in client context,
2862 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2865 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2866 disable just protocol X, but all protocols above X *if* there are
2867 protocols *below* X still enabled. In more practical terms it means
2868 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2869 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2870 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2874 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2876 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2877 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2878 in CRYPTO_realloc_clean.
2880 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2881 issue and to Adam Langley <agl@chromium.org> for fixing it.
2883 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2885 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2888 *) Workarounds for some broken servers that "hang" if a client hello
2889 record length exceeds 255 bytes.
2891 1. Do not use record version number > TLS 1.0 in initial client
2892 hello: some (but not all) hanging servers will now work.
2893 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2894 the number of ciphers sent in the client hello. This should be
2895 set to an even number, such as 50, for example by passing:
2896 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2897 Most broken servers should now work.
2898 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2899 TLS 1.2 client support entirely.
2902 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2905 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2907 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2908 STRING form instead of a DigestInfo.
2911 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2912 and the RSA_sign/RSA_verify functions. This was made more apparent when
2913 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2914 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2915 the correct format in RSA_verify so both forms transparently work.
2918 *) Some servers which support TLS 1.0 can choke if we initially indicate
2919 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2920 encrypted premaster secret. As a workaround use the maximum permitted
2921 client version in client hello, this should keep such servers happy
2922 and still work with previous versions of OpenSSL.
2925 *) Add support for TLS/DTLS heartbeats.
2926 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2928 *) Add support for SCTP.
2929 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2931 *) Improved PRNG seeding for VOS.
2932 [Paul Green <Paul.Green@stratus.com>]
2934 *) Extensive assembler packs updates, most notably:
2936 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2937 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2938 - x86_64: bit-sliced AES implementation;
2939 - ARM: NEON support, contemporary platforms optimizations;
2940 - s390x: z196 support;
2941 - *: GHASH and GF(2^m) multiplication implementations;
2945 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2946 (removal of unnecessary code)
2947 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2949 *) Add TLS key material exporter from RFC 5705.
2952 *) Add DTLS-SRTP negotiation from RFC 5764.
2955 *) Add Next Protocol Negotiation,
2956 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2957 disabled with a no-npn flag to config or Configure. Code donated
2959 [Adam Langley <agl@google.com> and Ben Laurie]
2961 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2962 NIST-P256, NIST-P521, with constant-time single point multiplication on
2963 typical inputs. Compiler support for the nonstandard type __uint128_t is
2964 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2965 Code made available under Apache License version 2.0.
2967 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2968 line to include this in your build of OpenSSL, and run "make depend" (or
2969 "make update"). This enables the following EC_METHODs:
2971 EC_GFp_nistp224_method()
2972 EC_GFp_nistp256_method()
2973 EC_GFp_nistp521_method()
2975 EC_GROUP_new_by_curve_name() will automatically use these (while
2976 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2978 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2980 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2981 all platforms. Move ssize_t definition from e_os.h to the public
2982 header file e_os2.h as it now appears in public header file cms.h
2985 *) New -sigopt option to the ca, req and x509 utilities. Additional
2986 signature parameters can be passed using this option and in
2990 *) Add RSA PSS signing function. This will generate and set the
2991 appropriate AlgorithmIdentifiers for PSS based on those in the
2992 corresponding EVP_MD_CTX structure. No application support yet.
2995 *) Support for companion algorithm specific ASN1 signing routines.
2996 New function ASN1_item_sign_ctx() signs a pre-initialised
2997 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2998 the appropriate parameters.
3001 *) Add new algorithm specific ASN1 verification initialisation function
3002 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3003 handling will be the same no matter what EVP_PKEY_METHOD is used.
3004 Add a PSS handler to support verification of PSS signatures: checked
3005 against a number of sample certificates.
3008 *) Add signature printing for PSS. Add PSS OIDs.
3009 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3011 *) Add algorithm specific signature printing. An individual ASN1 method
3012 can now print out signatures instead of the standard hex dump.
3014 More complex signatures (e.g. PSS) can print out more meaningful
3015 information. Include DSA version that prints out the signature
3019 *) Password based recipient info support for CMS library: implementing
3023 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3024 neatly separates the code into cipher and PBE sections and is required
3025 for some algorithms that split PBES2 into separate pieces (such as
3026 password based CMS).
3029 *) Session-handling fixes:
3030 - Fix handling of connections that are resuming with a session ID,
3031 but also support Session Tickets.
3032 - Fix a bug that suppressed issuing of a new ticket if the client
3033 presented a ticket with an expired session.
3034 - Try to set the ticket lifetime hint to something reasonable.
3035 - Make tickets shorter by excluding irrelevant information.
3036 - On the client side, don't ignore renewed tickets.
3037 [Adam Langley, Bodo Moeller (Google)]
3039 *) Fix PSK session representation.
3042 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3044 This work was sponsored by Intel.
3047 *) Add GCM support to TLS library. Some custom code is needed to split
3048 the IV between the fixed (from PRF) and explicit (from TLS record)
3049 portions. This adds all GCM ciphersuites supported by RFC5288 and
3050 RFC5289. Generalise some AES* cipherstrings to include GCM and
3051 add a special AESGCM string for GCM only.
3054 *) Expand range of ctrls for AES GCM. Permit setting invocation
3055 field on decrypt and retrieval of invocation field only on encrypt.
3058 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3059 As required by RFC5289 these ciphersuites cannot be used if for
3060 versions of TLS earlier than 1.2.
3063 *) For FIPS capable OpenSSL interpret a NULL default public key method
3064 as unset and return the appropriate default but do *not* set the default.
3065 This means we can return the appropriate method in applications that
3066 switch between FIPS and non-FIPS modes.
3069 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3070 ENGINE is used then we cannot handle that in the FIPS module so we
3071 keep original code iff non-FIPS operations are allowed.
3074 *) Add -attime option to openssl utilities.
3075 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3077 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3080 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3081 FIPS EC methods unconditionally for now.
3084 *) New build option no-ec2m to disable characteristic 2 code.
3087 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3088 all cases can be covered as some introduce binary incompatibilities.
3091 *) Redirect RSA operations to FIPS module including keygen,
3092 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3095 *) Add similar low level API blocking to ciphers.
3098 *) Low level digest APIs are not approved in FIPS mode: any attempt
3099 to use these will cause a fatal error. Applications that *really* want
3100 to use them can use the private_* version instead.
3103 *) Redirect cipher operations to FIPS module for FIPS builds.
3106 *) Redirect digest operations to FIPS module for FIPS builds.
3109 *) Update build system to add "fips" flag which will link in fipscanister.o
3110 for static and shared library builds embedding a signature if needed.
3113 *) Output TLS supported curves in preference order instead of numerical
3114 order. This is currently hardcoded for the highest order curves first.
3115 This should be configurable so applications can judge speed vs strength.
3118 *) Add TLS v1.2 server support for client authentication.
3121 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3125 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3126 FIPS modules versions.
3129 *) Add TLS v1.2 client side support for client authentication. Keep cache
3130 of handshake records longer as we don't know the hash algorithm to use
3131 until after the certificate request message is received.
3134 *) Initial TLS v1.2 client support. Add a default signature algorithms
3135 extension including all the algorithms we support. Parse new signature
3136 format in client key exchange. Relax some ECC signing restrictions for
3137 TLS v1.2 as indicated in RFC5246.
3140 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3141 to new signature format when needed using client digest preference.
3142 All server ciphersuites should now work correctly in TLS v1.2. No client
3143 support yet and no support for client certificates.
3146 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3147 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3148 ciphersuites. At present only RSA key exchange ciphersuites work with
3149 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3150 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3151 and version checking.
3154 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3155 with this defined it will not be affected by any changes to ssl internal
3156 structures. Add several utility functions to allow openssl application
3157 to work with OPENSSL_NO_SSL_INTERN defined.
3161 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
3163 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3166 *) Permit abbreviated handshakes when renegotiating using the function
3167 SSL_renegotiate_abbreviated().
3168 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3170 *) Add call to ENGINE_register_all_complete() to
3171 ENGINE_load_builtin_engines(), so some implementations get used
3172 automatically instead of needing explicit application support.
3175 *) Add support for TLS key exporter as described in RFC5705.
3176 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3178 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3179 a few changes are required:
3181 Add SSL_OP_NO_TLSv1_1 flag.
3182 Add TLSv1_1 methods.
3183 Update version checking logic to handle version 1.1.
3184 Add explicit IV handling (ported from DTLS code).
3185 Add command line options to s_client/s_server.
3188 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3190 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3191 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3192 content decryption and always return the same error. Note: this attack
3193 needs on average 2^20 messages so it only affects automated senders. The
3194 old behaviour can be re-enabled in the CMS code by setting the
3195 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3196 an MMA defence is not necessary.
3197 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3198 this issue. (CVE-2012-0884)
3201 *) Fix CVE-2011-4619: make sure we really are receiving a
3202 client hello before rejecting multiple SGC restarts. Thanks to
3203 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3206 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3208 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3209 Thanks to Antonio Martin, Enterprise Secure Access Research and
3210 Development, Cisco Systems, Inc. for discovering this bug and
3211 preparing a fix. (CVE-2012-0050)
3214 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3216 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3217 of the Vaudenay padding oracle attack on CBC mode encryption
3218 which enables an efficient plaintext recovery attack against
3219 the OpenSSL implementation of DTLS. Their attack exploits timing
3220 differences arising during decryption processing. A research
3221 paper describing this attack can be found at:
3222 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3223 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3224 Security Group at Royal Holloway, University of London
3225 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3226 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3227 for preparing the fix. (CVE-2011-4108)
3228 [Robin Seggelmann, Michael Tuexen]
3230 *) Clear bytes used for block padding of SSL 3.0 records.
3232 [Adam Langley (Google)]
3234 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3235 Kadianakis <desnacked@gmail.com> for discovering this issue and
3236 Adam Langley for preparing the fix. (CVE-2011-4619)
3237 [Adam Langley (Google)]
3239 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3240 [Andrey Kulikov <amdeich@gmail.com>]
3242 *) Prevent malformed RFC3779 data triggering an assertion failure.
3243 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3244 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3245 [Rob Austein <sra@hactrn.net>]
3247 *) Improved PRNG seeding for VOS.
3248 [Paul Green <Paul.Green@stratus.com>]
3250 *) Fix ssl_ciph.c set-up race.
3251 [Adam Langley (Google)]
3253 *) Fix spurious failures in ecdsatest.c.
3254 [Emilia Käsper (Google)]
3256 *) Fix the BIO_f_buffer() implementation (which was mixing different
3257 interpretations of the '..._len' fields).
3258 [Adam Langley (Google)]
3260 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3261 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3262 threads won't reuse the same blinding coefficients.
3264 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3265 lock to call BN_BLINDING_invert_ex, and avoids one use of
3266 BN_BLINDING_update for each BN_BLINDING structure (previously,
3267 the last update always remained unused).
3268 [Emilia Käsper (Google)]
3270 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3271 [Bob Buckholz (Google)]
3273 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3275 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3276 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3277 [Kaspar Brand <ossl@velox.ch>]
3279 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3280 for multi-threaded use of ECDH. (CVE-2011-3210)
3281 [Adam Langley (Google)]
3283 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3286 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3287 signature public key algorithm by using OID xref utilities instead.
3288 Before this you could only use some ECC ciphersuites with SHA1 only.
3291 *) Add protection against ECDSA timing attacks as mentioned in the paper
3292 by Billy Bob Brumley and Nicola Tuveri, see:
3294 http://eprint.iacr.org/2011/232.pdf
3296 [Billy Bob Brumley and Nicola Tuveri]
3298 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3300 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3301 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3303 *) Fix bug in string printing code: if *any* escaping is enabled we must
3304 escape the escape character (backslash) or the resulting string is
3308 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3310 *) Disable code workaround for ancient and obsolete Netscape browsers
3311 and servers: an attacker can use it in a ciphersuite downgrade attack.
3312 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3315 *) Fixed J-PAKE implementation error, originally discovered by
3316 Sebastien Martini, further info and confirmation from Stefan
3317 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3320 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3322 *) Fix extension code to avoid race conditions which can result in a buffer
3323 overrun vulnerability: resumed sessions must not be modified as they can
3324 be shared by multiple threads. CVE-2010-3864
3327 *) Fix WIN32 build system to correctly link an ENGINE directory into
3331 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3333 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3335 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3337 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3339 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3340 context. The operation can be customised via the ctrl mechanism in
3341 case ENGINEs want to include additional functionality.
3344 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3347 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3348 output hashes compatible with older versions of OpenSSL.
3349 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3351 *) Fix compression algorithm handling: if resuming a session use the
3352 compression algorithm of the resumed session instead of determining
3353 it from client hello again. Don't allow server to change algorithm.
3356 *) Add load_crls() function to apps tidying load_certs() too. Add option
3357 to verify utility to allow additional CRLs to be included.
3360 *) Update OCSP request code to permit adding custom headers to the request:
3361 some responders need this.
3364 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3366 [Julia Lawall <julia@diku.dk>]
3368 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3369 needlessly dereferenced structures, used obsolete functions and
3370 didn't handle all updated verify codes correctly.
3373 *) Disable MD2 in the default configuration.
3376 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3377 indicate the initial BIO being pushed or popped. This makes it possible
3378 to determine whether the BIO is the one explicitly called or as a result
3379 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3380 it handles reference counts correctly and doesn't zero out the I/O bio
3381 when it is not being explicitly popped. WARNING: applications which
3382 included workarounds for the old buggy behaviour will need to be modified
3383 or they could free up already freed BIOs.
3386 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3387 renaming to all platforms (within the 0.9.8 branch, this was
3388 done conditionally on Netware platforms to avoid a name clash).
3389 [Guenter <lists@gknw.net>]
3391 *) Add ECDHE and PSK support to DTLS.
3392 [Michael Tuexen <tuexen@fh-muenster.de>]
3394 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3398 *) Add "missing" function EVP_MD_flags() (without this the only way to
3399 retrieve a digest flags is by accessing the structure directly. Update
3400 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3401 or cipher is registered as in the "from" argument. Print out all
3402 registered digests in the dgst usage message instead of manually
3403 attempting to work them out.
3406 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3407 this allows the use of compression and extensions. Change default cipher
3408 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3409 by default unless an application cipher string requests it.
3412 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3413 key ids to find matching certificates and keys but some PKCS#12 files
3414 don't follow the (somewhat unwritten) rules and this strategy fails.
3415 Now just gather all certificates together and the first private key
3416 then look for the first certificate that matches the key.
3419 *) Support use of registered digest and cipher names for dgst and cipher
3420 commands instead of having to add each one as a special case. So now
3427 openssl dgst -sha256 foo
3429 and this works for ENGINE based algorithms too.
3433 *) Update Gost ENGINE to support parameter files.
3434 [Victor B. Wagner <vitus@cryptocom.ru>]
3436 *) Support GeneralizedTime in ca utility.
3437 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3439 *) Enhance the hash format used for certificate directory links. The new
3440 form uses the canonical encoding (meaning equivalent names will work
3441 even if they aren't identical) and uses SHA1 instead of MD5. This form
3442 is incompatible with the older format and as a result c_rehash should
3443 be used to rebuild symbolic links.
3446 *) Make PKCS#8 the default write format for private keys, replacing the
3447 traditional format. This form is standardised, more secure and doesn't
3448 include an implicit MD5 dependency.
3451 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3452 committed to OpenSSL should pass this lot as a minimum.
3455 *) Add session ticket override functionality for use by EAP-FAST.
3456 [Jouni Malinen <j@w1.fi>]
3458 *) Modify HMAC functions to return a value. Since these can be implemented
3459 in an ENGINE errors can occur.
3462 *) Type-checked OBJ_bsearch_ex.
3465 *) Type-checked OBJ_bsearch. Also some constification necessitated
3466 by type-checking. Still to come: TXT_DB, bsearch(?),
3467 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3471 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3472 seconds to a tm structure directly, instead of going through OS
3473 specific date routines. This avoids any issues with OS routines such
3474 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3475 and X509_time_adj_ex() to cover the extended range. The existing
3476 X509_time_adj() is still usable and will no longer have any date issues.
3479 *) Delta CRL support. New use deltas option which will attempt to locate
3480 and search any appropriate delta CRLs available.
3482 This work was sponsored by Google.
3485 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3486 code and add additional score elements. Validate alternate CRL paths
3487 as part of the CRL checking and indicate a new error "CRL path validation
3488 error" in this case. Applications wanting additional details can use
3489 the verify callback and check the new "parent" field. If this is not
3490 NULL CRL path validation is taking place. Existing applications won't
3491 see this because it requires extended CRL support which is off by
3494 This work was sponsored by Google.
3497 *) Support for freshest CRL extension.
3499 This work was sponsored by Google.
3502 *) Initial indirect CRL support. Currently only supported in the CRLs
3503 passed directly and not via lookup. Process certificate issuer
3504 CRL entry extension and lookup CRL entries by bother issuer name
3505 and serial number. Check and process CRL issuer entry in IDP extension.
3507 This work was sponsored by Google.
3510 *) Add support for distinct certificate and CRL paths. The CRL issuer
3511 certificate is validated separately in this case. Only enabled if
3512 an extended CRL support flag is set: this flag will enable additional
3513 CRL functionality in future.
3515 This work was sponsored by Google.
3518 *) Add support for policy mappings extension.
3520 This work was sponsored by Google.
3523 *) Fixes to pathlength constraint, self issued certificate handling,
3524 policy processing to align with RFC3280 and PKITS tests.
3526 This work was sponsored by Google.
3529 *) Support for name constraints certificate extension. DN, email, DNS
3530 and URI types are currently supported.
3532 This work was sponsored by Google.
3535 *) To cater for systems that provide a pointer-based thread ID rather
3536 than numeric, deprecate the current numeric thread ID mechanism and
3537 replace it with a structure and associated callback type. This
3538 mechanism allows a numeric "hash" to be extracted from a thread ID in
3539 either case, and on platforms where pointers are larger than 'long',
3540 mixing is done to help ensure the numeric 'hash' is usable even if it
3541 can't be guaranteed unique. The default mechanism is to use "&errno"
3542 as a pointer-based thread ID to distinguish between threads.
3544 Applications that want to provide their own thread IDs should now use
3545 CRYPTO_THREADID_set_callback() to register a callback that will call
3546 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3548 Note that ERR_remove_state() is now deprecated, because it is tied
3549 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3550 to free the current thread's error state should be replaced by
3551 ERR_remove_thread_state(NULL).
3553 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3554 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3555 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3556 application was previously providing a numeric thread callback that
3557 was inappropriate for distinguishing threads, then uniqueness might
3558 have been obtained with &errno that happened immediately in the
3559 intermediate development versions of OpenSSL; this is no longer the
3560 case, the numeric thread callback will now override the automatic use
3562 [Geoff Thorpe, with help from Bodo Moeller]
3564 *) Initial support for different CRL issuing certificates. This covers a
3565 simple case where the self issued certificates in the chain exist and
3566 the real CRL issuer is higher in the existing chain.
3568 This work was sponsored by Google.
3571 *) Removed effectively defunct crypto/store from the build.
3574 *) Revamp of STACK to provide stronger type-checking. Still to come:
3575 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3576 ASN1_STRING, CONF_VALUE.
3579 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3580 RAM on SSL connections. This option can save about 34k per idle SSL.
3583 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3584 STACK, TXT_DB, bsearch, qsort.
3587 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3588 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3589 support for data, signedData, compressedData, digestedData and
3590 encryptedData, envelopedData types included. Scripts to check against
3591 RFC4134 examples draft and interop and consistency checks of many
3592 content types and variants.
3595 *) Add options to enc utility to support use of zlib compression BIO.
3598 *) Extend mk1mf to support importing of options and assembly language
3599 files from Configure script, currently only included in VC-WIN32.
3600 The assembly language rules can now optionally generate the source
3601 files from the associated perl scripts.
3604 *) Implement remaining functionality needed to support GOST ciphersuites.
3605 Interop testing has been performed using CryptoPro implementations.
3606 [Victor B. Wagner <vitus@cryptocom.ru>]
3608 *) s390x assembler pack.
3611 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3615 *) Implement Opaque PRF Input TLS extension as specified in
3616 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3617 official specification yet and no extension type assignment by
3618 IANA exists, this extension (for now) will have to be explicitly
3619 enabled when building OpenSSL by providing the extension number
3620 to use. For example, specify an option
3622 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3624 to the "config" or "Configure" script to enable the extension,
3625 assuming extension number 0x9527 (which is a completely arbitrary
3626 and unofficial assignment based on the MD5 hash of the Internet
3627 Draft). Note that by doing so, you potentially lose
3628 interoperability with other TLS implementations since these might
3629 be using the same extension number for other purposes.
3631 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3632 opaque PRF input value to use in the handshake. This will create
3633 an interal copy of the length-'len' string at 'src', and will
3634 return non-zero for success.
3636 To get more control and flexibility, provide a callback function
3639 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3640 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3644 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3647 Callback function 'cb' will be called in handshakes, and is
3648 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3649 Argument 'arg' is for application purposes (the value as given to
3650 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3651 be provided to the callback function). The callback function
3652 has to return non-zero to report success: usually 1 to use opaque
3653 PRF input just if possible, or 2 to enforce use of the opaque PRF
3654 input. In the latter case, the library will abort the handshake
3655 if opaque PRF input is not successfully negotiated.
3657 Arguments 'peerinput' and 'len' given to the callback function
3658 will always be NULL and 0 in the case of a client. A server will
3659 see the client's opaque PRF input through these variables if
3660 available (NULL and 0 otherwise). Note that if the server
3661 provides an opaque PRF input, the length must be the same as the
3662 length of the client's opaque PRF input.
3664 Note that the callback function will only be called when creating
3665 a new session (session resumption can resume whatever was
3666 previously negotiated), and will not be called in SSL 2.0
3667 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3668 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3669 for applications that need to enforce opaque PRF input.
3673 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3676 [Victor B. Wagner <vitus@cryptocom.ru>]
3678 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3679 RFC4507bis. The encrypted ticket format is an encrypted encoded
3680 SSL_SESSION structure, that way new session features are automatically
3683 If a client application caches session in an SSL_SESSION structure
3684 support is transparent because tickets are now stored in the encoded
3687 The SSL_CTX structure automatically generates keys for ticket
3688 protection in servers so again support should be possible
3689 with no application modification.
3691 If a client or server wishes to disable RFC4507 support then the option
3692 SSL_OP_NO_TICKET can be set.
3694 Add a TLS extension debugging callback to allow the contents of any client
3695 or server extensions to be examined.
3697 This work was sponsored by Google.
3700 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3701 OpenSSL should now compile cleanly on gcc 4.2
3702 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3704 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3705 support including streaming MAC support: this is required for GOST
3706 ciphersuite support.
3707 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3709 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3710 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3711 to output in BER and PEM format.
3714 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3715 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3716 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3717 ENGINE support for HMAC keys which are unextractable. New -mac and
3718 -macopt options to dgst utility.
3721 *) New option -sigopt to dgst utility. Update dgst to use
3722 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3723 alternative signing parameters such as X9.31 or PSS in the dgst
3727 *) Change ssl_cipher_apply_rule(), the internal function that does
3728 the work each time a ciphersuite string requests enabling
3729 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3730 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3731 the order of disabled ciphersuites such that those ciphersuites
3732 that most recently went from enabled to disabled not only stay
3733 in order with respect to each other, but also have higher priority
3734 than other disabled ciphersuites the next time ciphersuites are
3737 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3738 the same ciphersuites as with "HIGH" alone, but in a specific
3739 order where the PSK ciphersuites come first (since they are the
3740 most recently disabled ciphersuites when "HIGH" is parsed).
3742 Also, change ssl_create_cipher_list() (using this new
3743 funcionality) such that between otherwise identical
3744 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3748 *) Change ssl_create_cipher_list() so that it automatically
3749 arranges the ciphersuites in reasonable order before starting
3750 to process the rule string. Thus, the definition for "DEFAULT"
3751 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3752 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3753 This makes it much easier to arrive at a reasonable default order
3754 in applications for which anonymous ciphers are OK (meaning
3755 that you can't actually use DEFAULT).
3756 [Bodo Moeller; suggested by Victor Duchovni]
3758 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3759 processing) into multiple integers instead of setting
3760 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3761 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3762 (These masks as well as the individual bit definitions are hidden
3763 away into the non-exported interface ssl/ssl_locl.h, so this
3764 change to the definition of the SSL_CIPHER structure shouldn't
3765 affect applications.) This give us more bits for each of these
3766 categories, so there is no longer a need to coagulate AES128 and
3767 AES256 into a single algorithm bit, and to coagulate Camellia128
3768 and Camellia256 into a single algorithm bit, which has led to all
3771 Thus, among other things, the kludge introduced in 0.9.7m and
3772 0.9.8e for masking out AES256 independently of AES128 or masking
3773 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3775 With the change, we also introduce new ciphersuite aliases that
3776 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3780 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3781 Use the leftmost N bytes of the signature input if the input is
3782 larger than the prime q (with N being the size in bytes of q).
3785 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3786 it yet and it is largely untested.
3789 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3792 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3793 some compilers (gcc 4.2 and later) reject their use. Safestack is
3794 reimplemented. Update ASN1 to avoid use of legacy functions.
3797 *) Win32/64 targets are linked with Winsock2.
3800 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3801 to external functions. This can be used to increase CRL handling
3802 efficiency especially when CRLs are very large by (for example) storing
3803 the CRL revoked certificates in a database.
3806 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3807 new CRLs added to a directory can be used. New command line option
3808 -verify_return_error to s_client and s_server. This causes real errors
3809 to be returned by the verify callback instead of carrying on no matter
3810 what. This reflects the way a "real world" verify callback would behave.
3813 *) GOST engine, supporting several GOST algorithms and public key formats.
3814 Kindly donated by Cryptocom.
3817 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3818 partitioned by DP are handled but no indirect CRL or reason partitioning
3819 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3820 selected via a scoring technique which handles IDP and AKID in CRLs.
3823 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3824 will ultimately be used for all verify operations: this will remove the
3825 X509_STORE dependency on certificate verification and allow alternative
3826 lookup methods. X509_STORE based implementations of these two callbacks.
3829 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3830 Modify get_crl() to find a valid (unexpired) CRL if possible.
3833 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3834 this would be called X509_CRL_cmp() but that name is already used by
3835 a function that just compares CRL issuer names. Cache several CRL
3836 extensions in X509_CRL structure and cache CRLDP in X509.
3839 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3840 this maps equivalent X509_NAME structures into a consistent structure.
3841 Name comparison can then be performed rapidly using memcmp().
3844 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3848 *) Allow digests to supply their own micalg string for S/MIME type using
3849 the ctrl EVP_MD_CTRL_MICALG.
3852 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3853 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3854 ctrl. It can then customise the structure before and/or after signing
3858 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3859 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3860 to free up any added signature OIDs.
3863 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3864 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3865 digest and cipher tables. New options added to openssl utility:
3866 list-message-digest-algorithms and list-cipher-algorithms.
3869 *) Change the array representation of binary polynomials: the list
3870 of degrees of non-zero coefficients is now terminated with -1.
3871 Previously it was terminated with 0, which was also part of the
3872 value; thus, the array representation was not applicable to
3873 polynomials where t^0 has coefficient zero. This change makes
3874 the array representation useful in a more general context.
3877 *) Various modifications and fixes to SSL/TLS cipher string
3878 handling. For ECC, the code now distinguishes between fixed ECDH
3879 with RSA certificates on the one hand and with ECDSA certificates
3880 on the other hand, since these are separate ciphersuites. The
3881 unused code for Fortezza ciphersuites has been removed.
3883 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3884 (not "ECDHE"). For consistency with the code for DH
3885 certificates, use of ECDH certificates is now considered ECDH
3886 authentication, not RSA or ECDSA authentication (the latter is
3887 merely the CA's signing algorithm and not actively used in the
3890 The temporary ciphersuite alias "ECCdraft" is no longer
3891 available, and ECC ciphersuites are no longer excluded from "ALL"
3892 and "DEFAULT". The following aliases now exist for RFC 4492
3893 ciphersuites, most of these by analogy with the DH case:
3895 kECDHr - ECDH cert, signed with RSA
3896 kECDHe - ECDH cert, signed with ECDSA
3897 kECDH - ECDH cert (signed with either RSA or ECDSA)
3898 kEECDH - ephemeral ECDH
3899 ECDH - ECDH cert or ephemeral ECDH
3905 AECDH - anonymous ECDH
3906 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3910 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3911 Use correct micalg parameters depending on digest(s) in signed message.
3914 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3915 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3918 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3919 an engine to register a method. Add ENGINE lookups for methods and
3920 functional reference processing.
3923 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3924 EVP_{Sign,Verify}* which allow an application to customise the signature
3928 *) New -resign option to smime utility. This adds one or more signers
3929 to an existing PKCS#7 signedData structure. Also -md option to use an
3930 alternative message digest algorithm for signing.
3933 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3934 create PKCS7 structures containing multiple signers. Update smime
3935 application to support multiple signers.
3938 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3942 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3943 Reorganize PBE internals to lookup from a static table using NIDs,
3944 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3945 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3946 PRF which will be automatically used with PBES2.
3949 *) Replace the algorithm specific calls to generate keys in "req" with the
3953 *) Update PKCS#7 enveloped data routines to use new API. This is now
3954 supported by any public key method supporting the encrypt operation. A
3955 ctrl is added to allow the public key algorithm to examine or modify
3956 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3960 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3961 a default digest type to use. In most cases this will be SHA1 but some
3962 algorithms (such as GOST) need to specify an alternative digest. The
3963 return value indicates how strong the preference is 1 means optional and
3964 2 is mandatory (that is it is the only supported type). Modify
3965 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3966 use the default md. Update openssl utilities to use the default digest
3967 type for signing if it is not explicitly indicated.
3970 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3971 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3972 signing method from the key type. This effectively removes the link
3973 between digests and public key types.
3976 *) Add an OID cross reference table and utility functions. Its purpose is to
3977 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3978 rsaEncryption. This will allow some of the algorithm specific hackery
3979 needed to use the correct OID to be removed.
3982 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3983 structures for PKCS7_sign(). They are now set up by the relevant public
3987 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3990 *) Add support for key derivation (agreement) in the API, DH method and
3994 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3995 public and private key formats. As a side effect these add additional
3996 command line functionality not previously available: DSA signatures can be
3997 generated and verified using pkeyutl and DH key support and generation in
4002 [Oliver Tappe <zooey@hirschkaefer.de>]
4004 *) New make target "install_html_docs" installs HTML renditions of the
4006 [Oliver Tappe <zooey@hirschkaefer.de>]
4008 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4009 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4010 support key and parameter generation and add initial key generation
4011 functionality for RSA.
4014 *) Add functions for main EVP_PKEY_method operations. The undocumented
4015 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4016 EVP_PKEY_{encrypt,decrypt}_old.
4019 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4020 key API, doesn't do much yet.
4023 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4024 public key algorithms. New option to openssl utility:
4025 "list-public-key-algorithms" to print out info.
4028 *) Implement the Supported Elliptic Curves Extension for
4029 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4032 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4033 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4036 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4037 utilities such as rsa, dsa, dsaparam etc except they process any key
4041 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4042 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4043 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4047 *) Initial support for pluggable public key ASN1.
4048 De-spaghettify the public key ASN1 handling. Move public and private
4049 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4050 algorithm specific handling to a single module within the relevant
4051 algorithm directory. Add functions to allow (near) opaque processing
4052 of public and private key structures.
4055 *) Implement the Supported Point Formats Extension for
4056 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4059 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4060 for the psk identity [hint] and the psk callback functions to the
4061 SSL_SESSION, SSL and SSL_CTX structure.
4064 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4068 SSL_CTX_use_psk_identity_hint
4069 SSL_get_psk_identity_hint
4070 SSL_get_psk_identity
4071 SSL_use_psk_identity_hint
4073 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4075 *) Add RFC 3161 compliant time stamp request creation, response generation
4076 and response verification functionality.
4077 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4079 *) Add initial support for TLS extensions, specifically for the server_name
4080 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4081 have new members for a host name. The SSL data structure has an
4082 additional member SSL_CTX *initial_ctx so that new sessions can be
4083 stored in that context to allow for session resumption, even after the
4084 SSL has been switched to a new SSL_CTX in reaction to a client's
4085 server_name extension.
4087 New functions (subject to change):
4089 SSL_get_servername()
4090 SSL_get_servername_type()
4093 New CTRL codes and macros (subject to change):
4095 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4096 - SSL_CTX_set_tlsext_servername_callback()
4097 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4098 - SSL_CTX_set_tlsext_servername_arg()
4099 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4101 openssl s_client has a new '-servername ...' option.
4103 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4104 '-key2 ...', '-servername_fatal' (subject to change). This allows
4105 testing the HostName extension for a specific single host name ('-cert'
4106 and '-key' remain fallbacks for handshakes without HostName
4107 negotiation). If the unrecognized_name alert has to be sent, this by
4108 default is a warning; it becomes fatal with the '-servername_fatal'
4111 [Peter Sylvester, Remy Allais, Christophe Renou]
4113 *) Whirlpool hash implementation is added.
4116 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4117 bn(64,32). Because of instruction set limitations it doesn't have
4118 any negative impact on performance. This was done mostly in order
4119 to make it possible to share assembler modules, such as bn_mul_mont
4120 implementations, between 32- and 64-bit builds without hassle.
4123 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4124 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4128 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4129 dedicated Montgomery multiplication procedure, is introduced.
4130 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4131 "64-bit" performance on certain 32-bit targets.
4134 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4135 in SSL structures. New SSL ctrl to set maximum send fragment size.
4136 Save memory by seeting the I/O buffer sizes dynamically instead of
4137 using the maximum available value.
4140 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4141 in addition to the text details.
4144 *) Very, very preliminary EXPERIMENTAL support for printing of general
4145 ASN1 structures. This currently produces rather ugly output and doesn't
4146 handle several customised structures at all.
4149 *) Integrated support for PVK file format and some related formats such
4150 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4151 these in the 'rsa' and 'dsa' utilities.
4154 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4157 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4158 place for the (very old) "NETSCAPE" format certificates which are now
4159 handled using new ASN1 code equivalents.
4162 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4163 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4164 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4167 *) Modify CRL distribution points extension code to print out previously
4168 unsupported fields. Enhance extension setting code to allow setting of
4172 *) Add print and set support for Issuing Distribution Point CRL extension.
4175 *) Change 'Configure' script to enable Camellia by default.
4178 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4180 *) When rejecting SSL/TLS records due to an incorrect version number, never
4181 update s->server with a new major version number. As of
4182 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4183 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4184 the previous behavior could result in a read attempt at NULL when
4185 receiving specific incorrect SSL/TLS records once record payload
4186 protection is active. (CVE-2010-0740)
4187 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4189 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4190 could be crashed if the relevant tables were not present (e.g. chrooted).
4191 [Tomas Hoger <thoger@redhat.com>]
4193 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4195 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
4196 [Martin Olsson, Neel Mehta]
4198 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4199 accommodate for stack sorting, always a write lock!).
4202 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4203 excessive delays in the RAND_poll(): over a minute. As a workaround
4204 include a time check in the inner Heap32Next loop too.
4207 *) The code that handled flushing of data in SSL/TLS originally used the
4208 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4209 the problem outlined in PR#1949. The fix suggested there however can
4210 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4211 of Apache). So instead simplify the code to flush unconditionally.
4212 This should be fine since flushing with no data to flush is a no op.
4215 *) Handle TLS versions 2.0 and later properly and correctly use the
4216 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4217 off ancient servers have a habit of sticking around for a while...
4220 *) Modify compression code so it frees up structures without using the
4221 ex_data callbacks. This works around a problem where some applications
4222 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4223 restarting) then use compression (e.g. SSL with compression) later.
4224 This results in significant per-connection memory leaks and
4225 has caused some security issues including CVE-2008-1678 and
4229 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4230 change when encrypting or decrypting.
4233 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4234 connect and renegotiate with servers which do not support RI.
4235 Until RI is more widely deployed this option is enabled by default.
4238 *) Add "missing" ssl ctrls to clear options and mode.
4241 *) If client attempts to renegotiate and doesn't support RI respond with
4242 a no_renegotiation alert as required by RFC5746. Some renegotiating
4243 TLS clients will continue a connection gracefully when they receive
4244 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4245 waiting for a server hello which it will never receive. Now we treat a
4246 received no_renegotiation alert as a fatal error. This is because
4247 applications requesting a renegotiation might well expect it to succeed
4248 and would have no code in place to handle the server denying it so the
4249 only safe thing to do is to terminate the connection.
4252 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4253 peer supports secure renegotiation and 0 otherwise. Print out peer
4254 renegotiation support in s_client/s_server.
4257 *) Replace the highly broken and deprecated SPKAC certification method with
4258 the updated NID creation version. This should correctly handle UTF8.
4261 *) Implement RFC5746. Re-enable renegotiation but require the extension
4262 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4263 turns out to be a bad idea. It has been replaced by
4264 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4265 SSL_CTX_set_options(). This is really not recommended unless you
4266 know what you are doing.
4267 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4269 *) Fixes to stateless session resumption handling. Use initial_ctx when
4270 issuing and attempting to decrypt tickets in case it has changed during
4271 servername handling. Use a non-zero length session ID when attempting
4272 stateless session resumption: this makes it possible to determine if
4273 a resumption has occurred immediately after receiving server hello
4274 (several places in OpenSSL subtly assume this) instead of later in
4278 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4279 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4280 fixes for a few places where the return code is not checked
4282 [Julia Lawall <julia@diku.dk>]
4284 *) Add --strict-warnings option to Configure script to include devteam
4285 warnings in other configurations.
4288 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4289 makes it possible to install openssl libraries in locations which
4290 have names other than "lib", for example "/usr/lib64" which some
4292 [Steve Henson, based on patch from Jeremy Utley]
4294 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4295 X690 8.9.12 and can produce some misleading textual output of OIDs.
4296 [Steve Henson, reported by Dan Kaminsky]
4298 *) Delete MD2 from algorithm tables. This follows the recommendation in
4299 several standards that it is not used in new applications due to
4300 several cryptographic weaknesses. For binary compatibility reasons
4301 the MD2 API is still compiled in by default.
4304 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4308 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4309 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4311 [Guenter <lists@gknw.net>]
4313 *) Fix the server certificate chain building code to use X509_verify_cert(),
4314 it used to have an ad-hoc builder which was unable to cope with anything
4315 other than a simple chain.
4316 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4318 *) Don't check self signed certificate signatures in X509_verify_cert()
4319 by default (a flag can override this): it just wastes time without
4320 adding any security. As a useful side effect self signed root CAs
4321 with non-FIPS digests are now usable in FIPS mode.
4324 *) In dtls1_process_out_of_seq_message() the check if the current message
4325 is already buffered was missing. For every new message was memory
4326 allocated, allowing an attacker to perform an denial of service attack
4327 with sending out of seq handshake messages until there is no memory
4328 left. Additionally every future messege was buffered, even if the
4329 sequence number made no sense and would be part of another handshake.
4330 So only messages with sequence numbers less than 10 in advance will be
4331 buffered. (CVE-2009-1378)
4332 [Robin Seggelmann, discovered by Daniel Mentz]
4334 *) Records are buffered if they arrive with a future epoch to be
4335 processed after finishing the corresponding handshake. There is
4336 currently no limitation to this buffer allowing an attacker to perform
4337 a DOS attack with sending records with future epochs until there is no
4338 memory left. This patch adds the pqueue_size() function to determine
4339 the size of a buffer and limits the record buffer to 100 entries.
4341 [Robin Seggelmann, discovered by Daniel Mentz]
4343 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4344 parent structure is freed. (CVE-2009-1379)
4347 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4348 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4351 [Ilya O. <vrghost@gmail.com>]
4353 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4355 *) Disable renegotiation completely - this fixes a severe security
4356 problem (CVE-2009-3555) at the cost of breaking all
4357 renegotiation. Renegotiation can be re-enabled by setting
4358 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4359 run-time. This is really not recommended unless you know what
4363 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4365 *) Don't set val to NULL when freeing up structures, it is freed up by
4366 underlying code. If sizeof(void *) > sizeof(long) this can result in
4367 zeroing past the valid field. (CVE-2009-0789)
4368 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4370 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4371 checked correctly. This would allow some invalid signed attributes to
4372 appear to verify correctly. (CVE-2009-0591)
4373 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4375 *) Reject UniversalString and BMPString types with invalid lengths. This
4376 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4377 a legal length. (CVE-2009-0590)
4380 *) Set S/MIME signing as the default purpose rather than setting it
4381 unconditionally. This allows applications to override it at the store
4385 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4386 to handle some structures.
4389 *) Improve efficiency of mem_gets: don't search whole buffer each time
4391 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4393 *) New -hex option for openssl rand.
4396 *) Print out UTF8String and NumericString when parsing ASN1.
4399 *) Support NumericString type for name components.
4402 *) Allow CC in the environment to override the automatically chosen
4403 compiler. Note that nothing is done to ensure flags work with the
4407 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4409 *) Properly check EVP_VerifyFinal() and similar return values
4411 [Ben Laurie, Bodo Moeller, Google Security Team]
4413 *) Enable TLS extensions by default.
4416 *) Allow the CHIL engine to be loaded, whether the application is
4417 multithreaded or not. (This does not release the developer from the
4418 obligation to set up the dynamic locking callbacks.)
4419 [Sander Temme <sander@temme.net>]
4421 *) Use correct exit code if there is an error in dgst command.
4422 [Steve Henson; problem pointed out by Roland Dirlewanger]
4424 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4425 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4428 *) Add experimental JPAKE support, including demo authentication in
4429 s_client and s_server.
4432 *) Set the comparison function in v3_addr_canonize().
4433 [Rob Austein <sra@hactrn.net>]
4435 *) Add support for XMPP STARTTLS in s_client.
4436 [Philip Paeps <philip@freebsd.org>]
4438 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4439 to ensure that even with this option, only ciphersuites in the
4440 server's preference list will be accepted. (Note that the option
4441 applies only when resuming a session, so the earlier behavior was
4442 just about the algorithm choice for symmetric cryptography.)
4445 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4447 *) Fix NULL pointer dereference if a DTLS server received
4448 ChangeCipherSpec as first record (CVE-2009-1386).
4451 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4452 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4455 *) The fix in 0.9.8c that supposedly got rid of unsafe
4456 double-checked locking was incomplete for RSA blinding,
4457 addressing just one layer of what turns out to have been
4458 doubly unsafe triple-checked locking.
4460 So now fix this for real by retiring the MONT_HELPER macro
4461 in crypto/rsa/rsa_eay.c.
4463 [Bodo Moeller; problem pointed out by Marius Schilder]
4465 *) Various precautionary measures:
4467 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4469 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4470 (NB: This would require knowledge of the secret session ticket key
4471 to exploit, in which case you'd be SOL either way.)
4473 - Change bn_nist.c so that it will properly handle input BIGNUMs
4474 outside the expected range.
4476 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4479 [Neel Mehta, Bodo Moeller]
4481 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4482 the load fails. Useful for distros.
4483 [Ben Laurie and the FreeBSD team]
4485 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4488 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4491 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4493 This work was sponsored by Logica.
4496 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4497 keystores. Support for SSL/TLS client authentication too.
4498 Not compiled unless enable-capieng specified to Configure.
4500 This work was sponsored by Logica.
4503 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4504 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4505 attribute creation routines such as certificate requests and PKCS#12
4509 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4511 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4512 handshake which could lead to a cilent crash as found using the
4513 Codenomicon TLS test suite (CVE-2008-1672)
4514 [Steve Henson, Mark Cox]
4516 *) Fix double free in TLS server name extensions which could lead to
4517 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4520 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4522 Clear the error queue to ensure that error entries left from
4523 older function calls do not interfere with the correct operation.
4524 [Lutz Jaenicke, Erik de Castro Lopo]
4526 *) Remove root CA certificates of commercial CAs:
4528 The OpenSSL project does not recommend any specific CA and does not
4529 have any policy with respect to including or excluding any CA.
4530 Therefore it does not make any sense to ship an arbitrary selection
4531 of root CA certificates with the OpenSSL software.
4534 *) RSA OAEP patches to fix two separate invalid memory reads.
4535 The first one involves inputs when 'lzero' is greater than
4536 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4537 before the beginning of from). The second one involves inputs where
4538 the 'db' section contains nothing but zeroes (there is a one-byte
4539 invalid read after the end of 'db').
4540 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4542 *) Partial backport from 0.9.9-dev:
4544 Introduce bn_mul_mont (dedicated Montgomery multiplication
4545 procedure) as a candidate for BIGNUM assembler implementation.
4546 While 0.9.9-dev uses assembler for various architectures, only
4547 x86_64 is available by default here in the 0.9.8 branch, and
4548 32-bit x86 is available through a compile-time setting.
4550 To try the 32-bit x86 assembler implementation, use Configure
4551 option "enable-montasm" (which exists only for this backport).
4553 As "enable-montasm" for 32-bit x86 disclaims code stability
4554 anyway, in this constellation we activate additional code
4555 backported from 0.9.9-dev for further performance improvements,
4556 namely BN_from_montgomery_word. (To enable this otherwise,
4557 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4559 [Andy Polyakov (backport partially by Bodo Moeller)]
4561 *) Add TLS session ticket callback. This allows an application to set
4562 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4563 values. This is useful for key rollover for example where several key
4564 sets may exist with different names.
4567 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4568 This was broken until now in 0.9.8 releases, such that the only way
4569 a registered ENGINE could be used (assuming it initialises
4570 successfully on the host) was to explicitly set it as the default
4571 for the relevant algorithms. This is in contradiction with 0.9.7
4572 behaviour and the documentation. With this fix, when an ENGINE is
4573 registered into a given algorithm's table of implementations, the
4574 'uptodate' flag is reset so that auto-discovery will be used next
4575 time a new context for that algorithm attempts to select an
4577 [Ian Lister (tweaked by Geoff Thorpe)]
4579 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4580 implementation in the following ways:
4582 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4585 Lack of BER streaming support means one pass streaming processing is
4586 only supported if data is detached: setting the streaming flag is
4587 ignored for embedded content.
4589 CMS support is disabled by default and must be explicitly enabled
4590 with the enable-cms configuration option.
4593 *) Update the GMP engine glue to do direct copies between BIGNUM and
4594 mpz_t when openssl and GMP use the same limb size. Otherwise the
4595 existing "conversion via a text string export" trick is still used.
4596 [Paul Sheer <paulsheer@gmail.com>]
4598 *) Zlib compression BIO. This is a filter BIO which compressed and
4599 uncompresses any data passed through it.
4602 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4603 RFC3394 compatible AES key wrapping.
4606 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4607 sets string data without copying. X509_ALGOR_set0() and
4608 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4609 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4610 from an X509_ATTRIBUTE structure optionally checking it occurs only
4611 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4615 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4616 to get the expected BN_FLG_CONSTTIME behavior.
4617 [Bodo Moeller (Google)]
4621 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4622 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4623 - added some more tests to do_tests.pl
4624 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4625 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4626 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4627 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4628 - various changes to netware.pl to enable gcc-cross builds on Win32
4630 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4631 - various changes to fix missing prototype warnings
4632 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4633 - added AES, WHIRLPOOL and CPUID assembler code to build files
4634 - added missing AES assembler make rules to mk1mf.pl
4635 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4636 [Guenter Knauf <eflash@gmx.net>]
4638 *) Implement certificate status request TLS extension defined in RFC3546.
4639 A client can set the appropriate parameters and receive the encoded
4640 OCSP response via a callback. A server can query the supplied parameters
4641 and set the encoded OCSP response in the callback. Add simplified examples
4642 to s_client and s_server.
4645 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4647 *) Fix various bugs:
4648 + Binary incompatibility of ssl_ctx_st structure
4649 + DTLS interoperation with non-compliant servers
4650 + Don't call get_session_cb() without proposed session
4651 + Fix ia64 assembler code
4652 [Andy Polyakov, Steve Henson]
4654 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4656 *) DTLS Handshake overhaul. There were longstanding issues with
4657 OpenSSL DTLS implementation, which were making it impossible for
4658 RFC 4347 compliant client to communicate with OpenSSL server.
4659 Unfortunately just fixing these incompatibilities would "cut off"
4660 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4661 server keeps tolerating non RFC compliant syntax. The opposite is
4662 not true, 0.9.8f client can not communicate with earlier server.
4663 This update even addresses CVE-2007-4995.
4666 *) Changes to avoid need for function casts in OpenSSL: some compilers
4667 (gcc 4.2 and later) reject their use.
4668 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4671 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4672 RFC4507bis. The encrypted ticket format is an encrypted encoded
4673 SSL_SESSION structure, that way new session features are automatically
4676 If a client application caches session in an SSL_SESSION structure
4677 support is transparent because tickets are now stored in the encoded
4680 The SSL_CTX structure automatically generates keys for ticket
4681 protection in servers so again support should be possible
4682 with no application modification.
4684 If a client or server wishes to disable RFC4507 support then the option
4685 SSL_OP_NO_TICKET can be set.
4687 Add a TLS extension debugging callback to allow the contents of any client
4688 or server extensions to be examined.
4690 This work was sponsored by Google.
4693 *) Add initial support for TLS extensions, specifically for the server_name
4694 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4695 have new members for a host name. The SSL data structure has an
4696 additional member SSL_CTX *initial_ctx so that new sessions can be
4697 stored in that context to allow for session resumption, even after the
4698 SSL has been switched to a new SSL_CTX in reaction to a client's
4699 server_name extension.
4701 New functions (subject to change):
4703 SSL_get_servername()
4704 SSL_get_servername_type()
4707 New CTRL codes and macros (subject to change):
4709 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4710 - SSL_CTX_set_tlsext_servername_callback()
4711 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4712 - SSL_CTX_set_tlsext_servername_arg()
4713 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4715 openssl s_client has a new '-servername ...' option.
4717 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4718 '-key2 ...', '-servername_fatal' (subject to change). This allows
4719 testing the HostName extension for a specific single host name ('-cert'
4720 and '-key' remain fallbacks for handshakes without HostName
4721 negotiation). If the unrecognized_name alert has to be sent, this by
4722 default is a warning; it becomes fatal with the '-servername_fatal'
4725 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4727 *) Add AES and SSE2 assembly language support to VC++ build.
4730 *) Mitigate attack on final subtraction in Montgomery reduction.
4733 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4734 (which previously caused an internal error).
4737 *) Squeeze another 10% out of IGE mode when in != out.
4740 *) AES IGE mode speedup.
4741 [Dean Gaudet (Google)]
4743 *) Add the Korean symmetric 128-bit cipher SEED (see
4744 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4745 add SEED ciphersuites from RFC 4162:
4747 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4748 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4749 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4750 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4752 To minimize changes between patchlevels in the OpenSSL 0.9.8
4753 series, SEED remains excluded from compilation unless OpenSSL
4754 is configured with 'enable-seed'.
4755 [KISA, Bodo Moeller]
4757 *) Mitigate branch prediction attacks, which can be practical if a
4758 single processor is shared, allowing a spy process to extract
4759 information. For detailed background information, see
4760 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4761 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4762 and Necessary Software Countermeasures"). The core of the change
4763 are new versions BN_div_no_branch() and
4764 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4765 respectively, which are slower, but avoid the security-relevant
4766 conditional branches. These are automatically called by BN_div()
4767 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4768 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4769 remove a conditional branch.
4771 BN_FLG_CONSTTIME is the new name for the previous
4772 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4773 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4774 in the exponent causes BN_mod_exp_mont() to use the alternative
4775 implementation in BN_mod_exp_mont_consttime().) The old name
4776 remains as a deprecated alias.
4778 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4779 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4780 constant-time implementations for more than just exponentiation.
4781 Here too the old name is kept as a deprecated alias.
4783 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4784 the BN_BLINDING structure gets an independent copy of the
4785 modulus. This means that the previous "BIGNUM *m" argument to
4786 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4787 essentially becomes "const BIGNUM *m", although we can't actually
4788 change this in the header file before 0.9.9. It allows
4789 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4790 enable BN_FLG_CONSTTIME.
4792 [Matthew D Wood (Intel Corp)]
4794 *) In the SSL/TLS server implementation, be strict about session ID
4795 context matching (which matters if an application uses a single
4796 external cache for different purposes). Previously,
4797 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4798 set. This did ensure strict client verification, but meant that,
4799 with applications using a single external cache for quite
4800 different requirements, clients could circumvent ciphersuite
4801 restrictions for a given session ID context by starting a session
4802 in a different context.
4805 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4806 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4807 authentication-only ciphersuites.
4810 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4811 not complete and could lead to a possible single byte overflow
4812 (CVE-2007-5135) [Ben Laurie]
4814 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4816 *) Since AES128 and AES256 (and similarly Camellia128 and
4817 Camellia256) share a single mask bit in the logic of
4818 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4819 kludge to work properly if AES128 is available and AES256 isn't
4820 (or if Camellia128 is available and Camellia256 isn't).
4823 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4824 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4825 When a point or a seed is encoded in a BIT STRING, we need to
4826 prevent the removal of trailing zero bits to get the proper DER
4827 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4828 of a NamedBitList, for which trailing 0 bits need to be removed.)
4831 *) Have SSL/TLS server implementation tolerate "mismatched" record
4832 protocol version while receiving ClientHello even if the
4833 ClientHello is fragmented. (The server can't insist on the
4834 particular protocol version it has chosen before the ServerHello
4835 message has informed the client about his choice.)
4838 *) Add RFC 3779 support.
4839 [Rob Austein for ARIN, Ben Laurie]
4841 *) Load error codes if they are not already present instead of using a
4842 static variable. This allows them to be cleanly unloaded and reloaded.
4843 Improve header file function name parsing.
4846 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4847 or CAPABILITY handshake as required by RFCs.
4850 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4852 *) Introduce limits to prevent malicious keys being able to
4853 cause a denial of service. (CVE-2006-2940)
4854 [Steve Henson, Bodo Moeller]
4856 *) Fix ASN.1 parsing of certain invalid structures that can result
4857 in a denial of service. (CVE-2006-2937) [Steve Henson]
4859 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4860 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4862 *) Fix SSL client code which could crash if connecting to a
4863 malicious SSLv2 server. (CVE-2006-4343)
4864 [Tavis Ormandy and Will Drewry, Google Security Team]
4866 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4867 match only those. Before that, "AES256-SHA" would be interpreted
4868 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4869 the same strength classification in 0.9.7h) as we currently only
4870 have a single AES bit in the ciphersuite description bitmap.
4871 That change, however, also applied to ciphersuite strings such as
4872 "RC4-MD5" that intentionally matched multiple ciphersuites --
4873 namely, SSL 2.0 ciphersuites in addition to the more common ones
4874 from SSL 3.0/TLS 1.0.
4876 So we change the selection algorithm again: Naming an explicit
4877 ciphersuite selects this one ciphersuite, and any other similar
4878 ciphersuite (same bitmap) from *other* protocol versions.
4879 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4880 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4882 Since SSL 2.0 does not have any ciphersuites for which the
4883 128/256 bit distinction would be relevant, this works for now.
4884 The proper fix will be to use different bits for AES128 and
4885 AES256, which would have avoided the problems from the beginning;
4886 however, bits are scarce, so we can only do this in a new release
4887 (not just a patchlevel) when we can change the SSL_CIPHER
4888 definition to split the single 'unsigned long mask' bitmap into
4889 multiple values to extend the available space.
4893 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4895 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4896 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4898 *) Add AES IGE and biIGE modes.
4901 *) Change the Unix randomness entropy gathering to use poll() when
4902 possible instead of select(), since the latter has some
4903 undesirable limitations.
4904 [Darryl Miles via Richard Levitte and Bodo Moeller]
4906 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4907 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4908 cannot be implicitly activated as part of, e.g., the "AES" alias.
4909 However, please upgrade to OpenSSL 0.9.9[-dev] for
4910 non-experimental use of the ECC ciphersuites to get TLS extension
4911 support, which is required for curve and point format negotiation
4912 to avoid potential handshake problems.
4915 *) Disable rogue ciphersuites:
4917 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4918 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4919 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4921 The latter two were purportedly from
4922 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4925 Also deactivate the remaining ciphersuites from
4926 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4927 unofficial, and the ID has long expired.
4930 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4931 dual-core machines) and other potential thread-safety issues.
4934 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4935 versions), which is now available for royalty-free use
4936 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4937 Also, add Camellia TLS ciphersuites from RFC 4132.
4939 To minimize changes between patchlevels in the OpenSSL 0.9.8
4940 series, Camellia remains excluded from compilation unless OpenSSL
4941 is configured with 'enable-camellia'.
4944 *) Disable the padding bug check when compression is in use. The padding
4945 bug check assumes the first packet is of even length, this is not
4946 necessarily true if compresssion is enabled and can result in false
4947 positives causing handshake failure. The actual bug test is ancient
4948 code so it is hoped that implementations will either have fixed it by
4949 now or any which still have the bug do not support compression.
4952 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4954 *) When applying a cipher rule check to see if string match is an explicit
4955 cipher suite and only match that one cipher suite if it is.
4958 *) Link in manifests for VC++ if needed.
4959 [Austin Ziegler <halostatue@gmail.com>]
4961 *) Update support for ECC-based TLS ciphersuites according to
4962 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4963 TLS extensions, which are supported starting with the 0.9.9
4964 branch, not in the OpenSSL 0.9.8 branch).
4967 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4968 opaque EVP_CIPHER_CTX handling.
4971 *) Fixes and enhancements to zlib compression code. We now only use
4972 "zlib1.dll" and use the default __cdecl calling convention on Win32
4973 to conform with the standards mentioned here:
4974 http://www.zlib.net/DLL_FAQ.txt
4975 Static zlib linking now works on Windows and the new --with-zlib-include
4976 --with-zlib-lib options to Configure can be used to supply the location
4977 of the headers and library. Gracefully handle case where zlib library
4981 *) Several fixes and enhancements to the OID generation code. The old code
4982 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4983 handle numbers larger than ULONG_MAX, truncated printing and had a
4984 non standard OBJ_obj2txt() behaviour.
4987 *) Add support for building of engines under engine/ as shared libraries
4988 under VC++ build system.
4991 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4992 Hopefully, we will not see any false combination of paths any more.
4995 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4997 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4998 (part of SSL_OP_ALL). This option used to disable the
4999 countermeasure against man-in-the-middle protocol-version
5000 rollback in the SSL 2.0 server implementation, which is a bad
5001 idea. (CVE-2005-2969)
5003 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5004 for Information Security, National Institute of Advanced Industrial
5005 Science and Technology [AIST], Japan)]
5007 *) Add two function to clear and return the verify parameter flags.
5010 *) Keep cipherlists sorted in the source instead of sorting them at
5011 runtime, thus removing the need for a lock.
5014 *) Avoid some small subgroup attacks in Diffie-Hellman.
5015 [Nick Mathewson and Ben Laurie]
5017 *) Add functions for well-known primes.
5020 *) Extended Windows CE support.
5021 [Satoshi Nakamura and Andy Polyakov]
5023 *) Initialize SSL_METHOD structures at compile time instead of during
5024 runtime, thus removing the need for a lock.
5027 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5028 attempting to decrypt each encrypted key in turn. Add support to
5032 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5034 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5037 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5040 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5041 key into the same file any more.
5044 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5047 *) Add -utf8 command line and config file option to 'ca'.
5048 [Stefan <stf@udoma.org]
5050 *) Removed the macro des_crypt(), as it seems to conflict with some
5051 libraries. Use DES_crypt().
5054 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5055 involves renaming the source and generated shared-libs for
5056 both. The engines will accept the corrected or legacy ids
5057 ('ncipher' and '4758_cca' respectively) when binding. NB,
5058 this only applies when building 'shared'.
5059 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5061 *) Add attribute functions to EVP_PKEY structure. Modify
5062 PKCS12_create() to recognize a CSP name attribute and
5063 use it. Make -CSP option work again in pkcs12 utility.
5066 *) Add new functionality to the bn blinding code:
5067 - automatic re-creation of the BN_BLINDING parameters after
5068 a fixed number of uses (currently 32)
5069 - add new function for parameter creation
5070 - introduce flags to control the update behaviour of the
5071 BN_BLINDING parameters
5072 - hide BN_BLINDING structure
5073 Add a second BN_BLINDING slot to the RSA structure to improve
5074 performance when a single RSA object is shared among several
5078 *) Add support for DTLS.
5079 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5081 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5082 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5085 *) Remove buggy and incomplete DH cert support from
5086 ssl/ssl_rsa.c and ssl/s3_both.c
5089 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5090 the apps/openssl applications.
5093 *) Compile clean with "-Wall -Wmissing-prototypes
5094 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5095 DEBUG_SAFESTACK must also be set.
5098 *) Change ./Configure so that certain algorithms can be disabled by default.
5099 The new counterpiece to "no-xxx" is "enable-xxx".
5101 The patented RC5 and MDC2 algorithms will now be disabled unless
5102 "enable-rc5" and "enable-mdc2", respectively, are specified.
5104 (IDEA remains enabled despite being patented. This is because IDEA
5105 is frequently required for interoperability, and there is no license
5106 fee for non-commercial use. As before, "no-idea" can be used to
5107 avoid this algorithm.)
5111 *) Add processing of proxy certificates (see RFC 3820). This work was
5112 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5113 EGEE (Enabling Grids for E-science in Europe).
5116 *) RC4 performance overhaul on modern architectures/implementations, such
5117 as Intel P4, IA-64 and AMD64.
5120 *) New utility extract-section.pl. This can be used specify an alternative
5121 section number in a pod file instead of having to treat each file as
5122 a separate case in Makefile. This can be done by adding two lines to the
5125 =for comment openssl_section:XXX
5127 The blank line is mandatory.
5131 *) New arguments -certform, -keyform and -pass for s_client and s_server
5132 to allow alternative format key and certificate files and passphrase
5136 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5137 update associated structures and add various utility functions.
5139 Add new policy related verify parameters, include policy checking in
5140 standard verify code. Enhance 'smime' application with extra parameters
5141 to support policy checking and print out.
5144 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5145 Nehemiah processors. These extensions support AES encryption in hardware
5146 as well as RNG (though RNG support is currently disabled).
5147 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5149 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5152 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5153 [Andy Polyakov and a number of other people]
5155 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5156 implementation contributed by IBM.
5157 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5159 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5160 exponent rather than 'unsigned long'. There is a corresponding change to
5161 the new 'rsa_keygen' element of the RSA_METHOD structure.
5162 [Jelte Jansen, Geoff Thorpe]
5164 *) Functionality for creating the initial serial number file is now
5165 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5167 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5168 number file to 1, which is bound to cause problems. To avoid
5169 the problems while respecting compatibility between different 0.9.7
5170 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5171 CA.pl for serial number initialization. With the new release 0.9.8,
5172 we can fix the problem directly in the 'ca' utility.)
5175 *) Reduced header interdepencies by declaring more opaque objects in
5176 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5177 give fewer recursive includes, which could break lazy source code - so
5178 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5179 developers should define this symbol when building and using openssl to
5180 ensure they track the recommended behaviour, interfaces, [etc], but
5181 backwards-compatible behaviour prevails when this isn't defined.
5184 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5187 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5188 This will generate a random key of the appropriate length based on the
5189 cipher context. The EVP_CIPHER can provide its own random key generation
5190 routine to support keys of a specific form. This is used in the des and
5191 3des routines to generate a key of the correct parity. Update S/MIME
5192 code to use new functions and hence generate correct parity DES keys.
5193 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5194 valid (weak or incorrect parity).
5197 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5198 as looking them up. This is useful when the verified structure may contain
5199 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5200 present unless the new PKCS7_NO_CRL flag is asserted.
5203 *) Extend ASN1 oid configuration module. It now additionally accepts the
5206 shortName = some long name, 1.2.3.4
5209 *) Reimplemented the BN_CTX implementation. There is now no more static
5210 limitation on the number of variables it can handle nor the depth of the
5211 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5212 information can now expand as required, and rather than having a single
5213 static array of bignums, BN_CTX now uses a linked-list of such arrays
5214 allowing it to expand on demand whilst maintaining the usefulness of
5215 BN_CTX's "bundling".
5218 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5219 to allow all RSA operations to function using a single BN_CTX.
5222 *) Preliminary support for certificate policy evaluation and checking. This
5223 is initially intended to pass the tests outlined in "Conformance Testing
5224 of Relying Party Client Certificate Path Processing Logic" v1.07.
5227 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5228 remained unused and not that useful. A variety of other little bignum
5229 tweaks and fixes have also been made continuing on from the audit (see
5233 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5234 associated ASN1, EVP and SSL functions and old ASN1 macros.
5237 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5238 and this should never fail. So the return value from the use of
5239 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5240 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5243 *) BN_CTX_get() should return zero-valued bignums, providing the same
5244 initialised value as BN_new().
5245 [Geoff Thorpe, suggested by Ulf Möller]
5247 *) Support for inhibitAnyPolicy certificate extension.
5250 *) An audit of the BIGNUM code is underway, for which debugging code is
5251 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5252 is considered valid when processing BIGNUMs, and causes execution to
5253 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5254 further steps are taken to deliberately pollute unused data in BIGNUM
5255 structures to try and expose faulty code further on. For now, openssl will
5256 (in its default mode of operation) continue to tolerate the inconsistent
5257 forms that it has tolerated in the past, but authors and packagers should
5258 consider trying openssl and their own applications when compiled with
5259 these debugging symbols defined. It will help highlight potential bugs in
5260 their own code, and will improve the test coverage for OpenSSL itself. At
5261 some point, these tighter rules will become openssl's default to improve
5262 maintainability, though the assert()s and other overheads will remain only
5263 in debugging configurations. See bn.h for more details.
5264 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5266 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5267 that can only be obtained through BN_CTX_new() (which implicitly
5268 initialises it). The presence of this function only made it possible
5269 to overwrite an existing structure (and cause memory leaks).
5272 *) Because of the callback-based approach for implementing LHASH as a
5273 template type, lh_insert() adds opaque objects to hash-tables and
5274 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5275 to clean up those corresponding objects before destroying the hash table
5276 (and losing the object pointers). So some over-zealous constifications in
5277 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5278 objects as "const" and the lh_doall[_arg] callback wrappers are not
5279 prototyped to have "const" restrictions on the object pointers they are
5280 given (and so aren't required to cast them away any more).
5283 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5284 (speed) prefers to use its own implementation. The two implementations
5285 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5286 its object type properly exposed (MS_TM) instead of casting to/from "char
5287 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5288 aren't necessarily the greatest nomenclatures - but this is what was used
5289 internally to the implementation so I've used that for now.
5292 *) Ensure that deprecated functions do not get compiled when
5293 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5294 the self-tests were still using deprecated key-generation functions so
5295 these have been updated also.
5298 *) Reorganise PKCS#7 code to separate the digest location functionality
5299 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5300 New function PKCS7_set_digest() to set the digest type for PKCS#7
5301 digestedData type. Add additional code to correctly generate the
5302 digestedData type and add support for this type in PKCS7 initialization
5306 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5307 structure of type "other".
5310 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5311 sure the loop does correctly stop and breaking ("division by zero")
5312 modulus operations are not performed. The (pre-generated) prime
5313 table crypto/bn/bn_prime.h was already correct, but it could not be
5314 re-generated on some platforms because of the "division by zero"
5315 situation in the script.
5316 [Ralf S. Engelschall]
5318 *) Update support for ECC-based TLS ciphersuites according to
5319 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5320 SHA-1 now is only used for "small" curves (where the
5321 representation of a field element takes up to 24 bytes); for
5322 larger curves, the field element resulting from ECDH is directly
5323 used as premaster secret.
5324 [Douglas Stebila (Sun Microsystems Laboratories)]
5326 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5327 curve secp160r1 to the tests.
5328 [Douglas Stebila (Sun Microsystems Laboratories)]
5330 *) Add the possibility to load symbols globally with DSO.
5331 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5333 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5334 control of the error stack.
5337 *) Add support for STORE in ENGINE.
5340 *) Add the STORE type. The intention is to provide a common interface
5341 to certificate and key stores, be they simple file-based stores, or
5342 HSM-type store, or LDAP stores, or...
5343 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5346 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5347 pass a list of arguments to any function as well as provide a way
5348 for a function to pass data back to the caller.
5351 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5352 works like BUF_strdup() but can be used to duplicate a portion of
5353 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5357 *) Add the function sk_find_ex() which works like sk_find(), but will
5358 return an index to an element even if an exact match couldn't be
5359 found. The index is guaranteed to point at the element where the
5360 searched-for key would be inserted to preserve sorting order.
5363 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5364 takes an extra flags argument for optional functionality. Currently,
5365 the following flags are defined:
5367 OBJ_BSEARCH_VALUE_ON_NOMATCH
5368 This one gets OBJ_bsearch_ex() to return a pointer to the first
5369 element where the comparing function returns a negative or zero
5372 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5373 This one gets OBJ_bsearch_ex() to return a pointer to the first
5374 element where the comparing function returns zero. This is useful
5375 if there are more than one element where the comparing function
5379 *) Make it possible to create self-signed certificates with 'openssl ca'
5380 in such a way that the self-signed certificate becomes part of the
5381 CA database and uses the same mechanisms for serial number generation
5382 as all other certificate signing. The new flag '-selfsign' enables
5383 this functionality. Adapt CA.sh and CA.pl.in.
5386 *) Add functionality to check the public key of a certificate request
5387 against a given private. This is useful to check that a certificate
5388 request can be signed by that key (self-signing).
5391 *) Make it possible to have multiple active certificates with the same
5392 subject in the CA index file. This is done only if the keyword
5393 'unique_subject' is set to 'no' in the main CA section (default
5394 if 'CA_default') of the configuration file. The value is saved
5395 with the database itself in a separate index attribute file,
5396 named like the index file with '.attr' appended to the name.
5399 *) Generate muti valued AVAs using '+' notation in config files for
5403 *) Support for nameConstraints certificate extension.
5406 *) Support for policyConstraints certificate extension.
5409 *) Support for policyMappings certificate extension.
5412 *) Make sure the default DSA_METHOD implementation only uses its
5413 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5414 and change its own handlers to be NULL so as to remove unnecessary
5415 indirection. This lets alternative implementations fallback to the
5416 default implementation more easily.
5419 *) Support for directoryName in GeneralName related extensions
5423 *) Make it possible to link applications using Makefile.shared.
5424 Make that possible even when linking against static libraries!
5427 *) Support for single pass processing for S/MIME signing. This now
5428 means that S/MIME signing can be done from a pipe, in addition
5429 cleartext signing (multipart/signed type) is effectively streaming
5430 and the signed data does not need to be all held in memory.
5432 This is done with a new flag PKCS7_STREAM. When this flag is set
5433 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5434 is done after the data is output (and digests calculated) in
5435 SMIME_write_PKCS7().
5438 *) Add full support for -rpath/-R, both in shared libraries and
5439 applications, at least on the platforms where it's known how
5443 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5444 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5445 will now compute a table of multiples of the generator that
5446 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5447 faster (notably in the case of a single point multiplication,
5448 scalar * generator).
5449 [Nils Larsch, Bodo Moeller]
5451 *) IPv6 support for certificate extensions. The various extensions
5452 which use the IP:a.b.c.d can now take IPv6 addresses using the
5453 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5457 *) Added an ENGINE that implements RSA by performing private key
5458 exponentiations with the GMP library. The conversions to and from
5459 GMP's mpz_t format aren't optimised nor are any montgomery forms
5460 cached, and on x86 it appears OpenSSL's own performance has caught up.
5461 However there are likely to be other architectures where GMP could
5462 provide a boost. This ENGINE is not built in by default, but it can be
5463 specified at Configure time and should be accompanied by the necessary
5464 linker additions, eg;
5465 ./config -DOPENSSL_USE_GMP -lgmp
5468 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5469 testing availability of engines with "-t" - the old behaviour is
5470 produced by increasing the feature's verbosity with "-tt".
5473 *) ECDSA routines: under certain error conditions uninitialized BN objects
5474 could be freed. Solution: make sure initialization is performed early
5475 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5479 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5480 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5481 software implementations. For DSA and DH, parameter generation can
5482 also be overridden by providing the appropriate method callbacks.
5485 *) Change the "progress" mechanism used in key-generation and
5486 primality testing to functions that take a new BN_GENCB pointer in
5487 place of callback/argument pairs. The new API functions have "_ex"
5488 postfixes and the older functions are reimplemented as wrappers for
5489 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5490 declarations of the old functions to help (graceful) attempts to
5491 migrate to the new functions. Also, the new key-generation API
5492 functions operate on a caller-supplied key-structure and return
5493 success/failure rather than returning a key or NULL - this is to
5494 help make "keygen" another member function of RSA_METHOD etc.
5496 Example for using the new callback interface:
5498 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5502 BN_GENCB_set(&my_cb, my_callback, my_arg);
5504 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5505 /* For the meaning of a, b in calls to my_callback(), see the
5506 * documentation of the function that calls the callback.
5507 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5508 * my_callback should return 1 if it wants BN_is_prime_ex()
5509 * to continue, or 0 to stop.
5514 *) Change the ZLIB compression method to be stateful, and make it
5515 available to TLS with the number defined in
5516 draft-ietf-tls-compression-04.txt.
5519 *) Add the ASN.1 structures and functions for CertificatePair, which
5520 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5522 CertificatePair ::= SEQUENCE {
5523 forward [0] Certificate OPTIONAL,
5524 reverse [1] Certificate OPTIONAL,
5525 -- at least one of the pair shall be present -- }
5527 Also implement the PEM functions to read and write certificate
5528 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5530 This needed to be defined, mostly for the sake of the LDAP
5531 attribute crossCertificatePair, but may prove useful elsewhere as
5535 *) Make it possible to inhibit symlinking of shared libraries in
5536 Makefile.shared, for Cygwin's sake.
5539 *) Extend the BIGNUM API by creating a function
5540 void BN_set_negative(BIGNUM *a, int neg);
5541 and a macro that behave like
5542 int BN_is_negative(const BIGNUM *a);
5544 to avoid the need to access 'a->neg' directly in applications.
5547 *) Implement fast modular reduction for pseudo-Mersenne primes
5548 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5549 EC_GROUP_new_curve_GFp() will now automatically use this
5551 [Nils Larsch <nla@trustcenter.de>]
5553 *) Add new lock type (CRYPTO_LOCK_BN).
5556 *) Change the ENGINE framework to automatically load engines
5557 dynamically from specific directories unless they could be
5558 found to already be built in or loaded. Move all the
5559 current engines except for the cryptodev one to a new
5561 The engines in engines/ are built as shared libraries if
5562 the "shared" options was given to ./Configure or ./config.
5563 Otherwise, they are inserted in libcrypto.a.
5564 /usr/local/ssl/engines is the default directory for dynamic
5565 engines, but that can be overridden at configure time through
5566 the usual use of --prefix and/or --openssldir, and at run
5567 time with the environment variable OPENSSL_ENGINES.
5568 [Geoff Thorpe and Richard Levitte]
5570 *) Add Makefile.shared, a helper makefile to build shared
5571 libraries. Adapt Makefile.org.
5574 *) Add version info to Win32 DLLs.
5575 [Peter 'Luna' Runestig" <peter@runestig.com>]
5577 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5578 can be added using this API to created arbitrary PKCS#12
5579 files while avoiding the low level API.
5581 New options to PKCS12_create(), key or cert can be NULL and
5582 will then be omitted from the output file. The encryption
5583 algorithm NIDs can be set to -1 for no encryption, the mac
5584 iteration count can be set to 0 to omit the mac.
5586 Enhance pkcs12 utility by making the -nokeys and -nocerts
5587 options work when creating a PKCS#12 file. New option -nomac
5588 to omit the mac, NONE can be set for an encryption algorithm.
5589 New code is modified to use the enhanced PKCS12_create()
5590 instead of the low level API.
5593 *) Extend ASN1 encoder to support indefinite length constructed
5594 encoding. This can output sequences tags and octet strings in
5595 this form. Modify pk7_asn1.c to support indefinite length
5596 encoding. This is experimental and needs additional code to
5597 be useful, such as an ASN1 bio and some enhanced streaming
5600 Extend template encode functionality so that tagging is passed
5601 down to the template encoder.
5604 *) Let 'openssl req' fail if an argument to '-newkey' is not
5605 recognized instead of using RSA as a default.
5608 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5609 As these are not official, they are not included in "ALL";
5610 the "ECCdraft" ciphersuite group alias can be used to select them.
5611 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5613 *) Add ECDH engine support.
5614 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5616 *) Add ECDH in new directory crypto/ecdh/.
5617 [Douglas Stebila (Sun Microsystems Laboratories)]
5619 *) Let BN_rand_range() abort with an error after 100 iterations
5620 without success (which indicates a broken PRNG).
5623 *) Change BN_mod_sqrt() so that it verifies that the input value
5624 is really the square of the return value. (Previously,
5625 BN_mod_sqrt would show GIGO behaviour.)
5628 *) Add named elliptic curves over binary fields from X9.62, SECG,
5629 and WAP/WTLS; add OIDs that were still missing.
5631 [Sheueling Chang Shantz and Douglas Stebila
5632 (Sun Microsystems Laboratories)]
5634 *) Extend the EC library for elliptic curves over binary fields
5635 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5638 EC_GF2m_simple_method
5642 EC_GROUP_new_curve_GF2m
5643 EC_GROUP_set_curve_GF2m
5644 EC_GROUP_get_curve_GF2m
5645 EC_POINT_set_affine_coordinates_GF2m
5646 EC_POINT_get_affine_coordinates_GF2m
5647 EC_POINT_set_compressed_coordinates_GF2m
5649 Point compression for binary fields is disabled by default for
5650 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5653 As binary polynomials are represented as BIGNUMs, various members
5654 of the EC_GROUP and EC_POINT data structures can be shared
5655 between the implementations for prime fields and binary fields;
5656 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5657 are essentially identical to their ..._GFp counterparts.
5658 (For simplicity, the '..._GFp' prefix has been dropped from
5659 various internal method names.)
5661 An internal 'field_div' method (similar to 'field_mul' and
5662 'field_sqr') has been added; this is used only for binary fields.
5664 [Sheueling Chang Shantz and Douglas Stebila
5665 (Sun Microsystems Laboratories)]
5667 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5668 through methods ('mul', 'precompute_mult').
5670 The generic implementations (now internally called 'ec_wNAF_mul'
5671 and 'ec_wNAF_precomputed_mult') remain the default if these
5672 methods are undefined.
5674 [Sheueling Chang Shantz and Douglas Stebila
5675 (Sun Microsystems Laboratories)]
5677 *) New function EC_GROUP_get_degree, which is defined through
5678 EC_METHOD. For curves over prime fields, this returns the bit
5679 length of the modulus.
5681 [Sheueling Chang Shantz and Douglas Stebila
5682 (Sun Microsystems Laboratories)]
5684 *) New functions EC_GROUP_dup, EC_POINT_dup.
5685 (These simply call ..._new and ..._copy).
5687 [Sheueling Chang Shantz and Douglas Stebila
5688 (Sun Microsystems Laboratories)]
5690 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5691 Polynomials are represented as BIGNUMs (where the sign bit is not
5692 used) in the following functions [macros]:
5695 BN_GF2m_sub [= BN_GF2m_add]
5696 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5697 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5698 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5700 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5701 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5702 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5703 BN_GF2m_cmp [= BN_ucmp]
5705 (Note that only the 'mod' functions are actually for fields GF(2^m).
5706 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5708 For some functions, an the irreducible polynomial defining a
5709 field can be given as an 'unsigned int[]' with strictly
5710 decreasing elements giving the indices of those bits that are set;
5711 i.e., p[] represents the polynomial
5712 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5714 p[0] > p[1] > ... > p[k] = 0.
5715 This applies to the following functions:
5720 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5721 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5723 BN_GF2m_mod_sqrt_arr
5724 BN_GF2m_mod_solve_quad_arr
5728 Conversion can be performed by the following functions:
5733 bntest.c has additional tests for binary polynomial arithmetic.
5735 Two implementations for BN_GF2m_mod_div() are available.
5736 The default algorithm simply uses BN_GF2m_mod_inv() and
5737 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5738 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5739 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5741 [Sheueling Chang Shantz and Douglas Stebila
5742 (Sun Microsystems Laboratories)]
5744 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5745 functionality is disabled at compile-time.
5746 [Douglas Stebila <douglas.stebila@sun.com>]
5748 *) Change default behaviour of 'openssl asn1parse' so that more
5749 information is visible when viewing, e.g., a certificate:
5751 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5752 mode the content of non-printable OCTET STRINGs is output in a
5753 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5754 avoid the appearance of a printable string.
5755 [Nils Larsch <nla@trustcenter.de>]
5757 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5759 EC_GROUP_set_asn1_flag()
5760 EC_GROUP_get_asn1_flag()
5761 EC_GROUP_set_point_conversion_form()
5762 EC_GROUP_get_point_conversion_form()
5763 These control ASN1 encoding details:
5764 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5765 has been set to OPENSSL_EC_NAMED_CURVE.
5766 - Points are encoded in uncompressed form by default; options for
5767 asn1_for are as for point2oct, namely
5768 POINT_CONVERSION_COMPRESSED
5769 POINT_CONVERSION_UNCOMPRESSED
5770 POINT_CONVERSION_HYBRID
5772 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5775 EC_GROUP_get0_seed()
5776 EC_GROUP_get_seed_len()
5777 This is used only for ASN1 purposes (so far).
5778 [Nils Larsch <nla@trustcenter.de>]
5780 *) Add 'field_type' member to EC_METHOD, which holds the NID
5781 of the appropriate field type OID. The new function
5782 EC_METHOD_get_field_type() returns this value.
5783 [Nils Larsch <nla@trustcenter.de>]
5788 EC_POINT_point2hex()
5789 EC_POINT_hex2point()
5790 providing useful interfaces to EC_POINT_point2oct() and
5791 EC_POINT_oct2point().
5792 [Nils Larsch <nla@trustcenter.de>]
5794 *) Change internals of the EC library so that the functions
5795 EC_GROUP_set_generator()
5796 EC_GROUP_get_generator()
5797 EC_GROUP_get_order()
5798 EC_GROUP_get_cofactor()
5799 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5800 to methods, which would lead to unnecessary code duplication when
5801 adding different types of curves.
5802 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5804 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5805 arithmetic, and such that modified wNAFs are generated
5806 (which avoid length expansion in many cases).
5809 *) Add a function EC_GROUP_check_discriminant() (defined via
5810 EC_METHOD) that verifies that the curve discriminant is non-zero.
5812 Add a function EC_GROUP_check() that makes some sanity tests
5813 on a EC_GROUP, its generator and order. This includes
5814 EC_GROUP_check_discriminant().
5815 [Nils Larsch <nla@trustcenter.de>]
5817 *) Add ECDSA in new directory crypto/ecdsa/.
5819 Add applications 'openssl ecparam' and 'openssl ecdsa'
5820 (these are based on 'openssl dsaparam' and 'openssl dsa').
5822 ECDSA support is also included in various other files across the
5823 library. Most notably,
5824 - 'openssl req' now has a '-newkey ecdsa:file' option;
5825 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5826 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5827 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5828 them suitable for ECDSA where domain parameters must be
5829 extracted before the specific public key;
5830 - ECDSA engine support has been added.
5831 [Nils Larsch <nla@trustcenter.de>]
5833 *) Include some named elliptic curves, and add OIDs from X9.62,
5834 SECG, and WAP/WTLS. Each curve can be obtained from the new
5836 EC_GROUP_new_by_curve_name(),
5837 and the list of available named curves can be obtained with
5838 EC_get_builtin_curves().
5839 Also add a 'curve_name' member to EC_GROUP objects, which can be
5841 EC_GROUP_set_curve_name()
5842 EC_GROUP_get_curve_name()
5843 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5845 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5846 was actually never needed) and in BN_mul(). The removal in BN_mul()
5847 required a small change in bn_mul_part_recursive() and the addition
5848 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5849 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5850 bn_sub_words() and bn_add_words() except they take arrays with
5854 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5856 *) Cleanse PEM buffers before freeing them since they may contain
5858 [Benjamin Bennett <ben@psc.edu>]
5860 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5861 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5862 authentication-only ciphersuites.
5865 *) Since AES128 and AES256 share a single mask bit in the logic of
5866 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5867 kludge to work properly if AES128 is available and AES256 isn't.
5870 *) Expand security boundary to match 1.1.1 module.
5873 *) Remove redundant features: hash file source, editing of test vectors
5874 modify fipsld to use external fips_premain.c signature.
5877 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5878 run algorithm test programs.
5881 *) Make algorithm test programs more tolerant of whitespace.
5884 *) Have SSL/TLS server implementation tolerate "mismatched" record
5885 protocol version while receiving ClientHello even if the
5886 ClientHello is fragmented. (The server can't insist on the
5887 particular protocol version it has chosen before the ServerHello
5888 message has informed the client about his choice.)
5891 *) Load error codes if they are not already present instead of using a
5892 static variable. This allows them to be cleanly unloaded and reloaded.
5895 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5897 *) Introduce limits to prevent malicious keys being able to
5898 cause a denial of service. (CVE-2006-2940)
5899 [Steve Henson, Bodo Moeller]
5901 *) Fix ASN.1 parsing of certain invalid structures that can result
5902 in a denial of service. (CVE-2006-2937) [Steve Henson]
5904 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5905 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5907 *) Fix SSL client code which could crash if connecting to a
5908 malicious SSLv2 server. (CVE-2006-4343)
5909 [Tavis Ormandy and Will Drewry, Google Security Team]
5911 *) Change ciphersuite string processing so that an explicit
5912 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5913 will no longer include "AES128-SHA"), and any other similar
5914 ciphersuite (same bitmap) from *other* protocol versions (so that
5915 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5916 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5917 changes from 0.9.8b and 0.9.8d.
5920 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5922 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5923 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5925 *) Change the Unix randomness entropy gathering to use poll() when
5926 possible instead of select(), since the latter has some
5927 undesirable limitations.
5928 [Darryl Miles via Richard Levitte and Bodo Moeller]
5930 *) Disable rogue ciphersuites:
5932 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5933 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5934 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5936 The latter two were purportedly from
5937 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5940 Also deactive the remaining ciphersuites from
5941 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5942 unofficial, and the ID has long expired.
5945 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5946 dual-core machines) and other potential thread-safety issues.
5949 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5951 *) Adapt fipsld and the build system to link against the validated FIPS
5952 module in FIPS mode.
5955 *) Fixes for VC++ 2005 build under Windows.
5958 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5959 from a Windows bash shell such as MSYS. It is autodetected from the
5960 "config" script when run from a VC++ environment. Modify standard VC++
5961 build to use fipscanister.o from the GNU make build.
5964 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5966 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5967 The value now differs depending on if you build for FIPS or not.
5968 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5969 safely run with a non-FIPSed libcrypto, as it may crash because of
5970 the difference induced by this change.
5973 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5975 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5976 (part of SSL_OP_ALL). This option used to disable the
5977 countermeasure against man-in-the-middle protocol-version
5978 rollback in the SSL 2.0 server implementation, which is a bad
5979 idea. (CVE-2005-2969)
5981 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5982 for Information Security, National Institute of Advanced Industrial
5983 Science and Technology [AIST], Japan)]
5985 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5986 mainly for FIPS compliance and not fully integrated at this stage.
5989 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5990 the exponentiation using a fixed-length exponent. (Otherwise,
5991 the information leaked through timing could expose the secret key
5992 after many signatures; cf. Bleichenbacher's attack on DSA with
5996 *) Make a new fixed-window mod_exp implementation the default for
5997 RSA, DSA, and DH private-key operations so that the sequence of
5998 squares and multiplies and the memory access pattern are
5999 independent of the particular secret key. This will mitigate
6000 cache-timing and potential related attacks.
6002 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6003 and this is automatically used by BN_mod_exp_mont() if the new flag
6004 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6005 will use this BN flag for private exponents unless the flag
6006 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6007 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6009 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6011 *) Change the client implementation for SSLv23_method() and
6012 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6013 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6014 (Previously, the SSL 2.0 backwards compatible Client Hello
6015 message format would be used even with SSL_OP_NO_SSLv2.)
6018 *) Add support for smime-type MIME parameter in S/MIME messages which some
6022 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6023 a threadsafe manner. Modify rsa code to use new function and add calls
6024 to dsa and dh code (which had race conditions before).
6027 *) Include the fixed error library code in the C error file definitions
6028 instead of fixing them up at runtime. This keeps the error code
6029 structures constant.
6032 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6034 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6037 *) Fixes for newer kerberos headers. NB: the casts are needed because
6038 the 'length' field is signed on one version and unsigned on another
6039 with no (?) obvious way to tell the difference, without these VC++
6040 complains. Also the "definition" of FAR (blank) is no longer included
6041 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6042 some needed definitions.
6045 *) Undo Cygwin change.
6048 *) Added support for proxy certificates according to RFC 3820.
6049 Because they may be a security thread to unaware applications,
6050 they must be explicitly allowed in run-time. See
6051 docs/HOWTO/proxy_certificates.txt for further information.
6054 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6056 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6057 server and client random values. Previously
6058 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6059 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6061 This change has negligible security impact because:
6063 1. Server and client random values still have 24 bytes of pseudo random
6066 2. Server and client random values are sent in the clear in the initial
6069 3. The master secret is derived using the premaster secret (48 bytes in
6070 size for static RSA ciphersuites) as well as client server and random
6073 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6076 [Stephen Henson, reported by UK NISCC]
6078 *) Use Windows randomness collection on Cygwin.
6081 *) Fix hang in EGD/PRNGD query when communication socket is closed
6082 prematurely by EGD/PRNGD.
6083 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6085 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6088 *) Back-port of selected performance improvements from development
6089 branch, as well as improved support for PowerPC platforms.
6092 *) Add lots of checks for memory allocation failure, error codes to indicate
6093 failure and freeing up memory if a failure occurs.
6094 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6096 *) Add new -passin argument to dgst.
6099 *) Perform some character comparisons of different types in X509_NAME_cmp:
6100 this is needed for some certificates that re-encode DNs into UTF8Strings
6101 (in violation of RFC3280) and can't or won't issue name rollover
6105 *) Make an explicit check during certificate validation to see that
6106 the CA setting in each certificate on the chain is correct. As a
6107 side effect always do the following basic checks on extensions,
6108 not just when there's an associated purpose to the check:
6110 - if there is an unhandled critical extension (unless the user
6111 has chosen to ignore this fault)
6112 - if the path length has been exceeded (if one is set at all)
6113 - that certain extensions fit the associated purpose (if one has
6117 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6119 *) Avoid a race condition when CRLs are checked in a multi threaded
6120 environment. This would happen due to the reordering of the revoked
6121 entries during signature checking and serial number lookup. Now the
6122 encoding is cached and the serial number sort performed under a lock.
6123 Add new STACK function sk_is_sorted().
6126 *) Add Delta CRL to the extension code.
6129 *) Various fixes to s3_pkt.c so alerts are sent properly.
6130 [David Holmes <d.holmes@f5.com>]
6132 *) Reduce the chances of duplicate issuer name and serial numbers (in
6133 violation of RFC3280) using the OpenSSL certificate creation utilities.
6134 This is done by creating a random 64 bit value for the initial serial
6135 number when a serial number file is created or when a self signed
6136 certificate is created using 'openssl req -x509'. The initial serial
6137 number file is created using 'openssl x509 -next_serial' in CA.pl
6138 rather than being initialized to 1.
6141 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6143 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6144 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6145 [Joe Orton, Steve Henson]
6147 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6149 [Joe Orton, Steve Henson]
6151 *) Make it possible to have multiple active certificates with the same
6152 subject in the CA index file. This is done only if the keyword
6153 'unique_subject' is set to 'no' in the main CA section (default
6154 if 'CA_default') of the configuration file. The value is saved
6155 with the database itself in a separate index attribute file,
6156 named like the index file with '.attr' appended to the name.
6159 *) X509 verify fixes. Disable broken certificate workarounds when
6160 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6161 keyUsage extension present. Don't accept CRLs with unhandled critical
6162 extensions: since verify currently doesn't process CRL extensions this
6163 rejects a CRL with *any* critical extensions. Add new verify error codes
6167 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6168 A clarification of RFC2560 will require the use of OCTET STRINGs and
6169 some implementations cannot handle the current raw format. Since OpenSSL
6170 copies and compares OCSP nonces as opaque blobs without any attempt at
6171 parsing them this should not create any compatibility issues.
6174 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6175 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6176 this HMAC (and other) operations are several times slower than OpenSSL
6180 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6181 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6183 *) Use the correct content when signing type "other".
6186 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6188 *) Fix various bugs revealed by running the NISCC test suite:
6190 Stop out of bounds reads in the ASN1 code when presented with
6191 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6193 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6195 If verify callback ignores invalid public key errors don't try to check
6196 certificate signature with the NULL public key.
6200 *) New -ignore_err option in ocsp application to stop the server
6201 exiting on the first error in a request.
6204 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6205 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6209 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6210 extra data after the compression methods not only for TLS 1.0
6211 but also for SSL 3.0 (as required by the specification).
6212 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6214 *) Change X509_certificate_type() to mark the key as exported/exportable
6215 when it's 512 *bits* long, not 512 bytes.
6218 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6219 blocks during encryption.
6222 *) Various fixes to base64 BIO and non blocking I/O. On write
6223 flushes were not handled properly if the BIO retried. On read
6224 data was not being buffered properly and had various logic bugs.
6225 This also affects blocking I/O when the data being decoded is a
6229 *) Various S/MIME bugfixes and compatibility changes:
6230 output correct application/pkcs7 MIME type if
6231 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6232 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6233 of files as .eml work). Correctly handle very long lines in MIME
6237 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6239 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6240 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6241 a protocol version number mismatch like a decryption error
6242 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6245 *) Turn on RSA blinding by default in the default implementation
6246 to avoid a timing attack. Applications that don't want it can call
6247 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6248 They would be ill-advised to do so in most cases.
6249 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6251 *) Change RSA blinding code so that it works when the PRNG is not
6252 seeded (in this case, the secret RSA exponent is abused as
6253 an unpredictable seed -- if it is not unpredictable, there
6254 is no point in blinding anyway). Make RSA blinding thread-safe
6255 by remembering the creator's thread ID in rsa->blinding and
6256 having all other threads use local one-time blinding factors
6257 (this requires more computation than sharing rsa->blinding, but
6258 avoids excessive locking; and if an RSA object is not shared
6259 between threads, blinding will still be very fast).
6262 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6263 ENGINE as defaults for all supported algorithms irrespective of
6264 the 'flags' parameter. 'flags' is now honoured, so applications
6265 should make sure they are passing it correctly.
6268 *) Target "mingw" now allows native Windows code to be generated in
6269 the Cygwin environment as well as with the MinGW compiler.
6272 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6274 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6275 via timing by performing a MAC computation even if incorrect
6276 block cipher padding has been found. This is a countermeasure
6277 against active attacks where the attacker has to distinguish
6278 between bad padding and a MAC verification error. (CVE-2003-0078)
6280 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6281 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6282 Martin Vuagnoux (EPFL, Ilion)]
6284 *) Make the no-err option work as intended. The intention with no-err
6285 is not to have the whole error stack handling routines removed from
6286 libcrypto, it's only intended to remove all the function name and
6287 reason texts, thereby removing some of the footprint that may not
6288 be interesting if those errors aren't displayed anyway.
6290 NOTE: it's still possible for any application or module to have it's
6291 own set of error texts inserted. The routines are there, just not
6292 used by default when no-err is given.
6295 *) Add support for FreeBSD on IA64.
6296 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6298 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6299 Kerberos function mit_des_cbc_cksum(). Before this change,
6300 the value returned by DES_cbc_cksum() was like the one from
6301 mit_des_cbc_cksum(), except the bytes were swapped.
6302 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6304 *) Allow an application to disable the automatic SSL chain building.
6305 Before this a rather primitive chain build was always performed in
6306 ssl3_output_cert_chain(): an application had no way to send the
6307 correct chain if the automatic operation produced an incorrect result.
6309 Now the chain builder is disabled if either:
6311 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6313 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6315 The reasoning behind this is that an application would not want the
6316 auto chain building to take place if extra chain certificates are
6317 present and it might also want a means of sending no additional
6318 certificates (for example the chain has two certificates and the
6322 *) Add the possibility to build without the ENGINE framework.
6323 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6325 *) Under Win32 gmtime() can return NULL: check return value in
6326 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6329 *) DSA routines: under certain error conditions uninitialized BN objects
6330 could be freed. Solution: make sure initialization is performed early
6331 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6332 Nils Larsch <nla@trustcenter.de> via PR#459)
6335 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6336 checked on reconnect on the client side, therefore session resumption
6337 could still fail with a "ssl session id is different" error. This
6338 behaviour is masked when SSL_OP_ALL is used due to
6339 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6340 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6341 followup to PR #377.
6344 *) IA-32 assembler support enhancements: unified ELF targets, support
6345 for SCO/Caldera platforms, fix for Cygwin shared build.
6348 *) Add support for FreeBSD on sparc64. As a consequence, support for
6349 FreeBSD on non-x86 processors is separate from x86 processors on
6350 the config script, much like the NetBSD support.
6351 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6353 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6355 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6358 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6359 code (06) was taken as the first octet of the session ID and the last
6360 octet was ignored consequently. As a result SSLv2 client side session
6361 caching could not have worked due to the session ID mismatch between
6363 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6367 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6368 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6372 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6373 seems that in spite of existing for more than a year, many application
6374 author have done nothing to provide the necessary callbacks, which
6375 means that this particular engine will not work properly anywhere.
6376 This is a very unfortunate situation which forces us, in the name
6377 of usability, to give the hw_ncipher.c a static lock, which is part
6379 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6380 appear in 0.9.8 or later. We EXPECT application authors to have
6381 dealt properly with this when 0.9.8 is released (unless we actually
6382 make such changes in the libcrypto locking code that changes will
6383 have to be made anyway).
6386 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6387 octets have been read, EOF or an error occurs. Without this change
6388 some truncated ASN1 structures will not produce an error.
6391 *) Disable Heimdal support, since it hasn't been fully implemented.
6392 Still give the possibility to force the use of Heimdal, but with
6393 warnings and a request that patches get sent to openssl-dev.
6396 *) Add the VC-CE target, introduce the WINCE sysname, and add
6397 INSTALL.WCE and appropriate conditionals to make it build.
6398 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6400 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6401 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6402 edit numbers of the version.
6403 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6405 *) Introduce safe string copy and catenation functions
6406 (BUF_strlcpy() and BUF_strlcat()).
6407 [Ben Laurie (CHATS) and Richard Levitte]
6409 *) Avoid using fixed-size buffers for one-line DNs.
6410 [Ben Laurie (CHATS)]
6412 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6413 resizing buffers containing secrets, and use where appropriate.
6414 [Ben Laurie (CHATS)]
6416 *) Avoid using fixed size buffers for configuration file location.
6417 [Ben Laurie (CHATS)]
6419 *) Avoid filename truncation for various CA files.
6420 [Ben Laurie (CHATS)]
6422 *) Use sizeof in preference to magic numbers.
6423 [Ben Laurie (CHATS)]
6425 *) Avoid filename truncation in cert requests.
6426 [Ben Laurie (CHATS)]
6428 *) Add assertions to check for (supposedly impossible) buffer
6430 [Ben Laurie (CHATS)]
6432 *) Don't cache truncated DNS entries in the local cache (this could
6433 potentially lead to a spoofing attack).
6434 [Ben Laurie (CHATS)]
6436 *) Fix various buffers to be large enough for hex/decimal
6437 representations in a platform independent manner.
6438 [Ben Laurie (CHATS)]
6440 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6441 resizing buffers containing secrets, and use where appropriate.
6442 [Ben Laurie (CHATS)]
6444 *) Add BIO_indent() to avoid much slightly worrying code to do
6446 [Ben Laurie (CHATS)]
6448 *) Convert sprintf()/BIO_puts() to BIO_printf().
6449 [Ben Laurie (CHATS)]
6451 *) buffer_gets() could terminate with the buffer only half
6453 [Ben Laurie (CHATS)]
6455 *) Add assertions to prevent user-supplied crypto functions from
6456 overflowing internal buffers by having large block sizes, etc.
6457 [Ben Laurie (CHATS)]
6459 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6461 [Ben Laurie (CHATS)]
6463 *) Eliminate unused copy of key in RC4.
6464 [Ben Laurie (CHATS)]
6466 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6467 [Ben Laurie (CHATS)]
6469 *) Fix off-by-one error in EGD path.
6470 [Ben Laurie (CHATS)]
6472 *) If RANDFILE path is too long, ignore instead of truncating.
6473 [Ben Laurie (CHATS)]
6475 *) Eliminate unused and incorrectly sized X.509 structure
6477 [Ben Laurie (CHATS)]
6479 *) Eliminate unused and dangerous function knumber().
6480 [Ben Laurie (CHATS)]
6482 *) Eliminate unused and dangerous structure, KSSL_ERR.
6483 [Ben Laurie (CHATS)]
6485 *) Protect against overlong session ID context length in an encoded
6486 session object. Since these are local, this does not appear to be
6488 [Ben Laurie (CHATS)]
6490 *) Change from security patch (see 0.9.6e below) that did not affect
6491 the 0.9.6 release series:
6493 Remote buffer overflow in SSL3 protocol - an attacker could
6494 supply an oversized master key in Kerberos-enabled versions.
6496 [Ben Laurie (CHATS)]
6498 *) Change the SSL kerb5 codes to match RFC 2712.
6501 *) Make -nameopt work fully for req and add -reqopt switch.
6502 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6504 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6505 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6507 *) Make sure tests can be performed even if the corresponding algorithms
6508 have been removed entirely. This was also the last step to make
6509 OpenSSL compilable with DJGPP under all reasonable conditions.
6510 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6512 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6513 to allow version independent disabling of normally unselected ciphers,
6514 which may be activated as a side-effect of selecting a single cipher.
6516 (E.g., cipher list string "RSA" enables ciphersuites that are left
6517 out of "ALL" because they do not provide symmetric encryption.
6518 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6519 [Lutz Jaenicke, Bodo Moeller]
6521 *) Add appropriate support for separate platform-dependent build
6522 directories. The recommended way to make a platform-dependent
6523 build directory is the following (tested on Linux), maybe with
6526 # Place yourself outside of the OpenSSL source tree. In
6527 # this example, the environment variable OPENSSL_SOURCE
6528 # is assumed to contain the absolute OpenSSL source directory.
6529 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6530 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6531 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6532 mkdir -p `dirname $F`
6533 ln -s $OPENSSL_SOURCE/$F $F
6536 To be absolutely sure not to disturb the source tree, a "make clean"
6537 is a good thing. If it isn't successful, don't worry about it,
6538 it probably means the source directory is very clean.
6541 *) Make sure any ENGINE control commands make local copies of string
6542 pointers passed to them whenever necessary. Otherwise it is possible
6543 the caller may have overwritten (or deallocated) the original string
6544 data when a later ENGINE operation tries to use the stored values.
6545 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6547 *) Improve diagnostics in file reading and command-line digests.
6548 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6550 *) Add AES modes CFB and OFB to the object database. Correct an
6551 error in AES-CFB decryption.
6554 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6555 allows existing EVP_CIPHER_CTX structures to be reused after
6556 calling EVP_*Final(). This behaviour is used by encryption
6557 BIOs and some applications. This has the side effect that
6558 applications must explicitly clean up cipher contexts with
6559 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6562 *) Check the values of dna and dnb in bn_mul_recursive before calling
6563 bn_mul_comba (a non zero value means the a or b arrays do not contain
6564 n2 elements) and fallback to bn_mul_normal if either is not zero.
6567 *) Fix escaping of non-ASCII characters when using the -subj option
6568 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6571 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6572 form for "surname", serialNumber has no short form.
6573 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6574 therefore remove "mail" short name for "internet 7".
6575 The OID for unique identifiers in X509 certificates is
6576 x500UniqueIdentifier, not uniqueIdentifier.
6577 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6580 *) Add an "init" command to the ENGINE config module and auto initialize
6581 ENGINEs. Without any "init" command the ENGINE will be initialized
6582 after all ctrl commands have been executed on it. If init=1 the
6583 ENGINE is initailized at that point (ctrls before that point are run
6584 on the uninitialized ENGINE and after on the initialized one). If
6585 init=0 then the ENGINE will not be iniatialized at all.
6588 *) Fix the 'app_verify_callback' interface so that the user-defined
6589 argument is actually passed to the callback: In the
6590 SSL_CTX_set_cert_verify_callback() prototype, the callback
6591 declaration has been changed from
6594 int (*cb)(X509_STORE_CTX *,void *);
6595 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6596 i=s->ctx->app_verify_callback(&ctx)
6597 has been changed into
6598 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6600 To update applications using SSL_CTX_set_cert_verify_callback(),
6601 a dummy argument can be added to their callback functions.
6602 [D. K. Smetters <smetters@parc.xerox.com>]
6604 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6605 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6607 *) Add and OPENSSL_LOAD_CONF define which will cause
6608 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6609 This allows older applications to transparently support certain
6610 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6611 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6612 load the config file and OPENSSL_add_all_algorithms_conf() which will
6613 always load it have also been added.
6616 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6617 Adjust NIDs and EVP layer.
6618 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6620 *) Config modules support in openssl utility.
6622 Most commands now load modules from the config file,
6623 though in a few (such as version) this isn't done
6624 because it couldn't be used for anything.
6626 In the case of ca and req the config file used is
6627 the same as the utility itself: that is the -config
6628 command line option can be used to specify an
6632 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6633 use "openssl_conf" if filename is NULL use default openssl config file.
6636 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6637 config section name. Add a new flag to tolerate a missing config file
6638 and move code to CONF_modules_load_file().
6641 *) Support for crypto accelerator cards from Accelerated Encryption
6642 Processing, www.aep.ie. (Use engine 'aep')
6643 The support was copied from 0.9.6c [engine] and adapted/corrected
6644 to work with the new engine framework.
6645 [AEP Inc. and Richard Levitte]
6647 *) Support for SureWare crypto accelerator cards from Baltimore
6648 Technologies. (Use engine 'sureware')
6649 The support was copied from 0.9.6c [engine] and adapted
6650 to work with the new engine framework.
6653 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6654 make the newer ENGINE framework commands for the CHIL engine work.
6655 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6657 *) Make it possible to produce shared libraries on ReliantUNIX.
6658 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6660 *) Add the configuration target debug-linux-ppro.
6661 Make 'openssl rsa' use the general key loading routines
6662 implemented in apps.c, and make those routines able to
6663 handle the key format FORMAT_NETSCAPE and the variant
6665 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6667 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6668 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6670 *) Add -keyform to rsautl, and document -engine.
6671 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6673 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6674 BIO_R_NO_SUCH_FILE error code rather than the generic
6675 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6678 *) Add new functions
6680 ERR_peek_last_error_line
6681 ERR_peek_last_error_line_data.
6682 These are similar to
6685 ERR_peek_error_line_data,
6686 but report on the latest error recorded rather than the first one
6687 still in the error queue.
6688 [Ben Laurie, Bodo Moeller]
6690 *) default_algorithms option in ENGINE config module. This allows things
6692 default_algorithms = ALL
6693 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6696 *) Preliminary ENGINE config module.
6699 *) New experimental application configuration code.
6702 *) Change the AES code to follow the same name structure as all other
6703 symmetric ciphers, and behave the same way. Move everything to
6704 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6705 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6707 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6708 [Ben Laurie and Theo de Raadt]
6710 *) Add option to output public keys in req command.
6711 [Massimiliano Pala madwolf@openca.org]
6713 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6714 (up to about 10% better than before for P-192 and P-224).
6717 *) New functions/macros
6719 SSL_CTX_set_msg_callback(ctx, cb)
6720 SSL_CTX_set_msg_callback_arg(ctx, arg)
6721 SSL_set_msg_callback(ssl, cb)
6722 SSL_set_msg_callback_arg(ssl, arg)
6724 to request calling a callback function
6726 void cb(int write_p, int version, int content_type,
6727 const void *buf, size_t len, SSL *ssl, void *arg)
6729 whenever a protocol message has been completely received
6730 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6731 protocol version according to which the SSL library interprets
6732 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6733 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6734 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6735 specification (change_cipher_spec(20), alert(21), handshake(22)).
6736 'buf' and 'len' point to the actual message, 'ssl' to the
6737 SSL object, and 'arg' is the application-defined value set by
6738 SSL[_CTX]_set_msg_callback_arg().
6740 'openssl s_client' and 'openssl s_server' have new '-msg' options
6741 to enable a callback that displays all protocol messages.
6744 *) Change the shared library support so shared libraries are built as
6745 soon as the corresponding static library is finished, and thereby get
6746 openssl and the test programs linked against the shared library.
6747 This still only happens when the keyword "shard" has been given to
6748 the configuration scripts.
6750 NOTE: shared library support is still an experimental thing, and
6751 backward binary compatibility is still not guaranteed.
6752 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6754 *) Add support for Subject Information Access extension.
6755 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6757 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6758 additional bytes when new memory had to be allocated, not just
6759 when reusing an existing buffer.
6762 *) New command line and configuration option 'utf8' for the req command.
6763 This allows field values to be specified as UTF8 strings.
6766 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6767 runs for the former and machine-readable output for the latter.
6770 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6771 of the e-mail address in the DN (i.e., it will go into a certificate
6772 extension only). The new configuration file option 'email_in_dn = no'
6773 has the same effect.
6774 [Massimiliano Pala madwolf@openca.org]
6776 *) Change all functions with names starting with des_ to be starting
6777 with DES_ instead. Add wrappers that are compatible with libdes,
6778 but are named _ossl_old_des_*. Finally, add macros that map the
6779 des_* symbols to the corresponding _ossl_old_des_* if libdes
6780 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6781 desired, the des_* symbols will be mapped to DES_*, with one
6784 Since we provide two compatibility mappings, the user needs to
6785 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6786 compatibility is desired. The default (i.e., when that macro
6787 isn't defined) is OpenSSL 0.9.6c compatibility.
6789 There are also macros that enable and disable the support of old
6790 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6791 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6792 are defined, the default will apply: to support the old des routines.
6794 In either case, one must include openssl/des.h to get the correct
6795 definitions. Do not try to just include openssl/des_old.h, that
6798 NOTE: This is a major break of an old API into a new one. Software
6799 authors are encouraged to switch to the DES_ style functions. Some
6800 time in the future, des_old.h and the libdes compatibility functions
6801 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6802 default), and then completely removed.
6805 *) Test for certificates which contain unsupported critical extensions.
6806 If such a certificate is found during a verify operation it is
6807 rejected by default: this behaviour can be overridden by either
6808 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6809 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6810 X509_supported_extension() has also been added which returns 1 if a
6811 particular extension is supported.
6814 *) Modify the behaviour of EVP cipher functions in similar way to digests
6815 to retain compatibility with existing code.
6818 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6819 compatibility with existing code. In particular the 'ctx' parameter does
6820 not have to be to be initialized before the call to EVP_DigestInit() and
6821 it is tidied up after a call to EVP_DigestFinal(). New function
6822 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6823 EVP_MD_CTX_copy() changed to not require the destination to be
6824 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6825 requires the destination to be valid.
6827 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6828 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6831 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6832 so that complete 'Handshake' protocol structures are kept in memory
6833 instead of overwriting 'msg_type' and 'length' with 'body' data.
6836 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6837 [Massimo Santin via Richard Levitte]
6839 *) Major restructuring to the underlying ENGINE code. This includes
6840 reduction of linker bloat, separation of pure "ENGINE" manipulation
6841 (initialisation, etc) from functionality dealing with implementations
6842 of specific crypto iterfaces. This change also introduces integrated
6843 support for symmetric ciphers and digest implementations - so ENGINEs
6844 can now accelerate these by providing EVP_CIPHER and EVP_MD
6845 implementations of their own. This is detailed in crypto/engine/README
6846 as it couldn't be adequately described here. However, there are a few
6847 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6848 were changed in the original introduction of ENGINE code have now
6849 reverted back - the hooking from this code to ENGINE is now a good
6850 deal more passive and at run-time, operations deal directly with
6851 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6852 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6853 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6854 they were not being used by the framework as there is no concept of a
6855 BIGNUM_METHOD and they could not be generalised to the new
6856 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6857 ENGINE_cpy() has been removed as it cannot be consistently defined in
6861 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6864 *) Change mkdef.pl to sort symbols that get the same entry number,
6865 and make sure the automatically generated functions ERR_load_*
6866 become part of libeay.num as well.
6869 *) New function SSL_renegotiate_pending(). This returns true once
6870 renegotiation has been requested (either SSL_renegotiate() call
6871 or HelloRequest/ClientHello received from the peer) and becomes
6872 false once a handshake has been completed.
6873 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6874 sends a HelloRequest, but does not ensure that a handshake takes
6875 place. SSL_renegotiate_pending() is useful for checking if the
6876 client has followed the request.)
6879 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6880 By default, clients may request session resumption even during
6881 renegotiation (if session ID contexts permit); with this option,
6882 session resumption is possible only in the first handshake.
6884 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6885 more bits available for options that should not be part of
6886 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6889 *) Add some demos for certificate and certificate request creation.
6892 *) Make maximum certificate chain size accepted from the peer application
6893 settable (SSL*_get/set_max_cert_list()), as proposed by
6894 "Douglas E. Engert" <deengert@anl.gov>.
6897 *) Add support for shared libraries for Unixware-7
6898 (Boyd Lynn Gerber <gerberb@zenez.com>).
6901 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6902 be done prior to destruction. Use this to unload error strings from
6903 ENGINEs that load their own error strings. NB: This adds two new API
6904 functions to "get" and "set" this destroy handler in an ENGINE.
6907 *) Alter all existing ENGINE implementations (except "openssl" and
6908 "openbsd") to dynamically instantiate their own error strings. This
6909 makes them more flexible to be built both as statically-linked ENGINEs
6910 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6911 Also, add stub code to each that makes building them as self-contained
6912 shared-libraries easier (see README.ENGINE).
6915 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6916 implementations into applications that are completely implemented in
6917 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6918 commands that can be used to configure what shared-library to load and
6919 to control aspects of the way it is handled. Also, made an update to
6920 the README.ENGINE file that brings its information up-to-date and
6921 provides some information and instructions on the "dynamic" ENGINE
6922 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6925 *) Make it possible to unload ranges of ERR strings with a new
6926 "ERR_unload_strings" function.
6929 *) Add a copy() function to EVP_MD.
6932 *) Make EVP_MD routines take a context pointer instead of just the
6933 md_data void pointer.
6936 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6937 that the digest can only process a single chunk of data
6938 (typically because it is provided by a piece of
6939 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6940 is only going to provide a single chunk of data, and hence the
6941 framework needn't accumulate the data for oneshot drivers.
6944 *) As with "ERR", make it possible to replace the underlying "ex_data"
6945 functions. This change also alters the storage and management of global
6946 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6947 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6948 index counters. The API functions that use this state have been changed
6949 to take a "class_index" rather than pointers to the class's local STACK
6950 and counter, and there is now an API function to dynamically create new
6951 classes. This centralisation allows us to (a) plug a lot of the
6952 thread-safety problems that existed, and (b) makes it possible to clean
6953 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6954 such data would previously have always leaked in application code and
6955 workarounds were in place to make the memory debugging turn a blind eye
6956 to it. Application code that doesn't use this new function will still
6957 leak as before, but their memory debugging output will announce it now
6958 rather than letting it slide.
6960 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6961 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6962 has a return value to indicate success or failure.
6965 *) Make it possible to replace the underlying "ERR" functions such that the
6966 global state (2 LHASH tables and 2 locks) is only used by the "default"
6967 implementation. This change also adds two functions to "get" and "set"
6968 the implementation prior to it being automatically set the first time
6969 any other ERR function takes place. Ie. an application can call "get",
6970 pass the return value to a module it has just loaded, and that module
6971 can call its own "set" function using that value. This means the
6972 module's "ERR" operations will use (and modify) the error state in the
6973 application and not in its own statically linked copy of OpenSSL code.
6976 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6977 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6978 the operation, and provides a more encapsulated way for external code
6979 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6980 to use these functions rather than manually incrementing the counts.
6982 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6985 *) Add EVP test program.
6988 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6991 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6992 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6993 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6994 These allow a CRL to be built without having to access X509_CRL fields
6995 directly. Modify 'ca' application to use new functions.
6998 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6999 bug workarounds. Rollback attack detection is a security feature.
7000 The problem will only arise on OpenSSL servers when TLSv1 is not
7001 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7002 Software authors not wanting to support TLSv1 will have special reasons
7003 for their choice and can explicitly enable this option.
7004 [Bodo Moeller, Lutz Jaenicke]
7006 *) Rationalise EVP so it can be extended: don't include a union of
7007 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7008 (similar to those existing for EVP_CIPHER_CTX).
7013 EVP_MD_CTX_init(&md); /* new function call */
7014 EVP_DigestInit(&md, EVP_sha1());
7015 EVP_DigestUpdate(&md, in, len);
7016 EVP_DigestFinal(&md, out, NULL);
7017 EVP_MD_CTX_cleanup(&md); /* new function call */
7021 *) Make DES key schedule conform to the usual scheme, as well as
7022 correcting its structure. This means that calls to DES functions
7023 now have to pass a pointer to a des_key_schedule instead of a
7024 plain des_key_schedule (which was actually always a pointer
7027 des_key_schedule ks;
7029 des_set_key_checked(..., &ks);
7030 des_ncbc_encrypt(..., &ks, ...);
7032 (Note that a later change renames 'des_...' into 'DES_...'.)
7035 *) Initial reduction of linker bloat: the use of some functions, such as
7036 PEM causes large amounts of unused functions to be linked in due to
7037 poor organisation. For example pem_all.c contains every PEM function
7038 which has a knock on effect of linking in large amounts of (unused)
7039 ASN1 code. Grouping together similar functions and splitting unrelated
7040 functions prevents this.
7043 *) Cleanup of EVP macros.
7046 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7047 correct _ecb suffix.
7050 *) Add initial OCSP responder support to ocsp application. The
7051 revocation information is handled using the text based index
7052 use by the ca application. The responder can either handle
7053 requests generated internally, supplied in files (for example
7054 via a CGI script) or using an internal minimal server.
7057 *) Add configuration choices to get zlib compression for TLS.
7060 *) Changes to Kerberos SSL for RFC 2712 compliance:
7061 1. Implemented real KerberosWrapper, instead of just using
7062 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7063 2. Implemented optional authenticator field of KerberosWrapper.
7065 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7066 and authenticator structs; see crypto/krb5/.
7068 Generalized Kerberos calls to support multiple Kerberos libraries.
7069 [Vern Staats <staatsvr@asc.hpc.mil>,
7070 Jeffrey Altman <jaltman@columbia.edu>
7071 via Richard Levitte]
7073 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7074 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7075 values for each of the key sizes rather than having just
7076 parameters (and 'speed' generating keys each time).
7079 *) Speed up EVP routines.
7082 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7083 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7084 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7085 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7087 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7088 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7089 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7092 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7094 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7097 *) Added the OS2-EMX target.
7098 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7100 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7101 to support NCONF routines in extension code. New function CONF_set_nconf()
7102 to allow functions which take an NCONF to also handle the old LHASH
7103 structure: this means that the old CONF compatible routines can be
7104 retained (in particular wrt extensions) without having to duplicate the
7105 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7108 *) Enhance the general user interface with mechanisms for inner control
7109 and with possibilities to have yes/no kind of prompts.
7112 *) Change all calls to low level digest routines in the library and
7113 applications to use EVP. Add missing calls to HMAC_cleanup() and
7114 don't assume HMAC_CTX can be copied using memcpy().
7115 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7117 *) Add the possibility to control engines through control names but with
7118 arbitrary arguments instead of just a string.
7119 Change the key loaders to take a UI_METHOD instead of a callback
7120 function pointer. NOTE: this breaks binary compatibility with earlier
7121 versions of OpenSSL [engine].
7122 Adapt the nCipher code for these new conditions and add a card insertion
7126 *) Enhance the general user interface with mechanisms to better support
7127 dialog box interfaces, application-defined prompts, the possibility
7128 to use defaults (for example default passwords from somewhere else)
7129 and interrupts/cancellations.
7132 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7133 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7136 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7137 tidy up some unnecessarily weird code in 'sk_new()').
7138 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7140 *) Change the key loading routines for ENGINEs to use the same kind
7141 callback (pem_password_cb) as all other routines that need this
7145 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7146 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7147 than this minimum value is recommended.
7150 *) New random seeder for OpenVMS, using the system process statistics
7151 that are easily reachable.
7154 *) Windows apparently can't transparently handle global
7155 variables defined in DLLs. Initialisations such as:
7157 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7159 won't compile. This is used by the any applications that need to
7160 declare their own ASN1 modules. This was fixed by adding the option
7161 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7162 needed for static libraries under Win32.
7165 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7166 setting of purpose and trust fields. New X509_STORE trust and
7167 purpose functions and tidy up setting in other SSL functions.
7170 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7171 structure. These are inherited by X509_STORE_CTX when it is
7172 initialised. This allows various defaults to be set in the
7173 X509_STORE structure (such as flags for CRL checking and custom
7174 purpose or trust settings) for functions which only use X509_STORE_CTX
7175 internally such as S/MIME.
7177 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7178 trust settings if they are not set in X509_STORE. This allows X509_STORE
7179 purposes and trust (in S/MIME for example) to override any set by default.
7181 Add command line options for CRL checking to smime, s_client and s_server
7185 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7186 are set then the CRL is looked up in the X509_STORE structure and
7187 its validity and signature checked, then if the certificate is found
7188 in the CRL the verify fails with a revoked error.
7190 Various new CRL related callbacks added to X509_STORE_CTX structure.
7192 Command line options added to 'verify' application to support this.
7194 This needs some additional work, such as being able to handle multiple
7195 CRLs with different times, extension based lookup (rather than just
7196 by subject name) and ultimately more complete V2 CRL extension
7200 *) Add a general user interface API (crypto/ui/). This is designed
7201 to replace things like des_read_password and friends (backward
7202 compatibility functions using this new API are provided).
7203 The purpose is to remove prompting functions from the DES code
7204 section as well as provide for prompting through dialog boxes in
7205 a window system and the like.
7208 *) Add "ex_data" support to ENGINE so implementations can add state at a
7209 per-structure level rather than having to store it globally.
7212 *) Make it possible for ENGINE structures to be copied when retrieved by
7213 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7214 This causes the "original" ENGINE structure to act like a template,
7215 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7216 operational state can be localised to each ENGINE structure, despite the
7217 fact they all share the same "methods". New ENGINE structures returned in
7218 this case have no functional references and the return value is the single
7219 structural reference. This matches the single structural reference returned
7220 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7224 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7225 needs to match any other type at all we need to manually clear the
7229 *) Changes to the "openssl engine" utility to include;
7230 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7231 about an ENGINE's available control commands.
7232 - executing control commands from command line arguments using the
7233 '-pre' and '-post' switches. '-post' is only used if '-t' is
7234 specified and the ENGINE is successfully initialised. The syntax for
7235 the individual commands are colon-separated, for example;
7236 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7239 *) New dynamic control command support for ENGINEs. ENGINEs can now
7240 declare their own commands (numbers), names (strings), descriptions,
7241 and input types for run-time discovery by calling applications. A
7242 subset of these commands are implicitly classed as "executable"
7243 depending on their input type, and only these can be invoked through
7244 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7245 can be based on user input, config files, etc). The distinction is
7246 that "executable" commands cannot return anything other than a boolean
7247 result and can only support numeric or string input, whereas some
7248 discoverable commands may only be for direct use through
7249 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7250 pointers, or other custom uses. The "executable" commands are to
7251 support parameterisations of ENGINE behaviour that can be
7252 unambiguously defined by ENGINEs and used consistently across any
7253 OpenSSL-based application. Commands have been added to all the
7254 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7255 control over shared-library paths without source code alterations.
7258 *) Changed all ENGINE implementations to dynamically allocate their
7259 ENGINEs rather than declaring them statically. Apart from this being
7260 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7261 this also allows the implementations to compile without using the
7262 internal engine_int.h header.
7265 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7266 'const' value. Any code that should be able to modify a RAND_METHOD
7267 should already have non-const pointers to it (ie. they should only
7268 modify their own ones).
7271 *) Made a variety of little tweaks to the ENGINE code.
7272 - "atalla" and "ubsec" string definitions were moved from header files
7273 to C code. "nuron" string definitions were placed in variables
7274 rather than hard-coded - allowing parameterisation of these values
7275 later on via ctrl() commands.
7276 - Removed unused "#if 0"'d code.
7277 - Fixed engine list iteration code so it uses ENGINE_free() to release
7278 structural references.
7279 - Constified the RAND_METHOD element of ENGINE structures.
7280 - Constified various get/set functions as appropriate and added
7281 missing functions (including a catch-all ENGINE_cpy that duplicates
7282 all ENGINE values onto a new ENGINE except reference counts/state).
7283 - Removed NULL parameter checks in get/set functions. Setting a method
7284 or function to NULL is a way of cancelling out a previously set
7285 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7286 and doesn't justify the extra error symbols and code.
7287 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7288 flags from engine_int.h to engine.h.
7289 - Changed prototypes for ENGINE handler functions (init(), finish(),
7290 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7293 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7294 to the algorithm using long division. The binary algorithm can be
7295 used only if the modulus is odd. On 32-bit systems, it is faster
7296 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7297 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7298 up to 450 bits. In 64-bit environments, the binary algorithm
7299 appears to be advantageous for much longer moduli; here we use it
7300 for moduli up to 2048 bits.
7303 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7304 could not support the combine flag in choice fields.
7307 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7308 extensions from a certificate request to the certificate.
7311 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7312 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7313 file: this allows the display of the certificate about to be
7314 signed to be customised, to allow certain fields to be included
7315 or excluded and extension details. The old system didn't display
7316 multicharacter strings properly, omitted fields not in the policy
7317 and couldn't display additional details such as extensions.
7320 *) Function EC_POINTs_mul for multiple scalar multiplication
7321 of an arbitrary number of elliptic curve points
7322 \sum scalars[i]*points[i],
7323 optionally including the generator defined for the EC_GROUP:
7324 scalar*generator + \sum scalars[i]*points[i].
7326 EC_POINT_mul is a simple wrapper function for the typical case
7327 that the point list has just one item (besides the optional
7331 *) First EC_METHODs for curves over GF(p):
7333 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7334 operations and provides various method functions that can also
7335 operate with faster implementations of modular arithmetic.
7337 EC_GFp_mont_method() reuses most functions that are part of
7338 EC_GFp_simple_method, but uses Montgomery arithmetic.
7340 [Bodo Moeller; point addition and point doubling
7341 implementation directly derived from source code provided by
7342 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7344 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7345 crypto/ec/ec_lib.c):
7347 Curves are EC_GROUP objects (with an optional group generator)
7348 based on EC_METHODs that are built into the library.
7350 Points are EC_POINT objects based on EC_GROUP objects.
7352 Most of the framework would be able to handle curves over arbitrary
7353 finite fields, but as there are no obvious types for fields other
7354 than GF(p), some functions are limited to that for now.
7357 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7358 that the file contains a complete HTTP response.
7361 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7362 change the def and num file printf format specifier from "%-40sXXX"
7363 to "%-39s XXX". The latter will always guarantee a space after the
7364 field while the former will cause them to run together if the field
7365 is 40 of more characters long.
7368 *) Constify the cipher and digest 'method' functions and structures
7369 and modify related functions to take constant EVP_MD and EVP_CIPHER
7373 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7374 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7377 *) Modify EVP_Digest*() routines so they now return values. Although the
7378 internal software routines can never fail additional hardware versions
7382 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7384 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7385 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7388 ERR_R_NESTED_ASN1_ERROR
7390 ERR_R_MISSING_ASN1_EOS
7391 were 4 .. 9, conflicting with
7392 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7394 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7395 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7397 Add new error code 'ERR_R_INTERNAL_ERROR'.
7400 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7404 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7405 sets the subject name for a new request or supersedes the
7406 subject name in a given request. Formats that can be parsed are
7407 'CN=Some Name, OU=myOU, C=IT'
7409 'CN=Some Name/OU=myOU/C=IT'.
7411 Add options '-batch' and '-verbose' to 'openssl req'.
7412 [Massimiliano Pala <madwolf@hackmasters.net>]
7414 *) Introduce the possibility to access global variables through
7415 functions on platform were that's the best way to handle exporting
7416 global variables in shared libraries. To enable this functionality,
7417 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7418 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7419 is normally done by Configure or something similar).
7421 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7422 in the source file (foo.c) like this:
7424 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7425 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7427 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7428 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7430 OPENSSL_DECLARE_GLOBAL(int,foo);
7431 #define foo OPENSSL_GLOBAL_REF(foo)
7432 OPENSSL_DECLARE_GLOBAL(double,bar);
7433 #define bar OPENSSL_GLOBAL_REF(bar)
7435 The #defines are very important, and therefore so is including the
7436 header file everywhere where the defined globals are used.
7438 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7439 of ASN.1 items, but that structure is a bit different.
7441 The largest change is in util/mkdef.pl which has been enhanced with
7442 better and easier to understand logic to choose which symbols should
7443 go into the Windows .def files as well as a number of fixes and code
7444 cleanup (among others, algorithm keywords are now sorted
7445 lexicographically to avoid constant rewrites).
7448 *) In BN_div() keep a copy of the sign of 'num' before writing the
7449 result to 'rm' because if rm==num the value will be overwritten
7450 and produce the wrong result if 'num' is negative: this caused
7451 problems with BN_mod() and BN_nnmod().
7454 *) Function OCSP_request_verify(). This checks the signature on an
7455 OCSP request and verifies the signer certificate. The signer
7456 certificate is just checked for a generic purpose and OCSP request
7460 *) Add OCSP_check_validity() function to check the validity of OCSP
7461 responses. OCSP responses are prepared in real time and may only
7462 be a few seconds old. Simply checking that the current time lies
7463 between thisUpdate and nextUpdate max reject otherwise valid responses
7464 caused by either OCSP responder or client clock inaccuracy. Instead
7465 we allow thisUpdate and nextUpdate to fall within a certain period of
7466 the current time. The age of the response can also optionally be
7467 checked. Two new options -validity_period and -status_age added to
7471 *) If signature or public key algorithm is unrecognized print out its
7472 OID rather that just UNKNOWN.
7475 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7476 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7477 ID to be generated from the issuer certificate alone which can then be
7478 passed to OCSP_id_issuer_cmp().
7481 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7482 ASN1 modules to export functions returning ASN1_ITEM pointers
7483 instead of the ASN1_ITEM structures themselves. This adds several
7484 new macros which allow the underlying ASN1 function/structure to
7485 be accessed transparently. As a result code should not use ASN1_ITEM
7486 references directly (such as &X509_it) but instead use the relevant
7487 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7488 use of the new ASN1 code on platforms where exporting structures
7489 is problematical (for example in shared libraries) but exporting
7490 functions returning pointers to structures is not.
7493 *) Add support for overriding the generation of SSL/TLS session IDs.
7494 These callbacks can be registered either in an SSL_CTX or per SSL.
7495 The purpose of this is to allow applications to control, if they wish,
7496 the arbitrary values chosen for use as session IDs, particularly as it
7497 can be useful for session caching in multiple-server environments. A
7498 command-line switch for testing this (and any client code that wishes
7499 to use such a feature) has been added to "s_server".
7500 [Geoff Thorpe, Lutz Jaenicke]
7502 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7503 of the form '#if defined(...) || defined(...) || ...' and
7504 '#if !defined(...) && !defined(...) && ...'. This also avoids
7505 the growing number of special cases it was previously handling.
7508 *) Make all configuration macros available for application by making
7509 sure they are available in opensslconf.h, by giving them names starting
7510 with "OPENSSL_" to avoid conflicts with other packages and by making
7511 sure e_os2.h will cover all platform-specific cases together with
7513 Additionally, it is now possible to define configuration/platform-
7514 specific names (called "system identities"). In the C code, these
7515 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7516 macro with the name beginning with "OPENSSL_SYS_", which is determined
7517 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7521 *) New option -set_serial to 'req' and 'x509' this allows the serial
7522 number to use to be specified on the command line. Previously self
7523 signed certificates were hard coded with serial number 0 and the
7524 CA options of 'x509' had to use a serial number in a file which was
7528 *) New options to 'ca' utility to support V2 CRL entry extensions.
7529 Currently CRL reason, invalidity date and hold instruction are
7530 supported. Add new CRL extensions to V3 code and some new objects.
7533 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7534 disable standard block padding (aka PKCS#5 padding) in the EVP
7535 API, which was previously mandatory. This means that the data is
7536 not padded in any way and so the total length much be a multiple
7537 of the block size, otherwise an error occurs.
7540 *) Initial (incomplete) OCSP SSL support.
7543 *) New function OCSP_parse_url(). This splits up a URL into its host,
7544 port and path components: primarily to parse OCSP URLs. New -url
7545 option to ocsp utility.
7548 *) New nonce behavior. The return value of OCSP_check_nonce() now
7549 reflects the various checks performed. Applications can decide
7550 whether to tolerate certain situations such as an absent nonce
7551 in a response when one was present in a request: the ocsp application
7552 just prints out a warning. New function OCSP_add1_basic_nonce()
7553 this is to allow responders to include a nonce in a response even if
7554 the request is nonce-less.
7557 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7558 skipped when using openssl x509 multiple times on a single input file,
7559 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7562 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7563 set string type: to handle setting ASN1_TIME structures. Fix ca
7564 utility to correctly initialize revocation date of CRLs.
7567 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7568 the clients preferred ciphersuites and rather use its own preferences.
7569 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7570 Internet Explorer by ensuring unchanged hash method during stepup.
7571 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7574 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7575 to aes and add a new 'exist' option to print out symbols that don't
7579 *) Additional options to ocsp utility to allow flags to be set and
7580 additional certificates supplied.
7583 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7584 OCSP client a number of certificate to only verify the response
7588 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7589 handle the new API. Currently only ECB, CBC modes supported. Add new
7592 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7593 Encryption Standard (AES) Ciphersuites for Transport Layer
7594 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7595 not enabled by default and were not part of the "ALL" ciphersuite
7596 alias because they were not yet official; they could be
7597 explicitly requested by specifying the "AESdraft" ciphersuite
7598 group alias. In the final release of OpenSSL 0.9.7, the group
7599 alias is called "AES" and is part of "ALL".)
7600 [Ben Laurie, Steve Henson, Bodo Moeller]
7602 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7603 request to response.
7606 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7607 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7608 extract information from a certificate request. OCSP_response_create()
7609 creates a response and optionally adds a basic response structure.
7610 OCSP_basic_add1_status() adds a complete single response to a basic
7611 response and returns the OCSP_SINGLERESP structure just added (to allow
7612 extensions to be included for example). OCSP_basic_add1_cert() adds a
7613 certificate to a basic response and OCSP_basic_sign() signs a basic
7614 response with various flags. New helper functions ASN1_TIME_check()
7615 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7616 (converts ASN1_TIME to GeneralizedTime).
7619 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7620 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7621 structure from a certificate. X509_pubkey_digest() digests the public_key
7622 contents: this is used in various key identifiers.
7625 *) Make sk_sort() tolerate a NULL argument.
7626 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7628 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7629 passed by the function are trusted implicitly. If any of them signed the
7630 response then it is assumed to be valid and is not verified.
7633 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7634 to data. This was previously part of the PKCS7 ASN1 code. This
7635 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7636 [Steve Henson, reported by Kenneth R. Robinette
7637 <support@securenetterm.com>]
7639 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7640 routines: without these tracing memory leaks is very painful.
7641 Fix leaks in PKCS12 and PKCS7 routines.
7644 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7645 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7646 effectively meant GeneralizedTime would never be used. Now it
7647 is initialised to -1 but X509_time_adj() now has to check the value
7648 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7649 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7650 [Steve Henson, reported by Kenneth R. Robinette
7651 <support@securenetterm.com>]
7653 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7654 result in a zero length in the ASN1_INTEGER structure which was
7655 not consistent with the structure when d2i_ASN1_INTEGER() was used
7656 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7657 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7658 where it did not print out a minus for negative ASN1_INTEGER.
7661 *) Add summary printout to ocsp utility. The various functions which
7662 convert status values to strings have been renamed to:
7663 OCSP_response_status_str(), OCSP_cert_status_str() and
7664 OCSP_crl_reason_str() and are no longer static. New options
7665 to verify nonce values and to disable verification. OCSP response
7666 printout format cleaned up.
7669 *) Add additional OCSP certificate checks. These are those specified
7670 in RFC2560. This consists of two separate checks: the CA of the
7671 certificate being checked must either be the OCSP signer certificate
7672 or the issuer of the OCSP signer certificate. In the latter case the
7673 OCSP signer certificate must contain the OCSP signing extended key
7674 usage. This check is performed by attempting to match the OCSP
7675 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7676 in the OCSP_CERTID structures of the response.
7679 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7680 and related routines. This uses the standard OpenSSL certificate
7681 verify routines to perform initial checks (just CA validity) and
7682 to obtain the certificate chain. Then additional checks will be
7683 performed on the chain. Currently the root CA is checked to see
7684 if it is explicitly trusted for OCSP signing. This is used to set
7685 a root CA as a global signing root: that is any certificate that
7686 chains to that CA is an acceptable OCSP signing certificate.
7689 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7690 extensions from a separate configuration file.
7691 As when reading extensions from the main configuration file,
7692 the '-extensions ...' option may be used for specifying the
7694 [Massimiliano Pala <madwolf@comune.modena.it>]
7696 *) New OCSP utility. Allows OCSP requests to be generated or
7697 read. The request can be sent to a responder and the output
7698 parsed, outputed or printed in text form. Not complete yet:
7699 still needs to check the OCSP response validity.
7702 *) New subcommands for 'openssl ca':
7703 'openssl ca -status <serial>' prints the status of the cert with
7704 the given serial number (according to the index file).
7705 'openssl ca -updatedb' updates the expiry status of certificates
7707 [Massimiliano Pala <madwolf@comune.modena.it>]
7709 *) New '-newreq-nodes' command option to CA.pl. This is like
7710 '-newreq', but calls 'openssl req' with the '-nodes' option
7711 so that the resulting key is not encrypted.
7712 [Damien Miller <djm@mindrot.org>]
7714 *) New configuration for the GNU Hurd.
7715 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7717 *) Initial code to implement OCSP basic response verify. This
7718 is currently incomplete. Currently just finds the signer's
7719 certificate and verifies the signature on the response.
7722 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7723 value of OPENSSLDIR. This is available via the new '-d' option
7724 to 'openssl version', and is also included in 'openssl version -a'.
7727 *) Allowing defining memory allocation callbacks that will be given
7728 file name and line number information in additional arguments
7729 (a const char* and an int). The basic functionality remains, as
7730 well as the original possibility to just replace malloc(),
7731 realloc() and free() by functions that do not know about these
7732 additional arguments. To register and find out the current
7733 settings for extended allocation functions, the following
7734 functions are provided:
7736 CRYPTO_set_mem_ex_functions
7737 CRYPTO_set_locked_mem_ex_functions
7738 CRYPTO_get_mem_ex_functions
7739 CRYPTO_get_locked_mem_ex_functions
7741 These work the same way as CRYPTO_set_mem_functions and friends.
7742 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7743 extended allocation function is enabled.
7744 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7745 a conventional allocation function is enabled.
7746 [Richard Levitte, Bodo Moeller]
7748 *) Finish off removing the remaining LHASH function pointer casts.
7749 There should no longer be any prototype-casting required when using
7750 the LHASH abstraction, and any casts that remain are "bugs". See
7751 the callback types and macros at the head of lhash.h for details
7752 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7755 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7756 If /dev/[u]random devices are not available or do not return enough
7757 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7759 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7760 /etc/entropy will be queried once each in this sequence, querying stops
7761 when enough entropy was collected without querying more sockets.
7764 *) Change the Unix RAND_poll() variant to be able to poll several
7765 random devices, as specified by DEVRANDOM, until a sufficient amount
7766 of data has been collected. We spend at most 10 ms on each file
7767 (select timeout) and read in non-blocking mode. DEVRANDOM now
7768 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7769 (previously it was just the string "/dev/urandom"), so on typical
7770 platforms the 10 ms delay will never occur.
7771 Also separate out the Unix variant to its own file, rand_unix.c.
7772 For VMS, there's a currently-empty rand_vms.c.
7775 *) Move OCSP client related routines to ocsp_cl.c. These
7776 provide utility functions which an application needing
7777 to issue a request to an OCSP responder and analyse the
7778 response will typically need: as opposed to those which an
7779 OCSP responder itself would need which will be added later.
7781 OCSP_request_sign() signs an OCSP request with an API similar
7782 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7783 response. OCSP_response_get1_basic() extracts basic response
7784 from response. OCSP_resp_find_status(): finds and extracts status
7785 information from an OCSP_CERTID structure (which will be created
7786 when the request structure is built). These are built from lower
7787 level functions which work on OCSP_SINGLERESP structures but
7788 won't normally be used unless the application wishes to examine
7789 extensions in the OCSP response for example.
7791 Replace nonce routines with a pair of functions.
7792 OCSP_request_add1_nonce() adds a nonce value and optionally
7793 generates a random value. OCSP_check_nonce() checks the
7794 validity of the nonce in an OCSP response.
7797 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7798 This doesn't copy the supplied OCSP_CERTID and avoids the
7799 need to free up the newly created id. Change return type
7800 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7801 This can then be used to add extensions to the request.
7802 Deleted OCSP_request_new(), since most of its functionality
7803 is now in OCSP_REQUEST_new() (and the case insensitive name
7804 clash) apart from the ability to set the request name which
7805 will be added elsewhere.
7808 *) Update OCSP API. Remove obsolete extensions argument from
7809 various functions. Extensions are now handled using the new
7810 OCSP extension code. New simple OCSP HTTP function which
7811 can be used to send requests and parse the response.
7814 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7815 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7816 uses the special reorder version of SET OF to sort the attributes
7817 and reorder them to match the encoded order. This resolves a long
7818 standing problem: a verify on a PKCS7 structure just after signing
7819 it used to fail because the attribute order did not match the
7820 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7821 it uses the received order. This is necessary to tolerate some broken
7822 software that does not order SET OF. This is handled by encoding
7823 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7824 to produce the required SET OF.
7827 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7828 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7829 files to get correct declarations of the ASN.1 item variables.
7832 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7833 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7834 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7835 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7836 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7837 ASN1_ITEM and no wrapper functions.
7840 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7841 replace the old function pointer based I/O routines. Change most of
7842 the *_d2i_bio() and *_d2i_fp() functions to use these.
7845 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7846 lines, recognice more "algorithms" that can be deselected, and make
7847 it complain about algorithm deselection that isn't recognised.
7850 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7851 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7852 to use new functions. Add NO_ASN1_OLD which can be set to remove
7853 some old style ASN1 functions: this can be used to determine if old
7854 code will still work when these eventually go away.
7857 *) New extension functions for OCSP structures, these follow the
7858 same conventions as certificates and CRLs.
7861 *) New function X509V3_add1_i2d(). This automatically encodes and
7862 adds an extension. Its behaviour can be customised with various
7863 flags to append, replace or delete. Various wrappers added for
7864 certificates and CRLs.
7867 *) Fix to avoid calling the underlying ASN1 print routine when
7868 an extension cannot be parsed. Correct a typo in the
7869 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7872 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7873 entries for variables.
7876 *) Add functionality to apps/openssl.c for detecting locking
7877 problems: As the program is single-threaded, all we have
7878 to do is register a locking callback using an array for
7879 storing which locks are currently held by the program.
7882 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7883 SSL_get_ex_data_X509_STORE_idx(), which is used in
7884 ssl_verify_cert_chain() and thus can be called at any time
7885 during TLS/SSL handshakes so that thread-safety is essential.
7886 Unfortunately, the ex_data design is not at all suited
7887 for multi-threaded use, so it probably should be abolished.
7890 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7891 [Broadcom, tweaked and integrated by Geoff Thorpe]
7893 *) Move common extension printing code to new function
7894 X509V3_print_extensions(). Reorganise OCSP print routines and
7895 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7898 *) New function X509_signature_print() to remove duplication in some
7902 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7903 set (this was treated exactly the same as SET OF previously). This
7904 is used to reorder the STACK representing the structure to match the
7905 encoding. This will be used to get round a problem where a PKCS7
7906 structure which was signed could not be verified because the STACK
7907 order did not reflect the encoded order.
7910 *) Reimplement the OCSP ASN1 module using the new code.
7913 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7914 for its ASN1 operations. The old style function pointers still exist
7915 for now but they will eventually go away.
7918 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7919 completely replaces the old ASN1 functionality with a table driven
7920 encoder and decoder which interprets an ASN1_ITEM structure describing
7921 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7922 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7923 has also been converted to the new form.
7926 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7927 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7928 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7929 for negative moduli.
7932 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7933 of not touching the result's sign bit.
7936 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7940 *) Changed the LHASH code to use prototypes for callbacks, and created
7941 macros to declare and implement thin (optionally static) functions
7942 that provide type-safety and avoid function pointer casting for the
7943 type-specific callbacks.
7946 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7948 [Veers Staats <staatsvr@asc.hpc.mil>,
7949 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7951 *) Reformat the FAQ so the different questions and answers can be divided
7952 in sections depending on the subject.
7955 *) Have the zlib compression code load ZLIB.DLL dynamically under
7959 *) New function BN_mod_sqrt for computing square roots modulo a prime
7960 (using the probabilistic Tonelli-Shanks algorithm unless
7961 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7962 be handled deterministically).
7963 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7965 *) Make BN_mod_inverse faster by explicitly handling small quotients
7966 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7967 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7970 *) New function BN_kronecker.
7973 *) Fix BN_gcd so that it works on negative inputs; the result is
7974 positive unless both parameters are zero.
7975 Previously something reasonably close to an infinite loop was
7976 possible because numbers could be growing instead of shrinking
7977 in the implementation of Euclid's algorithm.
7980 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7981 sign of the number in question.
7983 Fix BN_is_word(a,w) to work correctly for w == 0.
7985 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7986 because its test if the absolute value of 'a' equals 'w'.
7987 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7988 it exists mostly for use in the implementations of BN_is_zero(),
7989 BN_is_one(), and BN_is_word().
7992 *) New function BN_swap.
7995 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7996 the exponentiation functions are more likely to produce reasonable
7997 results on negative inputs.
8000 *) Change BN_mod_mul so that the result is always non-negative.
8001 Previously, it could be negative if one of the factors was negative;
8002 I don't think anyone really wanted that behaviour.
8005 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8006 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8007 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8008 and add new functions:
8017 BN_mod_lshift1_quick
8021 These functions always generate non-negative results.
8023 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8024 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8026 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8027 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8028 be reduced modulo m.
8029 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8032 The following entry accidentally appeared in the CHANGES file
8033 distributed with OpenSSL 0.9.7. The modifications described in
8034 it do *not* apply to OpenSSL 0.9.7.
8036 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8037 was actually never needed) and in BN_mul(). The removal in BN_mul()
8038 required a small change in bn_mul_part_recursive() and the addition
8039 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8040 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8041 bn_sub_words() and bn_add_words() except they take arrays with
8046 *) In 'openssl passwd', verify passwords read from the terminal
8047 unless the '-salt' option is used (which usually means that
8048 verification would just waste user's time since the resulting
8049 hash is going to be compared with some given password hash)
8050 or the new '-noverify' option is used.
8052 This is an incompatible change, but it does not affect
8053 non-interactive use of 'openssl passwd' (passwords on the command
8054 line, '-stdin' option, '-in ...' option) and thus should not
8058 *) Remove all references to RSAref, since there's no more need for it.
8061 *) Make DSO load along a path given through an environment variable
8062 (SHLIB_PATH) with shl_load().
8065 *) Constify the ENGINE code as a result of BIGNUM constification.
8066 Also constify the RSA code and most things related to it. In a
8067 few places, most notable in the depth of the ASN.1 code, ugly
8068 casts back to non-const were required (to be solved at a later
8072 *) Make it so the openssl application has all engines loaded by default.
8075 *) Constify the BIGNUM routines a little more.
8078 *) Add the following functions:
8080 ENGINE_load_cswift()
8082 ENGINE_load_atalla()
8084 ENGINE_load_builtin_engines()
8086 That way, an application can itself choose if external engines that
8087 are built-in in OpenSSL shall ever be used or not. The benefit is
8088 that applications won't have to be linked with libdl or other dso
8089 libraries unless it's really needed.
8091 Changed 'openssl engine' to load all engines on demand.
8092 Changed the engine header files to avoid the duplication of some
8093 declarations (they differed!).
8096 *) 'openssl engine' can now list capabilities.
8099 *) Better error reporting in 'openssl engine'.
8102 *) Never call load_dh_param(NULL) in s_server.
8105 *) Add engine application. It can currently list engines by name and
8106 identity, and test if they are actually available.
8109 *) Improve RPM specification file by forcing symbolic linking and making
8110 sure the installed documentation is also owned by root.root.
8111 [Damien Miller <djm@mindrot.org>]
8113 *) Give the OpenSSL applications more possibilities to make use of
8114 keys (public as well as private) handled by engines.
8117 *) Add OCSP code that comes from CertCo.
8120 *) Add VMS support for the Rijndael code.
8123 *) Added untested support for Nuron crypto accelerator.
8126 *) Add support for external cryptographic devices. This code was
8127 previously distributed separately as the "engine" branch.
8128 [Geoff Thorpe, Richard Levitte]
8130 *) Rework the filename-translation in the DSO code. It is now possible to
8131 have far greater control over how a "name" is turned into a filename
8132 depending on the operating environment and any oddities about the
8133 different shared library filenames on each system.
8136 *) Support threads on FreeBSD-elf in Configure.
8139 *) Fix for SHA1 assembly problem with MASM: it produces
8140 warnings about corrupt line number information when assembling
8141 with debugging information. This is caused by the overlapping
8143 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8146 NCONF_get_number() has no error checking at all. As a replacement,
8147 NCONF_get_number_e() is defined (_e for "error checking") and is
8148 promoted strongly. The old NCONF_get_number is kept around for
8149 binary backward compatibility.
8150 Make it possible for methods to load from something other than a BIO,
8151 by providing a function pointer that is given a name instead of a BIO.
8152 For example, this could be used to load configuration data from an
8156 *) Fix for non blocking accept BIOs. Added new I/O special reason
8157 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8158 with non blocking I/O was not possible because no retry code was
8159 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8163 *) Added the beginnings of Rijndael support.
8166 *) Fix for bug in DirectoryString mask setting. Add support for
8167 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8168 to allow certificate printing to more controllable, additional
8169 'certopt' option to 'x509' to allow new printing options to be
8173 *) Clean old EAY MD5 hack from e_os.h.
8176 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8178 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8179 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8180 [Joe Orton, Steve Henson]
8182 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8184 *) Fix additional bug revealed by the NISCC test suite:
8186 Stop bug triggering large recursion when presented with
8187 certain ASN.1 tags (CVE-2003-0851)
8190 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8192 *) Fix various bugs revealed by running the NISCC test suite:
8194 Stop out of bounds reads in the ASN1 code when presented with
8195 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8197 If verify callback ignores invalid public key errors don't try to check
8198 certificate signature with the NULL public key.
8202 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8203 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8207 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8208 extra data after the compression methods not only for TLS 1.0
8209 but also for SSL 3.0 (as required by the specification).
8210 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8212 *) Change X509_certificate_type() to mark the key as exported/exportable
8213 when it's 512 *bits* long, not 512 bytes.
8216 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8218 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8219 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8220 a protocol version number mismatch like a decryption error
8221 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8224 *) Turn on RSA blinding by default in the default implementation
8225 to avoid a timing attack. Applications that don't want it can call
8226 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8227 They would be ill-advised to do so in most cases.
8228 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8230 *) Change RSA blinding code so that it works when the PRNG is not
8231 seeded (in this case, the secret RSA exponent is abused as
8232 an unpredictable seed -- if it is not unpredictable, there
8233 is no point in blinding anyway). Make RSA blinding thread-safe
8234 by remembering the creator's thread ID in rsa->blinding and
8235 having all other threads use local one-time blinding factors
8236 (this requires more computation than sharing rsa->blinding, but
8237 avoids excessive locking; and if an RSA object is not shared
8238 between threads, blinding will still be very fast).
8241 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8243 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8244 via timing by performing a MAC computation even if incorrrect
8245 block cipher padding has been found. This is a countermeasure
8246 against active attacks where the attacker has to distinguish
8247 between bad padding and a MAC verification error. (CVE-2003-0078)
8249 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8250 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8251 Martin Vuagnoux (EPFL, Ilion)]
8253 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8255 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8256 memory from it's contents. This is done with a counter that will
8257 place alternating values in each byte. This can be used to solve
8258 two issues: 1) the removal of calls to memset() by highly optimizing
8259 compilers, and 2) cleansing with other values than 0, since those can
8260 be read through on certain media, for example a swap space on disk.
8263 *) Bugfix: client side session caching did not work with external caching,
8264 because the session->cipher setting was not restored when reloading
8265 from the external cache. This problem was masked, when
8266 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8267 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8270 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8271 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8272 [Zeev Lieber <zeev-l@yahoo.com>]
8274 *) Undo an undocumented change introduced in 0.9.6e which caused
8275 repeated calls to OpenSSL_add_all_ciphers() and
8276 OpenSSL_add_all_digests() to be ignored, even after calling
8280 *) Change the default configuration reader to deal with last line not
8281 being properly terminated.
8284 *) Change X509_NAME_cmp() so it applies the special rules on handling
8285 DN values that are of type PrintableString, as well as RDNs of type
8286 emailAddress where the value has the type ia5String.
8287 [stefank@valicert.com via Richard Levitte]
8289 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8290 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8291 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8292 the bitwise-OR of the two for use by the majority of applications
8293 wanting this behaviour, and update the docs. The documented
8294 behaviour and actual behaviour were inconsistent and had been
8295 changing anyway, so this is more a bug-fix than a behavioural
8297 [Geoff Thorpe, diagnosed by Nadav Har'El]
8299 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8300 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8303 *) Fix initialization code race conditions in
8304 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8305 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8306 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8307 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8308 ssl2_get_cipher_by_char(),
8309 ssl3_get_cipher_by_char().
8310 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8312 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8313 the cached sessions are flushed, as the remove_cb() might use ex_data
8314 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8315 (see [openssl.org #212]).
8316 [Geoff Thorpe, Lutz Jaenicke]
8318 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8319 length, instead of the encoding length to d2i_ASN1_OBJECT.
8322 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8324 *) [In 0.9.6g-engine release:]
8325 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8326 [Lynn Gazis <lgazis@rainbow.com>]
8328 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8330 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8331 and get fix the header length calculation.
8332 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8333 Alon Kantor <alonk@checkpoint.com> (and others),
8336 *) Use proper error handling instead of 'assertions' in buffer
8337 overflow checks added in 0.9.6e. This prevents DoS (the
8338 assertions could call abort()).
8339 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8341 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8343 *) Add various sanity checks to asn1_get_length() to reject
8344 the ASN1 length bytes if they exceed sizeof(long), will appear
8345 negative or the content length exceeds the length of the
8347 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8349 *) Fix cipher selection routines: ciphers without encryption had no flags
8350 for the cipher strength set and where therefore not handled correctly
8351 by the selection routines (PR #130).
8354 *) Fix EVP_dsa_sha macro.
8358 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8359 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8360 that was added in OpenSSL 0.9.6d.
8362 As the countermeasure turned out to be incompatible with some
8363 broken SSL implementations, the new option is part of SSL_OP_ALL.
8364 SSL_OP_ALL is usually employed when compatibility with weird SSL
8365 implementations is desired (e.g. '-bugs' option to 's_client' and
8366 's_server'), so the new option is automatically set in many
8370 *) Changes in security patch:
8372 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8373 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8374 Air Force Materiel Command, USAF, under agreement number
8377 *) Add various sanity checks to asn1_get_length() to reject
8378 the ASN1 length bytes if they exceed sizeof(long), will appear
8379 negative or the content length exceeds the length of the
8380 supplied buffer. (CVE-2002-0659)
8381 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8383 *) Assertions for various potential buffer overflows, not known to
8385 [Ben Laurie (CHATS)]
8387 *) Various temporary buffers to hold ASCII versions of integers were
8388 too small for 64 bit platforms. (CVE-2002-0655)
8389 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8391 *) Remote buffer overflow in SSL3 protocol - an attacker could
8392 supply an oversized session ID to a client. (CVE-2002-0656)
8393 [Ben Laurie (CHATS)]
8395 *) Remote buffer overflow in SSL2 protocol - an attacker could
8396 supply an oversized client master key. (CVE-2002-0656)
8397 [Ben Laurie (CHATS)]
8399 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8401 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8402 encoded as NULL) with id-dsa-with-sha1.
8403 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8405 *) Check various X509_...() return values in apps/req.c.
8406 [Nils Larsch <nla@trustcenter.de>]
8408 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8409 an end-of-file condition would erroneously be flagged, when the CRLF
8410 was just at the end of a processed block. The bug was discovered when
8411 processing data through a buffering memory BIO handing the data to a
8412 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8413 <ptsekov@syntrex.com> and Nedelcho Stanev.
8416 *) Implement a countermeasure against a vulnerability recently found
8417 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8418 before application data chunks to avoid the use of known IVs
8419 with data potentially chosen by the attacker.
8422 *) Fix length checks in ssl3_get_client_hello().
8425 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8426 to prevent ssl3_read_internal() from incorrectly assuming that
8427 ssl3_read_bytes() found application data while handshake
8428 processing was enabled when in fact s->s3->in_read_app_data was
8429 merely automatically cleared during the initial handshake.
8430 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8432 *) Fix object definitions for Private and Enterprise: they were not
8433 recognized in their shortname (=lowercase) representation. Extend
8434 obj_dat.pl to issue an error when using undefined keywords instead
8435 of silently ignoring the problem (Svenning Sorensen
8436 <sss@sss.dnsalias.net>).
8439 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8440 generators, i.e. generators other than 2 and 5. (Previously, the
8441 code did not properly initialise the 'add' and 'rem' values to
8442 BN_generate_prime().)
8444 In the new general case, we do not insist that 'generator' is
8445 actually a primitive root: This requirement is rather pointless;
8446 a generator of the order-q subgroup is just as good, if not
8450 *) Map new X509 verification errors to alerts. Discovered and submitted by
8451 Tom Wu <tom@arcot.com>.
8454 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8455 returning non-zero before the data has been completely received
8456 when using non-blocking I/O.
8457 [Bodo Moeller; problem pointed out by John Hughes]
8459 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8460 [Ben Laurie, Lutz Jaenicke]
8462 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8463 Yoram Zahavi <YoramZ@gilian.com>).
8466 *) Add information about CygWin 1.3 and on, and preserve proper
8467 configuration for the versions before that.
8468 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8470 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8471 check whether we deal with a copy of a session and do not delete from
8472 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8473 <izhar@checkpoint.com>.
8476 *) Do not store session data into the internal session cache, if it
8477 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8478 flag is set). Proposed by Aslam <aslam@funk.com>.
8481 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8485 *) [In 0.9.6d-engine release:]
8486 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8487 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8489 *) Add the configuration target linux-s390x.
8490 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8492 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8493 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8494 variable as an indication that a ClientHello message has been
8495 received. As the flag value will be lost between multiple
8496 invocations of ssl3_accept when using non-blocking I/O, the
8497 function may not be aware that a handshake has actually taken
8498 place, thus preventing a new session from being added to the
8501 To avoid this problem, we now set s->new_session to 2 instead of
8502 using a local variable.
8503 [Lutz Jaenicke, Bodo Moeller]
8505 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8506 if the SSL_R_LENGTH_MISMATCH error is detected.
8507 [Geoff Thorpe, Bodo Moeller]
8509 *) New 'shared_ldflag' column in Configure platform table.
8512 *) Fix EVP_CIPHER_mode macro.
8513 ["Dan S. Camper" <dan@bti.net>]
8515 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8516 type, we must throw them away by setting rr->length to 0.
8517 [D P Chang <dpc@qualys.com>]
8519 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8521 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8522 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8523 worked incorrectly for those cases where range = 10..._2 and
8524 3*range is two bits longer than range.)
8527 *) Only add signing time to PKCS7 structures if it is not already
8531 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8532 OBJ_ld_ce should be OBJ_id_ce.
8533 Also some ip-pda OIDs in crypto/objects/objects.txt were
8534 incorrect (cf. RFC 3039).
8535 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8537 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8538 returns early because it has nothing to do.
8539 [Andy Schneider <andy.schneider@bjss.co.uk>]
8541 *) [In 0.9.6c-engine release:]
8542 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8543 [Andy Schneider <andy.schneider@bjss.co.uk>]
8545 *) [In 0.9.6c-engine release:]
8546 Add support for Cryptographic Appliance's keyserver technology.
8547 (Use engine 'keyclient')
8548 [Cryptographic Appliances and Geoff Thorpe]
8550 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8551 is called via tools/c89.sh because arguments have to be
8552 rearranged (all '-L' options must appear before the first object
8554 [Richard Shapiro <rshapiro@abinitio.com>]
8556 *) [In 0.9.6c-engine release:]
8557 Add support for Broadcom crypto accelerator cards, backported
8559 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8561 *) [In 0.9.6c-engine release:]
8562 Add support for SureWare crypto accelerator cards from
8563 Baltimore Technologies. (Use engine 'sureware')
8564 [Baltimore Technologies and Mark Cox]
8566 *) [In 0.9.6c-engine release:]
8567 Add support for crypto accelerator cards from Accelerated
8568 Encryption Processing, www.aep.ie. (Use engine 'aep')
8569 [AEP Inc. and Mark Cox]
8571 *) Add a configuration entry for gcc on UnixWare.
8572 [Gary Benson <gbenson@redhat.com>]
8574 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8575 messages are stored in a single piece (fixed-length part and
8576 variable-length part combined) and fix various bugs found on the way.
8579 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8580 instead. BIO_gethostbyname() does not know what timeouts are
8581 appropriate, so entries would stay in cache even when they have
8583 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8585 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8586 faced with a pathologically small ClientHello fragment that does
8587 not contain client_version: Instead of aborting with an error,
8588 simply choose the highest available protocol version (i.e.,
8589 TLS 1.0 unless it is disabled). In practice, ClientHello
8590 messages are never sent like this, but this change gives us
8591 strictly correct behaviour at least for TLS.
8594 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8595 never resets s->method to s->ctx->method when called from within
8596 one of the SSL handshake functions.
8597 [Bodo Moeller; problem pointed out by Niko Baric]
8599 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8600 (sent using the client's version number) if client_version is
8601 smaller than the protocol version in use. Also change
8602 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8603 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8604 the client will at least see that alert.
8607 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8611 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8612 client receives HelloRequest while in a handshake.
8613 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8615 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8616 should end in 'break', not 'goto end' which circumvents various
8617 cleanups done in state SSL_ST_OK. But session related stuff
8618 must be disabled for SSL_ST_OK in the case that we just sent a
8621 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8622 before just sending a HelloRequest.
8623 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8625 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8626 reveal whether illegal block cipher padding was found or a MAC
8627 verification error occurred. (Neither SSLerr() codes nor alerts
8628 are directly visible to potential attackers, but the information
8629 may leak via logfiles.)
8631 Similar changes are not required for the SSL 2.0 implementation
8632 because the number of padding bytes is sent in clear for SSL 2.0,
8633 and the extra bytes are just ignored. However ssl/s2_pkt.c
8634 failed to verify that the purported number of padding bytes is in
8638 *) Add OpenUNIX-8 support including shared libraries
8639 (Boyd Lynn Gerber <gerberb@zenez.com>).
8642 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8643 'wristwatch attack' using huge encoding parameters (cf.
8644 James H. Manger's CRYPTO 2001 paper). Note that the
8645 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8646 encoding parameters and hence was not vulnerable.
8649 *) BN_sqr() bug fix.
8650 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8652 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8653 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8654 followed by modular reduction.
8655 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8657 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8658 equivalent based on BN_pseudo_rand() instead of BN_rand().
8661 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8662 This function was broken, as the check for a new client hello message
8663 to handle SGC did not allow these large messages.
8664 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8667 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8670 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8671 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8674 *) Rework the configuration and shared library support for Tru64 Unix.
8675 The configuration part makes use of modern compiler features and
8676 still retains old compiler behavior for those that run older versions
8677 of the OS. The shared library support part includes a variant that
8678 uses the RPATH feature, and is available through the special
8679 configuration target "alpha-cc-rpath", which will never be selected
8681 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8683 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8684 with the same message size as in ssl3_get_certificate_request().
8685 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8686 messages might inadvertently be reject as too long.
8687 [Petr Lampa <lampa@fee.vutbr.cz>]
8689 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8692 *) Modified SSL library such that the verify_callback that has been set
8693 specificly for an SSL object with SSL_set_verify() is actually being
8694 used. Before the change, a verify_callback set with this function was
8695 ignored and the verify_callback() set in the SSL_CTX at the time of
8696 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8697 to allow the necessary settings.
8700 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8701 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8702 done automatically (in contradiction to the requirements of the C
8703 standard). This made problems when used from OpenSSH.
8706 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8707 dh->length and always used
8709 BN_rand_range(priv_key, dh->p).
8711 BN_rand_range() is not necessary for Diffie-Hellman, and this
8712 specific range makes Diffie-Hellman unnecessarily inefficient if
8713 dh->length (recommended exponent length) is much smaller than the
8714 length of dh->p. We could use BN_rand_range() if the order of
8715 the subgroup was stored in the DH structure, but we only have
8720 BN_rand(priv_key, l, ...)
8722 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8728 RSA_eay_public_encrypt
8729 RSA_eay_private_decrypt
8730 RSA_eay_private_encrypt (signing)
8731 RSA_eay_public_decrypt (signature verification)
8733 (default implementations for RSA_public_encrypt,
8734 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8735 always reject numbers >= n.
8738 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8739 to synchronize access to 'locking_thread'. This is necessary on
8740 systems where access to 'locking_thread' (an 'unsigned long'
8741 variable) is not atomic.
8744 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8745 *before* setting the 'crypto_lock_rand' flag. The previous code had
8746 a race condition if 0 is a valid thread ID.
8747 [Travis Vitek <vitek@roguewave.com>]
8749 *) Add support for shared libraries under Irix.
8750 [Albert Chin-A-Young <china@thewrittenword.com>]
8752 *) Add configuration option to build on Linux on both big-endian and
8754 [Ralf Baechle <ralf@uni-koblenz.de>]
8756 *) Add the possibility to create shared libraries on HP-UX.
8759 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8761 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8762 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8763 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8764 PRNG state recovery was possible based on the output of
8765 one PRNG request appropriately sized to gain knowledge on
8766 'md' followed by enough consecutive 1-byte PRNG requests
8767 to traverse all of 'state'.
8769 1. When updating 'md_local' (the current thread's copy of 'md')
8770 during PRNG output generation, hash all of the previous
8771 'md_local' value, not just the half used for PRNG output.
8773 2. Make the number of bytes from 'state' included into the hash
8774 independent from the number of PRNG bytes requested.
8776 The first measure alone would be sufficient to avoid
8777 Markku-Juhani's attack. (Actually it had never occurred
8778 to me that the half of 'md_local' used for chaining was the
8779 half from which PRNG output bytes were taken -- I had always
8780 assumed that the secret half would be used.) The second
8781 measure makes sure that additional data from 'state' is never
8782 mixed into 'md_local' in small portions; this heuristically
8783 further strengthens the PRNG.
8786 *) Fix crypto/bn/asm/mips3.s.
8789 *) When only the key is given to "enc", the IV is undefined. Print out
8790 an error message in this case.
8793 *) Handle special case when X509_NAME is empty in X509 printing routines.
8796 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8797 positive and less than q.
8800 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8801 used: it isn't thread safe and the add_lock_callback should handle
8803 [Paul Rose <Paul.Rose@bridge.com>]
8805 *) Verify that incoming data obeys the block size in
8806 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8810 [Ulf Möller, Bodo Möller]
8812 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8813 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8814 when fixing the server behaviour for backwards-compatible 'client
8815 hello' messages. (Note that the attack is impractical against
8816 SSL 3.0 and TLS 1.0 anyway because length and version checking
8817 means that the probability of guessing a valid ciphertext is
8818 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8821 Before 0.9.5, the countermeasure (hide the error by generating a
8822 random 'decryption result') did not work properly because
8823 ERR_clear_error() was missing, meaning that SSL_get_error() would
8824 detect the supposedly ignored error.
8826 Both problems are now fixed.
8829 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8830 (previously it was 1024).
8833 *) Fix for compatibility mode trust settings: ignore trust settings
8834 unless some valid trust or reject settings are present.
8837 *) Fix for blowfish EVP: its a variable length cipher.
8840 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8841 parameters in DSA public key structures and return an error in the
8842 DSA routines if parameters are absent.
8845 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8846 in the current directory if neither $RANDFILE nor $HOME was set.
8847 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8848 caused some confusion to Windows users who haven't defined $HOME.
8849 Thus RAND_file_name() is changed again: e_os.h can define a
8850 DEFAULT_HOME, which will be used if $HOME is not set.
8851 For Windows, we use "C:"; on other platforms, we still require
8852 environment variables.
8854 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8855 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8856 having multiple threads call RAND_poll() concurrently.
8859 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8860 combination of a flag and a thread ID variable.
8861 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8862 flag), *other* threads can enter ssleay_add_bytes without obeying
8863 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8864 that they do not hold after the first thread unsets add_do_not_lock).
8867 *) Change bctest again: '-x' expressions are not available in all
8871 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8873 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8874 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8876 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8877 the default extension for executables, if any. Also, make the perl
8878 scripts that use symlink() to test if it really exists and use "cp"
8879 if it doesn't. All this made OpenSSL compilable and installable in
8883 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8884 If SEQUENCE is length is indefinite just set c->slen to the total
8885 amount of data available.
8886 [Steve Henson, reported by shige@FreeBSD.org]
8887 [This change does not apply to 0.9.7.]
8889 *) Change bctest to avoid here-documents inside command substitution
8890 (workaround for FreeBSD /bin/sh bug).
8891 For compatibility with Ultrix, avoid shell functions (introduced
8892 in the bctest version that searches along $PATH).
8895 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8896 with des_encrypt() defined on some operating systems, like Solaris
8900 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8901 On the Importance of Eliminating Errors in Cryptographic
8902 Computations, J. Cryptology 14 (2001) 2, 101-119,
8903 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8906 *) MIPS assembler BIGNUM division bug fix.
8909 *) Disabled incorrect Alpha assembler code.
8912 *) Fix PKCS#7 decode routines so they correctly update the length
8913 after reading an EOC for the EXPLICIT tag.
8915 [This change does not apply to 0.9.7.]
8917 *) Fix bug in PKCS#12 key generation routines. This was triggered
8918 if a 3DES key was generated with a 0 initial byte. Include
8919 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8920 (but broken) behaviour.
8923 *) Enhance bctest to search for a working bc along $PATH and print
8925 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8927 *) Fix memory leaks in err.c: free err_data string if necessary;
8928 don't write to the wrong index in ERR_set_error_data.
8931 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8935 *) Replace rdtsc with _emit statements for VC++ version 5.
8936 [Jeremy Cooper <jeremy@baymoo.org>]
8938 *) Make it possible to reuse SSLv2 sessions.
8941 *) In copy_email() check for >= 0 as a return value for
8942 X509_NAME_get_index_by_NID() since 0 is a valid index.
8943 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8945 *) Avoid coredump with unsupported or invalid public keys by checking if
8946 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8947 PKCS7_verify() fails with non detached data.
8950 *) Don't use getenv in library functions when run as setuid/setgid.
8951 New function OPENSSL_issetugid().
8954 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8955 due to incorrect handling of multi-threading:
8957 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8959 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8961 3. Count how many times MemCheck_off() has been called so that
8962 nested use can be treated correctly. This also avoids
8963 inband-signalling in the previous code (which relied on the
8964 assumption that thread ID 0 is impossible).
8967 *) Add "-rand" option also to s_client and s_server.
8970 *) Fix CPU detection on Irix 6.x.
8971 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8972 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8974 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8977 [This change does not apply to 0.9.7.]
8979 *) Use the cached encoding of an X509_NAME structure rather than
8980 copying it. This is apparently the reason for the libsafe "errors"
8981 but the code is actually correct.
8984 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8985 Bleichenbacher's DSA attack.
8986 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8987 to be set and top=0 forces the highest bit to be set; top=-1 is new
8988 and leaves the highest bit random.
8989 [Ulf Moeller, Bodo Moeller]
8991 *) In the NCONF_...-based implementations for CONF_... queries
8992 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8993 a temporary CONF structure with the data component set to NULL
8994 (which gives segmentation faults in lh_retrieve).
8995 Instead, use NULL for the CONF pointer in CONF_get_string and
8996 CONF_get_number (which may use environment variables) and directly
8997 return NULL from CONF_get_section.
9000 *) Fix potential buffer overrun for EBCDIC.
9003 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9004 keyUsage if basicConstraints absent for a CA.
9007 *) Make SMIME_write_PKCS7() write mail header values with a format that
9008 is more generally accepted (no spaces before the semicolon), since
9009 some programs can't parse those values properly otherwise. Also make
9010 sure BIO's that break lines after each write do not create invalid
9014 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9015 macros previously used would not encode an empty SEQUENCE OF
9016 and break the signature.
9018 [This change does not apply to 0.9.7.]
9020 *) Zero the premaster secret after deriving the master secret in
9024 *) Add some EVP_add_digest_alias registrations (as found in
9025 OpenSSL_add_all_digests()) to SSL_library_init()
9026 aka OpenSSL_add_ssl_algorithms(). This provides improved
9027 compatibility with peers using X.509 certificates
9028 with unconventional AlgorithmIdentifier OIDs.
9031 *) Fix for Irix with NO_ASM.
9032 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9034 *) ./config script fixes.
9035 [Ulf Moeller, Richard Levitte]
9037 *) Fix 'openssl passwd -1'.
9040 *) Change PKCS12_key_gen_asc() so it can cope with non null
9041 terminated strings whose length is passed in the passlen
9042 parameter, for example from PEM callbacks. This was done
9043 by adding an extra length parameter to asc2uni().
9044 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9046 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9047 call failed, free the DSA structure.
9050 *) Fix to uni2asc() to cope with zero length Unicode strings.
9051 These are present in some PKCS#12 files.
9054 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9055 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9056 when writing a 32767 byte record.
9057 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9059 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9060 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9062 (RSA objects have a reference count access to which is protected
9063 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9064 so they are meant to be shared between threads.)
9065 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9066 "Reddie, Steven" <Steven.Reddie@ca.com>]
9068 *) Fix a deadlock in CRYPTO_mem_leaks().
9071 *) Use better test patterns in bntest.
9074 *) rand_win.c fix for Borland C.
9077 *) BN_rshift bugfix for n == 0.
9080 *) Add a 'bctest' script that checks for some known 'bc' bugs
9081 so that 'make test' does not abort just because 'bc' is broken.
9084 *) Store verify_result within SSL_SESSION also for client side to
9085 avoid potential security hole. (Re-used sessions on the client side
9086 always resulted in verify_result==X509_V_OK, not using the original
9087 result of the server certificate verification.)
9090 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9091 SSL3_RT_APPLICATION_DATA, return 0.
9092 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9096 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9097 releases, have been re-implemented by renaming the previous
9098 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9099 and ssl3_read_internal, respectively, and adding 'peek' parameters
9100 to them. The new ssl[23]_{read,peek} functions are calls to
9101 ssl[23]_read_internal with the 'peek' flag set appropriately.
9102 A 'peek' parameter has also been added to ssl3_read_bytes, which
9103 does the actual work for ssl3_read_internal.
9106 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9107 the method-specific "init()" handler. Also clean up ex_data after
9108 calling the method-specific "finish()" handler. Previously, this was
9109 happening the other way round.
9112 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9113 The previous value, 12, was not always sufficient for BN_mod_exp().
9116 *) Make sure that shared libraries get the internal name engine with
9117 the full version number and not just 0. This should mark the
9118 shared libraries as not backward compatible. Of course, this should
9119 be changed again when we can guarantee backward binary compatibility.
9122 *) Fix typo in get_cert_by_subject() in by_dir.c
9123 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9125 *) Rework the system to generate shared libraries:
9127 - Make note of the expected extension for the shared libraries and
9128 if there is a need for symbolic links from for example libcrypto.so.0
9129 to libcrypto.so.0.9.7. There is extended info in Configure for
9132 - Make as few rebuilds of the shared libraries as possible.
9134 - Still avoid linking the OpenSSL programs with the shared libraries.
9136 - When installing, install the shared libraries separately from the
9140 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9142 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9143 and not in SSL_clear because the latter is also used by the
9144 accept/connect functions; previously, the settings made by
9145 SSL_set_read_ahead would be lost during the handshake.
9146 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9148 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9149 Previously, it would create entries for disabled algorithms no
9153 *) Added several new manual pages for SSL_* function.
9156 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9158 *) In ssl23_get_client_hello, generate an error message when faced
9159 with an initial SSL 3.0/TLS record that is too small to contain the
9160 first two bytes of the ClientHello message, i.e. client_version.
9161 (Note that this is a pathologic case that probably has never happened
9162 in real life.) The previous approach was to use the version number
9163 from the record header as a substitute; but our protocol choice
9164 should not depend on that one because it is not authenticated
9165 by the Finished messages.
9168 *) More robust randomness gathering functions for Windows.
9169 [Jeffrey Altman <jaltman@columbia.edu>]
9171 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9172 not set then we don't setup the error code for issuer check errors
9173 to avoid possibly overwriting other errors which the callback does
9174 handle. If an application does set the flag then we assume it knows
9175 what it is doing and can handle the new informational codes
9179 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9180 a general "ANY" type, as such it should be able to decode anything
9181 including tagged types. However it didn't check the class so it would
9182 wrongly interpret tagged types in the same way as their universal
9183 counterpart and unknown types were just rejected. Changed so that the
9184 tagged and unknown types are handled in the same way as a SEQUENCE:
9185 that is the encoding is stored intact. There is also a new type
9186 "V_ASN1_OTHER" which is used when the class is not universal, in this
9187 case we have no idea what the actual type is so we just lump them all
9191 *) On VMS, stdout may very well lead to a file that is written to
9192 in a record-oriented fashion. That means that every write() will
9193 write a separate record, which will be read separately by the
9194 programs trying to read from it. This can be very confusing.
9196 The solution is to put a BIO filter in the way that will buffer
9197 text until a linefeed is reached, and then write everything a
9198 line at a time, so every record written will be an actual line,
9199 not chunks of lines and not (usually doesn't happen, but I've
9200 seen it once) several lines in one record. BIO_f_linebuffer() is
9203 Currently, it's a VMS-only method, because that's where it has
9204 been tested well enough.
9207 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9208 it can return incorrect results.
9209 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9210 but it was in 0.9.6-beta[12].)
9213 *) Disable the check for content being present when verifying detached
9214 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9215 include zero length content when signing messages.
9218 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9219 BIO_ctrl (for BIO pairs).
9222 *) Add DSO method for VMS.
9225 *) Bug fix: Montgomery multiplication could produce results with the
9229 *) Add RPM specification openssl.spec and modify it to build three
9230 packages. The default package contains applications, application
9231 documentation and run-time libraries. The devel package contains
9232 include files, static libraries and function documentation. The
9233 doc package contains the contents of the doc directory. The original
9234 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9237 *) Add a large number of documentation files for many SSL routines.
9238 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9240 *) Add a configuration entry for Sony News 4.
9241 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9243 *) Don't set the two most significant bits to one when generating a
9244 random number < q in the DSA library.
9247 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9248 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9249 the underlying transport is blocking) if a handshake took place.
9250 (The default behaviour is needed by applications such as s_client
9251 and s_server that use select() to determine when to use SSL_read;
9252 but for applications that know in advance when to expect data, it
9253 just makes things more complicated.)
9256 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9260 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9261 work better on such systems.
9262 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9264 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9265 Update PKCS12_parse() so it copies the friendlyName and the
9266 keyid to the certificates aux info.
9269 *) Fix bug in PKCS7_verify() which caused an infinite loop
9270 if there was more than one signature.
9271 [Sven Uszpelkat <su@celocom.de>]
9273 *) Major change in util/mkdef.pl to include extra information
9274 about each symbol, as well as presenting variables as well
9275 as functions. This change means that there's n more need
9276 to rebuild the .num files when some algorithms are excluded.
9279 *) Allow the verify time to be set by an application,
9280 rather than always using the current time.
9283 *) Phase 2 verify code reorganisation. The certificate
9284 verify code now looks up an issuer certificate by a
9285 number of criteria: subject name, authority key id
9286 and key usage. It also verifies self signed certificates
9287 by the same criteria. The main comparison function is
9288 X509_check_issued() which performs these checks.
9290 Lot of changes were necessary in order to support this
9291 without completely rewriting the lookup code.
9293 Authority and subject key identifier are now cached.
9295 The LHASH 'certs' is X509_STORE has now been replaced
9296 by a STACK_OF(X509_OBJECT). This is mainly because an
9297 LHASH can't store or retrieve multiple objects with
9298 the same hash value.
9300 As a result various functions (which were all internal
9301 use only) have changed to handle the new X509_STORE
9302 structure. This will break anything that messed round
9303 with X509_STORE internally.
9305 The functions X509_STORE_add_cert() now checks for an
9306 exact match, rather than just subject name.
9308 The X509_STORE API doesn't directly support the retrieval
9309 of multiple certificates matching a given criteria, however
9310 this can be worked round by performing a lookup first
9311 (which will fill the cache with candidate certificates)
9312 and then examining the cache for matches. This is probably
9313 the best we can do without throwing out X509_LOOKUP
9314 entirely (maybe later...).
9316 The X509_VERIFY_CTX structure has been enhanced considerably.
9318 All certificate lookup operations now go via a get_issuer()
9319 callback. Although this currently uses an X509_STORE it
9320 can be replaced by custom lookups. This is a simple way
9321 to bypass the X509_STORE hackery necessary to make this
9322 work and makes it possible to use more efficient techniques
9323 in future. A very simple version which uses a simple
9324 STACK for its trusted certificate store is also provided
9325 using X509_STORE_CTX_trusted_stack().
9327 The verify_cb() and verify() callbacks now have equivalents
9328 in the X509_STORE_CTX structure.
9330 X509_STORE_CTX also has a 'flags' field which can be used
9331 to customise the verify behaviour.
9334 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9335 excludes S/MIME capabilities.
9338 *) When a certificate request is read in keep a copy of the
9339 original encoding of the signed data and use it when outputting
9340 again. Signatures then use the original encoding rather than
9341 a decoded, encoded version which may cause problems if the
9342 request is improperly encoded.
9345 *) For consistency with other BIO_puts implementations, call
9346 buffer_write(b, ...) directly in buffer_puts instead of calling
9349 In BIO_puts, increment b->num_write as in BIO_write.
9350 [Peter.Sylvester@EdelWeb.fr]
9352 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9353 BN_zero, we may not return a BIGNUM with an array consisting of
9357 *) Avoid calling abort() from within the library when problems are
9358 detected, except if preprocessor symbols have been defined
9359 (such as REF_CHECK, BN_DEBUG etc.).
9362 *) New openssl application 'rsautl'. This utility can be
9363 used for low level RSA operations. DER public key
9364 BIO/fp routines also added.
9367 *) New Configure entry and patches for compiling on QNX 4.
9368 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9370 *) A demo state-machine implementation was sponsored by
9371 Nuron (http://www.nuron.com/) and is now available in
9372 demos/state_machine.
9375 *) New options added to the 'dgst' utility for signature
9376 generation and verification.
9379 *) Unrecognized PKCS#7 content types are now handled via a
9380 catch all ASN1_TYPE structure. This allows unsupported
9381 types to be stored as a "blob" and an application can
9382 encode and decode it manually.
9385 *) Fix various signed/unsigned issues to make a_strex.c
9387 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9389 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9390 length if passed a buffer. ASN1_INTEGER_to_BN failed
9391 if passed a NULL BN and its argument was negative.
9392 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9394 *) Modification to PKCS#7 encoding routines to output definite
9395 length encoding. Since currently the whole structures are in
9396 memory there's not real point in using indefinite length
9397 constructed encoding. However if OpenSSL is compiled with
9398 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9401 *) Added BIO_vprintf() and BIO_vsnprintf().
9404 *) Added more prefixes to parse for in the the strings written
9405 through a logging bio, to cover all the levels that are available
9406 through syslog. The prefixes are now:
9408 PANIC, EMERG, EMR => LOG_EMERG
9409 ALERT, ALR => LOG_ALERT
9410 CRIT, CRI => LOG_CRIT
9411 ERROR, ERR => LOG_ERR
9412 WARNING, WARN, WAR => LOG_WARNING
9413 NOTICE, NOTE, NOT => LOG_NOTICE
9414 INFO, INF => LOG_INFO
9415 DEBUG, DBG => LOG_DEBUG
9417 and as before, if none of those prefixes are present at the
9418 beginning of the string, LOG_ERR is chosen.
9420 On Win32, the LOG_* levels are mapped according to this:
9422 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9423 LOG_WARNING => EVENTLOG_WARNING_TYPE
9424 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9428 *) Made it possible to reconfigure with just the configuration
9429 argument "reconf" or "reconfigure". The command line arguments
9430 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9431 and are retrieved from there when reconfiguring.
9435 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9437 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9440 *) The obj_dat.pl script was messing up the sorting of object
9441 names. The reason was that it compared the quoted version
9442 of strings as a result "OCSP" > "OCSP Signing" because
9443 " > SPACE. Changed script to store unquoted versions of
9444 names and add quotes on output. It was also omitting some
9445 names from the lookup table if they were given a default
9446 value (that is if SN is missing it is given the same
9447 value as LN and vice versa), these are now added on the
9448 grounds that if an object has a name we should be able to
9449 look it up. Finally added warning output when duplicate
9450 short or long names are found.
9453 *) Changes needed for Tandem NSK.
9454 [Scott Uroff <scott@xypro.com>]
9456 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9457 RSA_padding_check_SSLv23(), special padding was never detected
9458 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9459 version rollback attacks was not effective.
9461 In s23_clnt.c, don't use special rollback-attack detection padding
9462 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9463 client; similarly, in s23_srvr.c, don't do the rollback check if
9464 SSL 2.0 is the only protocol enabled in the server.
9467 *) Make it possible to get hexdumps of unprintable data with 'openssl
9468 asn1parse'. By implication, the functions ASN1_parse_dump() and
9469 BIO_dump_indent() are added.
9472 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9473 these print out strings and name structures based on various
9474 flags including RFC2253 support and proper handling of
9475 multibyte characters. Added options to the 'x509' utility
9476 to allow the various flags to be set.
9479 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9480 Also change the functions X509_cmp_current_time() and
9481 X509_gmtime_adj() work with an ASN1_TIME structure,
9482 this will enable certificates using GeneralizedTime in validity
9483 dates to be checked.
9486 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9487 negative public key encodings) on by default,
9488 NO_NEG_PUBKEY_BUG can be set to disable it.
9491 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9492 content octets. An i2c_ASN1_OBJECT is unnecessary because
9493 the encoding can be trivially obtained from the structure.
9496 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9497 not read locks (CRYPTO_r_[un]lock).
9500 *) A first attempt at creating official support for shared
9501 libraries through configuration. I've kept it so the
9502 default is static libraries only, and the OpenSSL programs
9503 are always statically linked for now, but there are
9504 preparations for dynamic linking in place.
9505 This has been tested on Linux and Tru64.
9508 *) Randomness polling function for Win9x, as described in:
9509 Peter Gutmann, Software Generation of Practically Strong
9513 *) Fix so PRNG is seeded in req if using an already existing
9517 *) New options to smime application. -inform and -outform
9518 allow alternative formats for the S/MIME message including
9519 PEM and DER. The -content option allows the content to be
9520 specified separately. This should allow things like Netscape
9521 form signing output easier to verify.
9524 *) Fix the ASN1 encoding of tags using the 'long form'.
9527 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9528 STRING types. These convert content octets to and from the
9529 underlying type. The actual tag and length octets are
9530 already assumed to have been read in and checked. These
9531 are needed because all other string types have virtually
9532 identical handling apart from the tag. By having versions
9533 of the ASN1 functions that just operate on content octets
9534 IMPLICIT tagging can be handled properly. It also allows
9535 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9536 and ASN1_INTEGER are identical apart from the tag.
9539 *) Change the handling of OID objects as follows:
9541 - New object identifiers are inserted in objects.txt, following
9542 the syntax given in objects.README.
9543 - objects.pl is used to process obj_mac.num and create a new
9545 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9548 This is currently kind of a hack, and the perl code in objects.pl
9549 isn't very elegant, but it works as I intended. The simplest way
9550 to check that it worked correctly is to look in obj_dat.h and
9551 check the array nid_objs and make sure the objects haven't moved
9552 around (this is important!). Additions are OK, as well as
9553 consistent name changes.
9556 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9559 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9560 The given file adds to whatever has already been seeded into the
9561 random pool through the RANDFILE configuration file option or
9562 environment variable, or the default random state file.
9565 *) mkstack.pl now sorts each macro group into lexical order.
9566 Previously the output order depended on the order the files
9567 appeared in the directory, resulting in needless rewriting
9571 *) Patches to make OpenSSL compile under Win32 again. Mostly
9572 work arounds for the VC++ problem that it treats func() as
9573 func(void). Also stripped out the parts of mkdef.pl that
9574 added extra typesafe functions: these no longer exist.
9577 *) Reorganisation of the stack code. The macros are now all
9578 collected in safestack.h . Each macro is defined in terms of
9579 a "stack macro" of the form SKM_<name>(type, a, b). The
9580 DEBUG_SAFESTACK is now handled in terms of function casts,
9581 this has the advantage of retaining type safety without the
9582 use of additional functions. If DEBUG_SAFESTACK is not defined
9583 then the non typesafe macros are used instead. Also modified the
9584 mkstack.pl script to handle the new form. Needs testing to see
9585 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9586 the default if no major problems. Similar behaviour for ASN1_SET_OF
9587 and PKCS12_STACK_OF.
9590 *) When some versions of IIS use the 'NET' form of private key the
9591 key derivation algorithm is different. Normally MD5(password) is
9592 used as a 128 bit RC4 key. In the modified case
9593 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9594 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9595 as the old Netscape_RSA functions except they have an additional
9596 'sgckey' parameter which uses the modified algorithm. Also added
9597 an -sgckey command line option to the rsa utility. Thanks to
9598 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9599 algorithm to openssl-dev.
9602 *) The evp_local.h macros were using 'c.##kname' which resulted in
9603 invalid expansion on some systems (SCO 5.0.5 for example).
9604 Corrected to 'c.kname'.
9605 [Phillip Porch <root@theporch.com>]
9607 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9608 a STACK of email addresses from a certificate or request, these look
9609 in the subject name and the subject alternative name extensions and
9610 omit any duplicate addresses.
9613 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9614 This makes DSA verification about 2 % faster.
9617 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9618 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9619 plus overhead for 1024 bit moduli).
9620 This makes exponentiations about 0.5 % faster for 1024 bit
9621 exponents (as measured by "openssl speed rsa2048").
9624 *) Rename memory handling macros to avoid conflicts with other
9626 Malloc => OPENSSL_malloc
9627 Malloc_locked => OPENSSL_malloc_locked
9628 Realloc => OPENSSL_realloc
9629 Free => OPENSSL_free
9632 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9633 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9636 *) CygWin32 support.
9637 [John Jarvie <jjarvie@newsguy.com>]
9639 *) The type-safe stack code has been rejigged. It is now only compiled
9640 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9641 by default all type-specific stack functions are "#define"d back to
9642 standard stack functions. This results in more streamlined output
9643 but retains the type-safety checking possibilities of the original
9647 *) The STACK code has been cleaned up, and certain type declarations
9648 that didn't make a lot of sense have been brought in line. This has
9649 also involved a cleanup of sorts in safestack.h to more correctly
9650 map type-safe stack functions onto their plain stack counterparts.
9651 This work has also resulted in a variety of "const"ifications of
9652 lots of the code, especially "_cmp" operations which should normally
9653 be prototyped with "const" parameters anyway.
9656 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9657 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9658 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9659 where all of 'md' is used each time the PRNG is used, but 'state'
9660 is used only indexed by a cyclic counter. As entropy may not be
9661 well distributed from the beginning, 'md' is important as a
9662 chaining variable. However, the output function chains only half
9663 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9664 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9665 in all of 'state' being rewritten, with the new values depending
9666 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9669 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9670 the handshake is continued after ssl_verify_cert_chain();
9671 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9672 can lead to 'unexplainable' connection aborts later.
9673 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9675 *) Major EVP API cipher revision.
9676 Add hooks for extra EVP features. This allows various cipher
9677 parameters to be set in the EVP interface. Support added for variable
9678 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9679 setting of RC2 and RC5 parameters.
9681 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9684 Remove lots of duplicated code from the EVP library. For example *every*
9685 cipher init() function handles the 'iv' in the same way according to the
9686 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9687 for CFB and OFB modes they zero ctx->num.
9689 New functionality allows removal of S/MIME code RC2 hack.
9691 Most of the routines have the same form and so can be declared in terms
9694 By shifting this to the top level EVP_CipherInit() it can be removed from
9695 all individual ciphers. If the cipher wants to handle IVs or keys
9696 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9699 Change lots of functions like EVP_EncryptUpdate() to now return a
9700 value: although software versions of the algorithms cannot fail
9701 any installed hardware versions can.
9704 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9705 this option is set, tolerate broken clients that send the negotiated
9706 protocol version number instead of the requested protocol version
9710 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9711 i.e. non-zero for export ciphersuites, zero otherwise.
9712 Previous versions had this flag inverted, inconsistent with
9713 rsa_tmp_cb (..._TMP_RSA_CB).
9714 [Bodo Moeller; problem reported by Amit Chopra]
9716 *) Add missing DSA library text string. Work around for some IIS
9717 key files with invalid SEQUENCE encoding.
9720 *) Add a document (doc/standards.txt) that list all kinds of standards
9721 and so on that are implemented in OpenSSL.
9724 *) Enhance c_rehash script. Old version would mishandle certificates
9725 with the same subject name hash and wouldn't handle CRLs at all.
9726 Added -fingerprint option to crl utility, to support new c_rehash
9730 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9733 *) Fix for SSL server purpose checking. Server checking was
9734 rejecting certificates which had extended key usage present
9735 but no ssl client purpose.
9736 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9738 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9739 is a little unclear about how a blank password is handled.
9740 Since the password in encoded as a BMPString with terminating
9741 double NULL a zero length password would end up as just the
9742 double NULL. However no password at all is different and is
9743 handled differently in the PKCS#12 key generation code. NS
9744 treats a blank password as zero length. MSIE treats it as no
9745 password on export: but it will try both on import. We now do
9746 the same: PKCS12_parse() tries zero length and no password if
9747 the password is set to "" or NULL (NULL is now a valid password:
9748 it wasn't before) as does the pkcs12 application.
9751 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9752 perror when PEM_read_bio_X509_REQ fails, the error message must
9753 be obtained from the error queue.
9756 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9757 it in ERR_remove_state if appropriate, and change ERR_get_state
9758 accordingly to avoid race conditions (this is necessary because
9759 thread_hash is no longer constant once set).
9762 *) Bugfix for linux-elf makefile.one.
9765 *) RSA_get_default_method() will now cause a default
9766 RSA_METHOD to be chosen if one doesn't exist already.
9767 Previously this was only set during a call to RSA_new()
9768 or RSA_new_method(NULL) meaning it was possible for
9769 RSA_get_default_method() to return NULL.
9772 *) Added native name translation to the existing DSO code
9773 that will convert (if the flag to do so is set) filenames
9774 that are sufficiently small and have no path information
9775 into a canonical native form. Eg. "blah" converted to
9776 "libblah.so" or "blah.dll" etc.
9779 *) New function ERR_error_string_n(e, buf, len) which is like
9780 ERR_error_string(e, buf), but writes at most 'len' bytes
9781 including the 0 terminator. For ERR_error_string_n, 'buf'
9783 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9785 *) CONF library reworked to become more general. A new CONF
9786 configuration file reader "class" is implemented as well as a
9787 new functions (NCONF_*, for "New CONF") to handle it. The now
9788 old CONF_* functions are still there, but are reimplemented to
9789 work in terms of the new functions. Also, a set of functions
9790 to handle the internal storage of the configuration data is
9791 provided to make it easier to write new configuration file
9792 reader "classes" (I can definitely see something reading a
9793 configuration file in XML format, for example), called _CONF_*,
9794 or "the configuration storage API"...
9796 The new configuration file reading functions are:
9798 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9799 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9801 NCONF_default, NCONF_WIN32
9803 NCONF_dump_fp, NCONF_dump_bio
9805 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9806 NCONF_new creates a new CONF object. This works in the same way
9807 as other interfaces in OpenSSL, like the BIO interface.
9808 NCONF_dump_* dump the internal storage of the configuration file,
9809 which is useful for debugging. All other functions take the same
9810 arguments as the old CONF_* functions wth the exception of the
9811 first that must be a `CONF *' instead of a `LHASH *'.
9813 To make it easer to use the new classes with the old CONF_* functions,
9814 the function CONF_set_default_method is provided.
9817 *) Add '-tls1' option to 'openssl ciphers', which was already
9818 mentioned in the documentation but had not been implemented.
9819 (This option is not yet really useful because even the additional
9820 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9823 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9824 OpenSSL-based applications) load shared libraries and bind to
9825 them in a portable way.
9826 [Geoff Thorpe, with contributions from Richard Levitte]
9828 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9830 *) Make sure _lrotl and _lrotr are only used with MSVC.
9832 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9833 (the default implementation of RAND_status).
9835 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9836 to '-clrext' (= clear extensions), as intended and documented.
9837 [Bodo Moeller; inconsistency pointed out by Michael Attili
9840 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9841 was larger than the MD block size.
9842 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9844 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9845 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9846 using the passed key: if the passed key was a private key the result
9847 of X509_print(), for example, would be to print out all the private key
9851 *) des_quad_cksum() byte order bug fix.
9852 [Ulf Möller, using the problem description in krb4-0.9.7, where
9853 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9855 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9857 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9859 *) For easily testing in shell scripts whether some command
9860 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9861 returns with exit code 0 iff no command of the given name is available.
9862 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9863 the output goes to stdout and nothing is printed to stderr.
9864 Additional arguments are always ignored.
9866 Since for each cipher there is a command of the same name,
9867 the 'no-cipher' compilation switches can be tested this way.
9869 ('openssl no-XXX' is not able to detect pseudo-commands such
9870 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9873 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9876 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9877 is set; it will be thrown away anyway because each handshake creates
9879 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9880 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9881 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9882 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9885 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9886 'Q' and 'R' lose their special meanings (quit/renegotiate).
9887 This is part of what -quiet does; unlike -quiet, -ign_eof
9888 does not suppress any output.
9891 *) Add compatibility options to the purpose and trust code. The
9892 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9893 accepts a certificate or CA, this was the previous behaviour,
9894 with all the associated security issues.
9896 X509_TRUST_COMPAT is the old trust behaviour: only and
9897 automatically trust self signed roots in certificate store. A
9898 new trust setting X509_TRUST_DEFAULT is used to specify that
9899 a purpose has no associated trust setting and it should instead
9900 use the value in the default purpose.
9903 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9904 and fix a memory leak.
9907 *) In util/mkerr.pl (which implements 'make errors'), preserve
9908 reason strings from the previous version of the .c file, as
9909 the default to have only downcase letters (and digits) in
9910 automatically generated reasons codes is not always appropriate.
9913 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9914 using strerror. Previously, ERR_reason_error_string() returned
9915 library names as reason strings for SYSerr; but SYSerr is a special
9916 case where small numbers are errno values, not library numbers.
9919 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9920 converts DSA parameters into DH parameters. (When creating parameters,
9921 DSA_generate_parameters is used.)
9924 *) Include 'length' (recommended exponent length) in C code generated
9925 by 'openssl dhparam -C'.
9928 *) The second argument to set_label in perlasm was already being used
9929 so couldn't be used as a "file scope" flag. Moved to third argument
9933 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9934 instead of RAND_bytes for encryption IVs and salts.
9937 *) Include RAND_status() into RAND_METHOD instead of implementing
9938 it only for md_rand.c Otherwise replacing the PRNG by calling
9939 RAND_set_rand_method would be impossible.
9942 *) Don't let DSA_generate_key() enter an infinite loop if the random
9943 number generation fails.
9946 *) New 'rand' application for creating pseudo-random output.
9949 *) Added configuration support for Linux/IA64
9950 [Rolf Haberrecker <rolf@suse.de>]
9952 *) Assembler module support for Mingw32.
9955 *) Shared library support for HPUX (in shlib/).
9956 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9958 *) Shared library support for Solaris gcc.
9959 [Lutz Behnke <behnke@trustcenter.de>]
9961 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9963 *) PKCS7_encrypt() was adding text MIME headers twice because they
9964 were added manually and by SMIME_crlf_copy().
9967 *) In bntest.c don't call BN_rand with zero bits argument.
9968 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9970 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9971 case was implemented. This caused BN_div_recp() to fail occasionally.
9974 *) Add an optional second argument to the set_label() in the perl
9975 assembly language builder. If this argument exists and is set
9976 to 1 it signals that the assembler should use a symbol whose
9977 scope is the entire file, not just the current function. This
9978 is needed with MASM which uses the format label:: for this scope.
9979 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9981 *) Change the ASN1 types so they are typedefs by default. Before
9982 almost all types were #define'd to ASN1_STRING which was causing
9983 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9987 *) Change names of new functions to the new get1/get0 naming
9988 convention: After 'get1', the caller owns a reference count
9989 and has to call ..._free; 'get0' returns a pointer to some
9990 data structure without incrementing reference counters.
9991 (Some of the existing 'get' functions increment a reference
9992 counter, some don't.)
9993 Similarly, 'set1' and 'add1' functions increase reference
9994 counters or duplicate objects.
9997 *) Allow for the possibility of temp RSA key generation failure:
9998 the code used to assume it always worked and crashed on failure.
10001 *) Fix potential buffer overrun problem in BIO_printf().
10002 [Ulf Möller, using public domain code by Patrick Powell; problem
10003 pointed out by David Sacerdote <das33@cornell.edu>]
10005 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10006 RAND_egd() and RAND_status(). In the command line application,
10007 the EGD socket can be specified like a seed file using RANDFILE
10011 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10012 Some CAs (e.g. Verisign) distribute certificates in this form.
10015 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10016 list to exclude them. This means that no special compilation option
10017 is needed to use anonymous DH: it just needs to be included in the
10021 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10022 EVP_MD_type. The old functionality is available in a new macro called
10023 EVP_MD_md(). Change code that uses it and update docs.
10026 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10027 where the 'void *' argument is replaced by a function pointer argument.
10028 Previously 'void *' was abused to point to functions, which works on
10029 many platforms, but is not correct. As these functions are usually
10030 called by macros defined in OpenSSL header files, most source code
10031 should work without changes.
10034 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10035 sections with information on -D... compiler switches used for
10036 compiling the library so that applications can see them. To enable
10037 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10038 must be defined. E.g.,
10039 #define OPENSSL_ALGORITHM_DEFINES
10040 #include <openssl/opensslconf.h>
10041 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10042 [Richard Levitte, Ulf and Bodo Möller]
10044 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10048 *) Change the 'other' type in certificate aux info to a STACK_OF
10049 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10050 the required ASN1 format: arbitrary types determined by an OID.
10053 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10054 argument to 'req'. This is not because the function is newer or
10055 better than others it just uses the work 'NEW' in the certificate
10056 request header lines. Some software needs this.
10059 *) Reorganise password command line arguments: now passwords can be
10060 obtained from various sources. Delete the PEM_cb function and make
10061 it the default behaviour: i.e. if the callback is NULL and the
10062 usrdata argument is not NULL interpret it as a null terminated pass
10063 phrase. If usrdata and the callback are NULL then the pass phrase
10064 is prompted for as usual.
10067 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10068 the support is automatically enabled. The resulting binaries will
10069 autodetect the card and use it if present.
10070 [Ben Laurie and Compaq Inc.]
10072 *) Work around for Netscape hang bug. This sends certificate request
10073 and server done in one record. Since this is perfectly legal in the
10074 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10075 the bugs/SSLv3 entry for more info.
10078 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10081 *) Add -rand argument to smime and pkcs12 applications and read/write
10085 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10088 *) Add command line password options to the remaining applications.
10091 *) Bug fix for BN_div_recp() for numerators with an even number of
10095 *) More tests in bntest.c, and changed test_bn output.
10098 *) ./config recognizes MacOS X now.
10101 *) Bug fix for BN_div() when the first words of num and divsor are
10102 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10105 *) Add support for various broken PKCS#8 formats, and command line
10106 options to produce them.
10109 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10110 get temporary BIGNUMs from a BN_CTX.
10113 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10117 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10118 include a #define from the old name to the new. The original intent
10119 was that statically linked binaries could for example just call
10120 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10121 link with digests. This never worked because SSLeay_add_all_digests()
10122 and SSLeay_add_all_ciphers() were in the same source file so calling
10123 one would link with the other. They are now in separate source files.
10126 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10129 *) Use a less unusual form of the Miller-Rabin primality test (it used
10130 a binary algorithm for exponentiation integrated into the Miller-Rabin
10131 loop, our standard modexp algorithms are faster).
10134 *) Support for the EBCDIC character set completed.
10135 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10137 *) Source code cleanups: use const where appropriate, eliminate casts,
10138 use void * instead of char * in lhash.
10141 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10142 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10143 this the server could overwrite ephemeral keys that the client
10147 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10148 using 50 iterations of the Rabin-Miller test.
10150 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10151 iterations of the Rabin-Miller test as required by the appendix
10152 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10153 As BN_is_prime_fasttest includes trial division, DSA parameter
10154 generation becomes much faster.
10156 This implies a change for the callback functions in DSA_is_prime
10157 and DSA_generate_parameters: The callback function is called once
10158 for each positive witness in the Rabin-Miller test, not just
10159 occasionally in the inner loop; and the parameters to the
10160 callback function now provide an iteration count for the outer
10161 loop rather than for the current invocation of the inner loop.
10162 DSA_generate_parameters additionally can call the callback
10163 function with an 'iteration count' of -1, meaning that a
10164 candidate has passed the trial division test (when q is generated
10165 from an application-provided seed, trial division is skipped).
10168 *) New function BN_is_prime_fasttest that optionally does trial
10169 division before starting the Rabin-Miller test and has
10170 an additional BN_CTX * argument (whereas BN_is_prime always
10171 has to allocate at least one BN_CTX).
10172 'callback(1, -1, cb_arg)' is called when a number has passed the
10173 trial division stage.
10176 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10180 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10183 *) New function BN_pseudo_rand().
10186 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10187 bignum version of BN_from_montgomery() with the working code from
10188 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10192 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10193 made it impossible to use the same SSL_SESSION data structure in
10194 SSL2 clients in multiple threads.
10197 *) The return value of RAND_load_file() no longer counts bytes obtained
10198 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10199 to seed the PRNG (previously an explicit byte count was required).
10200 [Ulf Möller, Bodo Möller]
10202 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10203 used (char *) instead of (void *) and had casts all over the place.
10206 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10209 *) Retain source code compatibility for BN_prime_checks macro:
10210 BN_is_prime(..., BN_prime_checks, ...) now uses
10211 BN_prime_checks_for_size to determine the appropriate number of
10212 Rabin-Miller iterations.
10215 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10216 DH_CHECK_P_NOT_SAFE_PRIME.
10217 (Check if this is true? OpenPGP calls them "strong".)
10220 *) Merge the functionality of "dh" and "gendh" programs into a new program
10221 "dhparam". The old programs are retained for now but will handle DH keys
10222 (instead of parameters) in future.
10225 *) Make the ciphers, s_server and s_client programs check the return values
10226 when a new cipher list is set.
10229 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10230 ciphers. Before when the 56bit ciphers were enabled the sorting was
10233 The syntax for the cipher sorting has been extended to support sorting by
10234 cipher-strength (using the strength_bits hard coded in the tables).
10235 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10237 Fix a bug in the cipher-command parser: when supplying a cipher command
10238 string with an "undefined" symbol (neither command nor alphanumeric
10239 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10240 an error is flagged.
10242 Due to the strength-sorting extension, the code of the
10243 ssl_create_cipher_list() function was completely rearranged. I hope that
10244 the readability was also increased :-)
10245 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10247 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10248 for the first serial number and places 2 in the serial number file. This
10249 avoids problems when the root CA is created with serial number zero and
10250 the first user certificate has the same issuer name and serial number
10254 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10255 the new code. Add documentation for this stuff.
10258 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10259 X509_*() to X509at_*() on the grounds that they don't handle X509
10260 structures and behave in an analogous way to the X509v3 functions:
10261 they shouldn't be called directly but wrapper functions should be used
10264 So we also now have some wrapper functions that call the X509at functions
10265 when passed certificate requests. (TO DO: similar things can be done with
10266 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10267 things. Some of these need some d2i or i2d and print functionality
10268 because they handle more complex structures.)
10271 *) Add missing #ifndefs that caused missing symbols when building libssl
10272 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10273 NO_RSA in ssl/s2*.c.
10274 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10276 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10277 has a return value which indicates the quality of the random data
10278 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10279 error queue. New function RAND_pseudo_bytes() generates output that is
10280 guaranteed to be unique but not unpredictable. RAND_add is like
10281 RAND_seed, but takes an extra argument for an entropy estimate
10282 (RAND_seed always assumes full entropy).
10285 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10286 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10287 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10288 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10289 false-positive rate of at most 2^-80 for random input.
10292 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10295 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10296 in the 0.9.5 release), this returns the chain
10297 from an X509_CTX structure with a dup of the stack and all
10298 the X509 reference counts upped: so the stack will exist
10299 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10302 Also make SSL_SESSION_print() print out the verify return
10306 *) Add manpage for the pkcs12 command. Also change the default
10307 behaviour so MAC iteration counts are used unless the new
10308 -nomaciter option is used. This improves file security and
10309 only older versions of MSIE (4.0 for example) need it.
10312 *) Honor the no-xxx Configure options when creating .DEF files.
10315 *) Add PKCS#10 attributes to field table: challengePassword,
10316 unstructuredName and unstructuredAddress. These are taken from
10317 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10318 international characters are used.
10320 More changes to X509_ATTRIBUTE code: allow the setting of types
10321 based on strings. Remove the 'loc' parameter when adding
10322 attributes because these will be a SET OF encoding which is sorted
10326 *) Initial changes to the 'req' utility to allow request generation
10327 automation. This will allow an application to just generate a template
10328 file containing all the field values and have req construct the
10331 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10332 used all over the place including certificate requests and PKCS#7
10333 structures. They are currently handled manually where necessary with
10334 some primitive wrappers for PKCS#7. The new functions behave in a
10335 manner analogous to the X509 extension functions: they allow
10336 attributes to be looked up by NID and added.
10338 Later something similar to the X509V3 code would be desirable to
10339 automatically handle the encoding, decoding and printing of the
10340 more complex types. The string types like challengePassword can
10341 be handled by the string table functions.
10343 Also modified the multi byte string table handling. Now there is
10344 a 'global mask' which masks out certain types. The table itself
10345 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10346 is useful when for example there is only one permissible type
10347 (as in countryName) and using the mask might result in no valid
10351 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10352 SSL_get_peer_finished to allow applications to obtain the latest
10353 Finished messages sent to the peer or expected from the peer,
10354 respectively. (SSL_get_peer_finished is usually the Finished message
10355 actually received from the peer, otherwise the protocol will be aborted.)
10357 As the Finished message are message digests of the complete handshake
10358 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10359 be used for external authentication procedures when the authentication
10360 provided by SSL/TLS is not desired or is not enough.
10363 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10364 the host supports BWX extension and if Compaq C is present on the
10365 $PATH. Just exploiting of the BWX extension results in 20-30%
10366 performance kick for some algorithms, e.g. DES and RC4 to mention
10367 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10371 *) Add support for MS "fast SGC". This is arguably a violation of the
10372 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10373 weak crypto and after checking the certificate is SGC a second one
10374 with strong crypto. MS SGC stops the first handshake after receiving
10375 the server certificate message and sends a second client hello. Since
10376 a server will typically do all the time consuming operations before
10377 expecting any further messages from the client (server key exchange
10378 is the most expensive) there is little difference between the two.
10380 To get OpenSSL to support MS SGC we have to permit a second client
10381 hello message after we have sent server done. In addition we have to
10382 reset the MAC if we do get this second client hello.
10385 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10386 if a DER encoded private key is RSA or DSA traditional format. Changed
10387 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10388 format DER encoded private key. Newer code should use PKCS#8 format which
10389 has the key type encoded in the ASN1 structure. Added DER private key
10390 support to pkcs8 application.
10393 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10394 ciphersuites has been selected (as required by the SSL 3/TLS 1
10395 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10396 is set, we interpret this as a request to violate the specification
10397 (the worst that can happen is a handshake failure, and 'correct'
10398 behaviour would result in a handshake failure anyway).
10401 *) In SSL_CTX_add_session, take into account that there might be multiple
10402 SSL_SESSION structures with the same session ID (e.g. when two threads
10403 concurrently obtain them from an external cache).
10404 The internal cache can handle only one SSL_SESSION with a given ID,
10405 so if there's a conflict, we now throw out the old one to achieve
10409 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10410 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10411 some routines that use cipher OIDs: some ciphers do not have OIDs
10412 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10416 *) Simplify the trust setting structure and code. Now we just have
10417 two sequences of OIDs for trusted and rejected settings. These will
10418 typically have values the same as the extended key usage extension
10419 and any application specific purposes.
10421 The trust checking code now has a default behaviour: it will just
10422 check for an object with the same NID as the passed id. Functions can
10423 be provided to override either the default behaviour or the behaviour
10424 for a given id. SSL client, server and email already have functions
10425 in place for compatibility: they check the NID and also return "trusted"
10426 if the certificate is self signed.
10429 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10430 traditional format into an EVP_PKEY structure.
10433 *) Add a password callback function PEM_cb() which either prompts for
10434 a password if usr_data is NULL or otherwise assumes it is a null
10435 terminated password. Allow passwords to be passed on command line
10436 environment or config files in a few more utilities.
10439 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10440 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10441 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10442 Update documentation.
10445 *) Support for ASN1 "NULL" type. This could be handled before by using
10446 ASN1_TYPE but there wasn't any function that would try to read a NULL
10447 and produce an error if it couldn't. For compatibility we also have
10448 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10449 don't allocate anything because they don't need to.
10452 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10454 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10456 *) Rebuild of the memory allocation routines used by OpenSSL code and
10457 possibly others as well. The purpose is to make an interface that
10458 provide hooks so anyone can build a separate set of allocation and
10459 deallocation routines to be used by OpenSSL, for example memory
10460 pool implementations, or something else, which was previously hard
10461 since Malloc(), Realloc() and Free() were defined as macros having
10462 the values malloc, realloc and free, respectively (except for Win32
10463 compilations). The same is provided for memory debugging code.
10464 OpenSSL already comes with functionality to find memory leaks, but
10465 this gives people a chance to debug other memory problems.
10467 With these changes, a new set of functions and macros have appeared:
10469 CRYPTO_set_mem_debug_functions() [F]
10470 CRYPTO_get_mem_debug_functions() [F]
10471 CRYPTO_dbg_set_options() [F]
10472 CRYPTO_dbg_get_options() [F]
10473 CRYPTO_malloc_debug_init() [M]
10475 The memory debug functions are NULL by default, unless the library
10476 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10477 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10478 gives the standard debugging functions that come with OpenSSL) or
10479 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10480 provided by the library user) must be used. When the standard
10481 debugging functions are used, CRYPTO_dbg_set_options can be used to
10482 request additional information:
10483 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10484 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10486 Also, things like CRYPTO_set_mem_functions will always give the
10487 expected result (the new set of functions is used for allocation
10488 and deallocation) at all times, regardless of platform and compiler
10491 To finish it up, some functions that were never use in any other
10492 way than through macros have a new API and new semantic:
10494 CRYPTO_dbg_malloc()
10495 CRYPTO_dbg_realloc()
10498 All macros of value have retained their old syntax.
10499 [Richard Levitte and Bodo Moeller]
10501 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10502 ordering of SMIMECapabilities wasn't in "strength order" and there
10503 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10507 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10508 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10509 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10511 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10512 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10513 functionality to handle multipart/signed properly) and a utility
10514 called 'smime' to call all this stuff. This is based on code I
10515 originally wrote for Celo who have kindly allowed it to be
10516 included in OpenSSL.
10519 *) Add variants des_set_key_checked and des_set_key_unchecked of
10520 des_set_key (aka des_key_sched). Global variable des_check_key
10521 decides which of these is called by des_set_key; this way
10522 des_check_key behaves as it always did, but applications and
10523 the library itself, which was buggy for des_check_key == 1,
10524 have a cleaner way to pick the version they need.
10527 *) New function PKCS12_newpass() which changes the password of a
10531 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10532 dynamic mix. In both cases the ids can be used as an index into the
10533 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10534 functions so they accept a list of the field values and the
10535 application doesn't need to directly manipulate the X509_TRUST
10539 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10543 *) Modify the way the V3 extension code looks up extensions. This now
10544 works in a similar way to the object code: we have some "standard"
10545 extensions in a static table which is searched with OBJ_bsearch()
10546 and the application can add dynamic ones if needed. The file
10547 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10548 updated whenever a new extension is added to the core code and kept
10549 in ext_nid order. There is a simple program 'tabtest.c' which checks
10550 this. New extensions are not added too often so this file can readily
10551 be maintained manually.
10553 There are two big advantages in doing things this way. The extensions
10554 can be looked up immediately and no longer need to be "added" using
10555 X509V3_add_standard_extensions(): this function now does nothing.
10556 [Side note: I get *lots* of email saying the extension code doesn't
10557 work because people forget to call this function]
10558 Also no dynamic allocation is done unless new extensions are added:
10559 so if we don't add custom extensions there is no need to call
10560 X509V3_EXT_cleanup().
10563 *) Modify enc utility's salting as follows: make salting the default. Add a
10564 magic header, so unsalted files fail gracefully instead of just decrypting
10565 to garbage. This is because not salting is a big security hole, so people
10566 should be discouraged from doing it.
10569 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10570 digest to be passed on the command line but it only used this
10571 parameter when signing a certificate. Modified so all relevant
10572 operations are affected by the digest parameter including the
10573 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10574 DSA key was used because it didn't fix the digest.
10577 *) Initial certificate chain verify code. Currently tests the untrusted
10578 certificates for consistency with the verify purpose (which is set
10579 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10581 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10582 this is because it will reject chains with invalid extensions whereas
10583 every previous version of OpenSSL and SSLeay made no checks at all.
10585 Trust code: checks the root CA for the relevant trust settings. Trust
10586 settings have an initial value consistent with the verify purpose: e.g.
10587 if the verify purpose is for SSL client use it expects the CA to be
10588 trusted for SSL client use. However the default value can be changed to
10589 permit custom trust settings: one example of this would be to only trust
10590 certificates from a specific "secure" set of CAs.
10592 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10593 which should be used for version portability: especially since the
10594 verify structure is likely to change more often now.
10596 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10597 to set them. If not set then assume SSL clients will verify SSL servers
10600 Two new options to the verify program: -untrusted allows a set of
10601 untrusted certificates to be passed in and -purpose which sets the
10602 intended purpose of the certificate. If a purpose is set then the
10603 new chain verify code is used to check extension consistency.
10606 *) Support for the authority information access extension.
10609 *) Modify RSA and DSA PEM read routines to transparently handle
10610 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10611 public keys in a format compatible with certificate
10612 SubjectPublicKeyInfo structures. Unfortunately there were already
10613 functions called *_PublicKey_* which used various odd formats so
10614 these are retained for compatibility: however the DSA variants were
10615 never in a public release so they have been deleted. Changed dsa/rsa
10616 utilities to handle the new format: note no releases ever handled public
10617 keys so we should be OK.
10619 The primary motivation for this change is to avoid the same fiasco
10620 that dogs private keys: there are several incompatible private key
10621 formats some of which are standard and some OpenSSL specific and
10622 require various evil hacks to allow partial transparent handling and
10623 even then it doesn't work with DER formats. Given the option anything
10624 other than PKCS#8 should be dumped: but the other formats have to
10625 stay in the name of compatibility.
10627 With public keys and the benefit of hindsight one standard format
10628 is used which works with EVP_PKEY, RSA or DSA structures: though
10629 it clearly returns an error if you try to read the wrong kind of key.
10631 Added a -pubkey option to the 'x509' utility to output the public key.
10632 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10633 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10634 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10635 that do the same as the EVP_PKEY_assign_*() except they up the
10636 reference count of the added key (they don't "swallow" the
10640 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10641 CRLs would fail if the file contained no certificates or no CRLs:
10642 added a new function to read in both types and return the number
10643 read: this means that if none are read it will be an error. The
10644 DER versions of the certificate and CRL reader would always fail
10645 because it isn't possible to mix certificates and CRLs in DER format
10646 without choking one or the other routine. Changed this to just read
10647 a certificate: this is the best we can do. Also modified the code
10648 in apps/verify.c to take notice of return codes: it was previously
10649 attempting to read in certificates from NULL pointers and ignoring
10650 any errors: this is one reason why the cert and CRL reader seemed
10651 to work. It doesn't check return codes from the default certificate
10652 routines: these may well fail if the certificates aren't installed.
10655 *) Code to support otherName option in GeneralName.
10658 *) First update to verify code. Change the verify utility
10659 so it warns if it is passed a self signed certificate:
10660 for consistency with the normal behaviour. X509_verify
10661 has been modified to it will now verify a self signed
10662 certificate if *exactly* the same certificate appears
10663 in the store: it was previously impossible to trust a
10664 single self signed certificate. This means that:
10665 openssl verify ss.pem
10666 now gives a warning about a self signed certificate but
10667 openssl verify -CAfile ss.pem ss.pem
10671 *) For servers, store verify_result in SSL_SESSION data structure
10672 (and add it to external session representation).
10673 This is needed when client certificate verifications fails,
10674 but an application-provided verification callback (set by
10675 SSL_CTX_set_cert_verify_callback) allows accepting the session
10676 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10677 but returns 1): When the session is reused, we have to set
10678 ssl->verify_result to the appropriate error code to avoid
10680 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10682 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10683 case in PKCS7_dataInit() where the signed PKCS7 structure
10684 didn't contain any existing data because it was being created.
10685 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10687 *) Add a salt to the key derivation routines in enc.c. This
10688 forms the first 8 bytes of the encrypted file. Also add a
10689 -S option to allow a salt to be input on the command line.
10692 *) New function X509_cmp(). Oddly enough there wasn't a function
10693 to compare two certificates. We do this by working out the SHA1
10694 hash and comparing that. X509_cmp() will be needed by the trust
10698 *) SSL_get1_session() is like SSL_get_session(), but increments
10699 the reference count in the SSL_SESSION returned.
10700 [Geoff Thorpe <geoff@eu.c2.net>]
10702 *) Fix for 'req': it was adding a null to request attributes.
10703 Also change the X509_LOOKUP and X509_INFO code to handle
10704 certificate auxiliary information.
10707 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10711 *) Add the possibility to add extra information to the memory leak
10712 detecting output, to form tracebacks, showing from where each
10713 allocation was originated: CRYPTO_push_info("constant string") adds
10714 the string plus current file name and line number to a per-thread
10715 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10716 is like calling CYRPTO_pop_info() until the stack is empty.
10717 Also updated memory leak detection code to be multi-thread-safe.
10720 *) Add options -text and -noout to pkcs7 utility and delete the
10721 encryption options which never did anything. Update docs.
10724 *) Add options to some of the utilities to allow the pass phrase
10725 to be included on either the command line (not recommended on
10726 OSes like Unix) or read from the environment. Update the
10727 manpages and fix a few bugs.
10730 *) Add a few manpages for some of the openssl commands.
10733 *) Fix the -revoke option in ca. It was freeing up memory twice,
10734 leaking and not finding already revoked certificates.
10737 *) Extensive changes to support certificate auxiliary information.
10738 This involves the use of X509_CERT_AUX structure and X509_AUX
10739 functions. An X509_AUX function such as PEM_read_X509_AUX()
10740 can still read in a certificate file in the usual way but it
10741 will also read in any additional "auxiliary information". By
10742 doing things this way a fair degree of compatibility can be
10743 retained: existing certificates can have this information added
10744 using the new 'x509' options.
10746 Current auxiliary information includes an "alias" and some trust
10747 settings. The trust settings will ultimately be used in enhanced
10748 certificate chain verification routines: currently a certificate
10749 can only be trusted if it is self signed and then it is trusted
10753 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10754 The problem was that one of the replacement routines had not been working
10755 since SSLeay releases. For now the offending routine has been replaced
10756 with non-optimised assembler. Even so, this now gives around 95%
10757 performance improvement for 1024 bit RSA signs.
10760 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10761 handling. Most clients have the effective key size in bits equal to
10762 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10763 A few however don't do this and instead use the size of the decrypted key
10764 to determine the RC2 key length and the AlgorithmIdentifier to determine
10765 the effective key length. In this case the effective key length can still
10766 be 40 bits but the key length can be 168 bits for example. This is fixed
10767 by manually forcing an RC2 key into the EVP_PKEY structure because the
10768 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10769 the key length and effective key length are equal.
10772 *) Add a bunch of functions that should simplify the creation of
10773 X509_NAME structures. Now you should be able to do:
10774 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10775 and have it automatically work out the correct field type and fill in
10776 the structures. The more adventurous can try:
10777 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10778 and it will (hopefully) work out the correct multibyte encoding.
10781 *) Change the 'req' utility to use the new field handling and multibyte
10782 copy routines. Before the DN field creation was handled in an ad hoc
10783 way in req, ca, and x509 which was rather broken and didn't support
10784 BMPStrings or UTF8Strings. Since some software doesn't implement
10785 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10786 using the dirstring_type option. See the new comment in the default
10787 openssl.cnf for more info.
10790 *) Make crypto/rand/md_rand.c more robust:
10791 - Assure unique random numbers after fork().
10792 - Make sure that concurrent threads access the global counter and
10793 md serializably so that we never lose entropy in them
10794 or use exactly the same state in multiple threads.
10795 Access to the large state is not always serializable because
10796 the additional locking could be a performance killer, and
10797 md should be large enough anyway.
10800 *) New file apps/app_rand.c with commonly needed functionality
10801 for handling the random seed file.
10803 Use the random seed file in some applications that previously did not:
10805 dsaparam -genkey (which also ignored its '-rand' option),
10808 x509 (when signing).
10809 Except on systems with /dev/urandom, it is crucial to have a random
10810 seed file at least for key creation, DSA signing, and for DH exchanges;
10811 for RSA signatures we could do without one.
10813 gendh and gendsa (unlike genrsa) used to read only the first byte
10814 of each file listed in the '-rand' option. The function as previously
10815 found in genrsa is now in app_rand.c and is used by all programs
10816 that support '-rand'.
10819 *) In RAND_write_file, use mode 0600 for creating files;
10820 don't just chmod when it may be too late.
10823 *) Report an error from X509_STORE_load_locations
10824 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10827 *) New function ASN1_mbstring_copy() this copies a string in either
10828 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10829 into an ASN1_STRING type. A mask of permissible types is passed
10830 and it chooses the "minimal" type to use or an error if not type
10834 *) Add function equivalents to the various macros in asn1.h. The old
10835 macros are retained with an M_ prefix. Code inside the library can
10836 use the M_ macros. External code (including the openssl utility)
10837 should *NOT* in order to be "shared library friendly".
10840 *) Add various functions that can check a certificate's extensions
10841 to see if it usable for various purposes such as SSL client,
10842 server or S/MIME and CAs of these types. This is currently
10843 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10844 verification. Also added a -purpose flag to x509 utility to
10845 print out all the purposes.
10848 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10852 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10853 for, obtain and decode and extension and obtain its critical flag.
10854 This allows all the necessary extension code to be handled in a
10855 single function call.
10858 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10859 platforms. See crypto/rc4/rc4_enc.c for further details.
10862 *) New -noout option to asn1parse. This causes no output to be produced
10863 its main use is when combined with -strparse and -out to extract data
10864 from a file (which may not be in ASN.1 format).
10867 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10868 when producing the local key id.
10869 [Richard Levitte <levitte@stacken.kth.se>]
10871 *) New option -dhparam in s_server. This allows a DH parameter file to be
10872 stated explicitly. If it is not stated then it tries the first server
10873 certificate file. The previous behaviour hard coded the filename
10877 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10878 a public key to be input or output. For example:
10879 openssl rsa -in key.pem -pubout -out pubkey.pem
10880 Also added necessary DSA public key functions to handle this.
10883 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10884 in the message. This was handled by allowing
10885 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10886 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10888 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10889 to the end of the strings whereas this didn't. This would cause problems
10890 if strings read with d2i_ASN1_bytes() were later modified.
10891 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10893 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10894 data and it contains EOF it will end up returning an error. This is
10895 caused by input 46 bytes long. The cause is due to the way base64
10896 BIOs find the start of base64 encoded data. They do this by trying a
10897 trial decode on each line until they find one that works. When they
10898 do a flag is set and it starts again knowing it can pass all the
10899 data directly through the decoder. Unfortunately it doesn't reset
10900 the context it uses. This means that if EOF is reached an attempt
10901 is made to pass two EOFs through the context and this causes the
10902 resulting error. This can also cause other problems as well. As is
10903 usual with these problems it takes *ages* to find and the fix is
10904 trivial: move one line.
10905 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10907 *) Ugly workaround to get s_client and s_server working under Windows. The
10908 old code wouldn't work because it needed to select() on sockets and the
10909 tty (for keypresses and to see if data could be written). Win32 only
10910 supports select() on sockets so we select() with a 1s timeout on the
10911 sockets and then see if any characters are waiting to be read, if none
10912 are present then we retry, we also assume we can always write data to
10913 the tty. This isn't nice because the code then blocks until we've
10914 received a complete line of data and it is effectively polling the
10915 keyboard at 1s intervals: however it's quite a bit better than not
10916 working at all :-) A dedicated Windows application might handle this
10917 with an event loop for example.
10920 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10921 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10922 will be called when RSA_sign() and RSA_verify() are used. This is useful
10923 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10924 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10925 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10926 This necessitated the support of an extra signature type NID_md5_sha1
10927 for SSL signatures and modifications to the SSL library to use it instead
10928 of calling RSA_public_decrypt() and RSA_private_encrypt().
10931 *) Add new -verify -CAfile and -CApath options to the crl program, these
10932 will lookup a CRL issuers certificate and verify the signature in a
10933 similar way to the verify program. Tidy up the crl program so it
10934 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10935 less strict. It will now permit CRL extensions even if it is not
10936 a V2 CRL: this will allow it to tolerate some broken CRLs.
10939 *) Initialize all non-automatic variables each time one of the openssl
10940 sub-programs is started (this is necessary as they may be started
10941 multiple times from the "OpenSSL>" prompt).
10942 [Lennart Bang, Bodo Moeller]
10944 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10945 removing all other RSA functionality (this is what NO_RSA does). This
10946 is so (for example) those in the US can disable those operations covered
10947 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10951 *) Non-copying interface to BIO pairs.
10952 (still largely untested)
10955 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10956 ASCII string. This was handled independently in various places before.
10959 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10960 UTF8 strings a character at a time.
10963 *) Use client_version from client hello to select the protocol
10964 (s23_srvr.c) and for RSA client key exchange verification
10965 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10968 *) Add various utility functions to handle SPKACs, these were previously
10969 handled by poking round in the structure internals. Added new function
10970 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10971 print, verify and generate SPKACs. Based on an original idea from
10972 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10975 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10978 *) Allow the config file extension section to be overwritten on the
10979 command line. Based on an original idea from Massimiliano Pala
10980 <madwolf@comune.modena.it>. The new option is called -extensions
10981 and can be applied to ca, req and x509. Also -reqexts to override
10982 the request extensions in req and -crlexts to override the crl extensions
10986 *) Add new feature to the SPKAC handling in ca. Now you can include
10987 the same field multiple times by preceding it by "XXXX." for example:
10990 this is the same syntax as used in the req config file.
10993 *) Allow certificate extensions to be added to certificate requests. These
10994 are specified in a 'req_extensions' option of the req section of the
10995 config file. They can be printed out with the -text option to req but
10996 are otherwise ignored at present.
10999 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11000 data read consists of only the final block it would not decrypted because
11001 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11002 A misplaced 'break' also meant the decrypted final block might not be
11003 copied until the next read.
11006 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11007 a few extra parameters to the DH structure: these will be useful if
11008 for example we want the value of 'q' or implement X9.42 DH.
11011 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11012 provides hooks that allow the default DSA functions or functions on a
11013 "per key" basis to be replaced. This allows hardware acceleration and
11014 hardware key storage to be handled without major modification to the
11015 library. Also added low level modexp hooks and CRYPTO_EX structure and
11016 associated functions.
11019 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11020 as "read only": it can't be written to and the buffer it points to will
11021 not be freed. Reading from a read only BIO is much more efficient than
11022 a normal memory BIO. This was added because there are several times when
11023 an area of memory needs to be read from a BIO. The previous method was
11024 to create a memory BIO and write the data to it, this results in two
11025 copies of the data and an O(n^2) reading algorithm. There is a new
11026 function BIO_new_mem_buf() which creates a read only memory BIO from
11027 an area of memory. Also modified the PKCS#7 routines to use read only
11031 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11032 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11033 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11034 but a retry condition occurred while trying to read the rest.
11037 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11038 NID_pkcs7_encrypted by default: this was wrong since this should almost
11039 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11040 the encrypted data type: this is a more sensible place to put it and it
11041 allows the PKCS#12 code to be tidied up that duplicated this
11045 *) Changed obj_dat.pl script so it takes its input and output files on
11046 the command line. This should avoid shell escape redirection problems
11050 *) Initial support for certificate extension requests, these are included
11051 in things like Xenroll certificate requests. Included functions to allow
11052 extensions to be obtained and added.
11055 *) -crlf option to s_client and s_server for sending newlines as
11056 CRLF (as required by many protocols).
11059 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11061 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11062 [Ralf S. Engelschall]
11064 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11065 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11067 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11071 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11072 DH parameters/keys (q is lost during that conversion, but the resulting
11073 DH parameters contain its length).
11075 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11076 much faster than DH_generate_parameters (which creates parameters
11077 where p = 2*q + 1), and also the smaller q makes DH computations
11078 much more efficient (160-bit exponentiation instead of 1024-bit
11079 exponentiation); so this provides a convenient way to support DHE
11080 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11081 utter importance to use
11082 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11084 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11085 when such DH parameters are used, because otherwise small subgroup
11086 attacks may become possible!
11089 *) Avoid memory leak in i2d_DHparams.
11092 *) Allow the -k option to be used more than once in the enc program:
11093 this allows the same encrypted message to be read by multiple recipients.
11096 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11097 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11098 it will always use the numerical form of the OID, even if it has a short
11102 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11103 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11104 otherwise bn_mod_exp was called. In the case of hardware keys for example
11105 no private key components need be present and it might store extra data
11106 in the RSA structure, which cannot be accessed from bn_mod_exp.
11107 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11108 private key operations.
11111 *) Added support for SPARC Linux.
11114 *) pem_password_cb function type incompatibly changed from
11115 typedef int pem_password_cb(char *buf, int size, int rwflag);
11117 ....(char *buf, int size, int rwflag, void *userdata);
11118 so that applications can pass data to their callbacks:
11119 The PEM[_ASN1]_{read,write}... functions and macros now take an
11120 additional void * argument, which is just handed through whenever
11121 the password callback is called.
11122 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11124 New function SSL_CTX_set_default_passwd_cb_userdata.
11126 Compatibility note: As many C implementations push function arguments
11127 onto the stack in reverse order, the new library version is likely to
11128 interoperate with programs that have been compiled with the old
11129 pem_password_cb definition (PEM_whatever takes some data that
11130 happens to be on the stack as its last argument, and the callback
11131 just ignores this garbage); but there is no guarantee whatsoever that
11134 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11135 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11136 problems not only on Windows, but also on some Unix platforms.
11137 To avoid problematic command lines, these definitions are now in an
11138 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11139 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11142 *) MIPS III/IV assembler module is reimplemented.
11145 *) More DES library cleanups: remove references to srand/rand and
11146 delete an unused file.
11149 *) Add support for the the free Netwide assembler (NASM) under Win32,
11150 since not many people have MASM (ml) and it can be hard to obtain.
11151 This is currently experimental but it seems to work OK and pass all
11152 the tests. Check out INSTALL.W32 for info.
11155 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11156 without temporary keys kept an extra copy of the server key,
11157 and connections with temporary keys did not free everything in case
11161 *) New function RSA_check_key and new openssl rsa option -check
11162 for verifying the consistency of RSA keys.
11163 [Ulf Moeller, Bodo Moeller]
11165 *) Various changes to make Win32 compile work:
11166 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11167 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11168 comparison" warnings.
11169 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11172 *) Add a debugging option to PKCS#5 v2 key generation function: when
11173 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11174 derived keys are printed to stderr.
11177 *) Copy the flags in ASN1_STRING_dup().
11178 [Roman E. Pavlov <pre@mo.msk.ru>]
11180 *) The x509 application mishandled signing requests containing DSA
11181 keys when the signing key was also DSA and the parameters didn't match.
11183 It was supposed to omit the parameters when they matched the signing key:
11184 the verifying software was then supposed to automatically use the CA's
11185 parameters if they were absent from the end user certificate.
11187 Omitting parameters is no longer recommended. The test was also
11188 the wrong way round! This was probably due to unusual behaviour in
11189 EVP_cmp_parameters() which returns 1 if the parameters match.
11190 This meant that parameters were omitted when they *didn't* match and
11191 the certificate was useless. Certificates signed with 'ca' didn't have
11193 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11195 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11196 The interface is as follows:
11197 Applications can use
11198 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11199 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11200 "off" is now the default.
11201 The library internally uses
11202 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11203 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11204 to disable memory-checking temporarily.
11206 Some inconsistent states that previously were possible (and were
11207 even the default) are now avoided.
11209 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11210 with each memory chunk allocated; this is occasionally more helpful
11211 than just having a counter.
11213 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11215 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11219 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11220 which largely parallels "options", but is for changing API behaviour,
11221 whereas "options" are about protocol behaviour.
11222 Initial "mode" flags are:
11224 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11225 a single record has been written.
11226 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11227 retries use the same buffer location.
11228 (But all of the contents must be
11232 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11235 *) Fix problems with no-hmac etc.
11236 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11238 *) New functions RSA_get_default_method(), RSA_set_method() and
11239 RSA_get_method(). These allows replacement of RSA_METHODs without having
11240 to mess around with the internals of an RSA structure.
11243 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11244 Also really enable memory leak checks in openssl.c and in some
11246 [Chad C. Mulligan, Bodo Moeller]
11248 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11249 up the length of negative integers. This has now been simplified to just
11250 store the length when it is first determined and use it later, rather
11251 than trying to keep track of where data is copied and updating it to
11253 [Steve Henson, reported by Brien Wheeler
11254 <bwheeler@authentica-security.com>]
11256 *) Add a new function PKCS7_signatureVerify. This allows the verification
11257 of a PKCS#7 signature but with the signing certificate passed to the
11258 function itself. This contrasts with PKCS7_dataVerify which assumes the
11259 certificate is present in the PKCS#7 structure. This isn't always the
11260 case: certificates can be omitted from a PKCS#7 structure and be
11261 distributed by "out of band" means (such as a certificate database).
11264 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11265 function prototypes in pem.h, also change util/mkdef.pl to add the
11266 necessary function names.
11269 *) mk1mf.pl (used by Windows builds) did not properly read the
11270 options set by Configure in the top level Makefile, and Configure
11271 was not even able to write more than one option correctly.
11272 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11275 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11276 file to be loaded from a BIO or FILE pointer. The BIO version will
11277 for example allow memory BIOs to contain config info.
11280 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11281 Whoever hopes to achieve shared-library compatibility across versions
11282 must use this, not the compile-time macro.
11283 (Exercise 0.9.4: Which is the minimum library version required by
11285 Note: All this applies only to multi-threaded programs, others don't
11289 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11290 through a BIO pair triggered the default case, i.e.
11291 SSLerr(...,SSL_R_UNKNOWN_STATE).
11294 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11295 can use the SSL library even if none of the specific BIOs is
11299 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11300 for the encoded length.
11301 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11303 *) Add initial documentation of the X509V3 functions.
11306 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11307 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11308 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11309 secure PKCS#8 private key format with a high iteration count.
11312 *) Fix determination of Perl interpreter: A perl or perl5
11313 _directory_ in $PATH was also accepted as the interpreter.
11314 [Ralf S. Engelschall]
11316 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11317 wrong with it but it was very old and did things like calling
11318 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11319 unusual formatting.
11322 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11323 to use the new extension code.
11326 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11327 with macros. This should make it easier to change their form, add extra
11328 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11332 *) Add to configuration table a new entry that can specify an alternative
11333 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11334 according to Mark Crispin <MRC@Panda.COM>.
11338 *) DES CBC did not update the IV. Weird.
11341 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11342 Changing the behaviour of the former might break existing programs --
11343 where IV updating is needed, des_ncbc_encrypt can be used.
11346 *) When bntest is run from "make test" it drives bc to check its
11347 calculations, as well as internally checking them. If an internal check
11348 fails, it needs to cause bc to give a non-zero result or make test carries
11349 on without noticing the failure. Fixed.
11352 *) DES library cleanups.
11355 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11356 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11357 ciphers. NOTE: although the key derivation function has been verified
11358 against some published test vectors it has not been extensively tested
11359 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11363 *) Instead of "mkdir -p", which is not fully portable, use new
11364 Perl script "util/mkdir-p.pl".
11367 *) Rewrite the way password based encryption (PBE) is handled. It used to
11368 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11369 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11370 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11371 the 'parameter' field of the AlgorithmIdentifier is passed to the
11372 underlying key generation function so it must do its own ASN1 parsing.
11373 This has also changed the EVP_PBE_CipherInit() function which now has a
11374 'parameter' argument instead of literal salt and iteration count values
11375 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11378 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11379 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11380 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11381 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11382 value was just used as a "magic string" and not used directly its
11383 value doesn't matter.
11386 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11390 *) "linux-sparc64" configuration (ultrapenguin).
11391 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11392 "linux-sparc" configuration.
11393 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11395 *) config now generates no-xxx options for missing ciphers.
11398 *) Support the EBCDIC character set (work in progress).
11399 File ebcdic.c not yet included because it has a different license.
11400 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11402 *) Support BS2000/OSD-POSIX.
11403 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11405 *) Make callbacks for key generation use void * instead of char *.
11408 *) Make S/MIME samples compile (not yet tested).
11411 *) Additional typesafe stacks.
11414 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11418 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11420 *) New configuration variant "sco5-gcc".
11422 *) Updated some demos.
11423 [Sean O Riordain, Wade Scholine]
11425 *) Add missing BIO_free at exit of pkcs12 application.
11428 *) Fix memory leak in conf.c.
11431 *) Updates for Win32 to assembler version of MD5.
11434 *) Set #! path to perl in apps/der_chop to where we found it
11435 instead of using a fixed path.
11438 *) SHA library changes for irix64-mips4-cc.
11441 *) Improvements for VMS support.
11445 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11447 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11448 This also avoids the problems with SC4.2 and unpatched SC5.
11449 [Andy Polyakov <appro@fy.chalmers.se>]
11451 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11452 These are required because of the typesafe stack would otherwise break
11453 existing code. If old code used a structure member which used to be STACK
11454 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11455 sk_num or sk_value it would produce an error because the num, data members
11456 are not present in STACK_OF. Now it just produces a warning. sk_set
11457 replaces the old method of assigning a value to sk_value
11458 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11459 that does this will no longer work (and should use sk_set instead) but
11460 this could be regarded as a "questionable" behaviour anyway.
11463 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11464 correctly handle encrypted S/MIME data.
11467 *) Change type of various DES function arguments from des_cblock
11468 (which means, in function argument declarations, pointer to char)
11469 to des_cblock * (meaning pointer to array with 8 char elements),
11470 which allows the compiler to do more typechecking; it was like
11471 that back in SSLeay, but with lots of ugly casts.
11473 Introduce new type const_des_cblock.
11476 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11477 problems: find RecipientInfo structure that matches recipient certificate
11478 and initialise the ASN1 structures properly based on passed cipher.
11481 *) Belatedly make the BN tests actually check the results.
11484 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11485 to and from BNs: it was completely broken. New compilation option
11486 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11487 key elements as negative integers.
11490 *) Reorganize and speed up MD5.
11491 [Andy Polyakov <appro@fy.chalmers.se>]
11494 [Richard Levitte <richard@levitte.org>]
11496 *) New option -out to asn1parse to allow the parsed structure to be
11497 output to a file. This is most useful when combined with the -strparse
11498 option to examine the output of things like OCTET STRINGS.
11501 *) Make SSL library a little more fool-proof by not requiring any longer
11502 that SSL_set_{accept,connect}_state be called before
11503 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11504 in many applications because usually everything *appeared* to work as
11505 intended anyway -- now it really works as intended).
11508 *) Move openssl.cnf out of lib/.
11511 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11512 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11513 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11514 [Ralf S. Engelschall]
11516 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11517 handle PKCS#7 enveloped data properly.
11518 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11520 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11521 copying pointers. The cert_st handling is changed by this in
11522 various ways (and thus what used to be known as ctx->default_cert
11523 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11524 any longer when s->cert does not give us what we need).
11525 ssl_cert_instantiate becomes obsolete by this change.
11526 As soon as we've got the new code right (possibly it already is?),
11527 we have solved a couple of bugs of the earlier code where s->cert
11528 was used as if it could not have been shared with other SSL structures.
11530 Note that using the SSL API in certain dirty ways now will result
11531 in different behaviour than observed with earlier library versions:
11532 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11533 does not influence s as it used to.
11535 In order to clean up things more thoroughly, inside SSL_SESSION
11536 we don't use CERT any longer, but a new structure SESS_CERT
11537 that holds per-session data (if available); currently, this is
11538 the peer's certificate chain and, for clients, the server's certificate
11539 and temporary key. CERT holds only those values that can have
11540 meaningful defaults in an SSL_CTX.
11543 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11544 from the internal representation. Various PKCS#7 fixes: remove some
11545 evil casts and set the enc_dig_alg field properly based on the signing
11549 *) Allow PKCS#12 password to be set from the command line or the
11550 environment. Let 'ca' get its config file name from the environment
11551 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11555 *) Allow certificate policies extension to use an IA5STRING for the
11556 organization field. This is contrary to the PKIX definition but
11557 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11561 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11562 without disallowing inline assembler and the like for non-pedantic builds.
11565 *) Support Borland C++ builder.
11566 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11568 *) Support Mingw32.
11571 *) SHA-1 cleanups and performance enhancements.
11572 [Andy Polyakov <appro@fy.chalmers.se>]
11574 *) Sparc v8plus assembler for the bignum library.
11575 [Andy Polyakov <appro@fy.chalmers.se>]
11577 *) Accept any -xxx and +xxx compiler options in Configure.
11580 *) Update HPUX configuration.
11583 *) Add missing sk_<type>_unshift() function to safestack.h
11584 [Ralf S. Engelschall]
11586 *) New function SSL_CTX_use_certificate_chain_file that sets the
11587 "extra_cert"s in addition to the certificate. (This makes sense
11588 only for "PEM" format files, as chains as a whole are not
11592 *) Support verify_depth from the SSL API.
11593 x509_vfy.c had what can be considered an off-by-one-error:
11594 Its depth (which was not part of the external interface)
11595 was actually counting the number of certificates in a chain;
11596 now it really counts the depth.
11599 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11600 instead of X509err, which often resulted in confusing error
11601 messages since the error codes are not globally unique
11602 (e.g. an alleged error in ssl3_accept when a certificate
11603 didn't match the private key).
11605 *) New function SSL_CTX_set_session_id_context that allows to set a default
11606 value (so that you don't need SSL_set_session_id_context for each
11607 connection using the SSL_CTX).
11610 *) OAEP decoding bug fix.
11613 *) Support INSTALL_PREFIX for package builders, as proposed by
11617 *) New Configure options "threads" and "no-threads". For systems
11618 where the proper compiler options are known (currently Solaris
11619 and Linux), "threads" is the default.
11622 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11625 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11626 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11627 such as /usr/local/bin.
11630 *) "make linux-shared" to build shared libraries.
11631 [Niels Poppe <niels@netbox.org>]
11633 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11636 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11637 extension adding in x509 utility.
11640 *) Remove NOPROTO sections and error code comments.
11643 *) Partial rewrite of the DEF file generator to now parse the ANSI
11647 *) New Configure options --prefix=DIR and --openssldir=DIR.
11650 *) Complete rewrite of the error code script(s). It is all now handled
11651 by one script at the top level which handles error code gathering,
11652 header rewriting and C source file generation. It should be much better
11653 than the old method: it now uses a modified version of Ulf's parser to
11654 read the ANSI prototypes in all header files (thus the old K&R definitions
11655 aren't needed for error creation any more) and do a better job of
11656 translating function codes into names. The old 'ASN1 error code imbedded
11657 in a comment' is no longer necessary and it doesn't use .err files which
11658 have now been deleted. Also the error code call doesn't have to appear all
11659 on one line (which resulted in some large lines...).
11662 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11665 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11666 0 (which usually indicates a closed connection), but continue reading.
11669 *) Fix some race conditions.
11672 *) Add support for CRL distribution points extension. Add Certificate
11673 Policies and CRL distribution points documentation.
11676 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11679 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11680 8 of keying material. Merlin has also confirmed interop with this fix
11681 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11682 [Merlin Hughes <merlin@baltimore.ie>]
11684 *) Fix lots of warnings.
11685 [Richard Levitte <levitte@stacken.kth.se>]
11687 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11688 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11689 [Richard Levitte <levitte@stacken.kth.se>]
11691 *) Fix problems with sizeof(long) == 8.
11692 [Andy Polyakov <appro@fy.chalmers.se>]
11694 *) Change functions to ANSI C.
11697 *) Fix typos in error codes.
11698 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11700 *) Remove defunct assembler files from Configure.
11703 *) SPARC v8 assembler BIGNUM implementation.
11704 [Andy Polyakov <appro@fy.chalmers.se>]
11706 *) Support for Certificate Policies extension: both print and set.
11707 Various additions to support the r2i method this uses.
11710 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11711 return a const string when you are expecting an allocated buffer.
11714 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11715 types DirectoryString and DisplayText.
11718 *) Add code to allow r2i extensions to access the configuration database,
11719 add an LHASH database driver and add several ctx helper functions.
11722 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11723 fail when they extended the size of a BIGNUM.
11726 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11727 support typesafe stack.
11730 *) Fix typo in SSL_[gs]et_options().
11731 [Nils Frostberg <nils@medcom.se>]
11733 *) Delete various functions and files that belonged to the (now obsolete)
11734 old X509V3 handling code.
11737 *) New Configure option "rsaref".
11740 *) Don't auto-generate pem.h.
11743 *) Introduce type-safe ASN.1 SETs.
11746 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11747 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11749 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11750 that links with OpenSSL (well at least cause lots of warnings), but fear
11751 not: the conversion is trivial, and it eliminates loads of evil casts. A
11752 few STACKed things have been converted already. Feel free to convert more.
11753 In the fullness of time, I'll do away with the STACK type altogether.
11756 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11757 specified in <certfile> by updating the entry in the index.txt file.
11758 This way one no longer has to edit the index.txt file manually for
11759 revoking a certificate. The -revoke option does the gory details now.
11760 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11762 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11763 `-text' option at all and this way the `-noout -text' combination was
11764 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11765 [Ralf S. Engelschall]
11767 *) Make sure a corresponding plain text error message exists for the
11768 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11769 verify callback function determined that a certificate was revoked.
11770 [Ralf S. Engelschall]
11772 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11773 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11774 all available cipers including rc5, which was forgotten until now.
11775 In order to let the testing shell script know which algorithms
11776 are available, a new (up to now undocumented) command
11777 "openssl list-cipher-commands" is used.
11780 *) Bugfix: s_client occasionally would sleep in select() when
11781 it should have checked SSL_pending() first.
11784 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11785 the raw DSA values prior to ASN.1 encoding.
11788 *) Tweaks to Configure
11789 [Niels Poppe <niels@netbox.org>]
11791 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11795 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11798 *) New config option to avoid instructions that are illegal on the 80386.
11799 The default code is faster, but requires at least a 486.
11802 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11803 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11804 same as SSL2_VERSION anyway.
11807 *) New "-showcerts" option for s_client.
11810 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11811 application. Various cleanups and fixes.
11814 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11815 modify error routines to work internally. Add error codes and PBE init
11816 to library startup routines.
11819 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11820 packing functions to asn1 and evp. Changed function names and error
11821 codes along the way.
11824 *) PKCS12 integration: and so it begins... First of several patches to
11825 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11826 objects to objects.h
11829 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11830 and display support for Thawte strong extranet extension.
11833 *) Add LinuxPPC support.
11834 [Jeff Dubrule <igor@pobox.org>]
11836 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11837 bn_div_words in alpha.s.
11838 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11840 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11841 OAEP isn't supported when OpenSSL is built with RSAref.
11842 [Ulf Moeller <ulf@fitug.de>]
11844 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11845 so they no longer are missing under -DNOPROTO.
11846 [Soren S. Jorvang <soren@t.dk>]
11849 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11851 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11852 doesn't work when the session is reused. Coming soon!
11855 *) Fix a security hole, that allows sessions to be reused in the wrong
11856 context thus bypassing client cert protection! All software that uses
11857 client certs and session caches in multiple contexts NEEDS PATCHING to
11858 allow session reuse! A fuller solution is in the works.
11859 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11861 *) Some more source tree cleanups (removed obsolete files
11862 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11863 permission on "config" script to be executable) and a fix for the INSTALL
11865 [Ulf Moeller <ulf@fitug.de>]
11867 *) Remove some legacy and erroneous uses of malloc, free instead of
11869 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11871 *) Make rsa_oaep_test return non-zero on error.
11872 [Ulf Moeller <ulf@fitug.de>]
11874 *) Add support for native Solaris shared libraries. Configure
11875 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11876 if someone would make that last step automatic.
11877 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11879 *) ctx_size was not built with the right compiler during "make links". Fixed.
11882 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11883 except NULL ciphers". This means the default cipher list will no longer
11884 enable NULL ciphers. They need to be specifically enabled e.g. with
11885 the string "DEFAULT:eNULL".
11888 *) Fix to RSA private encryption routines: if p < q then it would
11889 occasionally produce an invalid result. This will only happen with
11890 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11893 *) Be less restrictive and allow also `perl util/perlpath.pl
11894 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11895 because this way one can also use an interpreter named `perl5' (which is
11896 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11897 installed as `perl').
11898 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11900 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11901 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11903 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11904 advapi32.lib to Win32 build and change the pem test comparison
11905 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11906 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11907 and crypto/des/ede_cbcm_enc.c.
11910 *) DES quad checksum was broken on big-endian architectures. Fixed.
11913 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11914 Win32 test batch file so it (might) work again. The Win32 test batch file
11915 is horrible: I feel ill....
11918 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11919 in e_os.h. Audit of header files to check ANSI and non ANSI
11920 sections: 10 functions were absent from non ANSI section and not exported
11921 from Windows DLLs. Fixed up libeay.num for new functions.
11924 *) Make `openssl version' output lines consistent.
11925 [Ralf S. Engelschall]
11927 *) Fix Win32 symbol export lists for BIO functions: Added
11928 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11929 to ms/libeay{16,32}.def.
11930 [Ralf S. Engelschall]
11932 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11933 fine under Unix and passes some trivial tests I've now added. But the
11934 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11935 added to make sure no one expects that this stuff really works in the
11936 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11937 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11939 [Ralf S. Engelschall]
11941 *) Fix the generation of two part addresses in perl.
11942 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11944 *) Add config entry for Linux on MIPS.
11945 [John Tobey <jtobey@channel1.com>]
11947 *) Make links whenever Configure is run, unless we are on Windoze.
11950 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11951 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11955 *) Add a useful kludge to allow package maintainers to specify compiler and
11956 other platforms details on the command line without having to patch the
11957 Configure script everytime: One now can use ``perl Configure
11958 <id>:<details>'', i.e. platform ids are allowed to have details appended
11959 to them (separated by colons). This is treated as there would be a static
11960 pre-configured entry in Configure's %table under key <id> with value
11961 <details> and ``perl Configure <id>'' is called. So, when you want to
11962 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11963 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11964 now, which overrides the FreeBSD-elf entry on-the-fly.
11965 [Ralf S. Engelschall]
11967 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11970 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11971 on the `perl Configure ...' command line. This way one can compile
11972 OpenSSL libraries with Position Independent Code (PIC) which is needed
11973 for linking it into DSOs.
11974 [Ralf S. Engelschall]
11976 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11980 *) Cleaned up the LICENSE document: The official contact for any license
11981 questions now is the OpenSSL core team under openssl-core@openssl.org.
11982 And add a paragraph about the dual-license situation to make sure people
11983 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11984 to the OpenSSL toolkit.
11985 [Ralf S. Engelschall]
11987 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11988 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11989 Additionally cleaned up the `make links' target: Remove unnecessary
11990 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11991 to speed processing and no longer clutter the display with confusing
11992 stuff. Instead only the actually done links are displayed.
11993 [Ralf S. Engelschall]
11995 *) Permit null encryption ciphersuites, used for authentication only. It used
11996 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11997 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12001 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12002 signed attributes when verifying signatures (this would break them),
12003 the detached data encoding was wrong and public keys obtained using
12004 X509_get_pubkey() weren't freed.
12007 *) Add text documentation for the BUFFER functions. Also added a work around
12008 to a Win95 console bug. This was triggered by the password read stuff: the
12009 last character typed gets carried over to the next fread(). If you were
12010 generating a new cert request using 'req' for example then the last
12011 character of the passphrase would be CR which would then enter the first
12015 *) Added the new `Includes OpenSSL Cryptography Software' button as
12016 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12017 button and can be used by applications based on OpenSSL to show the
12018 relationship to the OpenSSL project.
12019 [Ralf S. Engelschall]
12021 *) Remove confusing variables in function signatures in files
12022 ssl/ssl_lib.c and ssl/ssl.h.
12023 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12025 *) Don't install bss_file.c under PREFIX/include/
12026 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12028 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12029 functions that return function pointers and has support for NT specific
12030 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12031 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12032 unsigned to signed types: this was killing the Win32 compile.
12035 *) Add new certificate file to stack functions,
12036 SSL_add_dir_cert_subjects_to_stack() and
12037 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12038 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12039 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12040 This means that Apache-SSL and similar packages don't have to mess around
12041 to add as many CAs as they want to the preferred list.
12044 *) Experiment with doxygen documentation. Currently only partially applied to
12046 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12047 openssl.doxy as the configuration file.
12050 *) Get rid of remaining C++-style comments which strict C compilers hate.
12051 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12053 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12054 compiled in by default: it has problems with large keys.
12057 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12058 DH private keys and/or callback functions which directly correspond to
12059 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12060 is needed for applications which have to configure certificates on a
12061 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12063 For the RSA certificate situation is makes no difference, but
12064 for the DSA certificate situation this fixes the "no shared cipher"
12065 problem where the OpenSSL cipher selection procedure failed because the
12066 temporary keys were not overtaken from the context and the API provided
12067 no way to reconfigure them.
12068 The new functions now let applications reconfigure the stuff and they
12069 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12070 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12071 non-public-API function ssl_cert_instantiate() is used as a helper
12072 function and also to reduce code redundancy inside ssl_rsa.c.
12073 [Ralf S. Engelschall]
12075 *) Move s_server -dcert and -dkey options out of the undocumented feature
12076 area because they are useful for the DSA situation and should be
12077 recognized by the users.
12078 [Ralf S. Engelschall]
12080 *) Fix the cipher decision scheme for export ciphers: the export bits are
12081 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12082 SSL_EXP_MASK. So, the original variable has to be used instead of the
12083 already masked variable.
12084 [Richard Levitte <levitte@stacken.kth.se>]
12086 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12087 [Richard Levitte <levitte@stacken.kth.se>]
12089 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12090 from `int' to `unsigned int' because it's a length and initialized by
12091 EVP_DigestFinal() which expects an `unsigned int *'.
12092 [Richard Levitte <levitte@stacken.kth.se>]
12094 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12095 script. Instead use the usual Shell->Perl transition trick.
12096 [Ralf S. Engelschall]
12098 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12099 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12100 -noout -modulus' as it's already the case for `openssl rsa -noout
12101 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12102 currently the public key is printed (a decision which was already done by
12103 `openssl dsa -modulus' in the past) which serves a similar purpose.
12104 Additionally the NO_RSA no longer completely removes the whole -modulus
12105 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12107 [Ralf S. Engelschall]
12109 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12110 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12111 [Arne Ansper <arne@ats.cyber.ee>]
12113 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12114 to be added. Now both 'req' and 'ca' can use new objects defined in the
12118 *) Add cool BIO that does syslog (or event log on NT).
12119 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12121 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12122 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12123 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12124 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12127 *) Add preliminary config info for new extension code.
12130 *) Make RSA_NO_PADDING really use no padding.
12131 [Ulf Moeller <ulf@fitug.de>]
12133 *) Generate errors when private/public key check is done.
12136 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12137 for some CRL extensions and new objects added.
12140 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12141 key usage extension and fuller support for authority key id.
12144 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12145 padding method for RSA, which is recommended for new applications in PKCS
12146 #1 v2.0 (RFC 2437, October 1998).
12147 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12148 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12149 against Bleichbacher's attack on RSA.
12150 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12153 *) Updates to the new SSL compression code
12154 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12156 *) Fix so that the version number in the master secret, when passed
12157 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12158 (because the server will not accept higher), that the version number
12159 is 0x03,0x01, not 0x03,0x00
12160 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12162 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12163 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12164 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12167 *) Support for RAW extensions where an arbitrary extension can be
12168 created by including its DER encoding. See apps/openssl.cnf for
12172 *) Make sure latest Perl versions don't interpret some generated C array
12173 code as Perl array code in the crypto/err/err_genc.pl script.
12174 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12176 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12177 not many people have the assembler. Various Win32 compilation fixes and
12178 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12179 build instructions.
12182 *) Modify configure script 'Configure' to automatically create crypto/date.h
12183 file under Win32 and also build pem.h from pem.org. New script
12184 util/mkfiles.pl to create the MINFO file on environments that can't do a
12185 'make files': perl util/mkfiles.pl >MINFO should work.
12188 *) Major rework of DES function declarations, in the pursuit of correctness
12189 and purity. As a result, many evil casts evaporated, and some weirdness,
12190 too. You may find this causes warnings in your code. Zapping your evil
12191 casts will probably fix them. Mostly.
12194 *) Fix for a typo in asn1.h. Bug fix to object creation script
12195 obj_dat.pl. It considered a zero in an object definition to mean
12196 "end of object": none of the objects in objects.h have any zeros
12197 so it wasn't spotted.
12198 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12200 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12201 Masking (CBCM). In the absence of test vectors, the best I have been able
12202 to do is check that the decrypt undoes the encrypt, so far. Send me test
12203 vectors if you have them.
12206 *) Correct calculation of key length for export ciphers (too much space was
12207 allocated for null ciphers). This has not been tested!
12210 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12211 message is now correct (it understands "crypto" and "ssl" on its
12212 command line). There is also now an "update" option. This will update
12213 the util/ssleay.num and util/libeay.num files with any new functions.
12215 perl util/mkdef.pl crypto ssl update
12216 it will update them.
12219 *) Overhauled the Perl interface (perl/*):
12220 - ported BN stuff to OpenSSL's different BN library
12221 - made the perl/ source tree CVS-aware
12222 - renamed the package from SSLeay to OpenSSL (the files still contain
12223 their history because I've copied them in the repository)
12224 - removed obsolete files (the test scripts will be replaced
12225 by better Test::Harness variants in the future)
12226 [Ralf S. Engelschall]
12228 *) First cut for a very conservative source tree cleanup:
12229 1. merge various obsolete readme texts into doc/ssleay.txt
12230 where we collect the old documents and readme texts.
12231 2. remove the first part of files where I'm already sure that we no
12232 longer need them because of three reasons: either they are just temporary
12233 files which were left by Eric or they are preserved original files where
12234 I've verified that the diff is also available in the CVS via "cvs diff
12235 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12236 the crypto/md/ stuff).
12237 [Ralf S. Engelschall]
12239 *) More extension code. Incomplete support for subject and issuer alt
12240 name, issuer and authority key id. Change the i2v function parameters
12241 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12242 what that's for :-) Fix to ASN1 macro which messed up
12243 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12246 *) Preliminary support for ENUMERATED type. This is largely copied from the
12250 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12251 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12253 *) Make sure `make rehash' target really finds the `openssl' program.
12254 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12256 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12257 like to hear about it if this slows down other processors.
12260 *) Add CygWin32 platform information to Configure script.
12261 [Alan Batie <batie@aahz.jf.intel.com>]
12263 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12264 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12266 *) New program nseq to manipulate netscape certificate sequences
12269 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12273 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12274 but the BN code had some problems that would cause failures when
12275 doing certificate verification and some other functions.
12276 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12278 *) Add ASN1 and PEM code to support netscape certificate sequences.
12281 *) Add ASN1 and PEM code to support netscape certificate sequences.
12284 *) Add several PKIX and private extended key usage OIDs.
12287 *) Modify the 'ca' program to handle the new extension code. Modify
12288 openssl.cnf for new extension format, add comments.
12291 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12292 and add a sample to openssl.cnf so req -x509 now adds appropriate
12296 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12297 error code, add initial support to X509_print() and x509 application.
12300 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12301 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12302 stuff is currently isolated and isn't even compiled yet.
12305 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12306 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12307 Removed the versions check from X509 routines when loading extensions:
12308 this allows certain broken certificates that don't set the version
12309 properly to be processed.
12312 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12313 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12314 can still be regenerated with "make depend".
12317 *) Spelling mistake in C version of CAST-128.
12318 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12320 *) Changes to the error generation code. The perl script err-code.pl
12321 now reads in the old error codes and retains the old numbers, only
12322 adding new ones if necessary. It also only changes the .err files if new
12323 codes are added. The makefiles have been modified to only insert errors
12324 when needed (to avoid needlessly modifying header files). This is done
12325 by only inserting errors if the .err file is newer than the auto generated
12326 C file. To rebuild all the error codes from scratch (the old behaviour)
12327 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12328 or delete all the .err files.
12331 *) CAST-128 was incorrectly implemented for short keys. The C version has
12332 been fixed, but is untested. The assembler versions are also fixed, but
12333 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12334 to regenerate it if needed.
12335 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12336 Hagino <itojun@kame.net>]
12338 *) File was opened incorrectly in randfile.c.
12339 [Ulf Möller <ulf@fitug.de>]
12341 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12342 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12343 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12344 al: it's just almost always a UTCTime. Note this patch adds new error
12345 codes so do a "make errors" if there are problems.
12348 *) Correct Linux 1 recognition in config.
12349 [Ulf Möller <ulf@fitug.de>]
12351 *) Remove pointless MD5 hash when using DSA keys in ca.
12352 [Anonymous <nobody@replay.com>]
12354 *) Generate an error if given an empty string as a cert directory. Also
12355 generate an error if handed NULL (previously returned 0 to indicate an
12356 error, but didn't set one).
12357 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12359 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12362 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12363 parameters. This was causing a warning which killed off the Win32 compile.
12366 *) Remove C++ style comments from crypto/bn/bn_local.h.
12367 [Neil Costigan <neil.costigan@celocom.com>]
12369 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12370 based on a text string, looking up short and long names and finally
12371 "dot" format. The "dot" format stuff didn't work. Added new function
12372 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12373 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12374 OID is not part of the table.
12377 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12378 X509_LOOKUP_by_alias().
12381 *) Sort openssl functions by name.
12384 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12385 encryption from sample DSA keys (in case anyone is interested the password
12389 *) Make _all_ *_free functions accept a NULL pointer.
12390 [Frans Heymans <fheymans@isaserver.be>]
12392 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12394 [Anonymous <nobody@replay.com>]
12396 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12397 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12399 *) Don't blow it for numeric -newkey arguments to apps/req.
12400 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12402 *) Temp key "for export" tests were wrong in s3_srvr.c.
12403 [Anonymous <nobody@replay.com>]
12405 *) Add prototype for temp key callback functions
12406 SSL_CTX_set_tmp_{rsa,dh}_callback().
12409 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12410 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12413 *) X509_name_add_entry() freed the wrong thing after an error.
12414 [Arne Ansper <arne@ats.cyber.ee>]
12416 *) rsa_eay.c would attempt to free a NULL context.
12417 [Arne Ansper <arne@ats.cyber.ee>]
12419 *) BIO_s_socket() had a broken should_retry() on Windoze.
12420 [Arne Ansper <arne@ats.cyber.ee>]
12422 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12423 [Arne Ansper <arne@ats.cyber.ee>]
12425 *) Make sure the already existing X509_STORE->depth variable is initialized
12426 in X509_STORE_new(), but document the fact that this variable is still
12427 unused in the certificate verification process.
12428 [Ralf S. Engelschall]
12430 *) Fix the various library and apps files to free up pkeys obtained from
12431 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12434 *) Fix reference counting in X509_PUBKEY_get(). This makes
12435 demos/maurice/example2.c work, amongst others, probably.
12436 [Steve Henson and Ben Laurie]
12438 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12439 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12440 are no longer created. This way we have a single and consistent command
12441 line interface `openssl <command>', similar to `cvs <command>'.
12442 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12444 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12445 BIT STRING wrapper always have zero unused bits.
12448 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12451 *) Make the top-level INSTALL documentation easier to understand.
12454 *) Makefiles updated to exit if an error occurs in a sub-directory
12455 make (including if user presses ^C) [Paul Sutton]
12457 *) Make Montgomery context stuff explicit in RSA data structure.
12460 *) Fix build order of pem and err to allow for generated pem.h.
12463 *) Fix renumbering bug in X509_NAME_delete_entry().
12466 *) Enhanced the err-ins.pl script so it makes the error library number
12467 global and can add a library name. This is needed for external ASN1 and
12468 other error libraries.
12471 *) Fixed sk_insert which never worked properly.
12474 *) Fix ASN1 macros so they can handle indefinite length constructed
12475 EXPLICIT tags. Some non standard certificates use these: they can now
12479 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12480 into a single doc/ssleay.txt bundle. This way the information is still
12481 preserved but no longer messes up this directory. Now it's new room for
12482 the new set of documentation files.
12483 [Ralf S. Engelschall]
12485 *) SETs were incorrectly DER encoded. This was a major pain, because they
12486 shared code with SEQUENCEs, which aren't coded the same. This means that
12487 almost everything to do with SETs or SEQUENCEs has either changed name or
12488 number of arguments.
12489 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12491 *) Fix test data to work with the above.
12494 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12495 was already fixed by Eric for 0.9.1 it seems.
12496 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12498 *) Autodetect FreeBSD3.
12501 *) Fix various bugs in Configure. This affects the following platforms:
12505 unixware-2.0-pentium
12509 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12510 before they are needed.
12513 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12517 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12519 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12520 changed SSLeay to OpenSSL in version strings.
12521 [Ralf S. Engelschall]
12523 *) Some fixups to the top-level documents.
12526 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12527 because the symlink to include/ was missing.
12528 [Ralf S. Engelschall]
12530 *) Incorporated the popular no-RSA/DSA-only patches
12531 which allow to compile a RSA-free SSLeay.
12532 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12534 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12535 when "ssleay" is still not found.
12536 [Ralf S. Engelschall]
12538 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12539 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12541 *) Updated the README file.
12542 [Ralf S. Engelschall]
12544 *) Added various .cvsignore files in the CVS repository subdirs
12545 to make a "cvs update" really silent.
12546 [Ralf S. Engelschall]
12548 *) Recompiled the error-definition header files and added
12549 missing symbols to the Win32 linker tables.
12550 [Ralf S. Engelschall]
12552 *) Cleaned up the top-level documents;
12553 o new files: CHANGES and LICENSE
12554 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12555 o merged COPYRIGHT into LICENSE
12556 o removed obsolete TODO file
12557 o renamed MICROSOFT to INSTALL.W32
12558 [Ralf S. Engelschall]
12560 *) Removed dummy files from the 0.9.1b source tree:
12561 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12562 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12563 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12564 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12565 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12566 [Ralf S. Engelschall]
12568 *) Added various platform portability fixes.
12571 *) The Genesis of the OpenSSL rpject:
12572 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12573 Young and Tim J. Hudson created while they were working for C2Net until
12575 [The OpenSSL Project]
12578 Changes between 0.9.0b and 0.9.1b [not released]
12580 *) Updated a few CA certificates under certs/
12583 *) Changed some BIGNUM api stuff.
12586 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12587 DGUX x86, Linux Alpha, etc.
12590 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12591 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12595 *) Add -strparse option to asn1pars program which parses nested
12597 [Dr Stephen Henson <shenson@bigfoot.com>]
12599 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12602 *) DSA fix for "ca" program.
12605 *) Added "-genkey" option to "dsaparam" program.
12608 *) Added RIPE MD160 (rmd160) message digest.
12611 *) Added -a (all) option to "ssleay version" command.
12614 *) Added PLATFORM define which is the id given to Configure.
12617 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12620 *) Extended the ASN.1 parser routines.
12623 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12626 *) Added a BN_CTX to the BN library.
12629 *) Fixed the weak key values in DES library
12632 *) Changed API in EVP library for cipher aliases.
12635 *) Added support for RC2/64bit cipher.
12638 *) Converted the lhash library to the crypto/mem.c functions.
12641 *) Added more recognized ASN.1 object ids.
12644 *) Added more RSA padding checks for SSL/TLS.
12647 *) Added BIO proxy/filter functionality.
12650 *) Added extra_certs to SSL_CTX which can be used
12651 send extra CA certificates to the client in the CA cert chain sending
12652 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12655 *) Now Fortezza is denied in the authentication phase because
12656 this is key exchange mechanism is not supported by SSLeay at all.
12659 *) Additional PKCS1 checks.
12662 *) Support the string "TLSv1" for all TLS v1 ciphers.
12665 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12666 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12669 *) Fixed a few memory leaks.
12672 *) Fixed various code and comment typos.
12675 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12676 bytes sent in the client random.
12677 [Edward Bishop <ebishop@spyglass.com>]