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 multi-prime RSA (RFC 8017) support.
15 *) Add SM3 implemented according to GB/T 32905-2016
16 [ Jack Lloyd <jack.lloyd@ribose.com>,
17 Ronald Tse <ronald.tse@ribose.com>,
18 Erick Borsboom <erick.borsboom@ribose.com> ]
20 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
21 as documented in RFC6066.
22 Based on a patch from Tomasz Moń
23 [Filipe Raimundo da Silva]
25 *) Add SM4 implemented according to GB/T 32907-2016.
26 [ Jack Lloyd <jack.lloyd@ribose.com>,
27 Ronald Tse <ronald.tse@ribose.com>,
28 Erick Borsboom <erick.borsboom@ribose.com> ]
30 *) Reimplement -newreq-nodes and ERR_error_string_n; the
31 original author does not agree with the license change.
34 *) Add ARIA AEAD TLS support.
37 *) Some macro definitions to support VS6 have been removed. Visual
38 Studio 6 has not worked since 1.1.0
41 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
42 without clearing the errors.
45 *) Add "atfork" functions. If building on a system that without
46 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
47 requirements. The RAND facility now uses/requires this.
53 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
54 not possible to disable entirely. However, it's still possible to
55 disable the console reading UI method, UI_OpenSSL() (use UI_null()
58 To disable, configure with 'no-ui-console'. 'no-ui' is still
59 possible to use as an alias. Check at compile time with the
60 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
61 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
64 *) Add a STORE module, which implements a uniform and URI based reader of
65 stores that can contain keys, certificates, CRLs and numerous other
66 objects. The main API is loosely based on a few stdio functions,
67 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
68 OSSL_STORE_error and OSSL_STORE_close.
69 The implementation uses backends called "loaders" to implement arbitrary
70 URI schemes. There is one built in "loader" for the 'file' scheme.
73 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
74 then adjusted to work on FreeBSD 8.4 as well.
75 Enable by configuring with 'enable-devcryptoeng'. This is done by default
76 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
79 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
80 util/mkerr.pl, which is adapted to allow those prefixes, leading to
81 error code calls like this:
83 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
85 With this change, we claim the namespaces OSSL and OPENSSL in a manner
86 that can be encoded in C. For the foreseeable future, this will only
88 [Richard Levitte and Tim Hudson]
90 *) Removed BSD cryptodev engine.
93 *) Add a build target 'build_all_generated', to build all generated files
94 and only that. This can be used to prepare everything that requires
95 things like perl for a system that lacks perl and then move everything
96 to that system and do the rest of the build there.
99 *) In the UI interface, make it possible to duplicate the user data. This
100 can be used by engines that need to retain the data for a longer time
101 than just the call where this user data is passed.
104 *) Ignore the '-named_curve auto' value for compatibility of applications
106 [Tomas Mraz <tmraz@fedoraproject.org>]
108 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
109 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
110 alerts across multiple records (some of which could be empty). In practice
111 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
112 prohibts this altogether and other libraries (BoringSSL, NSS) do not
113 support this at all. Supporting it adds significant complexity to the
114 record layer, and its removal is unlikely to cause inter-operability
118 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
119 with Z. These are meant to replace LONG and ZLONG and to be size safe.
120 The use of LONG and ZLONG is discouraged and scheduled for deprecation
124 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
125 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
126 [Richard Levitte, Andy Polyakov]
128 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
132 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
133 platform rather than 'mingw'.
136 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
137 success if they are asked to add an object which already exists
138 in the store. This change cascades to other functions which load
139 certificates and CRLs.
142 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
143 facilitate stack unwinding even from assembly subroutines.
146 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
147 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
150 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
151 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
152 which is the minimum version we support.
155 *) Certificate time validation (X509_cmp_time) enforces stricter
156 compliance with RFC 5280. Fractional seconds and timezone offsets
157 are no longer allowed.
160 *) Add support for ARIA
163 *) s_client will now send the Server Name Indication (SNI) extension by
164 default unless the new "-noservername" option is used. The server name is
165 based on the host provided to the "-connect" option unless overridden by
169 *) Add support for SipHash
172 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
173 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
174 prevent issues where no progress is being made and the peer continually
175 sends unrecognised record types, using up resources processing them.
178 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
179 using the algorithm defined in
180 https://www.akkadia.org/drepper/SHA-crypt.txt
183 *) Heartbeat support has been removed; the ABI is changed for now.
184 [Richard Levitte, Rich Salz]
186 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
189 *) The RSA "null" method, which was partially supported to avoid patent
190 issues, has been replaced to always returns NULL.
194 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
196 *) bn_sqrx8x_internal carry bug on x86_64
198 There is a carry propagating bug in the x86_64 Montgomery squaring
199 procedure. No EC algorithms are affected. Analysis suggests that attacks
200 against RSA and DSA as a result of this defect would be very difficult to
201 perform and are not believed likely. Attacks against DH are considered just
202 feasible (although very difficult) because most of the work necessary to
203 deduce information about a private key may be performed offline. The amount
204 of resources required for such an attack would be very significant and
205 likely only accessible to a limited number of attackers. An attacker would
206 additionally need online access to an unpatched system using the target
207 private key in a scenario with persistent DH parameters and a private
208 key that is shared between multiple clients.
210 This only affects processors that support the BMI1, BMI2 and ADX extensions
211 like Intel Broadwell (5th generation) and later or AMD Ryzen.
213 This issue was reported to OpenSSL by the OSS-Fuzz project.
217 *) Malformed X.509 IPAddressFamily could cause OOB read
219 If an X.509 certificate has a malformed IPAddressFamily extension,
220 OpenSSL could do a one-byte buffer overread. The most likely result
221 would be an erroneous display of the certificate in text format.
223 This issue was reported to OpenSSL by the OSS-Fuzz project.
227 Changes between 1.1.0e and 1.1.0f [25 May 2017]
229 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
230 platform rather than 'mingw'.
233 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
234 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
235 which is the minimum version we support.
238 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
240 *) Encrypt-Then-Mac renegotiation crash
242 During a renegotiation handshake if the Encrypt-Then-Mac extension is
243 negotiated where it was not in the original handshake (or vice-versa) then
244 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
245 and servers are affected.
247 This issue was reported to OpenSSL by Joe Orton (Red Hat).
251 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
253 *) Truncated packet could crash via OOB read
255 If one side of an SSL/TLS path is running on a 32-bit host and a specific
256 cipher is being used, then a truncated packet can cause that host to
257 perform an out-of-bounds read, usually resulting in a crash.
259 This issue was reported to OpenSSL by Robert Święcki of Google.
263 *) Bad (EC)DHE parameters cause a client crash
265 If a malicious server supplies bad parameters for a DHE or ECDHE key
266 exchange then this can result in the client attempting to dereference a
267 NULL pointer leading to a client crash. This could be exploited in a Denial
270 This issue was reported to OpenSSL by Guido Vranken.
274 *) BN_mod_exp may produce incorrect results on x86_64
276 There is a carry propagating bug in the x86_64 Montgomery squaring
277 procedure. No EC algorithms are affected. Analysis suggests that attacks
278 against RSA and DSA as a result of this defect would be very difficult to
279 perform and are not believed likely. Attacks against DH are considered just
280 feasible (although very difficult) because most of the work necessary to
281 deduce information about a private key may be performed offline. The amount
282 of resources required for such an attack would be very significant and
283 likely only accessible to a limited number of attackers. An attacker would
284 additionally need online access to an unpatched system using the target
285 private key in a scenario with persistent DH parameters and a private
286 key that is shared between multiple clients. For example this can occur by
287 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
288 similar to CVE-2015-3193 but must be treated as a separate problem.
290 This issue was reported to OpenSSL by the OSS-Fuzz project.
294 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
296 *) ChaCha20/Poly1305 heap-buffer-overflow
298 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
299 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
300 crash. This issue is not considered to be exploitable beyond a DoS.
302 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
306 *) CMS Null dereference
308 Applications parsing invalid CMS structures can crash with a NULL pointer
309 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
310 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
311 structure callback if an attempt is made to free certain invalid encodings.
312 Only CHOICE structures using a callback which do not handle NULL value are
315 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
319 *) Montgomery multiplication may produce incorrect results
321 There is a carry propagating bug in the Broadwell-specific Montgomery
322 multiplication procedure that handles input lengths divisible by, but
323 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
324 and DH private keys are impossible. This is because the subroutine in
325 question is not used in operations with the private key itself and an input
326 of the attacker's direct choice. Otherwise the bug can manifest itself as
327 transient authentication and key negotiation failures or reproducible
328 erroneous outcome of public-key operations with specially crafted input.
329 Among EC algorithms only Brainpool P-512 curves are affected and one
330 presumably can attack ECDH key negotiation. Impact was not analyzed in
331 detail, because pre-requisites for attack are considered unlikely. Namely
332 multiple clients have to choose the curve in question and the server has to
333 share the private key among them, neither of which is default behaviour.
334 Even then only clients that chose the curve will be affected.
336 This issue was publicly reported as transient failures and was not
337 initially recognized as a security issue. Thanks to Richard Morgan for
338 providing reproducible case.
342 *) Removed automatic addition of RPATH in shared libraries and executables,
343 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
346 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
348 *) Fix Use After Free for large message sizes
350 The patch applied to address CVE-2016-6307 resulted in an issue where if a
351 message larger than approx 16k is received then the underlying buffer to
352 store the incoming message is reallocated and moved. Unfortunately a
353 dangling pointer to the old location is left which results in an attempt to
354 write to the previously freed location. This is likely to result in a
355 crash, however it could potentially lead to execution of arbitrary code.
357 This issue only affects OpenSSL 1.1.0a.
359 This issue was reported to OpenSSL by Robert Święcki.
363 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
365 *) OCSP Status Request extension unbounded memory growth
367 A malicious client can send an excessively large OCSP Status Request
368 extension. If that client continually requests renegotiation, sending a
369 large OCSP Status Request extension each time, then there will be unbounded
370 memory growth on the server. This will eventually lead to a Denial Of
371 Service attack through memory exhaustion. Servers with a default
372 configuration are vulnerable even if they do not support OCSP. Builds using
373 the "no-ocsp" build time option are not affected.
375 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
379 *) SSL_peek() hang on empty record
381 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
382 sends an empty record. This could be exploited by a malicious peer in a
383 Denial Of Service attack.
385 This issue was reported to OpenSSL by Alex Gaynor.
389 *) Excessive allocation of memory in tls_get_message_header() and
390 dtls1_preprocess_fragment()
392 A (D)TLS message includes 3 bytes for its length in the header for the
393 message. This would allow for messages up to 16Mb in length. Messages of
394 this length are excessive and OpenSSL includes a check to ensure that a
395 peer is sending reasonably sized messages in order to avoid too much memory
396 being consumed to service a connection. A flaw in the logic of version
397 1.1.0 means that memory for the message is allocated too early, prior to
398 the excessive message length check. Due to way memory is allocated in
399 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
400 to service a connection. This could lead to a Denial of Service through
401 memory exhaustion. However, the excessive message length check still takes
402 place, and this would cause the connection to immediately fail. Assuming
403 that the application calls SSL_free() on the failed connection in a timely
404 manner then the 21Mb of allocated memory will then be immediately freed
405 again. Therefore the excessive memory allocation will be transitory in
406 nature. This then means that there is only a security impact if:
408 1) The application does not call SSL_free() in a timely manner in the event
409 that the connection fails
411 2) The application is working in a constrained environment where there is
412 very little free memory
414 3) The attacker initiates multiple connection attempts such that there are
415 multiple connections in a state where memory has been allocated for the
416 connection; SSL_free() has not yet been called; and there is insufficient
417 memory to service the multiple requests.
419 Except in the instance of (1) above any Denial Of Service is likely to be
420 transitory because as soon as the connection fails the memory is
421 subsequently freed again in the SSL_free() call. However there is an
422 increased risk during this period of application crashes due to the lack of
423 memory - which would then mean a more serious Denial of Service.
425 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
426 (CVE-2016-6307 and CVE-2016-6308)
429 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
430 had to be removed. Primary reason is that vendor assembler can't
431 assemble our modules with -KPIC flag. As result it, assembly
432 support, was not even available as option. But its lack means
433 lack of side-channel resistant code, which is incompatible with
434 security by todays standards. Fortunately gcc is readily available
435 prepackaged option, which we firmly point at...
438 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
440 *) Windows command-line tool supports UTF-8 opt-in option for arguments
441 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
442 (to any value) allows Windows user to access PKCS#12 file generated
443 with Windows CryptoAPI and protected with non-ASCII password, as well
444 as files generated under UTF-8 locale on Linux also protected with
448 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
449 have been disabled by default and removed from DEFAULT, just like RC4.
450 See the RC4 item below to re-enable both.
453 *) The method for finding the storage location for the Windows RAND seed file
454 has changed. First we check %RANDFILE%. If that is not set then we check
455 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
456 all else fails we fall back to C:\.
459 *) The EVP_EncryptUpdate() function has had its return type changed from void
460 to int. A return of 0 indicates and error while a return of 1 indicates
464 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
465 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
466 off the constant time implementation for RSA, DSA and DH have been made
467 no-ops and deprecated.
470 *) Windows RAND implementation was simplified to only get entropy by
471 calling CryptGenRandom(). Various other RAND-related tickets
473 [Joseph Wylie Yandle, Rich Salz]
475 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
476 and OPENSSL_LH_, respectively. The old names are available
477 with API compatibility. They new names are now completely documented.
480 *) Unify TYPE_up_ref(obj) methods signature.
481 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
482 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
483 int (instead of void) like all others TYPE_up_ref() methods.
484 So now these methods also check the return value of CRYPTO_atomic_add(),
485 and the validity of object reference counter.
486 [fdasilvayy@gmail.com]
488 *) With Windows Visual Studio builds, the .pdb files are installed
489 alongside the installed libraries and executables. For a static
490 library installation, ossl_static.pdb is the associate compiler
491 generated .pdb file to be used when linking programs.
494 *) Remove openssl.spec. Packaging files belong with the packagers.
497 *) Automatic Darwin/OSX configuration has had a refresh, it will now
498 recognise x86_64 architectures automatically. You can still decide
499 to build for a different bitness with the environment variable
500 KERNEL_BITS (can be 32 or 64), for example:
502 KERNEL_BITS=32 ./config
506 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
507 256 bit AES and HMAC with SHA256.
510 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
513 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
516 *) To enable users to have their own config files and build file templates,
517 Configure looks in the directory indicated by the environment variable
518 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
519 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
520 name and is used as is.
523 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
524 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
525 X509_CERT_FILE_CTX was removed.
528 *) "shared" builds are now the default. To create only static libraries use
529 the "no-shared" Configure option.
532 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
533 All of these option have not worked for some while and are fundamental
537 *) Make various cleanup routines no-ops and mark them as deprecated. Most
538 global cleanup functions are no longer required because they are handled
539 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
540 Explicitly de-initing can cause problems (e.g. where a library that uses
541 OpenSSL de-inits, but an application is still using it). The affected
542 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
543 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
544 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
548 *) --strict-warnings no longer enables runtime debugging options
549 such as REF_DEBUG. Instead, debug options are automatically
550 enabled with '--debug' builds.
551 [Andy Polyakov, Emilia Käsper]
553 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
554 have been moved out of the public header files. New functions for managing
555 these have been added.
558 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
559 objects have been moved out of the public header files. New
560 functions for managing these have been added.
563 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
564 have been moved out of the public header files. New functions for managing
565 these have been added.
568 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
569 moved out of the public header files. New functions for managing these
573 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
576 *) Removed the mk1mf build scripts.
579 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
580 it is always safe to #include a header now.
583 *) Removed the aged BC-32 config and all its supporting scripts
586 *) Removed support for Ultrix, Netware, and OS/2.
589 *) Add support for HKDF.
592 *) Add support for blake2b and blake2s
595 *) Added support for "pipelining". Ciphers that have the
596 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
597 encryptions/decryptions simultaneously. There are currently no built-in
598 ciphers with this property but the expectation is that engines will be able
599 to offer it to significantly improve throughput. Support has been extended
600 into libssl so that multiple records for a single connection can be
601 processed in one go (for >=TLS 1.1).
604 *) Added the AFALG engine. This is an async capable engine which is able to
605 offload work to the Linux kernel. In this initial version it only supports
606 AES128-CBC. The kernel must be version 4.1.0 or greater.
609 *) OpenSSL now uses a new threading API. It is no longer necessary to
610 set locking callbacks to use OpenSSL in a multi-threaded environment. There
611 are two supported threading models: pthreads and windows threads. It is
612 also possible to configure OpenSSL at compile time for "no-threads". The
613 old threading API should no longer be used. The functions have been
614 replaced with "no-op" compatibility macros.
615 [Alessandro Ghedini, Matt Caswell]
617 *) Modify behavior of ALPN to invoke callback after SNI/servername
618 callback, such that updates to the SSL_CTX affect ALPN.
621 *) Add SSL_CIPHER queries for authentication and key-exchange.
624 *) Changes to the DEFAULT cipherlist:
625 - Prefer (EC)DHE handshakes over plain RSA.
626 - Prefer AEAD ciphers over legacy ciphers.
627 - Prefer ECDSA over RSA when both certificates are available.
628 - Prefer TLSv1.2 ciphers/PRF.
629 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
633 *) Change the ECC default curve list to be this, in order: x25519,
634 secp256r1, secp521r1, secp384r1.
637 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
638 disabled by default. They can be re-enabled using the
639 enable-weak-ssl-ciphers option to Configure.
642 *) If the server has ALPN configured, but supports no protocols that the
643 client advertises, send a fatal "no_application_protocol" alert.
644 This behaviour is SHALL in RFC 7301, though it isn't universally
645 implemented by other servers.
648 *) Add X25519 support.
649 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
650 for public and private key encoding using the format documented in
651 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
652 key generation and key derivation.
654 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
658 *) Deprecate SRP_VBASE_get_by_user.
659 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
660 In order to fix an unavoidable memory leak (CVE-2016-0798),
661 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
662 seed, even if the seed is configured.
664 Users should use SRP_VBASE_get1_by_user instead. Note that in
665 SRP_VBASE_get1_by_user, caller must free the returned value. Note
666 also that even though configuring the SRP seed attempts to hide
667 invalid usernames by continuing the handshake with fake
668 credentials, this behaviour is not constant time and no strong
669 guarantees are made that the handshake is indistinguishable from
670 that of a valid user.
673 *) Configuration change; it's now possible to build dynamic engines
674 without having to build shared libraries and vice versa. This
675 only applies to the engines in engines/, those in crypto/engine/
676 will always be built into libcrypto (i.e. "static").
678 Building dynamic engines is enabled by default; to disable, use
679 the configuration option "disable-dynamic-engine".
681 The only requirements for building dynamic engines are the
682 presence of the DSO module and building with position independent
683 code, so they will also automatically be disabled if configuring
684 with "disable-dso" or "disable-pic".
686 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
687 are also taken away from openssl/opensslconf.h, as they are
691 *) Configuration change; if there is a known flag to compile
692 position independent code, it will always be applied on the
693 libcrypto and libssl object files, and never on the application
694 object files. This means other libraries that use routines from
695 libcrypto / libssl can be made into shared libraries regardless
696 of how OpenSSL was configured.
698 If this isn't desirable, the configuration options "disable-pic"
699 or "no-pic" can be used to disable the use of PIC. This will
700 also disable building shared libraries and dynamic engines.
703 *) Removed JPAKE code. It was experimental and has no wide use.
706 *) The INSTALL_PREFIX Makefile variable has been renamed to
707 DESTDIR. That makes for less confusion on what this variable
708 is for. Also, the configuration option --install_prefix is
712 *) Heartbeat for TLS has been removed and is disabled by default
713 for DTLS; configure with enable-heartbeats. Code that uses the
714 old #define's might need to be updated.
715 [Emilia Käsper, Rich Salz]
717 *) Rename REF_CHECK to REF_DEBUG.
720 *) New "unified" build system
722 The "unified" build system is aimed to be a common system for all
723 platforms we support. With it comes new support for VMS.
725 This system builds supports building in a different directory tree
726 than the source tree. It produces one Makefile (for unix family
727 or lookalikes), or one descrip.mms (for VMS).
729 The source of information to make the Makefile / descrip.mms is
730 small files called 'build.info', holding the necessary
731 information for each directory with source to compile, and a
732 template in Configurations, like unix-Makefile.tmpl or
735 With this change, the library names were also renamed on Windows
736 and on VMS. They now have names that are closer to the standard
737 on Unix, and include the major version number, and in certain
738 cases, the architecture they are built for. See "Notes on shared
739 libraries" in INSTALL.
741 We rely heavily on the perl module Text::Template.
744 *) Added support for auto-initialisation and de-initialisation of the library.
745 OpenSSL no longer requires explicit init or deinit routines to be called,
746 except in certain circumstances. See the OPENSSL_init_crypto() and
747 OPENSSL_init_ssl() man pages for further information.
750 *) The arguments to the DTLSv1_listen function have changed. Specifically the
751 "peer" argument is now expected to be a BIO_ADDR object.
753 *) Rewrite of BIO networking library. The BIO library lacked consistent
754 support of IPv6, and adding it required some more extensive
755 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
756 which hold all types of addresses and chains of address information.
757 It also introduces a new API, with functions like BIO_socket,
758 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
759 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
760 have been adapted accordingly.
763 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
767 *) CRIME protection: disable compression by default, even if OpenSSL is
768 compiled with zlib enabled. Applications can still enable compression
769 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
770 using the SSL_CONF library to configure compression.
773 *) The signature of the session callback configured with
774 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
775 was explicitly marked as 'const unsigned char*' instead of
779 *) Always DPURIFY. Remove the use of uninitialized memory in the
780 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
783 *) Removed many obsolete configuration items, including
784 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
785 MD2_CHAR, MD2_INT, MD2_LONG
787 IDEA_SHORT, IDEA_LONG
788 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
789 [Rich Salz, with advice from Andy Polyakov]
791 *) Many BN internals have been moved to an internal header file.
792 [Rich Salz with help from Andy Polyakov]
794 *) Configuration and writing out the results from it has changed.
795 Files such as Makefile include/openssl/opensslconf.h and are now
796 produced through general templates, such as Makefile.in and
797 crypto/opensslconf.h.in and some help from the perl module
800 Also, the center of configuration information is no longer
801 Makefile. Instead, Configure produces a perl module in
802 configdata.pm which holds most of the config data (in the hash
803 table %config), the target data that comes from the target
804 configuration in one of the Configurations/*.conf files (in
808 *) To clarify their intended purposes, the Configure options
809 --prefix and --openssldir change their semantics, and become more
810 straightforward and less interdependent.
812 --prefix shall be used exclusively to give the location INSTALLTOP
813 where programs, scripts, libraries, include files and manuals are
814 going to be installed. The default is now /usr/local.
816 --openssldir shall be used exclusively to give the default
817 location OPENSSLDIR where certificates, private keys, CRLs are
818 managed. This is also where the default openssl.cnf gets
820 If the directory given with this option is a relative path, the
821 values of both the --prefix value and the --openssldir value will
822 be combined to become OPENSSLDIR.
823 The default for --openssldir is INSTALLTOP/ssl.
825 Anyone who uses --openssldir to specify where OpenSSL is to be
826 installed MUST change to use --prefix instead.
829 *) The GOST engine was out of date and therefore it has been removed. An up
830 to date GOST engine is now being maintained in an external repository.
831 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
832 support for GOST ciphersuites (these are only activated if a GOST engine
836 *) EGD is no longer supported by default; use enable-egd when
838 [Ben Kaduk and Rich Salz]
840 *) The distribution now has Makefile.in files, which are used to
841 create Makefile's when Configure is run. *Configure must be run
842 before trying to build now.*
845 *) The return value for SSL_CIPHER_description() for error conditions
849 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
851 Obtaining and performing DNSSEC validation of TLSA records is
852 the application's responsibility. The application provides
853 the TLSA records of its choice to OpenSSL, and these are then
854 used to authenticate the peer.
856 The TLSA records need not even come from DNS. They can, for
857 example, be used to implement local end-entity certificate or
858 trust-anchor "pinning", where the "pin" data takes the form
859 of TLSA records, which can augment or replace verification
860 based on the usual WebPKI public certification authorities.
863 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
864 continues to support deprecated interfaces in default builds.
865 However, applications are strongly advised to compile their
866 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
867 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
868 or the 1.1.0 releases.
870 In environments in which all applications have been ported to
871 not use any deprecated interfaces OpenSSL's Configure script
872 should be used with the --api=1.1.0 option to entirely remove
873 support for the deprecated features from the library and
874 unconditionally disable them in the installed headers.
875 Essentially the same effect can be achieved with the "no-deprecated"
876 argument to Configure, except that this will always restrict
877 the build to just the latest API, rather than a fixed API
880 As applications are ported to future revisions of the API,
881 they should update their compile-time OPENSSL_API_COMPAT define
882 accordingly, but in most cases should be able to continue to
883 compile with later releases.
885 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
886 0x10000000L and 0x00908000L, respectively. However those
887 versions did not support the OPENSSL_API_COMPAT feature, and
888 so applications are not typically tested for explicit support
889 of just the undeprecated features of either release.
892 *) Add support for setting the minimum and maximum supported protocol.
893 It can bet set via the SSL_set_min_proto_version() and
894 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
895 MaxProtocol. It's recommended to use the new APIs to disable
896 protocols instead of disabling individual protocols using
897 SSL_set_options() or SSL_CONF's Protocol. This change also
898 removes support for disabling TLS 1.2 in the OpenSSL TLS
899 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
902 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
905 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
906 and integrates ECDSA and ECDH functionality into EC. Implementations can
907 now redirect key generation and no longer need to convert to or from
910 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
911 include the ec.h header file instead.
914 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
915 ciphers who are no longer supported and drops support the ephemeral RSA key
916 exchange. The LOW ciphers currently doesn't have any ciphers in it.
919 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
920 opaque. For HMAC_CTX, the following constructors and destructors
923 HMAC_CTX *HMAC_CTX_new(void);
924 void HMAC_CTX_free(HMAC_CTX *ctx);
926 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
927 destroy such methods has been added. See EVP_MD_meth_new(3) and
928 EVP_CIPHER_meth_new(3) for documentation.
931 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
932 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
933 EVP_MD_CTX_reset() should be called instead to reinitialise
934 an already created structure.
935 2) For consistency with the majority of our object creators and
936 destructors, EVP_MD_CTX_(create|destroy) were renamed to
937 EVP_MD_CTX_(new|free). The old names are retained as macros
938 for deprecated builds.
941 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
942 cryptographic operations to be performed asynchronously as long as an
943 asynchronous capable engine is used. See the ASYNC_start_job() man page for
944 further details. Libssl has also had this capability integrated with the
945 introduction of the new mode SSL_MODE_ASYNC and associated error
946 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
947 pages. This work was developed in partnership with Intel Corp.
950 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
951 always enabled now. If you want to disable the support you should
952 exclude it using the list of supported ciphers. This also means that the
953 "-no_ecdhe" option has been removed from s_server.
956 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
957 SSL_{CTX_}set1_curves() which can set a list.
960 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
961 curve you want to support using SSL_{CTX_}set1_curves().
964 *) State machine rewrite. The state machine code has been significantly
965 refactored in order to remove much duplication of code and solve issues
966 with the old code (see ssl/statem/README for further details). This change
967 does have some associated API changes. Notably the SSL_state() function
968 has been removed and replaced by SSL_get_state which now returns an
969 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
970 altogether. The previous handshake states defined in ssl.h and ssl3.h have
974 *) All instances of the string "ssleay" in the public API were replaced
975 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
976 Some error codes related to internal RSA_eay API's were renamed.
979 *) The demo files in crypto/threads were moved to demo/threads.
982 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
984 [Matt Caswell, Rich Salz]
986 *) New ASN.1 embed macro.
988 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
989 structure is not allocated: it is part of the parent. That is instead of
997 This reduces memory fragmentation and make it impossible to accidentally
998 set a mandatory field to NULL.
1000 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1001 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1002 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1006 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1009 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1010 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1011 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1012 DES and RC4 ciphersuites.
1015 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1016 This changes the decoding behaviour for some invalid messages,
1017 though the change is mostly in the more lenient direction, and
1018 legacy behaviour is preserved as much as possible.
1021 *) Fix no-stdio build.
1022 [ David Woodhouse <David.Woodhouse@intel.com> and also
1023 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1025 *) New testing framework
1026 The testing framework has been largely rewritten and is now using
1027 perl and the perl modules Test::Harness and an extended variant of
1028 Test::More called OpenSSL::Test to do its work. All test scripts in
1029 test/ have been rewritten into test recipes, and all direct calls to
1030 executables in test/Makefile have become individual recipes using the
1031 simplified testing OpenSSL::Test::Simple.
1033 For documentation on our testing modules, do:
1035 perldoc test/testlib/OpenSSL/Test/Simple.pm
1036 perldoc test/testlib/OpenSSL/Test.pm
1040 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1041 are used; the latter aborts on memory leaks (usually checked on exit).
1042 Some undocumented "set malloc, etc., hooks" functions were removed
1043 and others were changed. All are now documented.
1046 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1048 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1050 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1051 from RFC4279, RFC4785, RFC5487, RFC5489.
1053 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1054 original RSA_PSK patch.
1057 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1058 era flag was never set throughout the codebase (only read). Also removed
1059 SSL3_FLAGS_POP_BUFFER which was only used if
1060 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1063 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1064 to be "oneline" instead of "compat".
1067 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1068 not aware of clients that still exhibit this bug, and the workaround
1069 hasn't been working properly for a while.
1072 *) The return type of BIO_number_read() and BIO_number_written() as well as
1073 the corresponding num_read and num_write members in the BIO structure has
1074 changed from unsigned long to uint64_t. On platforms where an unsigned
1075 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1079 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1080 OpenSSL without support for them. It also means that maintaining
1081 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1082 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1085 *) Removed support for the two export grade static DH ciphersuites
1086 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1087 were newly added (along with a number of other static DH ciphersuites) to
1088 1.0.2. However the two export ones have *never* worked since they were
1089 introduced. It seems strange in any case to be adding new export
1090 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1093 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1094 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1095 and turned into macros which simply call the new preferred function names
1096 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1097 should use the new names instead. Also as part of this change the ssl23.h
1098 header file has been removed.
1101 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1102 code and the associated standard is no longer considered fit-for-purpose.
1105 *) RT2547 was closed. When generating a private key, try to make the
1106 output file readable only by the owner. This behavior change might
1107 be noticeable when interacting with other software.
1109 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1113 *) Added HTTP GET support to the ocsp command.
1116 *) Changed default digest for the dgst and enc commands from MD5 to
1120 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1123 *) Added support for TLS extended master secret from
1124 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1125 initial patch which was a great help during development.
1128 *) All libssl internal structures have been removed from the public header
1129 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1130 now redundant). Users should not attempt to access internal structures
1131 directly. Instead they should use the provided API functions.
1134 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1135 Access to deprecated functions can be re-enabled by running config with
1136 "enable-deprecated". In addition applications wishing to use deprecated
1137 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1138 will, by default, disable some transitive includes that previously existed
1139 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1142 *) Added support for OCB mode. OpenSSL has been granted a patent license
1143 compatible with the OpenSSL license for use of OCB. Details are available
1144 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1145 for OCB can be removed by calling config with no-ocb.
1148 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1149 compatible client hello.
1152 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1153 done while fixing the error code for the key-too-small case.
1154 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1156 *) CA.sh has been removed; use CA.pl instead.
1159 *) Removed old DES API.
1162 *) Remove various unsupported platforms:
1168 Sinix/ReliantUNIX RM400
1173 16-bit platforms such as WIN16
1176 *) Clean up OPENSSL_NO_xxx #define's
1177 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1178 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1179 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1180 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1181 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1182 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1183 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1184 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1185 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1186 Remove MS_STATIC; it's a relic from platforms <32 bits.
1189 *) Cleaned up dead code
1190 Remove all but one '#ifdef undef' which is to be looked at.
1193 *) Clean up calling of xxx_free routines.
1194 Just like free(), fix most of the xxx_free routines to accept
1195 NULL. Remove the non-null checks from callers. Save much code.
1198 *) Add secure heap for storage of private keys (when possible).
1199 Add BIO_s_secmem(), CBIGNUM, etc.
1200 Contributed by Akamai Technologies under our Corporate CLA.
1203 *) Experimental support for a new, fast, unbiased prime candidate generator,
1204 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1205 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1207 *) New output format NSS in the sess_id command line tool. This allows
1208 exporting the session id and the master key in NSS keylog format.
1209 [Martin Kaiser <martin@kaiser.cx>]
1211 *) Harmonize version and its documentation. -f flag is used to display
1213 [mancha <mancha1@zoho.com>]
1215 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1216 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1217 [mancha <mancha1@zoho.com>]
1219 *) Fix some double frees. These are not thought to be exploitable.
1220 [mancha <mancha1@zoho.com>]
1222 *) A missing bounds check in the handling of the TLS heartbeat extension
1223 can be used to reveal up to 64k of memory to a connected client or
1226 Thanks for Neel Mehta of Google Security for discovering this bug and to
1227 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1228 preparing the fix (CVE-2014-0160)
1229 [Adam Langley, Bodo Moeller]
1231 *) Fix for the attack described in the paper "Recovering OpenSSL
1232 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1233 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1234 http://eprint.iacr.org/2014/140
1236 Thanks to Yuval Yarom and Naomi Benger for discovering this
1237 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1238 [Yuval Yarom and Naomi Benger]
1240 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1241 this fixes a limitation in previous versions of OpenSSL.
1244 *) Experimental encrypt-then-mac support.
1246 Experimental support for encrypt then mac from
1247 draft-gutmann-tls-encrypt-then-mac-02.txt
1249 To enable it set the appropriate extension number (0x42 for the test
1250 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1252 For non-compliant peers (i.e. just about everything) this should have no
1255 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1259 *) Add EVP support for key wrapping algorithms, to avoid problems with
1260 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1261 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1262 algorithms and include tests cases.
1265 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1269 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1270 MGF1 digest and OAEP label.
1273 *) Make openssl verify return errors.
1274 [Chris Palmer <palmer@google.com> and Ben Laurie]
1276 *) New function ASN1_TIME_diff to calculate the difference between two
1277 ASN1_TIME structures or one structure and the current time.
1280 *) Update fips_test_suite to support multiple command line options. New
1281 test to induce all self test errors in sequence and check expected
1285 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1286 sign or verify all in one operation.
1289 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1290 test programs and fips_test_suite. Includes functionality to parse
1291 the minimal script output of fipsalgest.pl directly.
1294 *) Add authorisation parameter to FIPS_module_mode_set().
1297 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1300 *) Use separate DRBG fields for internal and external flags. New function
1301 FIPS_drbg_health_check() to perform on demand health checking. Add
1302 generation tests to fips_test_suite with reduced health check interval to
1303 demonstrate periodic health checking. Add "nodh" option to
1304 fips_test_suite to skip very slow DH test.
1307 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1311 *) More extensive health check for DRBG checking many more failure modes.
1312 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1313 combination: call this in fips_test_suite.
1316 *) Add support for canonical generation of DSA parameter 'g'. See
1319 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1320 POST to handle HMAC cases.
1323 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1324 to return numerical and string versions of the FIPS module number.
1327 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1328 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1329 outside the validated module in the FIPS capable OpenSSL.
1332 *) Minor change to DRBG entropy callback semantics. In some cases
1333 there is no multiple of the block length between min_len and
1334 max_len. Allow the callback to return more than max_len bytes
1335 of entropy but discard any extra: it is the callback's responsibility
1336 to ensure that the extra data discarded does not impact the
1337 requested amount of entropy.
1340 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1341 information in FIPS186-3, SP800-57 and SP800-131A.
1344 *) CCM support via EVP. Interface is very similar to GCM case except we
1345 must supply all data in one chunk (i.e. no update, final) and the
1346 message length must be supplied if AAD is used. Add algorithm test
1350 *) Initial version of POST overhaul. Add POST callback to allow the status
1351 of POST to be monitored and/or failures induced. Modify fips_test_suite
1352 to use callback. Always run all selftests even if one fails.
1355 *) XTS support including algorithm test driver in the fips_gcmtest program.
1356 Note: this does increase the maximum key length from 32 to 64 bytes but
1357 there should be no binary compatibility issues as existing applications
1358 will never use XTS mode.
1361 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1362 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1363 performs algorithm blocking for unapproved PRNG types. Also do not
1364 set PRNG type in FIPS_mode_set(): leave this to the application.
1365 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1366 the standard OpenSSL PRNG: set additional data to a date time vector.
1369 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1370 This shouldn't present any incompatibility problems because applications
1371 shouldn't be using these directly and any that are will need to rethink
1372 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1375 *) Extensive self tests and health checking required by SP800-90 DRBG.
1376 Remove strength parameter from FIPS_drbg_instantiate and always
1377 instantiate at maximum supported strength.
1380 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1383 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1386 *) New function DH_compute_key_padded() to compute a DH key and pad with
1387 leading zeroes if needed: this complies with SP800-56A et al.
1390 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1391 anything, incomplete, subject to change and largely untested at present.
1394 *) Modify fipscanisteronly build option to only build the necessary object
1395 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1398 *) Add experimental option FIPSSYMS to give all symbols in
1399 fipscanister.o and FIPS or fips prefix. This will avoid
1400 conflicts with future versions of OpenSSL. Add perl script
1401 util/fipsas.pl to preprocess assembly language source files
1402 and rename any affected symbols.
1405 *) Add selftest checks and algorithm block of non-fips algorithms in
1406 FIPS mode. Remove DES2 from selftests.
1409 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1410 return internal method without any ENGINE dependencies. Add new
1411 tiny fips sign and verify functions.
1414 *) New build option no-ec2m to disable characteristic 2 code.
1417 *) New build option "fipscanisteronly". This only builds fipscanister.o
1418 and (currently) associated fips utilities. Uses the file Makefile.fips
1419 instead of Makefile.org as the prototype.
1422 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1423 Update fips_gcmtest to use IV generator.
1426 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1427 setting output buffer to NULL. The *Final function must be
1428 called although it will not retrieve any additional data. The tag
1429 can be set or retrieved with a ctrl. The IV length is by default 12
1430 bytes (96 bits) but can be set to an alternative value. If the IV
1431 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1435 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1436 underlying do_cipher function handles all cipher semantics itself
1437 including padding and finalisation. This is useful if (for example)
1438 an ENGINE cipher handles block padding itself. The behaviour of
1439 do_cipher is subtly changed if this flag is set: the return value
1440 is the number of characters written to the output buffer (zero is
1441 no longer an error code) or a negative error code. Also if the
1442 input buffer is NULL and length 0 finalisation should be performed.
1445 *) If a candidate issuer certificate is already part of the constructed
1446 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1449 *) Improve forward-security support: add functions
1451 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1452 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1454 for use by SSL/TLS servers; the callback function will be called whenever a
1455 new session is created, and gets to decide whether the session may be
1456 cached to make it resumable (return 0) or not (return 1). (As by the
1457 SSL/TLS protocol specifications, the session_id sent by the server will be
1458 empty to indicate that the session is not resumable; also, the server will
1459 not generate RFC 4507 (RFC 5077) session tickets.)
1461 A simple reasonable callback implementation is to return is_forward_secure.
1462 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1463 by the SSL/TLS server library, indicating whether it can provide forward
1465 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1467 *) New -verify_name option in command line utilities to set verification
1471 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1472 Add CMAC pkey methods.
1475 *) Experimental renegotiation in s_server -www mode. If the client
1476 browses /reneg connection is renegotiated. If /renegcert it is
1477 renegotiated requesting a certificate.
1480 *) Add an "external" session cache for debugging purposes to s_server. This
1481 should help trace issues which normally are only apparent in deployed
1482 multi-process servers.
1485 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1486 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1487 BIO_set_cipher() and some obscure PEM functions were changed so they
1488 can now return an error. The RAND changes required a change to the
1489 RAND_METHOD structure.
1492 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1493 a gcc attribute to warn if the result of a function is ignored. This
1494 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1495 whose return value is often ignored.
1498 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1499 These allow SCTs (signed certificate timestamps) to be requested and
1500 validated when establishing a connection.
1501 [Rob Percival <robpercival@google.com>]
1503 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1505 *) Prevent padding oracle in AES-NI CBC MAC check
1507 A MITM attacker can use a padding oracle attack to decrypt traffic
1508 when the connection uses an AES CBC cipher and the server support
1511 This issue was introduced as part of the fix for Lucky 13 padding
1512 attack (CVE-2013-0169). The padding check was rewritten to be in
1513 constant time by making sure that always the same bytes are read and
1514 compared against either the MAC or padding bytes. But it no longer
1515 checked that there was enough data to have both the MAC and padding
1518 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1522 *) Fix EVP_EncodeUpdate overflow
1524 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1525 Base64 encoding of binary data. If an attacker is able to supply very large
1526 amounts of input data then a length check can overflow resulting in a heap
1529 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1530 the PEM_write_bio* family of functions. These are mainly used within the
1531 OpenSSL command line applications, so any application which processes data
1532 from an untrusted source and outputs it as a PEM file should be considered
1533 vulnerable to this issue. User applications that call these APIs directly
1534 with large amounts of untrusted data may also be vulnerable.
1536 This issue was reported by Guido Vranken.
1540 *) Fix EVP_EncryptUpdate overflow
1542 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1543 is able to supply very large amounts of input data after a previous call to
1544 EVP_EncryptUpdate() with a partial block then a length check can overflow
1545 resulting in a heap corruption. Following an analysis of all OpenSSL
1546 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1547 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1548 the first called function after an EVP_EncryptInit(), and therefore that
1549 specific call must be safe. The second form is where the length passed to
1550 EVP_EncryptUpdate() can be seen from the code to be some small value and
1551 therefore there is no possibility of an overflow. Since all instances are
1552 one of these two forms, it is believed that there can be no overflows in
1553 internal code due to this problem. It should be noted that
1554 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1555 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1556 of these calls have also been analysed too and it is believed there are no
1557 instances in internal usage where an overflow could occur.
1559 This issue was reported by Guido Vranken.
1563 *) Prevent ASN.1 BIO excessive memory allocation
1565 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1566 a short invalid encoding can cause allocation of large amounts of memory
1567 potentially consuming excessive resources or exhausting memory.
1569 Any application parsing untrusted data through d2i BIO functions is
1570 affected. The memory based functions such as d2i_X509() are *not* affected.
1571 Since the memory based functions are used by the TLS library, TLS
1572 applications are not affected.
1574 This issue was reported by Brian Carpenter.
1580 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1581 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1582 in arbitrary stack data being returned in the buffer.
1584 This issue was reported by Guido Vranken.
1588 *) Modify behavior of ALPN to invoke callback after SNI/servername
1589 callback, such that updates to the SSL_CTX affect ALPN.
1592 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1596 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1597 methods are enabled and ssl2 is disabled the methods return NULL.
1600 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1602 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1603 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1604 provide any "EXPORT" or "LOW" strength ciphers.
1607 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1608 is by default disabled at build-time. Builds that are not configured with
1609 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1610 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1611 will need to explicitly call either of:
1613 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1615 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1617 as appropriate. Even if either of those is used, or the application
1618 explicitly uses the version-specific SSLv2_method() or its client and
1619 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1620 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1621 ciphers, and SSLv2 56-bit DES are no longer available.
1625 *) Fix a double-free in DSA code
1627 A double free bug was discovered when OpenSSL parses malformed DSA private
1628 keys and could lead to a DoS attack or memory corruption for applications
1629 that receive DSA private keys from untrusted sources. This scenario is
1632 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1637 *) Disable SRP fake user seed to address a server memory leak.
1639 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1641 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1642 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1643 was changed to ignore the "fake user" SRP seed, even if the seed
1646 Users should use SRP_VBASE_get1_by_user instead. Note that in
1647 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1648 also that even though configuring the SRP seed attempts to hide
1649 invalid usernames by continuing the handshake with fake
1650 credentials, this behaviour is not constant time and no strong
1651 guarantees are made that the handshake is indistinguishable from
1652 that of a valid user.
1656 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1658 In the BN_hex2bn function the number of hex digits is calculated using an
1659 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1660 large values of |i| this can result in |bn_expand| not allocating any
1661 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1662 field as NULL leading to a subsequent NULL ptr deref. For very large values
1663 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1664 In this case memory is allocated to the internal BIGNUM data field, but it
1665 is insufficiently sized leading to heap corruption. A similar issue exists
1666 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1667 is ever called by user applications with very large untrusted hex/dec data.
1668 This is anticipated to be a rare occurrence.
1670 All OpenSSL internal usage of these functions use data that is not expected
1671 to be untrusted, e.g. config file data or application command line
1672 arguments. If user developed applications generate config file data based
1673 on untrusted data then it is possible that this could also lead to security
1674 consequences. This is also anticipated to be rare.
1676 This issue was reported to OpenSSL by Guido Vranken.
1680 *) Fix memory issues in BIO_*printf functions
1682 The internal |fmtstr| function used in processing a "%s" format string in
1683 the BIO_*printf functions could overflow while calculating the length of a
1684 string and cause an OOB read when printing very long strings.
1686 Additionally the internal |doapr_outch| function can attempt to write to an
1687 OOB memory location (at an offset from the NULL pointer) in the event of a
1688 memory allocation failure. In 1.0.2 and below this could be caused where
1689 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1690 could be in processing a very long "%s" format string. Memory leaks can
1693 The first issue may mask the second issue dependent on compiler behaviour.
1694 These problems could enable attacks where large amounts of untrusted data
1695 is passed to the BIO_*printf functions. If applications use these functions
1696 in this way then they could be vulnerable. OpenSSL itself uses these
1697 functions when printing out human-readable dumps of ASN.1 data. Therefore
1698 applications that print this data could be vulnerable if the data is from
1699 untrusted sources. OpenSSL command line applications could also be
1700 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1701 as command line arguments.
1703 Libssl is not considered directly vulnerable. Additionally certificates etc
1704 received via remote connections via libssl are also unlikely to be able to
1705 trigger these issues because of message size limits enforced within libssl.
1707 This issue was reported to OpenSSL Guido Vranken.
1711 *) Side channel attack on modular exponentiation
1713 A side-channel attack was found which makes use of cache-bank conflicts on
1714 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1715 of RSA keys. The ability to exploit this issue is limited as it relies on
1716 an attacker who has control of code in a thread running on the same
1717 hyper-threaded core as the victim thread which is performing decryptions.
1719 This issue was reported to OpenSSL by Yuval Yarom, The University of
1720 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1721 Nadia Heninger, University of Pennsylvania with more information at
1722 http://cachebleed.info.
1726 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1727 if no keysize is specified with default_bits. This fixes an
1728 omission in an earlier change that changed all RSA/DSA key generation
1729 apps to use 2048 bits by default.
1732 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1733 *) DH small subgroups
1735 Historically OpenSSL only ever generated DH parameters based on "safe"
1736 primes. More recently (in version 1.0.2) support was provided for
1737 generating X9.42 style parameter files such as those required for RFC 5114
1738 support. The primes used in such files may not be "safe". Where an
1739 application is using DH configured with parameters based on primes that are
1740 not "safe" then an attacker could use this fact to find a peer's private
1741 DH exponent. This attack requires that the attacker complete multiple
1742 handshakes in which the peer uses the same private DH exponent. For example
1743 this could be used to discover a TLS server's private DH exponent if it's
1744 reusing the private DH exponent or it's using a static DH ciphersuite.
1746 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1747 TLS. It is not on by default. If the option is not set then the server
1748 reuses the same private DH exponent for the life of the server process and
1749 would be vulnerable to this attack. It is believed that many popular
1750 applications do set this option and would therefore not be at risk.
1752 The fix for this issue adds an additional check where a "q" parameter is
1753 available (as is the case in X9.42 based parameters). This detects the
1754 only known attack, and is the only possible defense for static DH
1755 ciphersuites. This could have some performance impact.
1757 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1758 default and cannot be disabled. This could have some performance impact.
1760 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1764 *) SSLv2 doesn't block disabled ciphers
1766 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1767 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1768 been disabled, provided that the SSLv2 protocol was not also disabled via
1771 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1772 and Sebastian Schinzel.
1776 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1778 *) BN_mod_exp may produce incorrect results on x86_64
1780 There is a carry propagating bug in the x86_64 Montgomery squaring
1781 procedure. No EC algorithms are affected. Analysis suggests that attacks
1782 against RSA and DSA as a result of this defect would be very difficult to
1783 perform and are not believed likely. Attacks against DH are considered just
1784 feasible (although very difficult) because most of the work necessary to
1785 deduce information about a private key may be performed offline. The amount
1786 of resources required for such an attack would be very significant and
1787 likely only accessible to a limited number of attackers. An attacker would
1788 additionally need online access to an unpatched system using the target
1789 private key in a scenario with persistent DH parameters and a private
1790 key that is shared between multiple clients. For example this can occur by
1791 default in OpenSSL DHE based SSL/TLS ciphersuites.
1793 This issue was reported to OpenSSL by Hanno Böck.
1797 *) Certificate verify crash with missing PSS parameter
1799 The signature verification routines will crash with a NULL pointer
1800 dereference if presented with an ASN.1 signature using the RSA PSS
1801 algorithm and absent mask generation function parameter. Since these
1802 routines are used to verify certificate signature algorithms this can be
1803 used to crash any certificate verification operation and exploited in a
1804 DoS attack. Any application which performs certificate verification is
1805 vulnerable including OpenSSL clients and servers which enable client
1808 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1812 *) X509_ATTRIBUTE memory leak
1814 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1815 memory. This structure is used by the PKCS#7 and CMS routines so any
1816 application which reads PKCS#7 or CMS data from untrusted sources is
1817 affected. SSL/TLS is not affected.
1819 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1824 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1825 This changes the decoding behaviour for some invalid messages,
1826 though the change is mostly in the more lenient direction, and
1827 legacy behaviour is preserved as much as possible.
1830 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1832 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1834 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1836 *) Alternate chains certificate forgery
1838 During certificate verification, OpenSSL will attempt to find an
1839 alternative certificate chain if the first attempt to build such a chain
1840 fails. An error in the implementation of this logic can mean that an
1841 attacker could cause certain checks on untrusted certificates to be
1842 bypassed, such as the CA flag, enabling them to use a valid leaf
1843 certificate to act as a CA and "issue" an invalid certificate.
1845 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1849 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1851 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1852 incompatibility in the handling of HMAC. The previous ABI has now been
1856 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1858 *) Malformed ECParameters causes infinite loop
1860 When processing an ECParameters structure OpenSSL enters an infinite loop
1861 if the curve specified is over a specially malformed binary polynomial
1864 This can be used to perform denial of service against any
1865 system which processes public keys, certificate requests or
1866 certificates. This includes TLS clients and TLS servers with
1867 client authentication enabled.
1869 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1873 *) Exploitable out-of-bounds read in X509_cmp_time
1875 X509_cmp_time does not properly check the length of the ASN1_TIME
1876 string and can read a few bytes out of bounds. In addition,
1877 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1880 An attacker can use this to craft malformed certificates and CRLs of
1881 various sizes and potentially cause a segmentation fault, resulting in
1882 a DoS on applications that verify certificates or CRLs. TLS clients
1883 that verify CRLs are affected. TLS clients and servers with client
1884 authentication enabled may be affected if they use custom verification
1887 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1888 independently by Hanno Böck.
1892 *) PKCS7 crash with missing EnvelopedContent
1894 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1895 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1896 with missing content and trigger a NULL pointer dereference on parsing.
1898 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1899 structures from untrusted sources are affected. OpenSSL clients and
1900 servers are not affected.
1902 This issue was reported to OpenSSL by Michal Zalewski (Google).
1906 *) CMS verify infinite loop with unknown hash function
1908 When verifying a signedData message the CMS code can enter an infinite loop
1909 if presented with an unknown hash function OID. This can be used to perform
1910 denial of service against any system which verifies signedData messages using
1912 This issue was reported to OpenSSL by Johannes Bauer.
1916 *) Race condition handling NewSessionTicket
1918 If a NewSessionTicket is received by a multi-threaded client when attempting to
1919 reuse a previous ticket then a race condition can occur potentially leading to
1920 a double free of the ticket data.
1924 *) Only support 256-bit or stronger elliptic curves with the
1925 'ecdh_auto' setting (server) or by default (client). Of supported
1926 curves, prefer P-256 (both).
1929 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1931 *) ClientHello sigalgs DoS fix
1933 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1934 invalid signature algorithms extension a NULL pointer dereference will
1935 occur. This can be exploited in a DoS attack against the server.
1937 This issue was was reported to OpenSSL by David Ramos of Stanford
1940 [Stephen Henson and Matt Caswell]
1942 *) Multiblock corrupted pointer fix
1944 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1945 feature only applies on 64 bit x86 architecture platforms that support AES
1946 NI instructions. A defect in the implementation of "multiblock" can cause
1947 OpenSSL's internal write buffer to become incorrectly set to NULL when
1948 using non-blocking IO. Typically, when the user application is using a
1949 socket BIO for writing, this will only result in a failed connection.
1950 However if some other BIO is used then it is likely that a segmentation
1951 fault will be triggered, thus enabling a potential DoS attack.
1953 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1957 *) Segmentation fault in DTLSv1_listen fix
1959 The DTLSv1_listen function is intended to be stateless and processes the
1960 initial ClientHello from many peers. It is common for user code to loop
1961 over the call to DTLSv1_listen until a valid ClientHello is received with
1962 an associated cookie. A defect in the implementation of DTLSv1_listen means
1963 that state is preserved in the SSL object from one invocation to the next
1964 that can lead to a segmentation fault. Errors processing the initial
1965 ClientHello can trigger this scenario. An example of such an error could be
1966 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1969 This issue was reported to OpenSSL by Per Allansson.
1973 *) Segmentation fault in ASN1_TYPE_cmp fix
1975 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1976 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1977 certificate signature algorithm consistency this can be used to crash any
1978 certificate verification operation and exploited in a DoS attack. Any
1979 application which performs certificate verification is vulnerable including
1980 OpenSSL clients and servers which enable client authentication.
1984 *) Segmentation fault for invalid PSS parameters fix
1986 The signature verification routines will crash with a NULL pointer
1987 dereference if presented with an ASN.1 signature using the RSA PSS
1988 algorithm and invalid parameters. Since these routines are used to verify
1989 certificate signature algorithms this can be used to crash any
1990 certificate verification operation and exploited in a DoS attack. Any
1991 application which performs certificate verification is vulnerable including
1992 OpenSSL clients and servers which enable client authentication.
1994 This issue was was reported to OpenSSL by Brian Carpenter.
1998 *) ASN.1 structure reuse memory corruption fix
2000 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2001 memory corruption via an invalid write. Such reuse is and has been
2002 strongly discouraged and is believed to be rare.
2004 Applications that parse structures containing CHOICE or ANY DEFINED BY
2005 components may be affected. Certificate parsing (d2i_X509 and related
2006 functions) are however not affected. OpenSSL clients and servers are
2011 *) PKCS7 NULL pointer dereferences fix
2013 The PKCS#7 parsing code does not handle missing outer ContentInfo
2014 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2015 missing content and trigger a NULL pointer dereference on parsing.
2017 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2018 otherwise parse PKCS#7 structures from untrusted sources are
2019 affected. OpenSSL clients and servers are not affected.
2021 This issue was reported to OpenSSL by Michal Zalewski (Google).
2025 *) DoS via reachable assert in SSLv2 servers fix
2027 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2028 servers that both support SSLv2 and enable export cipher suites by sending
2029 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2031 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2032 (OpenSSL development team).
2036 *) Empty CKE with client auth and DHE fix
2038 If client auth is used then a server can seg fault in the event of a DHE
2039 ciphersuite being selected and a zero length ClientKeyExchange message
2040 being sent by the client. This could be exploited in a DoS attack.
2044 *) Handshake with unseeded PRNG fix
2046 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2047 with an unseeded PRNG. The conditions are:
2048 - The client is on a platform where the PRNG has not been seeded
2049 automatically, and the user has not seeded manually
2050 - A protocol specific client method version has been used (i.e. not
2051 SSL_client_methodv23)
2052 - A ciphersuite is used that does not require additional random data from
2053 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2055 If the handshake succeeds then the client random that has been used will
2056 have been generated from a PRNG with insufficient entropy and therefore the
2057 output may be predictable.
2059 For example using the following command with an unseeded openssl will
2060 succeed on an unpatched platform:
2062 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2066 *) Use After Free following d2i_ECPrivatekey error fix
2068 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2069 could cause a use after free condition. This, in turn, could cause a double
2070 free in several private key parsing functions (such as d2i_PrivateKey
2071 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2072 for applications that receive EC private keys from untrusted
2073 sources. This scenario is considered rare.
2075 This issue was discovered by the BoringSSL project and fixed in their
2080 *) X509_to_X509_REQ NULL pointer deref fix
2082 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2083 the certificate key is invalid. This function is rarely used in practice.
2085 This issue was discovered by Brian Carpenter.
2089 *) Removed the export ciphers from the DEFAULT ciphers
2092 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2094 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2095 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2096 So far those who have to target multiple platforms would compromise
2097 and argue that binary targeting say ARMv5 would still execute on
2098 ARMv8. "Universal" build resolves this compromise by providing
2099 near-optimal performance even on newer platforms.
2102 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2103 (other platforms pending).
2104 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2106 *) Add support for the SignedCertificateTimestampList certificate and
2107 OCSP response extensions from RFC6962.
2110 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2111 for corner cases. (Certain input points at infinity could lead to
2112 bogus results, with non-infinity inputs mapped to infinity too.)
2115 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2116 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2117 common cases are optimized and there still is room for further
2118 improvements. Vector Permutation AES for Altivec is also added.
2121 *) Add support for little-endian ppc64 Linux target.
2122 [Marcelo Cerri (IBM)]
2124 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2125 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2126 are optimized and there still is room for further improvements.
2127 Both 32- and 64-bit modes are supported.
2128 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2130 *) Improved ARMv7 NEON support.
2133 *) Support for SPARC Architecture 2011 crypto extensions, first
2134 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2135 SHA256/512, MD5, GHASH and modular exponentiation.
2136 [Andy Polyakov, David Miller]
2138 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2140 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2142 *) Support for new and upcoming Intel processors, including AVX2,
2143 BMI and SHA ISA extensions. This includes additional "stitched"
2144 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2147 This work was sponsored by Intel Corp.
2150 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2151 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2152 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2155 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2156 this fixes a limitation in previous versions of OpenSSL.
2159 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2160 MGF1 digest and OAEP label.
2163 *) Add EVP support for key wrapping algorithms, to avoid problems with
2164 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2165 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2166 algorithms and include tests cases.
2169 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2171 [Douglas E. Engert, Steve Henson]
2173 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2174 difference in days and seconds between two tm or ASN1_TIME structures.
2177 *) Add -rev test option to s_server to just reverse order of characters
2178 received by client and send back to server. Also prints an abbreviated
2179 summary of the connection parameters.
2182 *) New option -brief for s_client and s_server to print out a brief summary
2183 of connection parameters.
2186 *) Add callbacks for arbitrary TLS extensions.
2187 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2189 *) New option -crl_download in several openssl utilities to download CRLs
2190 from CRLDP extension in certificates.
2193 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2196 *) New function X509_CRL_diff to generate a delta CRL from the difference
2197 of two full CRLs. Add support to "crl" utility.
2200 *) New functions to set lookup_crls function and to retrieve
2201 X509_STORE from X509_STORE_CTX.
2204 *) Print out deprecated issuer and subject unique ID fields in
2208 *) Extend OCSP I/O functions so they can be used for simple general purpose
2209 HTTP as well as OCSP. New wrapper function which can be used to download
2210 CRLs using the OCSP API.
2213 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2216 *) SSL_CONF* functions. These provide a common framework for application
2217 configuration using configuration files or command lines.
2220 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2221 message callback and prints the results. Needs compile time option
2222 "enable-ssl-trace". New options to s_client and s_server to enable
2226 *) New ctrl and macro to retrieve supported points extensions.
2227 Print out extension in s_server and s_client.
2230 *) New functions to retrieve certificate signature and signature
2234 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2238 *) New Suite B modes for TLS code. These use and enforce the requirements
2239 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2240 only use Suite B curves. The Suite B modes can be set by using the
2241 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2244 *) New chain verification flags for Suite B levels of security. Check
2245 algorithms are acceptable when flags are set in X509_verify_cert.
2248 *) Make tls1_check_chain return a set of flags indicating checks passed
2249 by a certificate chain. Add additional tests to handle client
2250 certificates: checks for matching certificate type and issuer name
2254 *) If an attempt is made to use a signature algorithm not in the peer
2255 preference list abort the handshake. If client has no suitable
2256 signature algorithms in response to a certificate request do not
2257 use the certificate.
2260 *) If server EC tmp key is not in client preference list abort handshake.
2263 *) Add support for certificate stores in CERT structure. This makes it
2264 possible to have different stores per SSL structure or one store in
2265 the parent SSL_CTX. Include distinct stores for certificate chain
2266 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2267 to build and store a certificate chain in CERT structure: returning
2268 an error if the chain cannot be built: this will allow applications
2269 to test if a chain is correctly configured.
2271 Note: if the CERT based stores are not set then the parent SSL_CTX
2272 store is used to retain compatibility with existing behaviour.
2276 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2277 mask based on the current session, check mask when sending client
2278 hello and checking the requested ciphersuite.
2281 *) New ctrls to retrieve and set certificate types in a certificate
2282 request message. Print out received values in s_client. If certificate
2283 types is not set with custom values set sensible values based on
2284 supported signature algorithms.
2287 *) Support for distinct client and server supported signature algorithms.
2290 *) Add certificate callback. If set this is called whenever a certificate
2291 is required by client or server. An application can decide which
2292 certificate chain to present based on arbitrary criteria: for example
2293 supported signature algorithms. Add very simple example to s_server.
2294 This fixes many of the problems and restrictions of the existing client
2295 certificate callback: for example you can now clear an existing
2296 certificate and specify the whole chain.
2299 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2300 the certificate can be used for (if anything). Set valid_flags field
2301 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2302 to have similar checks in it.
2304 Add new "cert_flags" field to CERT structure and include a "strict mode".
2305 This enforces some TLS certificate requirements (such as only permitting
2306 certificate signature algorithms contained in the supported algorithms
2307 extension) which some implementations ignore: this option should be used
2308 with caution as it could cause interoperability issues.
2311 *) Update and tidy signature algorithm extension processing. Work out
2312 shared signature algorithms based on preferences and peer algorithms
2313 and print them out in s_client and s_server. Abort handshake if no
2314 shared signature algorithms.
2317 *) Add new functions to allow customised supported signature algorithms
2318 for SSL and SSL_CTX structures. Add options to s_client and s_server
2322 *) New function SSL_certs_clear() to delete all references to certificates
2323 from an SSL structure. Before this once a certificate had been added
2324 it couldn't be removed.
2327 *) Integrate hostname, email address and IP address checking with certificate
2328 verification. New verify options supporting checking in openssl utility.
2331 *) Fixes and wildcard matching support to hostname and email checking
2332 functions. Add manual page.
2333 [Florian Weimer (Red Hat Product Security Team)]
2335 *) New functions to check a hostname email or IP address against a
2336 certificate. Add options x509 utility to print results of checks against
2340 *) Fix OCSP checking.
2341 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2343 *) Initial experimental support for explicitly trusted non-root CAs.
2344 OpenSSL still tries to build a complete chain to a root but if an
2345 intermediate CA has a trust setting included that is used. The first
2346 setting is used: whether to trust (e.g., -addtrust option to the x509
2350 *) Add -trusted_first option which attempts to find certificates in the
2351 trusted store even if an untrusted chain is also supplied.
2354 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2355 platform support for Linux and Android.
2358 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2361 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2362 When in FIPS mode the approved implementations are used as normal,
2363 when not in FIPS mode the internal unapproved versions are used instead.
2364 This means that the FIPS capable OpenSSL isn't forced to use the
2365 (often lower performance) FIPS implementations outside FIPS mode.
2368 *) Transparently support X9.42 DH parameters when calling
2369 PEM_read_bio_DHparameters. This means existing applications can handle
2370 the new parameter format automatically.
2373 *) Initial experimental support for X9.42 DH parameter format: mainly
2374 to support use of 'q' parameter for RFC5114 parameters.
2377 *) Add DH parameters from RFC5114 including test data to dhtest.
2380 *) Support for automatic EC temporary key parameter selection. If enabled
2381 the most preferred EC parameters are automatically used instead of
2382 hardcoded fixed parameters. Now a server just has to call:
2383 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2384 support ECDH and use the most appropriate parameters.
2387 *) Enhance and tidy EC curve and point format TLS extension code. Use
2388 static structures instead of allocation if default values are used.
2389 New ctrls to set curves we wish to support and to retrieve shared curves.
2390 Print out shared curves in s_server. New options to s_server and s_client
2391 to set list of supported curves.
2394 *) New ctrls to retrieve supported signature algorithms and
2395 supported curve values as an array of NIDs. Extend openssl utility
2396 to print out received values.
2399 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2400 between NIDs and the more common NIST names such as "P-256". Enhance
2401 ecparam utility and ECC method to recognise the NIST names for curves.
2404 *) Enhance SSL/TLS certificate chain handling to support different
2405 chains for each certificate instead of one chain in the parent SSL_CTX.
2408 *) Support for fixed DH ciphersuite client authentication: where both
2409 server and client use DH certificates with common parameters.
2412 *) Support for fixed DH ciphersuites: those requiring DH server
2416 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2418 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2419 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2420 X509_CINF_get_signature were reverted post internal team review.
2422 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2424 *) Build fixes for the Windows and OpenVMS platforms
2425 [Matt Caswell and Richard Levitte]
2427 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2429 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2430 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2431 dereference. This could lead to a Denial Of Service attack. Thanks to
2432 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2436 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2437 dtls1_buffer_record function under certain conditions. In particular this
2438 could occur if an attacker sent repeated DTLS records with the same
2439 sequence number but for the next epoch. The memory leak could be exploited
2440 by an attacker in a Denial of Service attack through memory exhaustion.
2441 Thanks to Chris Mueller for reporting this issue.
2445 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2446 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2447 method would be set to NULL which could later result in a NULL pointer
2448 dereference. Thanks to Frank Schmirler for reporting this issue.
2452 *) Abort handshake if server key exchange message is omitted for ephemeral
2455 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2456 reporting this issue.
2460 *) Remove non-export ephemeral RSA code on client and server. This code
2461 violated the TLS standard by allowing the use of temporary RSA keys in
2462 non-export ciphersuites and could be used by a server to effectively
2463 downgrade the RSA key length used to a value smaller than the server
2464 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2465 INRIA or reporting this issue.
2469 *) Fixed issue where DH client certificates are accepted without verification.
2470 An OpenSSL server will accept a DH certificate for client authentication
2471 without the certificate verify message. This effectively allows a client to
2472 authenticate without the use of a private key. This only affects servers
2473 which trust a client certificate authority which issues certificates
2474 containing DH keys: these are extremely rare and hardly ever encountered.
2475 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2480 *) Ensure that the session ID context of an SSL is updated when its
2481 SSL_CTX is updated via SSL_set_SSL_CTX.
2483 The session ID context is typically set from the parent SSL_CTX,
2484 and can vary with the CTX.
2487 *) Fix various certificate fingerprint issues.
2489 By using non-DER or invalid encodings outside the signed portion of a
2490 certificate the fingerprint can be changed without breaking the signature.
2491 Although no details of the signed portion of the certificate can be changed
2492 this can cause problems with some applications: e.g. those using the
2493 certificate fingerprint for blacklists.
2495 1. Reject signatures with non zero unused bits.
2497 If the BIT STRING containing the signature has non zero unused bits reject
2498 the signature. All current signature algorithms require zero unused bits.
2500 2. Check certificate algorithm consistency.
2502 Check the AlgorithmIdentifier inside TBS matches the one in the
2503 certificate signature. NB: this will result in signature failure
2504 errors for some broken certificates.
2506 Thanks to Konrad Kraszewski from Google for reporting this issue.
2508 3. Check DSA/ECDSA signatures use DER.
2510 Re-encode DSA/ECDSA signatures and compare with the original received
2511 signature. Return an error if there is a mismatch.
2513 This will reject various cases including garbage after signature
2514 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2515 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2516 (negative or with leading zeroes).
2518 Further analysis was conducted and fixes were developed by Stephen Henson
2519 of the OpenSSL core team.
2524 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2525 results on some platforms, including x86_64. This bug occurs at random
2526 with a very low probability, and is not known to be exploitable in any
2527 way, though its exact impact is difficult to determine. Thanks to Pieter
2528 Wuille (Blockstream) who reported this issue and also suggested an initial
2529 fix. Further analysis was conducted by the OpenSSL development team and
2530 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2531 the OpenSSL core team.
2535 *) Do not resume sessions on the server if the negotiated protocol
2536 version does not match the session's version. Resuming with a different
2537 version, while not strictly forbidden by the RFC, is of questionable
2538 sanity and breaks all known clients.
2539 [David Benjamin, Emilia Käsper]
2541 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2542 early CCS messages during renegotiation. (Note that because
2543 renegotiation is encrypted, this early CCS was not exploitable.)
2546 *) Tighten client-side session ticket handling during renegotiation:
2547 ensure that the client only accepts a session ticket if the server sends
2548 the extension anew in the ServerHello. Previously, a TLS client would
2549 reuse the old extension state and thus accept a session ticket if one was
2550 announced in the initial ServerHello.
2552 Similarly, ensure that the client requires a session ticket if one
2553 was advertised in the ServerHello. Previously, a TLS client would
2554 ignore a missing NewSessionTicket message.
2557 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2559 *) SRTP Memory Leak.
2561 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2562 sends a carefully crafted handshake message, to cause OpenSSL to fail
2563 to free up to 64k of memory causing a memory leak. This could be
2564 exploited in a Denial Of Service attack. This issue affects OpenSSL
2565 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2566 whether SRTP is used or configured. Implementations of OpenSSL that
2567 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2569 The fix was developed by the OpenSSL team.
2573 *) Session Ticket Memory Leak.
2575 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2576 integrity of that ticket is first verified. In the event of a session
2577 ticket integrity check failing, OpenSSL will fail to free memory
2578 causing a memory leak. By sending a large number of invalid session
2579 tickets an attacker could exploit this issue in a Denial Of Service
2584 *) Build option no-ssl3 is incomplete.
2586 When OpenSSL is configured with "no-ssl3" as a build option, servers
2587 could accept and complete a SSL 3.0 handshake, and clients could be
2588 configured to send them.
2590 [Akamai and the OpenSSL team]
2592 *) Add support for TLS_FALLBACK_SCSV.
2593 Client applications doing fallback retries should call
2594 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2596 [Adam Langley, Bodo Moeller]
2598 *) Add additional DigestInfo checks.
2600 Re-encode DigestInto in DER and check against the original when
2601 verifying RSA signature: this will reject any improperly encoded
2602 DigestInfo structures.
2604 Note: this is a precautionary measure and no attacks are currently known.
2608 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2610 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2611 SRP code can be overrun an internal buffer. Add sanity check that
2612 g, A, B < N to SRP code.
2614 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2615 Group for discovering this issue.
2619 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2620 TLS 1.0 instead of higher protocol versions when the ClientHello message
2621 is badly fragmented. This allows a man-in-the-middle attacker to force a
2622 downgrade to TLS 1.0 even if both the server and the client support a
2623 higher protocol version, by modifying the client's TLS records.
2625 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2626 researching this issue.
2630 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2631 to a denial of service attack. A malicious server can crash the client
2632 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2633 ciphersuite and sending carefully crafted handshake messages.
2635 Thanks to Felix Gröbert (Google) for discovering and researching this
2640 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2641 to leak memory. This can be exploited through a Denial of Service attack.
2642 Thanks to Adam Langley for discovering and researching this issue.
2646 *) An attacker can force openssl to consume large amounts of memory whilst
2647 processing DTLS handshake messages. This can be exploited through a
2648 Denial of Service attack.
2649 Thanks to Adam Langley for discovering and researching this issue.
2653 *) An attacker can force an error condition which causes openssl to crash
2654 whilst processing DTLS packets due to memory being freed twice. This
2655 can be exploited through a Denial of Service attack.
2656 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2661 *) If a multithreaded client connects to a malicious server using a resumed
2662 session and the server sends an ec point format extension it could write
2663 up to 255 bytes to freed memory.
2665 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2670 *) A malicious server can crash an OpenSSL client with a null pointer
2671 dereference (read) by specifying an SRP ciphersuite even though it was not
2672 properly negotiated with the client. This can be exploited through a
2673 Denial of Service attack.
2675 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2676 discovering and researching this issue.
2680 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2681 X509_name_oneline, X509_name_print_ex et al. to leak some information
2682 from the stack. Applications may be affected if they echo pretty printing
2683 output to the attacker.
2685 Thanks to Ivan Fratric (Google) for discovering this issue.
2687 [Emilia Käsper, and Steve Henson]
2689 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2690 for corner cases. (Certain input points at infinity could lead to
2691 bogus results, with non-infinity inputs mapped to infinity too.)
2694 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2696 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2697 handshake can force the use of weak keying material in OpenSSL
2698 SSL/TLS clients and servers.
2700 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2701 researching this issue. (CVE-2014-0224)
2702 [KIKUCHI Masashi, Steve Henson]
2704 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2705 OpenSSL DTLS client the code can be made to recurse eventually crashing
2708 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2710 [Imre Rad, Steve Henson]
2712 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2713 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2714 client or server. This is potentially exploitable to run arbitrary
2715 code on a vulnerable client or server.
2717 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2718 [Jüri Aedla, Steve Henson]
2720 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2721 are subject to a denial of service attack.
2723 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2724 this issue. (CVE-2014-3470)
2725 [Felix Gröbert, Ivan Fratric, Steve Henson]
2727 *) Harmonize version and its documentation. -f flag is used to display
2729 [mancha <mancha1@zoho.com>]
2731 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2732 in i2d_ECPrivateKey.
2733 [mancha <mancha1@zoho.com>]
2735 *) Fix some double frees. These are not thought to be exploitable.
2736 [mancha <mancha1@zoho.com>]
2738 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2740 *) A missing bounds check in the handling of the TLS heartbeat extension
2741 can be used to reveal up to 64k of memory to a connected client or
2744 Thanks for Neel Mehta of Google Security for discovering this bug and to
2745 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2746 preparing the fix (CVE-2014-0160)
2747 [Adam Langley, Bodo Moeller]
2749 *) Fix for the attack described in the paper "Recovering OpenSSL
2750 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2751 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2752 http://eprint.iacr.org/2014/140
2754 Thanks to Yuval Yarom and Naomi Benger for discovering this
2755 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2756 [Yuval Yarom and Naomi Benger]
2758 *) TLS pad extension: draft-agl-tls-padding-03
2760 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2761 TLS client Hello record length value would otherwise be > 255 and
2762 less that 512 pad with a dummy extension containing zeroes so it
2763 is at least 512 bytes long.
2765 [Adam Langley, Steve Henson]
2767 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2769 *) Fix for TLS record tampering bug. A carefully crafted invalid
2770 handshake could crash OpenSSL with a NULL pointer exception.
2771 Thanks to Anton Johansson for reporting this issues.
2774 *) Keep original DTLS digest and encryption contexts in retransmission
2775 structures so we can use the previous session parameters if they need
2776 to be resent. (CVE-2013-6450)
2779 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2780 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2781 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2782 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2783 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2784 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2785 [Rob Stradling, Adam Langley]
2787 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2789 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2790 supporting platforms or when small records were transferred.
2791 [Andy Polyakov, Steve Henson]
2793 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2795 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2797 This addresses the flaw in CBC record processing discovered by
2798 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2799 at: http://www.isg.rhul.ac.uk/tls/
2801 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2802 Security Group at Royal Holloway, University of London
2803 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2804 Emilia Käsper for the initial patch.
2806 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2808 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2809 ciphersuites which can be exploited in a denial of service attack.
2810 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2811 and detecting this bug and to Wolfgang Ettlinger
2812 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2816 *) Return an error when checking OCSP signatures when key is NULL.
2817 This fixes a DoS attack. (CVE-2013-0166)
2820 *) Make openssl verify return errors.
2821 [Chris Palmer <palmer@google.com> and Ben Laurie]
2823 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2824 the right response is stapled. Also change SSL_get_certificate()
2825 so it returns the certificate actually sent.
2826 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2827 [Rob Stradling <rob.stradling@comodo.com>]
2829 *) Fix possible deadlock when decoding public keys.
2832 *) Don't use TLS 1.0 record version number in initial client hello
2836 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2838 *) Sanity check record length before skipping explicit IV in TLS
2839 1.2, 1.1 and DTLS to fix DoS attack.
2841 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2842 fuzzing as a service testing platform.
2846 *) Initialise tkeylen properly when encrypting CMS messages.
2847 Thanks to Solar Designer of Openwall for reporting this issue.
2850 *) In FIPS mode don't try to use composite ciphers as they are not
2854 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2856 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2857 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2858 mean any application compiled against OpenSSL 1.0.0 headers setting
2859 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
2860 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2861 0x10000000L Any application which was previously compiled against
2862 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2863 will need to be recompiled as a result. Letting be results in
2864 inability to disable specifically TLS 1.1 and in client context,
2865 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2868 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
2869 disable just protocol X, but all protocols above X *if* there are
2870 protocols *below* X still enabled. In more practical terms it means
2871 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2872 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2873 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2877 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2879 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2880 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2881 in CRYPTO_realloc_clean.
2883 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2884 issue and to Adam Langley <agl@chromium.org> for fixing it.
2886 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2888 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2891 *) Workarounds for some broken servers that "hang" if a client hello
2892 record length exceeds 255 bytes.
2894 1. Do not use record version number > TLS 1.0 in initial client
2895 hello: some (but not all) hanging servers will now work.
2896 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2897 the number of ciphers sent in the client hello. This should be
2898 set to an even number, such as 50, for example by passing:
2899 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2900 Most broken servers should now work.
2901 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2902 TLS 1.2 client support entirely.
2905 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2908 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2910 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2911 STRING form instead of a DigestInfo.
2914 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2915 and the RSA_sign/RSA_verify functions. This was made more apparent when
2916 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2917 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2918 the correct format in RSA_verify so both forms transparently work.
2921 *) Some servers which support TLS 1.0 can choke if we initially indicate
2922 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2923 encrypted premaster secret. As a workaround use the maximum permitted
2924 client version in client hello, this should keep such servers happy
2925 and still work with previous versions of OpenSSL.
2928 *) Add support for TLS/DTLS heartbeats.
2929 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2931 *) Add support for SCTP.
2932 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2934 *) Improved PRNG seeding for VOS.
2935 [Paul Green <Paul.Green@stratus.com>]
2937 *) Extensive assembler packs updates, most notably:
2939 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2940 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2941 - x86_64: bit-sliced AES implementation;
2942 - ARM: NEON support, contemporary platforms optimizations;
2943 - s390x: z196 support;
2944 - *: GHASH and GF(2^m) multiplication implementations;
2948 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2949 (removal of unnecessary code)
2950 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2952 *) Add TLS key material exporter from RFC 5705.
2955 *) Add DTLS-SRTP negotiation from RFC 5764.
2958 *) Add Next Protocol Negotiation,
2959 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2960 disabled with a no-npn flag to config or Configure. Code donated
2962 [Adam Langley <agl@google.com> and Ben Laurie]
2964 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2965 NIST-P256, NIST-P521, with constant-time single point multiplication on
2966 typical inputs. Compiler support for the nonstandard type __uint128_t is
2967 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2968 Code made available under Apache License version 2.0.
2970 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2971 line to include this in your build of OpenSSL, and run "make depend" (or
2972 "make update"). This enables the following EC_METHODs:
2974 EC_GFp_nistp224_method()
2975 EC_GFp_nistp256_method()
2976 EC_GFp_nistp521_method()
2978 EC_GROUP_new_by_curve_name() will automatically use these (while
2979 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2981 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2983 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2984 all platforms. Move ssize_t definition from e_os.h to the public
2985 header file e_os2.h as it now appears in public header file cms.h
2988 *) New -sigopt option to the ca, req and x509 utilities. Additional
2989 signature parameters can be passed using this option and in
2993 *) Add RSA PSS signing function. This will generate and set the
2994 appropriate AlgorithmIdentifiers for PSS based on those in the
2995 corresponding EVP_MD_CTX structure. No application support yet.
2998 *) Support for companion algorithm specific ASN1 signing routines.
2999 New function ASN1_item_sign_ctx() signs a pre-initialised
3000 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3001 the appropriate parameters.
3004 *) Add new algorithm specific ASN1 verification initialisation function
3005 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3006 handling will be the same no matter what EVP_PKEY_METHOD is used.
3007 Add a PSS handler to support verification of PSS signatures: checked
3008 against a number of sample certificates.
3011 *) Add signature printing for PSS. Add PSS OIDs.
3012 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3014 *) Add algorithm specific signature printing. An individual ASN1 method
3015 can now print out signatures instead of the standard hex dump.
3017 More complex signatures (e.g. PSS) can print out more meaningful
3018 information. Include DSA version that prints out the signature
3022 *) Password based recipient info support for CMS library: implementing
3026 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3027 neatly separates the code into cipher and PBE sections and is required
3028 for some algorithms that split PBES2 into separate pieces (such as
3029 password based CMS).
3032 *) Session-handling fixes:
3033 - Fix handling of connections that are resuming with a session ID,
3034 but also support Session Tickets.
3035 - Fix a bug that suppressed issuing of a new ticket if the client
3036 presented a ticket with an expired session.
3037 - Try to set the ticket lifetime hint to something reasonable.
3038 - Make tickets shorter by excluding irrelevant information.
3039 - On the client side, don't ignore renewed tickets.
3040 [Adam Langley, Bodo Moeller (Google)]
3042 *) Fix PSK session representation.
3045 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3047 This work was sponsored by Intel.
3050 *) Add GCM support to TLS library. Some custom code is needed to split
3051 the IV between the fixed (from PRF) and explicit (from TLS record)
3052 portions. This adds all GCM ciphersuites supported by RFC5288 and
3053 RFC5289. Generalise some AES* cipherstrings to include GCM and
3054 add a special AESGCM string for GCM only.
3057 *) Expand range of ctrls for AES GCM. Permit setting invocation
3058 field on decrypt and retrieval of invocation field only on encrypt.
3061 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3062 As required by RFC5289 these ciphersuites cannot be used if for
3063 versions of TLS earlier than 1.2.
3066 *) For FIPS capable OpenSSL interpret a NULL default public key method
3067 as unset and return the appropriate default but do *not* set the default.
3068 This means we can return the appropriate method in applications that
3069 switch between FIPS and non-FIPS modes.
3072 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3073 ENGINE is used then we cannot handle that in the FIPS module so we
3074 keep original code iff non-FIPS operations are allowed.
3077 *) Add -attime option to openssl utilities.
3078 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3080 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3083 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3084 FIPS EC methods unconditionally for now.
3087 *) New build option no-ec2m to disable characteristic 2 code.
3090 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3091 all cases can be covered as some introduce binary incompatibilities.
3094 *) Redirect RSA operations to FIPS module including keygen,
3095 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3098 *) Add similar low level API blocking to ciphers.
3101 *) Low level digest APIs are not approved in FIPS mode: any attempt
3102 to use these will cause a fatal error. Applications that *really* want
3103 to use them can use the private_* version instead.
3106 *) Redirect cipher operations to FIPS module for FIPS builds.
3109 *) Redirect digest operations to FIPS module for FIPS builds.
3112 *) Update build system to add "fips" flag which will link in fipscanister.o
3113 for static and shared library builds embedding a signature if needed.
3116 *) Output TLS supported curves in preference order instead of numerical
3117 order. This is currently hardcoded for the highest order curves first.
3118 This should be configurable so applications can judge speed vs strength.
3121 *) Add TLS v1.2 server support for client authentication.
3124 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3128 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3129 FIPS modules versions.
3132 *) Add TLS v1.2 client side support for client authentication. Keep cache
3133 of handshake records longer as we don't know the hash algorithm to use
3134 until after the certificate request message is received.
3137 *) Initial TLS v1.2 client support. Add a default signature algorithms
3138 extension including all the algorithms we support. Parse new signature
3139 format in client key exchange. Relax some ECC signing restrictions for
3140 TLS v1.2 as indicated in RFC5246.
3143 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3144 to new signature format when needed using client digest preference.
3145 All server ciphersuites should now work correctly in TLS v1.2. No client
3146 support yet and no support for client certificates.
3149 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3150 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3151 ciphersuites. At present only RSA key exchange ciphersuites work with
3152 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3153 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3154 and version checking.
3157 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3158 with this defined it will not be affected by any changes to ssl internal
3159 structures. Add several utility functions to allow openssl application
3160 to work with OPENSSL_NO_SSL_INTERN defined.
3164 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
3166 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3169 *) Permit abbreviated handshakes when renegotiating using the function
3170 SSL_renegotiate_abbreviated().
3171 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3173 *) Add call to ENGINE_register_all_complete() to
3174 ENGINE_load_builtin_engines(), so some implementations get used
3175 automatically instead of needing explicit application support.
3178 *) Add support for TLS key exporter as described in RFC5705.
3179 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3181 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3182 a few changes are required:
3184 Add SSL_OP_NO_TLSv1_1 flag.
3185 Add TLSv1_1 methods.
3186 Update version checking logic to handle version 1.1.
3187 Add explicit IV handling (ported from DTLS code).
3188 Add command line options to s_client/s_server.
3191 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3193 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3194 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3195 content decryption and always return the same error. Note: this attack
3196 needs on average 2^20 messages so it only affects automated senders. The
3197 old behaviour can be re-enabled in the CMS code by setting the
3198 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3199 an MMA defence is not necessary.
3200 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3201 this issue. (CVE-2012-0884)
3204 *) Fix CVE-2011-4619: make sure we really are receiving a
3205 client hello before rejecting multiple SGC restarts. Thanks to
3206 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3209 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3211 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3212 Thanks to Antonio Martin, Enterprise Secure Access Research and
3213 Development, Cisco Systems, Inc. for discovering this bug and
3214 preparing a fix. (CVE-2012-0050)
3217 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3219 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3220 of the Vaudenay padding oracle attack on CBC mode encryption
3221 which enables an efficient plaintext recovery attack against
3222 the OpenSSL implementation of DTLS. Their attack exploits timing
3223 differences arising during decryption processing. A research
3224 paper describing this attack can be found at:
3225 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3226 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3227 Security Group at Royal Holloway, University of London
3228 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3229 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3230 for preparing the fix. (CVE-2011-4108)
3231 [Robin Seggelmann, Michael Tuexen]
3233 *) Clear bytes used for block padding of SSL 3.0 records.
3235 [Adam Langley (Google)]
3237 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3238 Kadianakis <desnacked@gmail.com> for discovering this issue and
3239 Adam Langley for preparing the fix. (CVE-2011-4619)
3240 [Adam Langley (Google)]
3242 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3243 [Andrey Kulikov <amdeich@gmail.com>]
3245 *) Prevent malformed RFC3779 data triggering an assertion failure.
3246 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3247 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3248 [Rob Austein <sra@hactrn.net>]
3250 *) Improved PRNG seeding for VOS.
3251 [Paul Green <Paul.Green@stratus.com>]
3253 *) Fix ssl_ciph.c set-up race.
3254 [Adam Langley (Google)]
3256 *) Fix spurious failures in ecdsatest.c.
3257 [Emilia Käsper (Google)]
3259 *) Fix the BIO_f_buffer() implementation (which was mixing different
3260 interpretations of the '..._len' fields).
3261 [Adam Langley (Google)]
3263 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3264 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3265 threads won't reuse the same blinding coefficients.
3267 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3268 lock to call BN_BLINDING_invert_ex, and avoids one use of
3269 BN_BLINDING_update for each BN_BLINDING structure (previously,
3270 the last update always remained unused).
3271 [Emilia Käsper (Google)]
3273 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3274 [Bob Buckholz (Google)]
3276 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3278 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3279 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3280 [Kaspar Brand <ossl@velox.ch>]
3282 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3283 for multi-threaded use of ECDH. (CVE-2011-3210)
3284 [Adam Langley (Google)]
3286 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3289 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3290 signature public key algorithm by using OID xref utilities instead.
3291 Before this you could only use some ECC ciphersuites with SHA1 only.
3294 *) Add protection against ECDSA timing attacks as mentioned in the paper
3295 by Billy Bob Brumley and Nicola Tuveri, see:
3297 http://eprint.iacr.org/2011/232.pdf
3299 [Billy Bob Brumley and Nicola Tuveri]
3301 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3303 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3304 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3306 *) Fix bug in string printing code: if *any* escaping is enabled we must
3307 escape the escape character (backslash) or the resulting string is
3311 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3313 *) Disable code workaround for ancient and obsolete Netscape browsers
3314 and servers: an attacker can use it in a ciphersuite downgrade attack.
3315 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3318 *) Fixed J-PAKE implementation error, originally discovered by
3319 Sebastien Martini, further info and confirmation from Stefan
3320 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3323 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3325 *) Fix extension code to avoid race conditions which can result in a buffer
3326 overrun vulnerability: resumed sessions must not be modified as they can
3327 be shared by multiple threads. CVE-2010-3864
3330 *) Fix WIN32 build system to correctly link an ENGINE directory into
3334 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3336 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3338 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3340 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3342 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3343 context. The operation can be customised via the ctrl mechanism in
3344 case ENGINEs want to include additional functionality.
3347 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3350 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3351 output hashes compatible with older versions of OpenSSL.
3352 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3354 *) Fix compression algorithm handling: if resuming a session use the
3355 compression algorithm of the resumed session instead of determining
3356 it from client hello again. Don't allow server to change algorithm.
3359 *) Add load_crls() function to apps tidying load_certs() too. Add option
3360 to verify utility to allow additional CRLs to be included.
3363 *) Update OCSP request code to permit adding custom headers to the request:
3364 some responders need this.
3367 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3369 [Julia Lawall <julia@diku.dk>]
3371 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3372 needlessly dereferenced structures, used obsolete functions and
3373 didn't handle all updated verify codes correctly.
3376 *) Disable MD2 in the default configuration.
3379 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3380 indicate the initial BIO being pushed or popped. This makes it possible
3381 to determine whether the BIO is the one explicitly called or as a result
3382 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3383 it handles reference counts correctly and doesn't zero out the I/O bio
3384 when it is not being explicitly popped. WARNING: applications which
3385 included workarounds for the old buggy behaviour will need to be modified
3386 or they could free up already freed BIOs.
3389 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3390 renaming to all platforms (within the 0.9.8 branch, this was
3391 done conditionally on Netware platforms to avoid a name clash).
3392 [Guenter <lists@gknw.net>]
3394 *) Add ECDHE and PSK support to DTLS.
3395 [Michael Tuexen <tuexen@fh-muenster.de>]
3397 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3401 *) Add "missing" function EVP_MD_flags() (without this the only way to
3402 retrieve a digest flags is by accessing the structure directly. Update
3403 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3404 or cipher is registered as in the "from" argument. Print out all
3405 registered digests in the dgst usage message instead of manually
3406 attempting to work them out.
3409 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3410 this allows the use of compression and extensions. Change default cipher
3411 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3412 by default unless an application cipher string requests it.
3415 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3416 key ids to find matching certificates and keys but some PKCS#12 files
3417 don't follow the (somewhat unwritten) rules and this strategy fails.
3418 Now just gather all certificates together and the first private key
3419 then look for the first certificate that matches the key.
3422 *) Support use of registered digest and cipher names for dgst and cipher
3423 commands instead of having to add each one as a special case. So now
3430 openssl dgst -sha256 foo
3432 and this works for ENGINE based algorithms too.
3436 *) Update Gost ENGINE to support parameter files.
3437 [Victor B. Wagner <vitus@cryptocom.ru>]
3439 *) Support GeneralizedTime in ca utility.
3440 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3442 *) Enhance the hash format used for certificate directory links. The new
3443 form uses the canonical encoding (meaning equivalent names will work
3444 even if they aren't identical) and uses SHA1 instead of MD5. This form
3445 is incompatible with the older format and as a result c_rehash should
3446 be used to rebuild symbolic links.
3449 *) Make PKCS#8 the default write format for private keys, replacing the
3450 traditional format. This form is standardised, more secure and doesn't
3451 include an implicit MD5 dependency.
3454 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3455 committed to OpenSSL should pass this lot as a minimum.
3458 *) Add session ticket override functionality for use by EAP-FAST.
3459 [Jouni Malinen <j@w1.fi>]
3461 *) Modify HMAC functions to return a value. Since these can be implemented
3462 in an ENGINE errors can occur.
3465 *) Type-checked OBJ_bsearch_ex.
3468 *) Type-checked OBJ_bsearch. Also some constification necessitated
3469 by type-checking. Still to come: TXT_DB, bsearch(?),
3470 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3474 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3475 seconds to a tm structure directly, instead of going through OS
3476 specific date routines. This avoids any issues with OS routines such
3477 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3478 and X509_time_adj_ex() to cover the extended range. The existing
3479 X509_time_adj() is still usable and will no longer have any date issues.
3482 *) Delta CRL support. New use deltas option which will attempt to locate
3483 and search any appropriate delta CRLs available.
3485 This work was sponsored by Google.
3488 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3489 code and add additional score elements. Validate alternate CRL paths
3490 as part of the CRL checking and indicate a new error "CRL path validation
3491 error" in this case. Applications wanting additional details can use
3492 the verify callback and check the new "parent" field. If this is not
3493 NULL CRL path validation is taking place. Existing applications won't
3494 see this because it requires extended CRL support which is off by
3497 This work was sponsored by Google.
3500 *) Support for freshest CRL extension.
3502 This work was sponsored by Google.
3505 *) Initial indirect CRL support. Currently only supported in the CRLs
3506 passed directly and not via lookup. Process certificate issuer
3507 CRL entry extension and lookup CRL entries by bother issuer name
3508 and serial number. Check and process CRL issuer entry in IDP extension.
3510 This work was sponsored by Google.
3513 *) Add support for distinct certificate and CRL paths. The CRL issuer
3514 certificate is validated separately in this case. Only enabled if
3515 an extended CRL support flag is set: this flag will enable additional
3516 CRL functionality in future.
3518 This work was sponsored by Google.
3521 *) Add support for policy mappings extension.
3523 This work was sponsored by Google.
3526 *) Fixes to pathlength constraint, self issued certificate handling,
3527 policy processing to align with RFC3280 and PKITS tests.
3529 This work was sponsored by Google.
3532 *) Support for name constraints certificate extension. DN, email, DNS
3533 and URI types are currently supported.
3535 This work was sponsored by Google.
3538 *) To cater for systems that provide a pointer-based thread ID rather
3539 than numeric, deprecate the current numeric thread ID mechanism and
3540 replace it with a structure and associated callback type. This
3541 mechanism allows a numeric "hash" to be extracted from a thread ID in
3542 either case, and on platforms where pointers are larger than 'long',
3543 mixing is done to help ensure the numeric 'hash' is usable even if it
3544 can't be guaranteed unique. The default mechanism is to use "&errno"
3545 as a pointer-based thread ID to distinguish between threads.
3547 Applications that want to provide their own thread IDs should now use
3548 CRYPTO_THREADID_set_callback() to register a callback that will call
3549 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3551 Note that ERR_remove_state() is now deprecated, because it is tied
3552 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3553 to free the current thread's error state should be replaced by
3554 ERR_remove_thread_state(NULL).
3556 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3557 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3558 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3559 application was previously providing a numeric thread callback that
3560 was inappropriate for distinguishing threads, then uniqueness might
3561 have been obtained with &errno that happened immediately in the
3562 intermediate development versions of OpenSSL; this is no longer the
3563 case, the numeric thread callback will now override the automatic use
3565 [Geoff Thorpe, with help from Bodo Moeller]
3567 *) Initial support for different CRL issuing certificates. This covers a
3568 simple case where the self issued certificates in the chain exist and
3569 the real CRL issuer is higher in the existing chain.
3571 This work was sponsored by Google.
3574 *) Removed effectively defunct crypto/store from the build.
3577 *) Revamp of STACK to provide stronger type-checking. Still to come:
3578 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3579 ASN1_STRING, CONF_VALUE.
3582 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3583 RAM on SSL connections. This option can save about 34k per idle SSL.
3586 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3587 STACK, TXT_DB, bsearch, qsort.
3590 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3591 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3592 support for data, signedData, compressedData, digestedData and
3593 encryptedData, envelopedData types included. Scripts to check against
3594 RFC4134 examples draft and interop and consistency checks of many
3595 content types and variants.
3598 *) Add options to enc utility to support use of zlib compression BIO.
3601 *) Extend mk1mf to support importing of options and assembly language
3602 files from Configure script, currently only included in VC-WIN32.
3603 The assembly language rules can now optionally generate the source
3604 files from the associated perl scripts.
3607 *) Implement remaining functionality needed to support GOST ciphersuites.
3608 Interop testing has been performed using CryptoPro implementations.
3609 [Victor B. Wagner <vitus@cryptocom.ru>]
3611 *) s390x assembler pack.
3614 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3618 *) Implement Opaque PRF Input TLS extension as specified in
3619 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3620 official specification yet and no extension type assignment by
3621 IANA exists, this extension (for now) will have to be explicitly
3622 enabled when building OpenSSL by providing the extension number
3623 to use. For example, specify an option
3625 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3627 to the "config" or "Configure" script to enable the extension,
3628 assuming extension number 0x9527 (which is a completely arbitrary
3629 and unofficial assignment based on the MD5 hash of the Internet
3630 Draft). Note that by doing so, you potentially lose
3631 interoperability with other TLS implementations since these might
3632 be using the same extension number for other purposes.
3634 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3635 opaque PRF input value to use in the handshake. This will create
3636 an internal copy of the length-'len' string at 'src', and will
3637 return non-zero for success.
3639 To get more control and flexibility, provide a callback function
3642 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3643 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3647 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3650 Callback function 'cb' will be called in handshakes, and is
3651 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3652 Argument 'arg' is for application purposes (the value as given to
3653 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3654 be provided to the callback function). The callback function
3655 has to return non-zero to report success: usually 1 to use opaque
3656 PRF input just if possible, or 2 to enforce use of the opaque PRF
3657 input. In the latter case, the library will abort the handshake
3658 if opaque PRF input is not successfully negotiated.
3660 Arguments 'peerinput' and 'len' given to the callback function
3661 will always be NULL and 0 in the case of a client. A server will
3662 see the client's opaque PRF input through these variables if
3663 available (NULL and 0 otherwise). Note that if the server
3664 provides an opaque PRF input, the length must be the same as the
3665 length of the client's opaque PRF input.
3667 Note that the callback function will only be called when creating
3668 a new session (session resumption can resume whatever was
3669 previously negotiated), and will not be called in SSL 2.0
3670 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3671 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3672 for applications that need to enforce opaque PRF input.
3676 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3679 [Victor B. Wagner <vitus@cryptocom.ru>]
3681 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3682 RFC4507bis. The encrypted ticket format is an encrypted encoded
3683 SSL_SESSION structure, that way new session features are automatically
3686 If a client application caches session in an SSL_SESSION structure
3687 support is transparent because tickets are now stored in the encoded
3690 The SSL_CTX structure automatically generates keys for ticket
3691 protection in servers so again support should be possible
3692 with no application modification.
3694 If a client or server wishes to disable RFC4507 support then the option
3695 SSL_OP_NO_TICKET can be set.
3697 Add a TLS extension debugging callback to allow the contents of any client
3698 or server extensions to be examined.
3700 This work was sponsored by Google.
3703 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3704 OpenSSL should now compile cleanly on gcc 4.2
3705 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3707 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3708 support including streaming MAC support: this is required for GOST
3709 ciphersuite support.
3710 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3712 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3713 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3714 to output in BER and PEM format.
3717 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3718 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3719 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3720 ENGINE support for HMAC keys which are unextractable. New -mac and
3721 -macopt options to dgst utility.
3724 *) New option -sigopt to dgst utility. Update dgst to use
3725 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3726 alternative signing parameters such as X9.31 or PSS in the dgst
3730 *) Change ssl_cipher_apply_rule(), the internal function that does
3731 the work each time a ciphersuite string requests enabling
3732 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3733 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3734 the order of disabled ciphersuites such that those ciphersuites
3735 that most recently went from enabled to disabled not only stay
3736 in order with respect to each other, but also have higher priority
3737 than other disabled ciphersuites the next time ciphersuites are
3740 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3741 the same ciphersuites as with "HIGH" alone, but in a specific
3742 order where the PSK ciphersuites come first (since they are the
3743 most recently disabled ciphersuites when "HIGH" is parsed).
3745 Also, change ssl_create_cipher_list() (using this new
3746 functionality) such that between otherwise identical
3747 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
3751 *) Change ssl_create_cipher_list() so that it automatically
3752 arranges the ciphersuites in reasonable order before starting
3753 to process the rule string. Thus, the definition for "DEFAULT"
3754 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3755 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3756 This makes it much easier to arrive at a reasonable default order
3757 in applications for which anonymous ciphers are OK (meaning
3758 that you can't actually use DEFAULT).
3759 [Bodo Moeller; suggested by Victor Duchovni]
3761 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3762 processing) into multiple integers instead of setting
3763 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3764 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3765 (These masks as well as the individual bit definitions are hidden
3766 away into the non-exported interface ssl/ssl_locl.h, so this
3767 change to the definition of the SSL_CIPHER structure shouldn't
3768 affect applications.) This give us more bits for each of these
3769 categories, so there is no longer a need to coagulate AES128 and
3770 AES256 into a single algorithm bit, and to coagulate Camellia128
3771 and Camellia256 into a single algorithm bit, which has led to all
3774 Thus, among other things, the kludge introduced in 0.9.7m and
3775 0.9.8e for masking out AES256 independently of AES128 or masking
3776 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3778 With the change, we also introduce new ciphersuite aliases that
3779 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3783 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3784 Use the leftmost N bytes of the signature input if the input is
3785 larger than the prime q (with N being the size in bytes of q).
3788 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3789 it yet and it is largely untested.
3792 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3795 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3796 some compilers (gcc 4.2 and later) reject their use. Safestack is
3797 reimplemented. Update ASN1 to avoid use of legacy functions.
3800 *) Win32/64 targets are linked with Winsock2.
3803 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3804 to external functions. This can be used to increase CRL handling
3805 efficiency especially when CRLs are very large by (for example) storing
3806 the CRL revoked certificates in a database.
3809 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3810 new CRLs added to a directory can be used. New command line option
3811 -verify_return_error to s_client and s_server. This causes real errors
3812 to be returned by the verify callback instead of carrying on no matter
3813 what. This reflects the way a "real world" verify callback would behave.
3816 *) GOST engine, supporting several GOST algorithms and public key formats.
3817 Kindly donated by Cryptocom.
3820 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3821 partitioned by DP are handled but no indirect CRL or reason partitioning
3822 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3823 selected via a scoring technique which handles IDP and AKID in CRLs.
3826 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3827 will ultimately be used for all verify operations: this will remove the
3828 X509_STORE dependency on certificate verification and allow alternative
3829 lookup methods. X509_STORE based implementations of these two callbacks.
3832 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3833 Modify get_crl() to find a valid (unexpired) CRL if possible.
3836 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3837 this would be called X509_CRL_cmp() but that name is already used by
3838 a function that just compares CRL issuer names. Cache several CRL
3839 extensions in X509_CRL structure and cache CRLDP in X509.
3842 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3843 this maps equivalent X509_NAME structures into a consistent structure.
3844 Name comparison can then be performed rapidly using memcmp().
3847 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3851 *) Allow digests to supply their own micalg string for S/MIME type using
3852 the ctrl EVP_MD_CTRL_MICALG.
3855 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3856 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3857 ctrl. It can then customise the structure before and/or after signing
3861 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3862 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3863 to free up any added signature OIDs.
3866 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3867 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3868 digest and cipher tables. New options added to openssl utility:
3869 list-message-digest-algorithms and list-cipher-algorithms.
3872 *) Change the array representation of binary polynomials: the list
3873 of degrees of non-zero coefficients is now terminated with -1.
3874 Previously it was terminated with 0, which was also part of the
3875 value; thus, the array representation was not applicable to
3876 polynomials where t^0 has coefficient zero. This change makes
3877 the array representation useful in a more general context.
3880 *) Various modifications and fixes to SSL/TLS cipher string
3881 handling. For ECC, the code now distinguishes between fixed ECDH
3882 with RSA certificates on the one hand and with ECDSA certificates
3883 on the other hand, since these are separate ciphersuites. The
3884 unused code for Fortezza ciphersuites has been removed.
3886 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3887 (not "ECDHE"). For consistency with the code for DH
3888 certificates, use of ECDH certificates is now considered ECDH
3889 authentication, not RSA or ECDSA authentication (the latter is
3890 merely the CA's signing algorithm and not actively used in the
3893 The temporary ciphersuite alias "ECCdraft" is no longer
3894 available, and ECC ciphersuites are no longer excluded from "ALL"
3895 and "DEFAULT". The following aliases now exist for RFC 4492
3896 ciphersuites, most of these by analogy with the DH case:
3898 kECDHr - ECDH cert, signed with RSA
3899 kECDHe - ECDH cert, signed with ECDSA
3900 kECDH - ECDH cert (signed with either RSA or ECDSA)
3901 kEECDH - ephemeral ECDH
3902 ECDH - ECDH cert or ephemeral ECDH
3908 AECDH - anonymous ECDH
3909 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3913 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3914 Use correct micalg parameters depending on digest(s) in signed message.
3917 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3918 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3921 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3922 an engine to register a method. Add ENGINE lookups for methods and
3923 functional reference processing.
3926 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
3927 EVP_{Sign,Verify}* which allow an application to customise the signature
3931 *) New -resign option to smime utility. This adds one or more signers
3932 to an existing PKCS#7 signedData structure. Also -md option to use an
3933 alternative message digest algorithm for signing.
3936 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3937 create PKCS7 structures containing multiple signers. Update smime
3938 application to support multiple signers.
3941 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3945 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3946 Reorganize PBE internals to lookup from a static table using NIDs,
3947 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3948 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3949 PRF which will be automatically used with PBES2.
3952 *) Replace the algorithm specific calls to generate keys in "req" with the
3956 *) Update PKCS#7 enveloped data routines to use new API. This is now
3957 supported by any public key method supporting the encrypt operation. A
3958 ctrl is added to allow the public key algorithm to examine or modify
3959 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3963 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3964 a default digest type to use. In most cases this will be SHA1 but some
3965 algorithms (such as GOST) need to specify an alternative digest. The
3966 return value indicates how strong the preference is 1 means optional and
3967 2 is mandatory (that is it is the only supported type). Modify
3968 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3969 use the default md. Update openssl utilities to use the default digest
3970 type for signing if it is not explicitly indicated.
3973 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3974 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3975 signing method from the key type. This effectively removes the link
3976 between digests and public key types.
3979 *) Add an OID cross reference table and utility functions. Its purpose is to
3980 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3981 rsaEncryption. This will allow some of the algorithm specific hackery
3982 needed to use the correct OID to be removed.
3985 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3986 structures for PKCS7_sign(). They are now set up by the relevant public
3990 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3993 *) Add support for key derivation (agreement) in the API, DH method and
3997 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3998 public and private key formats. As a side effect these add additional
3999 command line functionality not previously available: DSA signatures can be
4000 generated and verified using pkeyutl and DH key support and generation in
4005 [Oliver Tappe <zooey@hirschkaefer.de>]
4007 *) New make target "install_html_docs" installs HTML renditions of the
4009 [Oliver Tappe <zooey@hirschkaefer.de>]
4011 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4012 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4013 support key and parameter generation and add initial key generation
4014 functionality for RSA.
4017 *) Add functions for main EVP_PKEY_method operations. The undocumented
4018 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4019 EVP_PKEY_{encrypt,decrypt}_old.
4022 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4023 key API, doesn't do much yet.
4026 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4027 public key algorithms. New option to openssl utility:
4028 "list-public-key-algorithms" to print out info.
4031 *) Implement the Supported Elliptic Curves Extension for
4032 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4035 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4036 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4039 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4040 utilities such as rsa, dsa, dsaparam etc except they process any key
4044 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4045 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4046 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4050 *) Initial support for pluggable public key ASN1.
4051 De-spaghettify the public key ASN1 handling. Move public and private
4052 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4053 algorithm specific handling to a single module within the relevant
4054 algorithm directory. Add functions to allow (near) opaque processing
4055 of public and private key structures.
4058 *) Implement the Supported Point Formats Extension for
4059 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4062 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4063 for the psk identity [hint] and the psk callback functions to the
4064 SSL_SESSION, SSL and SSL_CTX structure.
4067 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4071 SSL_CTX_use_psk_identity_hint
4072 SSL_get_psk_identity_hint
4073 SSL_get_psk_identity
4074 SSL_use_psk_identity_hint
4076 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4078 *) Add RFC 3161 compliant time stamp request creation, response generation
4079 and response verification functionality.
4080 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4082 *) Add initial support for TLS extensions, specifically for the server_name
4083 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4084 have new members for a host name. The SSL data structure has an
4085 additional member SSL_CTX *initial_ctx so that new sessions can be
4086 stored in that context to allow for session resumption, even after the
4087 SSL has been switched to a new SSL_CTX in reaction to a client's
4088 server_name extension.
4090 New functions (subject to change):
4092 SSL_get_servername()
4093 SSL_get_servername_type()
4096 New CTRL codes and macros (subject to change):
4098 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4099 - SSL_CTX_set_tlsext_servername_callback()
4100 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4101 - SSL_CTX_set_tlsext_servername_arg()
4102 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4104 openssl s_client has a new '-servername ...' option.
4106 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4107 '-key2 ...', '-servername_fatal' (subject to change). This allows
4108 testing the HostName extension for a specific single host name ('-cert'
4109 and '-key' remain fallbacks for handshakes without HostName
4110 negotiation). If the unrecognized_name alert has to be sent, this by
4111 default is a warning; it becomes fatal with the '-servername_fatal'
4114 [Peter Sylvester, Remy Allais, Christophe Renou]
4116 *) Whirlpool hash implementation is added.
4119 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4120 bn(64,32). Because of instruction set limitations it doesn't have
4121 any negative impact on performance. This was done mostly in order
4122 to make it possible to share assembler modules, such as bn_mul_mont
4123 implementations, between 32- and 64-bit builds without hassle.
4126 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4127 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4131 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4132 dedicated Montgomery multiplication procedure, is introduced.
4133 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4134 "64-bit" performance on certain 32-bit targets.
4137 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4138 in SSL structures. New SSL ctrl to set maximum send fragment size.
4139 Save memory by setting the I/O buffer sizes dynamically instead of
4140 using the maximum available value.
4143 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4144 in addition to the text details.
4147 *) Very, very preliminary EXPERIMENTAL support for printing of general
4148 ASN1 structures. This currently produces rather ugly output and doesn't
4149 handle several customised structures at all.
4152 *) Integrated support for PVK file format and some related formats such
4153 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4154 these in the 'rsa' and 'dsa' utilities.
4157 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4160 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4161 place for the (very old) "NETSCAPE" format certificates which are now
4162 handled using new ASN1 code equivalents.
4165 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4166 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4167 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4170 *) Modify CRL distribution points extension code to print out previously
4171 unsupported fields. Enhance extension setting code to allow setting of
4175 *) Add print and set support for Issuing Distribution Point CRL extension.
4178 *) Change 'Configure' script to enable Camellia by default.
4181 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4183 *) When rejecting SSL/TLS records due to an incorrect version number, never
4184 update s->server with a new major version number. As of
4185 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4186 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4187 the previous behavior could result in a read attempt at NULL when
4188 receiving specific incorrect SSL/TLS records once record payload
4189 protection is active. (CVE-2010-0740)
4190 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4192 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4193 could be crashed if the relevant tables were not present (e.g. chrooted).
4194 [Tomas Hoger <thoger@redhat.com>]
4196 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4198 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4199 [Martin Olsson, Neel Mehta]
4201 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4202 accommodate for stack sorting, always a write lock!).
4205 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4206 excessive delays in the RAND_poll(): over a minute. As a workaround
4207 include a time check in the inner Heap32Next loop too.
4210 *) The code that handled flushing of data in SSL/TLS originally used the
4211 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4212 the problem outlined in PR#1949. The fix suggested there however can
4213 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4214 of Apache). So instead simplify the code to flush unconditionally.
4215 This should be fine since flushing with no data to flush is a no op.
4218 *) Handle TLS versions 2.0 and later properly and correctly use the
4219 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4220 off ancient servers have a habit of sticking around for a while...
4223 *) Modify compression code so it frees up structures without using the
4224 ex_data callbacks. This works around a problem where some applications
4225 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4226 restarting) then use compression (e.g. SSL with compression) later.
4227 This results in significant per-connection memory leaks and
4228 has caused some security issues including CVE-2008-1678 and
4232 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4233 change when encrypting or decrypting.
4236 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4237 connect and renegotiate with servers which do not support RI.
4238 Until RI is more widely deployed this option is enabled by default.
4241 *) Add "missing" ssl ctrls to clear options and mode.
4244 *) If client attempts to renegotiate and doesn't support RI respond with
4245 a no_renegotiation alert as required by RFC5746. Some renegotiating
4246 TLS clients will continue a connection gracefully when they receive
4247 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4248 waiting for a server hello which it will never receive. Now we treat a
4249 received no_renegotiation alert as a fatal error. This is because
4250 applications requesting a renegotiation might well expect it to succeed
4251 and would have no code in place to handle the server denying it so the
4252 only safe thing to do is to terminate the connection.
4255 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4256 peer supports secure renegotiation and 0 otherwise. Print out peer
4257 renegotiation support in s_client/s_server.
4260 *) Replace the highly broken and deprecated SPKAC certification method with
4261 the updated NID creation version. This should correctly handle UTF8.
4264 *) Implement RFC5746. Re-enable renegotiation but require the extension
4265 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4266 turns out to be a bad idea. It has been replaced by
4267 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4268 SSL_CTX_set_options(). This is really not recommended unless you
4269 know what you are doing.
4270 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4272 *) Fixes to stateless session resumption handling. Use initial_ctx when
4273 issuing and attempting to decrypt tickets in case it has changed during
4274 servername handling. Use a non-zero length session ID when attempting
4275 stateless session resumption: this makes it possible to determine if
4276 a resumption has occurred immediately after receiving server hello
4277 (several places in OpenSSL subtly assume this) instead of later in
4281 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4282 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4283 fixes for a few places where the return code is not checked
4285 [Julia Lawall <julia@diku.dk>]
4287 *) Add --strict-warnings option to Configure script to include devteam
4288 warnings in other configurations.
4291 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4292 makes it possible to install openssl libraries in locations which
4293 have names other than "lib", for example "/usr/lib64" which some
4295 [Steve Henson, based on patch from Jeremy Utley]
4297 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4298 X690 8.9.12 and can produce some misleading textual output of OIDs.
4299 [Steve Henson, reported by Dan Kaminsky]
4301 *) Delete MD2 from algorithm tables. This follows the recommendation in
4302 several standards that it is not used in new applications due to
4303 several cryptographic weaknesses. For binary compatibility reasons
4304 the MD2 API is still compiled in by default.
4307 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4311 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4312 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4314 [Guenter <lists@gknw.net>]
4316 *) Fix the server certificate chain building code to use X509_verify_cert(),
4317 it used to have an ad-hoc builder which was unable to cope with anything
4318 other than a simple chain.
4319 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4321 *) Don't check self signed certificate signatures in X509_verify_cert()
4322 by default (a flag can override this): it just wastes time without
4323 adding any security. As a useful side effect self signed root CAs
4324 with non-FIPS digests are now usable in FIPS mode.
4327 *) In dtls1_process_out_of_seq_message() the check if the current message
4328 is already buffered was missing. For every new message was memory
4329 allocated, allowing an attacker to perform an denial of service attack
4330 with sending out of seq handshake messages until there is no memory
4331 left. Additionally every future message was buffered, even if the
4332 sequence number made no sense and would be part of another handshake.
4333 So only messages with sequence numbers less than 10 in advance will be
4334 buffered. (CVE-2009-1378)
4335 [Robin Seggelmann, discovered by Daniel Mentz]
4337 *) Records are buffered if they arrive with a future epoch to be
4338 processed after finishing the corresponding handshake. There is
4339 currently no limitation to this buffer allowing an attacker to perform
4340 a DOS attack with sending records with future epochs until there is no
4341 memory left. This patch adds the pqueue_size() function to determine
4342 the size of a buffer and limits the record buffer to 100 entries.
4344 [Robin Seggelmann, discovered by Daniel Mentz]
4346 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4347 parent structure is freed. (CVE-2009-1379)
4350 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4351 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4354 [Ilya O. <vrghost@gmail.com>]
4356 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4358 *) Disable renegotiation completely - this fixes a severe security
4359 problem (CVE-2009-3555) at the cost of breaking all
4360 renegotiation. Renegotiation can be re-enabled by setting
4361 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4362 run-time. This is really not recommended unless you know what
4366 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4368 *) Don't set val to NULL when freeing up structures, it is freed up by
4369 underlying code. If sizeof(void *) > sizeof(long) this can result in
4370 zeroing past the valid field. (CVE-2009-0789)
4371 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4373 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4374 checked correctly. This would allow some invalid signed attributes to
4375 appear to verify correctly. (CVE-2009-0591)
4376 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4378 *) Reject UniversalString and BMPString types with invalid lengths. This
4379 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4380 a legal length. (CVE-2009-0590)
4383 *) Set S/MIME signing as the default purpose rather than setting it
4384 unconditionally. This allows applications to override it at the store
4388 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4389 to handle some structures.
4392 *) Improve efficiency of mem_gets: don't search whole buffer each time
4394 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4396 *) New -hex option for openssl rand.
4399 *) Print out UTF8String and NumericString when parsing ASN1.
4402 *) Support NumericString type for name components.
4405 *) Allow CC in the environment to override the automatically chosen
4406 compiler. Note that nothing is done to ensure flags work with the
4410 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4412 *) Properly check EVP_VerifyFinal() and similar return values
4414 [Ben Laurie, Bodo Moeller, Google Security Team]
4416 *) Enable TLS extensions by default.
4419 *) Allow the CHIL engine to be loaded, whether the application is
4420 multithreaded or not. (This does not release the developer from the
4421 obligation to set up the dynamic locking callbacks.)
4422 [Sander Temme <sander@temme.net>]
4424 *) Use correct exit code if there is an error in dgst command.
4425 [Steve Henson; problem pointed out by Roland Dirlewanger]
4427 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4428 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4431 *) Add experimental JPAKE support, including demo authentication in
4432 s_client and s_server.
4435 *) Set the comparison function in v3_addr_canonize().
4436 [Rob Austein <sra@hactrn.net>]
4438 *) Add support for XMPP STARTTLS in s_client.
4439 [Philip Paeps <philip@freebsd.org>]
4441 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4442 to ensure that even with this option, only ciphersuites in the
4443 server's preference list will be accepted. (Note that the option
4444 applies only when resuming a session, so the earlier behavior was
4445 just about the algorithm choice for symmetric cryptography.)
4448 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4450 *) Fix NULL pointer dereference if a DTLS server received
4451 ChangeCipherSpec as first record (CVE-2009-1386).
4454 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4455 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4458 *) The fix in 0.9.8c that supposedly got rid of unsafe
4459 double-checked locking was incomplete for RSA blinding,
4460 addressing just one layer of what turns out to have been
4461 doubly unsafe triple-checked locking.
4463 So now fix this for real by retiring the MONT_HELPER macro
4464 in crypto/rsa/rsa_eay.c.
4466 [Bodo Moeller; problem pointed out by Marius Schilder]
4468 *) Various precautionary measures:
4470 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4472 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4473 (NB: This would require knowledge of the secret session ticket key
4474 to exploit, in which case you'd be SOL either way.)
4476 - Change bn_nist.c so that it will properly handle input BIGNUMs
4477 outside the expected range.
4479 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4482 [Neel Mehta, Bodo Moeller]
4484 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4485 the load fails. Useful for distros.
4486 [Ben Laurie and the FreeBSD team]
4488 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4491 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4494 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4496 This work was sponsored by Logica.
4499 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4500 keystores. Support for SSL/TLS client authentication too.
4501 Not compiled unless enable-capieng specified to Configure.
4503 This work was sponsored by Logica.
4506 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4507 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4508 attribute creation routines such as certificate requests and PKCS#12
4512 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4514 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4515 handshake which could lead to a client crash as found using the
4516 Codenomicon TLS test suite (CVE-2008-1672)
4517 [Steve Henson, Mark Cox]
4519 *) Fix double free in TLS server name extensions which could lead to
4520 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4523 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4525 Clear the error queue to ensure that error entries left from
4526 older function calls do not interfere with the correct operation.
4527 [Lutz Jaenicke, Erik de Castro Lopo]
4529 *) Remove root CA certificates of commercial CAs:
4531 The OpenSSL project does not recommend any specific CA and does not
4532 have any policy with respect to including or excluding any CA.
4533 Therefore it does not make any sense to ship an arbitrary selection
4534 of root CA certificates with the OpenSSL software.
4537 *) RSA OAEP patches to fix two separate invalid memory reads.
4538 The first one involves inputs when 'lzero' is greater than
4539 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4540 before the beginning of from). The second one involves inputs where
4541 the 'db' section contains nothing but zeroes (there is a one-byte
4542 invalid read after the end of 'db').
4543 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4545 *) Partial backport from 0.9.9-dev:
4547 Introduce bn_mul_mont (dedicated Montgomery multiplication
4548 procedure) as a candidate for BIGNUM assembler implementation.
4549 While 0.9.9-dev uses assembler for various architectures, only
4550 x86_64 is available by default here in the 0.9.8 branch, and
4551 32-bit x86 is available through a compile-time setting.
4553 To try the 32-bit x86 assembler implementation, use Configure
4554 option "enable-montasm" (which exists only for this backport).
4556 As "enable-montasm" for 32-bit x86 disclaims code stability
4557 anyway, in this constellation we activate additional code
4558 backported from 0.9.9-dev for further performance improvements,
4559 namely BN_from_montgomery_word. (To enable this otherwise,
4560 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4562 [Andy Polyakov (backport partially by Bodo Moeller)]
4564 *) Add TLS session ticket callback. This allows an application to set
4565 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4566 values. This is useful for key rollover for example where several key
4567 sets may exist with different names.
4570 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4571 This was broken until now in 0.9.8 releases, such that the only way
4572 a registered ENGINE could be used (assuming it initialises
4573 successfully on the host) was to explicitly set it as the default
4574 for the relevant algorithms. This is in contradiction with 0.9.7
4575 behaviour and the documentation. With this fix, when an ENGINE is
4576 registered into a given algorithm's table of implementations, the
4577 'uptodate' flag is reset so that auto-discovery will be used next
4578 time a new context for that algorithm attempts to select an
4580 [Ian Lister (tweaked by Geoff Thorpe)]
4582 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4583 implementation in the following ways:
4585 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4588 Lack of BER streaming support means one pass streaming processing is
4589 only supported if data is detached: setting the streaming flag is
4590 ignored for embedded content.
4592 CMS support is disabled by default and must be explicitly enabled
4593 with the enable-cms configuration option.
4596 *) Update the GMP engine glue to do direct copies between BIGNUM and
4597 mpz_t when openssl and GMP use the same limb size. Otherwise the
4598 existing "conversion via a text string export" trick is still used.
4599 [Paul Sheer <paulsheer@gmail.com>]
4601 *) Zlib compression BIO. This is a filter BIO which compressed and
4602 uncompresses any data passed through it.
4605 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4606 RFC3394 compatible AES key wrapping.
4609 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4610 sets string data without copying. X509_ALGOR_set0() and
4611 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4612 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4613 from an X509_ATTRIBUTE structure optionally checking it occurs only
4614 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4618 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4619 to get the expected BN_FLG_CONSTTIME behavior.
4620 [Bodo Moeller (Google)]
4624 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4625 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4626 - added some more tests to do_tests.pl
4627 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4628 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4629 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4630 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4631 - various changes to netware.pl to enable gcc-cross builds on Win32
4633 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4634 - various changes to fix missing prototype warnings
4635 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4636 - added AES, WHIRLPOOL and CPUID assembler code to build files
4637 - added missing AES assembler make rules to mk1mf.pl
4638 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4639 [Guenter Knauf <eflash@gmx.net>]
4641 *) Implement certificate status request TLS extension defined in RFC3546.
4642 A client can set the appropriate parameters and receive the encoded
4643 OCSP response via a callback. A server can query the supplied parameters
4644 and set the encoded OCSP response in the callback. Add simplified examples
4645 to s_client and s_server.
4648 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4650 *) Fix various bugs:
4651 + Binary incompatibility of ssl_ctx_st structure
4652 + DTLS interoperation with non-compliant servers
4653 + Don't call get_session_cb() without proposed session
4654 + Fix ia64 assembler code
4655 [Andy Polyakov, Steve Henson]
4657 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4659 *) DTLS Handshake overhaul. There were longstanding issues with
4660 OpenSSL DTLS implementation, which were making it impossible for
4661 RFC 4347 compliant client to communicate with OpenSSL server.
4662 Unfortunately just fixing these incompatibilities would "cut off"
4663 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4664 server keeps tolerating non RFC compliant syntax. The opposite is
4665 not true, 0.9.8f client can not communicate with earlier server.
4666 This update even addresses CVE-2007-4995.
4669 *) Changes to avoid need for function casts in OpenSSL: some compilers
4670 (gcc 4.2 and later) reject their use.
4671 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4674 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4675 RFC4507bis. The encrypted ticket format is an encrypted encoded
4676 SSL_SESSION structure, that way new session features are automatically
4679 If a client application caches session in an SSL_SESSION structure
4680 support is transparent because tickets are now stored in the encoded
4683 The SSL_CTX structure automatically generates keys for ticket
4684 protection in servers so again support should be possible
4685 with no application modification.
4687 If a client or server wishes to disable RFC4507 support then the option
4688 SSL_OP_NO_TICKET can be set.
4690 Add a TLS extension debugging callback to allow the contents of any client
4691 or server extensions to be examined.
4693 This work was sponsored by Google.
4696 *) Add initial support for TLS extensions, specifically for the server_name
4697 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4698 have new members for a host name. The SSL data structure has an
4699 additional member SSL_CTX *initial_ctx so that new sessions can be
4700 stored in that context to allow for session resumption, even after the
4701 SSL has been switched to a new SSL_CTX in reaction to a client's
4702 server_name extension.
4704 New functions (subject to change):
4706 SSL_get_servername()
4707 SSL_get_servername_type()
4710 New CTRL codes and macros (subject to change):
4712 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4713 - SSL_CTX_set_tlsext_servername_callback()
4714 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4715 - SSL_CTX_set_tlsext_servername_arg()
4716 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4718 openssl s_client has a new '-servername ...' option.
4720 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4721 '-key2 ...', '-servername_fatal' (subject to change). This allows
4722 testing the HostName extension for a specific single host name ('-cert'
4723 and '-key' remain fallbacks for handshakes without HostName
4724 negotiation). If the unrecognized_name alert has to be sent, this by
4725 default is a warning; it becomes fatal with the '-servername_fatal'
4728 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4730 *) Add AES and SSE2 assembly language support to VC++ build.
4733 *) Mitigate attack on final subtraction in Montgomery reduction.
4736 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4737 (which previously caused an internal error).
4740 *) Squeeze another 10% out of IGE mode when in != out.
4743 *) AES IGE mode speedup.
4744 [Dean Gaudet (Google)]
4746 *) Add the Korean symmetric 128-bit cipher SEED (see
4747 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4748 add SEED ciphersuites from RFC 4162:
4750 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4751 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4752 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4753 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4755 To minimize changes between patchlevels in the OpenSSL 0.9.8
4756 series, SEED remains excluded from compilation unless OpenSSL
4757 is configured with 'enable-seed'.
4758 [KISA, Bodo Moeller]
4760 *) Mitigate branch prediction attacks, which can be practical if a
4761 single processor is shared, allowing a spy process to extract
4762 information. For detailed background information, see
4763 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4764 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4765 and Necessary Software Countermeasures"). The core of the change
4766 are new versions BN_div_no_branch() and
4767 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4768 respectively, which are slower, but avoid the security-relevant
4769 conditional branches. These are automatically called by BN_div()
4770 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4771 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4772 remove a conditional branch.
4774 BN_FLG_CONSTTIME is the new name for the previous
4775 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4776 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4777 in the exponent causes BN_mod_exp_mont() to use the alternative
4778 implementation in BN_mod_exp_mont_consttime().) The old name
4779 remains as a deprecated alias.
4781 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4782 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4783 constant-time implementations for more than just exponentiation.
4784 Here too the old name is kept as a deprecated alias.
4786 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4787 the BN_BLINDING structure gets an independent copy of the
4788 modulus. This means that the previous "BIGNUM *m" argument to
4789 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4790 essentially becomes "const BIGNUM *m", although we can't actually
4791 change this in the header file before 0.9.9. It allows
4792 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4793 enable BN_FLG_CONSTTIME.
4795 [Matthew D Wood (Intel Corp)]
4797 *) In the SSL/TLS server implementation, be strict about session ID
4798 context matching (which matters if an application uses a single
4799 external cache for different purposes). Previously,
4800 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4801 set. This did ensure strict client verification, but meant that,
4802 with applications using a single external cache for quite
4803 different requirements, clients could circumvent ciphersuite
4804 restrictions for a given session ID context by starting a session
4805 in a different context.
4808 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4809 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4810 authentication-only ciphersuites.
4813 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4814 not complete and could lead to a possible single byte overflow
4815 (CVE-2007-5135) [Ben Laurie]
4817 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4819 *) Since AES128 and AES256 (and similarly Camellia128 and
4820 Camellia256) share a single mask bit in the logic of
4821 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4822 kludge to work properly if AES128 is available and AES256 isn't
4823 (or if Camellia128 is available and Camellia256 isn't).
4826 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4827 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4828 When a point or a seed is encoded in a BIT STRING, we need to
4829 prevent the removal of trailing zero bits to get the proper DER
4830 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4831 of a NamedBitList, for which trailing 0 bits need to be removed.)
4834 *) Have SSL/TLS server implementation tolerate "mismatched" record
4835 protocol version while receiving ClientHello even if the
4836 ClientHello is fragmented. (The server can't insist on the
4837 particular protocol version it has chosen before the ServerHello
4838 message has informed the client about his choice.)
4841 *) Add RFC 3779 support.
4842 [Rob Austein for ARIN, Ben Laurie]
4844 *) Load error codes if they are not already present instead of using a
4845 static variable. This allows them to be cleanly unloaded and reloaded.
4846 Improve header file function name parsing.
4849 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4850 or CAPABILITY handshake as required by RFCs.
4853 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4855 *) Introduce limits to prevent malicious keys being able to
4856 cause a denial of service. (CVE-2006-2940)
4857 [Steve Henson, Bodo Moeller]
4859 *) Fix ASN.1 parsing of certain invalid structures that can result
4860 in a denial of service. (CVE-2006-2937) [Steve Henson]
4862 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4863 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4865 *) Fix SSL client code which could crash if connecting to a
4866 malicious SSLv2 server. (CVE-2006-4343)
4867 [Tavis Ormandy and Will Drewry, Google Security Team]
4869 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4870 match only those. Before that, "AES256-SHA" would be interpreted
4871 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4872 the same strength classification in 0.9.7h) as we currently only
4873 have a single AES bit in the ciphersuite description bitmap.
4874 That change, however, also applied to ciphersuite strings such as
4875 "RC4-MD5" that intentionally matched multiple ciphersuites --
4876 namely, SSL 2.0 ciphersuites in addition to the more common ones
4877 from SSL 3.0/TLS 1.0.
4879 So we change the selection algorithm again: Naming an explicit
4880 ciphersuite selects this one ciphersuite, and any other similar
4881 ciphersuite (same bitmap) from *other* protocol versions.
4882 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4883 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4885 Since SSL 2.0 does not have any ciphersuites for which the
4886 128/256 bit distinction would be relevant, this works for now.
4887 The proper fix will be to use different bits for AES128 and
4888 AES256, which would have avoided the problems from the beginning;
4889 however, bits are scarce, so we can only do this in a new release
4890 (not just a patchlevel) when we can change the SSL_CIPHER
4891 definition to split the single 'unsigned long mask' bitmap into
4892 multiple values to extend the available space.
4896 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4898 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4899 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4901 *) Add AES IGE and biIGE modes.
4904 *) Change the Unix randomness entropy gathering to use poll() when
4905 possible instead of select(), since the latter has some
4906 undesirable limitations.
4907 [Darryl Miles via Richard Levitte and Bodo Moeller]
4909 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4910 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4911 cannot be implicitly activated as part of, e.g., the "AES" alias.
4912 However, please upgrade to OpenSSL 0.9.9[-dev] for
4913 non-experimental use of the ECC ciphersuites to get TLS extension
4914 support, which is required for curve and point format negotiation
4915 to avoid potential handshake problems.
4918 *) Disable rogue ciphersuites:
4920 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4921 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4922 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4924 The latter two were purportedly from
4925 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4928 Also deactivate the remaining ciphersuites from
4929 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4930 unofficial, and the ID has long expired.
4933 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4934 dual-core machines) and other potential thread-safety issues.
4937 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4938 versions), which is now available for royalty-free use
4939 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4940 Also, add Camellia TLS ciphersuites from RFC 4132.
4942 To minimize changes between patchlevels in the OpenSSL 0.9.8
4943 series, Camellia remains excluded from compilation unless OpenSSL
4944 is configured with 'enable-camellia'.
4947 *) Disable the padding bug check when compression is in use. The padding
4948 bug check assumes the first packet is of even length, this is not
4949 necessarily true if compression is enabled and can result in false
4950 positives causing handshake failure. The actual bug test is ancient
4951 code so it is hoped that implementations will either have fixed it by
4952 now or any which still have the bug do not support compression.
4955 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4957 *) When applying a cipher rule check to see if string match is an explicit
4958 cipher suite and only match that one cipher suite if it is.
4961 *) Link in manifests for VC++ if needed.
4962 [Austin Ziegler <halostatue@gmail.com>]
4964 *) Update support for ECC-based TLS ciphersuites according to
4965 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4966 TLS extensions, which are supported starting with the 0.9.9
4967 branch, not in the OpenSSL 0.9.8 branch).
4970 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4971 opaque EVP_CIPHER_CTX handling.
4974 *) Fixes and enhancements to zlib compression code. We now only use
4975 "zlib1.dll" and use the default __cdecl calling convention on Win32
4976 to conform with the standards mentioned here:
4977 http://www.zlib.net/DLL_FAQ.txt
4978 Static zlib linking now works on Windows and the new --with-zlib-include
4979 --with-zlib-lib options to Configure can be used to supply the location
4980 of the headers and library. Gracefully handle case where zlib library
4984 *) Several fixes and enhancements to the OID generation code. The old code
4985 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4986 handle numbers larger than ULONG_MAX, truncated printing and had a
4987 non standard OBJ_obj2txt() behaviour.
4990 *) Add support for building of engines under engine/ as shared libraries
4991 under VC++ build system.
4994 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4995 Hopefully, we will not see any false combination of paths any more.
4998 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5000 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5001 (part of SSL_OP_ALL). This option used to disable the
5002 countermeasure against man-in-the-middle protocol-version
5003 rollback in the SSL 2.0 server implementation, which is a bad
5004 idea. (CVE-2005-2969)
5006 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5007 for Information Security, National Institute of Advanced Industrial
5008 Science and Technology [AIST], Japan)]
5010 *) Add two function to clear and return the verify parameter flags.
5013 *) Keep cipherlists sorted in the source instead of sorting them at
5014 runtime, thus removing the need for a lock.
5017 *) Avoid some small subgroup attacks in Diffie-Hellman.
5018 [Nick Mathewson and Ben Laurie]
5020 *) Add functions for well-known primes.
5023 *) Extended Windows CE support.
5024 [Satoshi Nakamura and Andy Polyakov]
5026 *) Initialize SSL_METHOD structures at compile time instead of during
5027 runtime, thus removing the need for a lock.
5030 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5031 attempting to decrypt each encrypted key in turn. Add support to
5035 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5037 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5040 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5043 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5044 key into the same file any more.
5047 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5050 *) Add -utf8 command line and config file option to 'ca'.
5051 [Stefan <stf@udoma.org]
5053 *) Removed the macro des_crypt(), as it seems to conflict with some
5054 libraries. Use DES_crypt().
5057 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5058 involves renaming the source and generated shared-libs for
5059 both. The engines will accept the corrected or legacy ids
5060 ('ncipher' and '4758_cca' respectively) when binding. NB,
5061 this only applies when building 'shared'.
5062 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5064 *) Add attribute functions to EVP_PKEY structure. Modify
5065 PKCS12_create() to recognize a CSP name attribute and
5066 use it. Make -CSP option work again in pkcs12 utility.
5069 *) Add new functionality to the bn blinding code:
5070 - automatic re-creation of the BN_BLINDING parameters after
5071 a fixed number of uses (currently 32)
5072 - add new function for parameter creation
5073 - introduce flags to control the update behaviour of the
5074 BN_BLINDING parameters
5075 - hide BN_BLINDING structure
5076 Add a second BN_BLINDING slot to the RSA structure to improve
5077 performance when a single RSA object is shared among several
5081 *) Add support for DTLS.
5082 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5084 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5085 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5088 *) Remove buggy and incomplete DH cert support from
5089 ssl/ssl_rsa.c and ssl/s3_both.c
5092 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5093 the apps/openssl applications.
5096 *) Compile clean with "-Wall -Wmissing-prototypes
5097 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5098 DEBUG_SAFESTACK must also be set.
5101 *) Change ./Configure so that certain algorithms can be disabled by default.
5102 The new counterpiece to "no-xxx" is "enable-xxx".
5104 The patented RC5 and MDC2 algorithms will now be disabled unless
5105 "enable-rc5" and "enable-mdc2", respectively, are specified.
5107 (IDEA remains enabled despite being patented. This is because IDEA
5108 is frequently required for interoperability, and there is no license
5109 fee for non-commercial use. As before, "no-idea" can be used to
5110 avoid this algorithm.)
5114 *) Add processing of proxy certificates (see RFC 3820). This work was
5115 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5116 EGEE (Enabling Grids for E-science in Europe).
5119 *) RC4 performance overhaul on modern architectures/implementations, such
5120 as Intel P4, IA-64 and AMD64.
5123 *) New utility extract-section.pl. This can be used specify an alternative
5124 section number in a pod file instead of having to treat each file as
5125 a separate case in Makefile. This can be done by adding two lines to the
5128 =for comment openssl_section:XXX
5130 The blank line is mandatory.
5134 *) New arguments -certform, -keyform and -pass for s_client and s_server
5135 to allow alternative format key and certificate files and passphrase
5139 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5140 update associated structures and add various utility functions.
5142 Add new policy related verify parameters, include policy checking in
5143 standard verify code. Enhance 'smime' application with extra parameters
5144 to support policy checking and print out.
5147 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5148 Nehemiah processors. These extensions support AES encryption in hardware
5149 as well as RNG (though RNG support is currently disabled).
5150 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5152 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5155 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5156 [Andy Polyakov and a number of other people]
5158 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5159 implementation contributed by IBM.
5160 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5162 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5163 exponent rather than 'unsigned long'. There is a corresponding change to
5164 the new 'rsa_keygen' element of the RSA_METHOD structure.
5165 [Jelte Jansen, Geoff Thorpe]
5167 *) Functionality for creating the initial serial number file is now
5168 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5170 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5171 number file to 1, which is bound to cause problems. To avoid
5172 the problems while respecting compatibility between different 0.9.7
5173 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5174 CA.pl for serial number initialization. With the new release 0.9.8,
5175 we can fix the problem directly in the 'ca' utility.)
5178 *) Reduced header interdependencies by declaring more opaque objects in
5179 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5180 give fewer recursive includes, which could break lazy source code - so
5181 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5182 developers should define this symbol when building and using openssl to
5183 ensure they track the recommended behaviour, interfaces, [etc], but
5184 backwards-compatible behaviour prevails when this isn't defined.
5187 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5190 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5191 This will generate a random key of the appropriate length based on the
5192 cipher context. The EVP_CIPHER can provide its own random key generation
5193 routine to support keys of a specific form. This is used in the des and
5194 3des routines to generate a key of the correct parity. Update S/MIME
5195 code to use new functions and hence generate correct parity DES keys.
5196 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5197 valid (weak or incorrect parity).
5200 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5201 as looking them up. This is useful when the verified structure may contain
5202 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5203 present unless the new PKCS7_NO_CRL flag is asserted.
5206 *) Extend ASN1 oid configuration module. It now additionally accepts the
5209 shortName = some long name, 1.2.3.4
5212 *) Reimplemented the BN_CTX implementation. There is now no more static
5213 limitation on the number of variables it can handle nor the depth of the
5214 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5215 information can now expand as required, and rather than having a single
5216 static array of bignums, BN_CTX now uses a linked-list of such arrays
5217 allowing it to expand on demand whilst maintaining the usefulness of
5218 BN_CTX's "bundling".
5221 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5222 to allow all RSA operations to function using a single BN_CTX.
5225 *) Preliminary support for certificate policy evaluation and checking. This
5226 is initially intended to pass the tests outlined in "Conformance Testing
5227 of Relying Party Client Certificate Path Processing Logic" v1.07.
5230 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5231 remained unused and not that useful. A variety of other little bignum
5232 tweaks and fixes have also been made continuing on from the audit (see
5236 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5237 associated ASN1, EVP and SSL functions and old ASN1 macros.
5240 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5241 and this should never fail. So the return value from the use of
5242 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5243 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5246 *) BN_CTX_get() should return zero-valued bignums, providing the same
5247 initialised value as BN_new().
5248 [Geoff Thorpe, suggested by Ulf Möller]
5250 *) Support for inhibitAnyPolicy certificate extension.
5253 *) An audit of the BIGNUM code is underway, for which debugging code is
5254 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5255 is considered valid when processing BIGNUMs, and causes execution to
5256 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5257 further steps are taken to deliberately pollute unused data in BIGNUM
5258 structures to try and expose faulty code further on. For now, openssl will
5259 (in its default mode of operation) continue to tolerate the inconsistent
5260 forms that it has tolerated in the past, but authors and packagers should
5261 consider trying openssl and their own applications when compiled with
5262 these debugging symbols defined. It will help highlight potential bugs in
5263 their own code, and will improve the test coverage for OpenSSL itself. At
5264 some point, these tighter rules will become openssl's default to improve
5265 maintainability, though the assert()s and other overheads will remain only
5266 in debugging configurations. See bn.h for more details.
5267 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5269 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5270 that can only be obtained through BN_CTX_new() (which implicitly
5271 initialises it). The presence of this function only made it possible
5272 to overwrite an existing structure (and cause memory leaks).
5275 *) Because of the callback-based approach for implementing LHASH as a
5276 template type, lh_insert() adds opaque objects to hash-tables and
5277 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5278 to clean up those corresponding objects before destroying the hash table
5279 (and losing the object pointers). So some over-zealous constifications in
5280 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5281 objects as "const" and the lh_doall[_arg] callback wrappers are not
5282 prototyped to have "const" restrictions on the object pointers they are
5283 given (and so aren't required to cast them away any more).
5286 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5287 (speed) prefers to use its own implementation. The two implementations
5288 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5289 its object type properly exposed (MS_TM) instead of casting to/from "char
5290 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5291 aren't necessarily the greatest nomenclatures - but this is what was used
5292 internally to the implementation so I've used that for now.
5295 *) Ensure that deprecated functions do not get compiled when
5296 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5297 the self-tests were still using deprecated key-generation functions so
5298 these have been updated also.
5301 *) Reorganise PKCS#7 code to separate the digest location functionality
5302 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5303 New function PKCS7_set_digest() to set the digest type for PKCS#7
5304 digestedData type. Add additional code to correctly generate the
5305 digestedData type and add support for this type in PKCS7 initialization
5309 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5310 structure of type "other".
5313 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5314 sure the loop does correctly stop and breaking ("division by zero")
5315 modulus operations are not performed. The (pre-generated) prime
5316 table crypto/bn/bn_prime.h was already correct, but it could not be
5317 re-generated on some platforms because of the "division by zero"
5318 situation in the script.
5319 [Ralf S. Engelschall]
5321 *) Update support for ECC-based TLS ciphersuites according to
5322 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5323 SHA-1 now is only used for "small" curves (where the
5324 representation of a field element takes up to 24 bytes); for
5325 larger curves, the field element resulting from ECDH is directly
5326 used as premaster secret.
5327 [Douglas Stebila (Sun Microsystems Laboratories)]
5329 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5330 curve secp160r1 to the tests.
5331 [Douglas Stebila (Sun Microsystems Laboratories)]
5333 *) Add the possibility to load symbols globally with DSO.
5334 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5336 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5337 control of the error stack.
5340 *) Add support for STORE in ENGINE.
5343 *) Add the STORE type. The intention is to provide a common interface
5344 to certificate and key stores, be they simple file-based stores, or
5345 HSM-type store, or LDAP stores, or...
5346 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5349 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5350 pass a list of arguments to any function as well as provide a way
5351 for a function to pass data back to the caller.
5354 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5355 works like BUF_strdup() but can be used to duplicate a portion of
5356 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5360 *) Add the function sk_find_ex() which works like sk_find(), but will
5361 return an index to an element even if an exact match couldn't be
5362 found. The index is guaranteed to point at the element where the
5363 searched-for key would be inserted to preserve sorting order.
5366 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5367 takes an extra flags argument for optional functionality. Currently,
5368 the following flags are defined:
5370 OBJ_BSEARCH_VALUE_ON_NOMATCH
5371 This one gets OBJ_bsearch_ex() to return a pointer to the first
5372 element where the comparing function returns a negative or zero
5375 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5376 This one gets OBJ_bsearch_ex() to return a pointer to the first
5377 element where the comparing function returns zero. This is useful
5378 if there are more than one element where the comparing function
5382 *) Make it possible to create self-signed certificates with 'openssl ca'
5383 in such a way that the self-signed certificate becomes part of the
5384 CA database and uses the same mechanisms for serial number generation
5385 as all other certificate signing. The new flag '-selfsign' enables
5386 this functionality. Adapt CA.sh and CA.pl.in.
5389 *) Add functionality to check the public key of a certificate request
5390 against a given private. This is useful to check that a certificate
5391 request can be signed by that key (self-signing).
5394 *) Make it possible to have multiple active certificates with the same
5395 subject in the CA index file. This is done only if the keyword
5396 'unique_subject' is set to 'no' in the main CA section (default
5397 if 'CA_default') of the configuration file. The value is saved
5398 with the database itself in a separate index attribute file,
5399 named like the index file with '.attr' appended to the name.
5402 *) Generate multi-valued AVAs using '+' notation in config files for
5406 *) Support for nameConstraints certificate extension.
5409 *) Support for policyConstraints certificate extension.
5412 *) Support for policyMappings certificate extension.
5415 *) Make sure the default DSA_METHOD implementation only uses its
5416 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5417 and change its own handlers to be NULL so as to remove unnecessary
5418 indirection. This lets alternative implementations fallback to the
5419 default implementation more easily.
5422 *) Support for directoryName in GeneralName related extensions
5426 *) Make it possible to link applications using Makefile.shared.
5427 Make that possible even when linking against static libraries!
5430 *) Support for single pass processing for S/MIME signing. This now
5431 means that S/MIME signing can be done from a pipe, in addition
5432 cleartext signing (multipart/signed type) is effectively streaming
5433 and the signed data does not need to be all held in memory.
5435 This is done with a new flag PKCS7_STREAM. When this flag is set
5436 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5437 is done after the data is output (and digests calculated) in
5438 SMIME_write_PKCS7().
5441 *) Add full support for -rpath/-R, both in shared libraries and
5442 applications, at least on the platforms where it's known how
5446 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5447 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5448 will now compute a table of multiples of the generator that
5449 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5450 faster (notably in the case of a single point multiplication,
5451 scalar * generator).
5452 [Nils Larsch, Bodo Moeller]
5454 *) IPv6 support for certificate extensions. The various extensions
5455 which use the IP:a.b.c.d can now take IPv6 addresses using the
5456 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5460 *) Added an ENGINE that implements RSA by performing private key
5461 exponentiations with the GMP library. The conversions to and from
5462 GMP's mpz_t format aren't optimised nor are any montgomery forms
5463 cached, and on x86 it appears OpenSSL's own performance has caught up.
5464 However there are likely to be other architectures where GMP could
5465 provide a boost. This ENGINE is not built in by default, but it can be
5466 specified at Configure time and should be accompanied by the necessary
5467 linker additions, eg;
5468 ./config -DOPENSSL_USE_GMP -lgmp
5471 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5472 testing availability of engines with "-t" - the old behaviour is
5473 produced by increasing the feature's verbosity with "-tt".
5476 *) ECDSA routines: under certain error conditions uninitialized BN objects
5477 could be freed. Solution: make sure initialization is performed early
5478 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5482 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5483 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5484 software implementations. For DSA and DH, parameter generation can
5485 also be overridden by providing the appropriate method callbacks.
5488 *) Change the "progress" mechanism used in key-generation and
5489 primality testing to functions that take a new BN_GENCB pointer in
5490 place of callback/argument pairs. The new API functions have "_ex"
5491 postfixes and the older functions are reimplemented as wrappers for
5492 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5493 declarations of the old functions to help (graceful) attempts to
5494 migrate to the new functions. Also, the new key-generation API
5495 functions operate on a caller-supplied key-structure and return
5496 success/failure rather than returning a key or NULL - this is to
5497 help make "keygen" another member function of RSA_METHOD etc.
5499 Example for using the new callback interface:
5501 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5505 BN_GENCB_set(&my_cb, my_callback, my_arg);
5507 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5508 /* For the meaning of a, b in calls to my_callback(), see the
5509 * documentation of the function that calls the callback.
5510 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5511 * my_callback should return 1 if it wants BN_is_prime_ex()
5512 * to continue, or 0 to stop.
5517 *) Change the ZLIB compression method to be stateful, and make it
5518 available to TLS with the number defined in
5519 draft-ietf-tls-compression-04.txt.
5522 *) Add the ASN.1 structures and functions for CertificatePair, which
5523 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5525 CertificatePair ::= SEQUENCE {
5526 forward [0] Certificate OPTIONAL,
5527 reverse [1] Certificate OPTIONAL,
5528 -- at least one of the pair shall be present -- }
5530 Also implement the PEM functions to read and write certificate
5531 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5533 This needed to be defined, mostly for the sake of the LDAP
5534 attribute crossCertificatePair, but may prove useful elsewhere as
5538 *) Make it possible to inhibit symlinking of shared libraries in
5539 Makefile.shared, for Cygwin's sake.
5542 *) Extend the BIGNUM API by creating a function
5543 void BN_set_negative(BIGNUM *a, int neg);
5544 and a macro that behave like
5545 int BN_is_negative(const BIGNUM *a);
5547 to avoid the need to access 'a->neg' directly in applications.
5550 *) Implement fast modular reduction for pseudo-Mersenne primes
5551 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5552 EC_GROUP_new_curve_GFp() will now automatically use this
5554 [Nils Larsch <nla@trustcenter.de>]
5556 *) Add new lock type (CRYPTO_LOCK_BN).
5559 *) Change the ENGINE framework to automatically load engines
5560 dynamically from specific directories unless they could be
5561 found to already be built in or loaded. Move all the
5562 current engines except for the cryptodev one to a new
5564 The engines in engines/ are built as shared libraries if
5565 the "shared" options was given to ./Configure or ./config.
5566 Otherwise, they are inserted in libcrypto.a.
5567 /usr/local/ssl/engines is the default directory for dynamic
5568 engines, but that can be overridden at configure time through
5569 the usual use of --prefix and/or --openssldir, and at run
5570 time with the environment variable OPENSSL_ENGINES.
5571 [Geoff Thorpe and Richard Levitte]
5573 *) Add Makefile.shared, a helper makefile to build shared
5574 libraries. Adapt Makefile.org.
5577 *) Add version info to Win32 DLLs.
5578 [Peter 'Luna' Runestig" <peter@runestig.com>]
5580 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5581 can be added using this API to created arbitrary PKCS#12
5582 files while avoiding the low level API.
5584 New options to PKCS12_create(), key or cert can be NULL and
5585 will then be omitted from the output file. The encryption
5586 algorithm NIDs can be set to -1 for no encryption, the mac
5587 iteration count can be set to 0 to omit the mac.
5589 Enhance pkcs12 utility by making the -nokeys and -nocerts
5590 options work when creating a PKCS#12 file. New option -nomac
5591 to omit the mac, NONE can be set for an encryption algorithm.
5592 New code is modified to use the enhanced PKCS12_create()
5593 instead of the low level API.
5596 *) Extend ASN1 encoder to support indefinite length constructed
5597 encoding. This can output sequences tags and octet strings in
5598 this form. Modify pk7_asn1.c to support indefinite length
5599 encoding. This is experimental and needs additional code to
5600 be useful, such as an ASN1 bio and some enhanced streaming
5603 Extend template encode functionality so that tagging is passed
5604 down to the template encoder.
5607 *) Let 'openssl req' fail if an argument to '-newkey' is not
5608 recognized instead of using RSA as a default.
5611 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5612 As these are not official, they are not included in "ALL";
5613 the "ECCdraft" ciphersuite group alias can be used to select them.
5614 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5616 *) Add ECDH engine support.
5617 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5619 *) Add ECDH in new directory crypto/ecdh/.
5620 [Douglas Stebila (Sun Microsystems Laboratories)]
5622 *) Let BN_rand_range() abort with an error after 100 iterations
5623 without success (which indicates a broken PRNG).
5626 *) Change BN_mod_sqrt() so that it verifies that the input value
5627 is really the square of the return value. (Previously,
5628 BN_mod_sqrt would show GIGO behaviour.)
5631 *) Add named elliptic curves over binary fields from X9.62, SECG,
5632 and WAP/WTLS; add OIDs that were still missing.
5634 [Sheueling Chang Shantz and Douglas Stebila
5635 (Sun Microsystems Laboratories)]
5637 *) Extend the EC library for elliptic curves over binary fields
5638 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5641 EC_GF2m_simple_method
5645 EC_GROUP_new_curve_GF2m
5646 EC_GROUP_set_curve_GF2m
5647 EC_GROUP_get_curve_GF2m
5648 EC_POINT_set_affine_coordinates_GF2m
5649 EC_POINT_get_affine_coordinates_GF2m
5650 EC_POINT_set_compressed_coordinates_GF2m
5652 Point compression for binary fields is disabled by default for
5653 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5656 As binary polynomials are represented as BIGNUMs, various members
5657 of the EC_GROUP and EC_POINT data structures can be shared
5658 between the implementations for prime fields and binary fields;
5659 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5660 are essentially identical to their ..._GFp counterparts.
5661 (For simplicity, the '..._GFp' prefix has been dropped from
5662 various internal method names.)
5664 An internal 'field_div' method (similar to 'field_mul' and
5665 'field_sqr') has been added; this is used only for binary fields.
5667 [Sheueling Chang Shantz and Douglas Stebila
5668 (Sun Microsystems Laboratories)]
5670 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5671 through methods ('mul', 'precompute_mult').
5673 The generic implementations (now internally called 'ec_wNAF_mul'
5674 and 'ec_wNAF_precomputed_mult') remain the default if these
5675 methods are undefined.
5677 [Sheueling Chang Shantz and Douglas Stebila
5678 (Sun Microsystems Laboratories)]
5680 *) New function EC_GROUP_get_degree, which is defined through
5681 EC_METHOD. For curves over prime fields, this returns the bit
5682 length of the modulus.
5684 [Sheueling Chang Shantz and Douglas Stebila
5685 (Sun Microsystems Laboratories)]
5687 *) New functions EC_GROUP_dup, EC_POINT_dup.
5688 (These simply call ..._new and ..._copy).
5690 [Sheueling Chang Shantz and Douglas Stebila
5691 (Sun Microsystems Laboratories)]
5693 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5694 Polynomials are represented as BIGNUMs (where the sign bit is not
5695 used) in the following functions [macros]:
5698 BN_GF2m_sub [= BN_GF2m_add]
5699 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5700 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5701 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5703 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5704 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5705 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5706 BN_GF2m_cmp [= BN_ucmp]
5708 (Note that only the 'mod' functions are actually for fields GF(2^m).
5709 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5711 For some functions, an the irreducible polynomial defining a
5712 field can be given as an 'unsigned int[]' with strictly
5713 decreasing elements giving the indices of those bits that are set;
5714 i.e., p[] represents the polynomial
5715 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5717 p[0] > p[1] > ... > p[k] = 0.
5718 This applies to the following functions:
5723 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5724 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5726 BN_GF2m_mod_sqrt_arr
5727 BN_GF2m_mod_solve_quad_arr
5731 Conversion can be performed by the following functions:
5736 bntest.c has additional tests for binary polynomial arithmetic.
5738 Two implementations for BN_GF2m_mod_div() are available.
5739 The default algorithm simply uses BN_GF2m_mod_inv() and
5740 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5741 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5742 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5744 [Sheueling Chang Shantz and Douglas Stebila
5745 (Sun Microsystems Laboratories)]
5747 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5748 functionality is disabled at compile-time.
5749 [Douglas Stebila <douglas.stebila@sun.com>]
5751 *) Change default behaviour of 'openssl asn1parse' so that more
5752 information is visible when viewing, e.g., a certificate:
5754 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5755 mode the content of non-printable OCTET STRINGs is output in a
5756 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5757 avoid the appearance of a printable string.
5758 [Nils Larsch <nla@trustcenter.de>]
5760 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5762 EC_GROUP_set_asn1_flag()
5763 EC_GROUP_get_asn1_flag()
5764 EC_GROUP_set_point_conversion_form()
5765 EC_GROUP_get_point_conversion_form()
5766 These control ASN1 encoding details:
5767 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5768 has been set to OPENSSL_EC_NAMED_CURVE.
5769 - Points are encoded in uncompressed form by default; options for
5770 asn1_for are as for point2oct, namely
5771 POINT_CONVERSION_COMPRESSED
5772 POINT_CONVERSION_UNCOMPRESSED
5773 POINT_CONVERSION_HYBRID
5775 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5778 EC_GROUP_get0_seed()
5779 EC_GROUP_get_seed_len()
5780 This is used only for ASN1 purposes (so far).
5781 [Nils Larsch <nla@trustcenter.de>]
5783 *) Add 'field_type' member to EC_METHOD, which holds the NID
5784 of the appropriate field type OID. The new function
5785 EC_METHOD_get_field_type() returns this value.
5786 [Nils Larsch <nla@trustcenter.de>]
5791 EC_POINT_point2hex()
5792 EC_POINT_hex2point()
5793 providing useful interfaces to EC_POINT_point2oct() and
5794 EC_POINT_oct2point().
5795 [Nils Larsch <nla@trustcenter.de>]
5797 *) Change internals of the EC library so that the functions
5798 EC_GROUP_set_generator()
5799 EC_GROUP_get_generator()
5800 EC_GROUP_get_order()
5801 EC_GROUP_get_cofactor()
5802 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5803 to methods, which would lead to unnecessary code duplication when
5804 adding different types of curves.
5805 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5807 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5808 arithmetic, and such that modified wNAFs are generated
5809 (which avoid length expansion in many cases).
5812 *) Add a function EC_GROUP_check_discriminant() (defined via
5813 EC_METHOD) that verifies that the curve discriminant is non-zero.
5815 Add a function EC_GROUP_check() that makes some sanity tests
5816 on a EC_GROUP, its generator and order. This includes
5817 EC_GROUP_check_discriminant().
5818 [Nils Larsch <nla@trustcenter.de>]
5820 *) Add ECDSA in new directory crypto/ecdsa/.
5822 Add applications 'openssl ecparam' and 'openssl ecdsa'
5823 (these are based on 'openssl dsaparam' and 'openssl dsa').
5825 ECDSA support is also included in various other files across the
5826 library. Most notably,
5827 - 'openssl req' now has a '-newkey ecdsa:file' option;
5828 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5829 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5830 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5831 them suitable for ECDSA where domain parameters must be
5832 extracted before the specific public key;
5833 - ECDSA engine support has been added.
5834 [Nils Larsch <nla@trustcenter.de>]
5836 *) Include some named elliptic curves, and add OIDs from X9.62,
5837 SECG, and WAP/WTLS. Each curve can be obtained from the new
5839 EC_GROUP_new_by_curve_name(),
5840 and the list of available named curves can be obtained with
5841 EC_get_builtin_curves().
5842 Also add a 'curve_name' member to EC_GROUP objects, which can be
5844 EC_GROUP_set_curve_name()
5845 EC_GROUP_get_curve_name()
5846 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5848 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5849 was actually never needed) and in BN_mul(). The removal in BN_mul()
5850 required a small change in bn_mul_part_recursive() and the addition
5851 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5852 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5853 bn_sub_words() and bn_add_words() except they take arrays with
5857 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5859 *) Cleanse PEM buffers before freeing them since they may contain
5861 [Benjamin Bennett <ben@psc.edu>]
5863 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5864 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5865 authentication-only ciphersuites.
5868 *) Since AES128 and AES256 share a single mask bit in the logic of
5869 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5870 kludge to work properly if AES128 is available and AES256 isn't.
5873 *) Expand security boundary to match 1.1.1 module.
5876 *) Remove redundant features: hash file source, editing of test vectors
5877 modify fipsld to use external fips_premain.c signature.
5880 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5881 run algorithm test programs.
5884 *) Make algorithm test programs more tolerant of whitespace.
5887 *) Have SSL/TLS server implementation tolerate "mismatched" record
5888 protocol version while receiving ClientHello even if the
5889 ClientHello is fragmented. (The server can't insist on the
5890 particular protocol version it has chosen before the ServerHello
5891 message has informed the client about his choice.)
5894 *) Load error codes if they are not already present instead of using a
5895 static variable. This allows them to be cleanly unloaded and reloaded.
5898 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5900 *) Introduce limits to prevent malicious keys being able to
5901 cause a denial of service. (CVE-2006-2940)
5902 [Steve Henson, Bodo Moeller]
5904 *) Fix ASN.1 parsing of certain invalid structures that can result
5905 in a denial of service. (CVE-2006-2937) [Steve Henson]
5907 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5908 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5910 *) Fix SSL client code which could crash if connecting to a
5911 malicious SSLv2 server. (CVE-2006-4343)
5912 [Tavis Ormandy and Will Drewry, Google Security Team]
5914 *) Change ciphersuite string processing so that an explicit
5915 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5916 will no longer include "AES128-SHA"), and any other similar
5917 ciphersuite (same bitmap) from *other* protocol versions (so that
5918 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5919 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5920 changes from 0.9.8b and 0.9.8d.
5923 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5925 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5926 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5928 *) Change the Unix randomness entropy gathering to use poll() when
5929 possible instead of select(), since the latter has some
5930 undesirable limitations.
5931 [Darryl Miles via Richard Levitte and Bodo Moeller]
5933 *) Disable rogue ciphersuites:
5935 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5936 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5937 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5939 The latter two were purportedly from
5940 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5943 Also deactivate the remaining ciphersuites from
5944 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5945 unofficial, and the ID has long expired.
5948 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5949 dual-core machines) and other potential thread-safety issues.
5952 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5954 *) Adapt fipsld and the build system to link against the validated FIPS
5955 module in FIPS mode.
5958 *) Fixes for VC++ 2005 build under Windows.
5961 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5962 from a Windows bash shell such as MSYS. It is autodetected from the
5963 "config" script when run from a VC++ environment. Modify standard VC++
5964 build to use fipscanister.o from the GNU make build.
5967 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5969 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5970 The value now differs depending on if you build for FIPS or not.
5971 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5972 safely run with a non-FIPSed libcrypto, as it may crash because of
5973 the difference induced by this change.
5976 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5978 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5979 (part of SSL_OP_ALL). This option used to disable the
5980 countermeasure against man-in-the-middle protocol-version
5981 rollback in the SSL 2.0 server implementation, which is a bad
5982 idea. (CVE-2005-2969)
5984 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5985 for Information Security, National Institute of Advanced Industrial
5986 Science and Technology [AIST], Japan)]
5988 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5989 mainly for FIPS compliance and not fully integrated at this stage.
5992 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5993 the exponentiation using a fixed-length exponent. (Otherwise,
5994 the information leaked through timing could expose the secret key
5995 after many signatures; cf. Bleichenbacher's attack on DSA with
5999 *) Make a new fixed-window mod_exp implementation the default for
6000 RSA, DSA, and DH private-key operations so that the sequence of
6001 squares and multiplies and the memory access pattern are
6002 independent of the particular secret key. This will mitigate
6003 cache-timing and potential related attacks.
6005 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6006 and this is automatically used by BN_mod_exp_mont() if the new flag
6007 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6008 will use this BN flag for private exponents unless the flag
6009 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6010 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6012 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6014 *) Change the client implementation for SSLv23_method() and
6015 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6016 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6017 (Previously, the SSL 2.0 backwards compatible Client Hello
6018 message format would be used even with SSL_OP_NO_SSLv2.)
6021 *) Add support for smime-type MIME parameter in S/MIME messages which some
6025 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6026 a threadsafe manner. Modify rsa code to use new function and add calls
6027 to dsa and dh code (which had race conditions before).
6030 *) Include the fixed error library code in the C error file definitions
6031 instead of fixing them up at runtime. This keeps the error code
6032 structures constant.
6035 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6037 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6040 *) Fixes for newer kerberos headers. NB: the casts are needed because
6041 the 'length' field is signed on one version and unsigned on another
6042 with no (?) obvious way to tell the difference, without these VC++
6043 complains. Also the "definition" of FAR (blank) is no longer included
6044 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6045 some needed definitions.
6048 *) Undo Cygwin change.
6051 *) Added support for proxy certificates according to RFC 3820.
6052 Because they may be a security thread to unaware applications,
6053 they must be explicitly allowed in run-time. See
6054 docs/HOWTO/proxy_certificates.txt for further information.
6057 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6059 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6060 server and client random values. Previously
6061 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6062 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6064 This change has negligible security impact because:
6066 1. Server and client random values still have 24 bytes of pseudo random
6069 2. Server and client random values are sent in the clear in the initial
6072 3. The master secret is derived using the premaster secret (48 bytes in
6073 size for static RSA ciphersuites) as well as client server and random
6076 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6079 [Stephen Henson, reported by UK NISCC]
6081 *) Use Windows randomness collection on Cygwin.
6084 *) Fix hang in EGD/PRNGD query when communication socket is closed
6085 prematurely by EGD/PRNGD.
6086 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6088 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6091 *) Back-port of selected performance improvements from development
6092 branch, as well as improved support for PowerPC platforms.
6095 *) Add lots of checks for memory allocation failure, error codes to indicate
6096 failure and freeing up memory if a failure occurs.
6097 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6099 *) Add new -passin argument to dgst.
6102 *) Perform some character comparisons of different types in X509_NAME_cmp:
6103 this is needed for some certificates that re-encode DNs into UTF8Strings
6104 (in violation of RFC3280) and can't or won't issue name rollover
6108 *) Make an explicit check during certificate validation to see that
6109 the CA setting in each certificate on the chain is correct. As a
6110 side effect always do the following basic checks on extensions,
6111 not just when there's an associated purpose to the check:
6113 - if there is an unhandled critical extension (unless the user
6114 has chosen to ignore this fault)
6115 - if the path length has been exceeded (if one is set at all)
6116 - that certain extensions fit the associated purpose (if one has
6120 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6122 *) Avoid a race condition when CRLs are checked in a multi threaded
6123 environment. This would happen due to the reordering of the revoked
6124 entries during signature checking and serial number lookup. Now the
6125 encoding is cached and the serial number sort performed under a lock.
6126 Add new STACK function sk_is_sorted().
6129 *) Add Delta CRL to the extension code.
6132 *) Various fixes to s3_pkt.c so alerts are sent properly.
6133 [David Holmes <d.holmes@f5.com>]
6135 *) Reduce the chances of duplicate issuer name and serial numbers (in
6136 violation of RFC3280) using the OpenSSL certificate creation utilities.
6137 This is done by creating a random 64 bit value for the initial serial
6138 number when a serial number file is created or when a self signed
6139 certificate is created using 'openssl req -x509'. The initial serial
6140 number file is created using 'openssl x509 -next_serial' in CA.pl
6141 rather than being initialized to 1.
6144 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6146 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6147 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6148 [Joe Orton, Steve Henson]
6150 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6152 [Joe Orton, Steve Henson]
6154 *) Make it possible to have multiple active certificates with the same
6155 subject in the CA index file. This is done only if the keyword
6156 'unique_subject' is set to 'no' in the main CA section (default
6157 if 'CA_default') of the configuration file. The value is saved
6158 with the database itself in a separate index attribute file,
6159 named like the index file with '.attr' appended to the name.
6162 *) X509 verify fixes. Disable broken certificate workarounds when
6163 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6164 keyUsage extension present. Don't accept CRLs with unhandled critical
6165 extensions: since verify currently doesn't process CRL extensions this
6166 rejects a CRL with *any* critical extensions. Add new verify error codes
6170 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6171 A clarification of RFC2560 will require the use of OCTET STRINGs and
6172 some implementations cannot handle the current raw format. Since OpenSSL
6173 copies and compares OCSP nonces as opaque blobs without any attempt at
6174 parsing them this should not create any compatibility issues.
6177 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6178 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6179 this HMAC (and other) operations are several times slower than OpenSSL
6183 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6184 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6186 *) Use the correct content when signing type "other".
6189 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6191 *) Fix various bugs revealed by running the NISCC test suite:
6193 Stop out of bounds reads in the ASN1 code when presented with
6194 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6196 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6198 If verify callback ignores invalid public key errors don't try to check
6199 certificate signature with the NULL public key.
6203 *) New -ignore_err option in ocsp application to stop the server
6204 exiting on the first error in a request.
6207 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6208 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6212 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6213 extra data after the compression methods not only for TLS 1.0
6214 but also for SSL 3.0 (as required by the specification).
6215 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6217 *) Change X509_certificate_type() to mark the key as exported/exportable
6218 when it's 512 *bits* long, not 512 bytes.
6221 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6222 blocks during encryption.
6225 *) Various fixes to base64 BIO and non blocking I/O. On write
6226 flushes were not handled properly if the BIO retried. On read
6227 data was not being buffered properly and had various logic bugs.
6228 This also affects blocking I/O when the data being decoded is a
6232 *) Various S/MIME bugfixes and compatibility changes:
6233 output correct application/pkcs7 MIME type if
6234 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6235 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6236 of files as .eml work). Correctly handle very long lines in MIME
6240 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6242 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6243 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6244 a protocol version number mismatch like a decryption error
6245 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6248 *) Turn on RSA blinding by default in the default implementation
6249 to avoid a timing attack. Applications that don't want it can call
6250 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6251 They would be ill-advised to do so in most cases.
6252 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6254 *) Change RSA blinding code so that it works when the PRNG is not
6255 seeded (in this case, the secret RSA exponent is abused as
6256 an unpredictable seed -- if it is not unpredictable, there
6257 is no point in blinding anyway). Make RSA blinding thread-safe
6258 by remembering the creator's thread ID in rsa->blinding and
6259 having all other threads use local one-time blinding factors
6260 (this requires more computation than sharing rsa->blinding, but
6261 avoids excessive locking; and if an RSA object is not shared
6262 between threads, blinding will still be very fast).
6265 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6266 ENGINE as defaults for all supported algorithms irrespective of
6267 the 'flags' parameter. 'flags' is now honoured, so applications
6268 should make sure they are passing it correctly.
6271 *) Target "mingw" now allows native Windows code to be generated in
6272 the Cygwin environment as well as with the MinGW compiler.
6275 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6277 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6278 via timing by performing a MAC computation even if incorrect
6279 block cipher padding has been found. This is a countermeasure
6280 against active attacks where the attacker has to distinguish
6281 between bad padding and a MAC verification error. (CVE-2003-0078)
6283 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6284 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6285 Martin Vuagnoux (EPFL, Ilion)]
6287 *) Make the no-err option work as intended. The intention with no-err
6288 is not to have the whole error stack handling routines removed from
6289 libcrypto, it's only intended to remove all the function name and
6290 reason texts, thereby removing some of the footprint that may not
6291 be interesting if those errors aren't displayed anyway.
6293 NOTE: it's still possible for any application or module to have it's
6294 own set of error texts inserted. The routines are there, just not
6295 used by default when no-err is given.
6298 *) Add support for FreeBSD on IA64.
6299 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6301 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6302 Kerberos function mit_des_cbc_cksum(). Before this change,
6303 the value returned by DES_cbc_cksum() was like the one from
6304 mit_des_cbc_cksum(), except the bytes were swapped.
6305 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6307 *) Allow an application to disable the automatic SSL chain building.
6308 Before this a rather primitive chain build was always performed in
6309 ssl3_output_cert_chain(): an application had no way to send the
6310 correct chain if the automatic operation produced an incorrect result.
6312 Now the chain builder is disabled if either:
6314 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6316 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6318 The reasoning behind this is that an application would not want the
6319 auto chain building to take place if extra chain certificates are
6320 present and it might also want a means of sending no additional
6321 certificates (for example the chain has two certificates and the
6325 *) Add the possibility to build without the ENGINE framework.
6326 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6328 *) Under Win32 gmtime() can return NULL: check return value in
6329 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6332 *) DSA routines: under certain error conditions uninitialized BN objects
6333 could be freed. Solution: make sure initialization is performed early
6334 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6335 Nils Larsch <nla@trustcenter.de> via PR#459)
6338 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6339 checked on reconnect on the client side, therefore session resumption
6340 could still fail with a "ssl session id is different" error. This
6341 behaviour is masked when SSL_OP_ALL is used due to
6342 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6343 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6344 followup to PR #377.
6347 *) IA-32 assembler support enhancements: unified ELF targets, support
6348 for SCO/Caldera platforms, fix for Cygwin shared build.
6351 *) Add support for FreeBSD on sparc64. As a consequence, support for
6352 FreeBSD on non-x86 processors is separate from x86 processors on
6353 the config script, much like the NetBSD support.
6354 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6356 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6358 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6361 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6362 code (06) was taken as the first octet of the session ID and the last
6363 octet was ignored consequently. As a result SSLv2 client side session
6364 caching could not have worked due to the session ID mismatch between
6366 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6370 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6371 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6375 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6376 seems that in spite of existing for more than a year, many application
6377 author have done nothing to provide the necessary callbacks, which
6378 means that this particular engine will not work properly anywhere.
6379 This is a very unfortunate situation which forces us, in the name
6380 of usability, to give the hw_ncipher.c a static lock, which is part
6382 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6383 appear in 0.9.8 or later. We EXPECT application authors to have
6384 dealt properly with this when 0.9.8 is released (unless we actually
6385 make such changes in the libcrypto locking code that changes will
6386 have to be made anyway).
6389 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6390 octets have been read, EOF or an error occurs. Without this change
6391 some truncated ASN1 structures will not produce an error.
6394 *) Disable Heimdal support, since it hasn't been fully implemented.
6395 Still give the possibility to force the use of Heimdal, but with
6396 warnings and a request that patches get sent to openssl-dev.
6399 *) Add the VC-CE target, introduce the WINCE sysname, and add
6400 INSTALL.WCE and appropriate conditionals to make it build.
6401 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6403 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6404 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6405 edit numbers of the version.
6406 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6408 *) Introduce safe string copy and catenation functions
6409 (BUF_strlcpy() and BUF_strlcat()).
6410 [Ben Laurie (CHATS) and Richard Levitte]
6412 *) Avoid using fixed-size buffers for one-line DNs.
6413 [Ben Laurie (CHATS)]
6415 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6416 resizing buffers containing secrets, and use where appropriate.
6417 [Ben Laurie (CHATS)]
6419 *) Avoid using fixed size buffers for configuration file location.
6420 [Ben Laurie (CHATS)]
6422 *) Avoid filename truncation for various CA files.
6423 [Ben Laurie (CHATS)]
6425 *) Use sizeof in preference to magic numbers.
6426 [Ben Laurie (CHATS)]
6428 *) Avoid filename truncation in cert requests.
6429 [Ben Laurie (CHATS)]
6431 *) Add assertions to check for (supposedly impossible) buffer
6433 [Ben Laurie (CHATS)]
6435 *) Don't cache truncated DNS entries in the local cache (this could
6436 potentially lead to a spoofing attack).
6437 [Ben Laurie (CHATS)]
6439 *) Fix various buffers to be large enough for hex/decimal
6440 representations in a platform independent manner.
6441 [Ben Laurie (CHATS)]
6443 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6444 resizing buffers containing secrets, and use where appropriate.
6445 [Ben Laurie (CHATS)]
6447 *) Add BIO_indent() to avoid much slightly worrying code to do
6449 [Ben Laurie (CHATS)]
6451 *) Convert sprintf()/BIO_puts() to BIO_printf().
6452 [Ben Laurie (CHATS)]
6454 *) buffer_gets() could terminate with the buffer only half
6456 [Ben Laurie (CHATS)]
6458 *) Add assertions to prevent user-supplied crypto functions from
6459 overflowing internal buffers by having large block sizes, etc.
6460 [Ben Laurie (CHATS)]
6462 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6464 [Ben Laurie (CHATS)]
6466 *) Eliminate unused copy of key in RC4.
6467 [Ben Laurie (CHATS)]
6469 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6470 [Ben Laurie (CHATS)]
6472 *) Fix off-by-one error in EGD path.
6473 [Ben Laurie (CHATS)]
6475 *) If RANDFILE path is too long, ignore instead of truncating.
6476 [Ben Laurie (CHATS)]
6478 *) Eliminate unused and incorrectly sized X.509 structure
6480 [Ben Laurie (CHATS)]
6482 *) Eliminate unused and dangerous function knumber().
6483 [Ben Laurie (CHATS)]
6485 *) Eliminate unused and dangerous structure, KSSL_ERR.
6486 [Ben Laurie (CHATS)]
6488 *) Protect against overlong session ID context length in an encoded
6489 session object. Since these are local, this does not appear to be
6491 [Ben Laurie (CHATS)]
6493 *) Change from security patch (see 0.9.6e below) that did not affect
6494 the 0.9.6 release series:
6496 Remote buffer overflow in SSL3 protocol - an attacker could
6497 supply an oversized master key in Kerberos-enabled versions.
6499 [Ben Laurie (CHATS)]
6501 *) Change the SSL kerb5 codes to match RFC 2712.
6504 *) Make -nameopt work fully for req and add -reqopt switch.
6505 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6507 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6508 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6510 *) Make sure tests can be performed even if the corresponding algorithms
6511 have been removed entirely. This was also the last step to make
6512 OpenSSL compilable with DJGPP under all reasonable conditions.
6513 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6515 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6516 to allow version independent disabling of normally unselected ciphers,
6517 which may be activated as a side-effect of selecting a single cipher.
6519 (E.g., cipher list string "RSA" enables ciphersuites that are left
6520 out of "ALL" because they do not provide symmetric encryption.
6521 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6522 [Lutz Jaenicke, Bodo Moeller]
6524 *) Add appropriate support for separate platform-dependent build
6525 directories. The recommended way to make a platform-dependent
6526 build directory is the following (tested on Linux), maybe with
6529 # Place yourself outside of the OpenSSL source tree. In
6530 # this example, the environment variable OPENSSL_SOURCE
6531 # is assumed to contain the absolute OpenSSL source directory.
6532 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6533 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6534 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6535 mkdir -p `dirname $F`
6536 ln -s $OPENSSL_SOURCE/$F $F
6539 To be absolutely sure not to disturb the source tree, a "make clean"
6540 is a good thing. If it isn't successful, don't worry about it,
6541 it probably means the source directory is very clean.
6544 *) Make sure any ENGINE control commands make local copies of string
6545 pointers passed to them whenever necessary. Otherwise it is possible
6546 the caller may have overwritten (or deallocated) the original string
6547 data when a later ENGINE operation tries to use the stored values.
6548 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6550 *) Improve diagnostics in file reading and command-line digests.
6551 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6553 *) Add AES modes CFB and OFB to the object database. Correct an
6554 error in AES-CFB decryption.
6557 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6558 allows existing EVP_CIPHER_CTX structures to be reused after
6559 calling EVP_*Final(). This behaviour is used by encryption
6560 BIOs and some applications. This has the side effect that
6561 applications must explicitly clean up cipher contexts with
6562 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6565 *) Check the values of dna and dnb in bn_mul_recursive before calling
6566 bn_mul_comba (a non zero value means the a or b arrays do not contain
6567 n2 elements) and fallback to bn_mul_normal if either is not zero.
6570 *) Fix escaping of non-ASCII characters when using the -subj option
6571 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6574 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6575 form for "surname", serialNumber has no short form.
6576 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6577 therefore remove "mail" short name for "internet 7".
6578 The OID for unique identifiers in X509 certificates is
6579 x500UniqueIdentifier, not uniqueIdentifier.
6580 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6583 *) Add an "init" command to the ENGINE config module and auto initialize
6584 ENGINEs. Without any "init" command the ENGINE will be initialized
6585 after all ctrl commands have been executed on it. If init=1 the
6586 ENGINE is initialized at that point (ctrls before that point are run
6587 on the uninitialized ENGINE and after on the initialized one). If
6588 init=0 then the ENGINE will not be initialized at all.
6591 *) Fix the 'app_verify_callback' interface so that the user-defined
6592 argument is actually passed to the callback: In the
6593 SSL_CTX_set_cert_verify_callback() prototype, the callback
6594 declaration has been changed from
6597 int (*cb)(X509_STORE_CTX *,void *);
6598 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6599 i=s->ctx->app_verify_callback(&ctx)
6600 has been changed into
6601 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6603 To update applications using SSL_CTX_set_cert_verify_callback(),
6604 a dummy argument can be added to their callback functions.
6605 [D. K. Smetters <smetters@parc.xerox.com>]
6607 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6608 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6610 *) Add and OPENSSL_LOAD_CONF define which will cause
6611 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6612 This allows older applications to transparently support certain
6613 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6614 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6615 load the config file and OPENSSL_add_all_algorithms_conf() which will
6616 always load it have also been added.
6619 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6620 Adjust NIDs and EVP layer.
6621 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6623 *) Config modules support in openssl utility.
6625 Most commands now load modules from the config file,
6626 though in a few (such as version) this isn't done
6627 because it couldn't be used for anything.
6629 In the case of ca and req the config file used is
6630 the same as the utility itself: that is the -config
6631 command line option can be used to specify an
6635 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6636 use "openssl_conf" if filename is NULL use default openssl config file.
6639 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6640 config section name. Add a new flag to tolerate a missing config file
6641 and move code to CONF_modules_load_file().
6644 *) Support for crypto accelerator cards from Accelerated Encryption
6645 Processing, www.aep.ie. (Use engine 'aep')
6646 The support was copied from 0.9.6c [engine] and adapted/corrected
6647 to work with the new engine framework.
6648 [AEP Inc. and Richard Levitte]
6650 *) Support for SureWare crypto accelerator cards from Baltimore
6651 Technologies. (Use engine 'sureware')
6652 The support was copied from 0.9.6c [engine] and adapted
6653 to work with the new engine framework.
6656 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6657 make the newer ENGINE framework commands for the CHIL engine work.
6658 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6660 *) Make it possible to produce shared libraries on ReliantUNIX.
6661 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6663 *) Add the configuration target debug-linux-ppro.
6664 Make 'openssl rsa' use the general key loading routines
6665 implemented in apps.c, and make those routines able to
6666 handle the key format FORMAT_NETSCAPE and the variant
6668 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6670 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6671 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6673 *) Add -keyform to rsautl, and document -engine.
6674 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6676 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6677 BIO_R_NO_SUCH_FILE error code rather than the generic
6678 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6681 *) Add new functions
6683 ERR_peek_last_error_line
6684 ERR_peek_last_error_line_data.
6685 These are similar to
6688 ERR_peek_error_line_data,
6689 but report on the latest error recorded rather than the first one
6690 still in the error queue.
6691 [Ben Laurie, Bodo Moeller]
6693 *) default_algorithms option in ENGINE config module. This allows things
6695 default_algorithms = ALL
6696 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6699 *) Preliminary ENGINE config module.
6702 *) New experimental application configuration code.
6705 *) Change the AES code to follow the same name structure as all other
6706 symmetric ciphers, and behave the same way. Move everything to
6707 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6708 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6710 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6711 [Ben Laurie and Theo de Raadt]
6713 *) Add option to output public keys in req command.
6714 [Massimiliano Pala madwolf@openca.org]
6716 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6717 (up to about 10% better than before for P-192 and P-224).
6720 *) New functions/macros
6722 SSL_CTX_set_msg_callback(ctx, cb)
6723 SSL_CTX_set_msg_callback_arg(ctx, arg)
6724 SSL_set_msg_callback(ssl, cb)
6725 SSL_set_msg_callback_arg(ssl, arg)
6727 to request calling a callback function
6729 void cb(int write_p, int version, int content_type,
6730 const void *buf, size_t len, SSL *ssl, void *arg)
6732 whenever a protocol message has been completely received
6733 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6734 protocol version according to which the SSL library interprets
6735 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6736 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6737 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6738 specification (change_cipher_spec(20), alert(21), handshake(22)).
6739 'buf' and 'len' point to the actual message, 'ssl' to the
6740 SSL object, and 'arg' is the application-defined value set by
6741 SSL[_CTX]_set_msg_callback_arg().
6743 'openssl s_client' and 'openssl s_server' have new '-msg' options
6744 to enable a callback that displays all protocol messages.
6747 *) Change the shared library support so shared libraries are built as
6748 soon as the corresponding static library is finished, and thereby get
6749 openssl and the test programs linked against the shared library.
6750 This still only happens when the keyword "shard" has been given to
6751 the configuration scripts.
6753 NOTE: shared library support is still an experimental thing, and
6754 backward binary compatibility is still not guaranteed.
6755 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6757 *) Add support for Subject Information Access extension.
6758 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6760 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6761 additional bytes when new memory had to be allocated, not just
6762 when reusing an existing buffer.
6765 *) New command line and configuration option 'utf8' for the req command.
6766 This allows field values to be specified as UTF8 strings.
6769 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6770 runs for the former and machine-readable output for the latter.
6773 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6774 of the e-mail address in the DN (i.e., it will go into a certificate
6775 extension only). The new configuration file option 'email_in_dn = no'
6776 has the same effect.
6777 [Massimiliano Pala madwolf@openca.org]
6779 *) Change all functions with names starting with des_ to be starting
6780 with DES_ instead. Add wrappers that are compatible with libdes,
6781 but are named _ossl_old_des_*. Finally, add macros that map the
6782 des_* symbols to the corresponding _ossl_old_des_* if libdes
6783 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6784 desired, the des_* symbols will be mapped to DES_*, with one
6787 Since we provide two compatibility mappings, the user needs to
6788 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6789 compatibility is desired. The default (i.e., when that macro
6790 isn't defined) is OpenSSL 0.9.6c compatibility.
6792 There are also macros that enable and disable the support of old
6793 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6794 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6795 are defined, the default will apply: to support the old des routines.
6797 In either case, one must include openssl/des.h to get the correct
6798 definitions. Do not try to just include openssl/des_old.h, that
6801 NOTE: This is a major break of an old API into a new one. Software
6802 authors are encouraged to switch to the DES_ style functions. Some
6803 time in the future, des_old.h and the libdes compatibility functions
6804 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6805 default), and then completely removed.
6808 *) Test for certificates which contain unsupported critical extensions.
6809 If such a certificate is found during a verify operation it is
6810 rejected by default: this behaviour can be overridden by either
6811 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6812 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6813 X509_supported_extension() has also been added which returns 1 if a
6814 particular extension is supported.
6817 *) Modify the behaviour of EVP cipher functions in similar way to digests
6818 to retain compatibility with existing code.
6821 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6822 compatibility with existing code. In particular the 'ctx' parameter does
6823 not have to be to be initialized before the call to EVP_DigestInit() and
6824 it is tidied up after a call to EVP_DigestFinal(). New function
6825 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6826 EVP_MD_CTX_copy() changed to not require the destination to be
6827 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6828 requires the destination to be valid.
6830 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6831 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6834 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6835 so that complete 'Handshake' protocol structures are kept in memory
6836 instead of overwriting 'msg_type' and 'length' with 'body' data.
6839 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6840 [Massimo Santin via Richard Levitte]
6842 *) Major restructuring to the underlying ENGINE code. This includes
6843 reduction of linker bloat, separation of pure "ENGINE" manipulation
6844 (initialisation, etc) from functionality dealing with implementations
6845 of specific crypto interfaces. This change also introduces integrated
6846 support for symmetric ciphers and digest implementations - so ENGINEs
6847 can now accelerate these by providing EVP_CIPHER and EVP_MD
6848 implementations of their own. This is detailed in crypto/engine/README
6849 as it couldn't be adequately described here. However, there are a few
6850 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6851 were changed in the original introduction of ENGINE code have now
6852 reverted back - the hooking from this code to ENGINE is now a good
6853 deal more passive and at run-time, operations deal directly with
6854 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6855 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6856 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6857 they were not being used by the framework as there is no concept of a
6858 BIGNUM_METHOD and they could not be generalised to the new
6859 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6860 ENGINE_cpy() has been removed as it cannot be consistently defined in
6864 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6867 *) Change mkdef.pl to sort symbols that get the same entry number,
6868 and make sure the automatically generated functions ERR_load_*
6869 become part of libeay.num as well.
6872 *) New function SSL_renegotiate_pending(). This returns true once
6873 renegotiation has been requested (either SSL_renegotiate() call
6874 or HelloRequest/ClientHello received from the peer) and becomes
6875 false once a handshake has been completed.
6876 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6877 sends a HelloRequest, but does not ensure that a handshake takes
6878 place. SSL_renegotiate_pending() is useful for checking if the
6879 client has followed the request.)
6882 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6883 By default, clients may request session resumption even during
6884 renegotiation (if session ID contexts permit); with this option,
6885 session resumption is possible only in the first handshake.
6887 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6888 more bits available for options that should not be part of
6889 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6892 *) Add some demos for certificate and certificate request creation.
6895 *) Make maximum certificate chain size accepted from the peer application
6896 settable (SSL*_get/set_max_cert_list()), as proposed by
6897 "Douglas E. Engert" <deengert@anl.gov>.
6900 *) Add support for shared libraries for Unixware-7
6901 (Boyd Lynn Gerber <gerberb@zenez.com>).
6904 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6905 be done prior to destruction. Use this to unload error strings from
6906 ENGINEs that load their own error strings. NB: This adds two new API
6907 functions to "get" and "set" this destroy handler in an ENGINE.
6910 *) Alter all existing ENGINE implementations (except "openssl" and
6911 "openbsd") to dynamically instantiate their own error strings. This
6912 makes them more flexible to be built both as statically-linked ENGINEs
6913 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6914 Also, add stub code to each that makes building them as self-contained
6915 shared-libraries easier (see README.ENGINE).
6918 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6919 implementations into applications that are completely implemented in
6920 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6921 commands that can be used to configure what shared-library to load and
6922 to control aspects of the way it is handled. Also, made an update to
6923 the README.ENGINE file that brings its information up-to-date and
6924 provides some information and instructions on the "dynamic" ENGINE
6925 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6928 *) Make it possible to unload ranges of ERR strings with a new
6929 "ERR_unload_strings" function.
6932 *) Add a copy() function to EVP_MD.
6935 *) Make EVP_MD routines take a context pointer instead of just the
6936 md_data void pointer.
6939 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6940 that the digest can only process a single chunk of data
6941 (typically because it is provided by a piece of
6942 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6943 is only going to provide a single chunk of data, and hence the
6944 framework needn't accumulate the data for oneshot drivers.
6947 *) As with "ERR", make it possible to replace the underlying "ex_data"
6948 functions. This change also alters the storage and management of global
6949 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6950 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6951 index counters. The API functions that use this state have been changed
6952 to take a "class_index" rather than pointers to the class's local STACK
6953 and counter, and there is now an API function to dynamically create new
6954 classes. This centralisation allows us to (a) plug a lot of the
6955 thread-safety problems that existed, and (b) makes it possible to clean
6956 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6957 such data would previously have always leaked in application code and
6958 workarounds were in place to make the memory debugging turn a blind eye
6959 to it. Application code that doesn't use this new function will still
6960 leak as before, but their memory debugging output will announce it now
6961 rather than letting it slide.
6963 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6964 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6965 has a return value to indicate success or failure.
6968 *) Make it possible to replace the underlying "ERR" functions such that the
6969 global state (2 LHASH tables and 2 locks) is only used by the "default"
6970 implementation. This change also adds two functions to "get" and "set"
6971 the implementation prior to it being automatically set the first time
6972 any other ERR function takes place. Ie. an application can call "get",
6973 pass the return value to a module it has just loaded, and that module
6974 can call its own "set" function using that value. This means the
6975 module's "ERR" operations will use (and modify) the error state in the
6976 application and not in its own statically linked copy of OpenSSL code.
6979 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6980 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6981 the operation, and provides a more encapsulated way for external code
6982 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6983 to use these functions rather than manually incrementing the counts.
6985 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6988 *) Add EVP test program.
6991 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6994 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6995 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6996 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6997 These allow a CRL to be built without having to access X509_CRL fields
6998 directly. Modify 'ca' application to use new functions.
7001 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7002 bug workarounds. Rollback attack detection is a security feature.
7003 The problem will only arise on OpenSSL servers when TLSv1 is not
7004 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7005 Software authors not wanting to support TLSv1 will have special reasons
7006 for their choice and can explicitly enable this option.
7007 [Bodo Moeller, Lutz Jaenicke]
7009 *) Rationalise EVP so it can be extended: don't include a union of
7010 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7011 (similar to those existing for EVP_CIPHER_CTX).
7016 EVP_MD_CTX_init(&md); /* new function call */
7017 EVP_DigestInit(&md, EVP_sha1());
7018 EVP_DigestUpdate(&md, in, len);
7019 EVP_DigestFinal(&md, out, NULL);
7020 EVP_MD_CTX_cleanup(&md); /* new function call */
7024 *) Make DES key schedule conform to the usual scheme, as well as
7025 correcting its structure. This means that calls to DES functions
7026 now have to pass a pointer to a des_key_schedule instead of a
7027 plain des_key_schedule (which was actually always a pointer
7030 des_key_schedule ks;
7032 des_set_key_checked(..., &ks);
7033 des_ncbc_encrypt(..., &ks, ...);
7035 (Note that a later change renames 'des_...' into 'DES_...'.)
7038 *) Initial reduction of linker bloat: the use of some functions, such as
7039 PEM causes large amounts of unused functions to be linked in due to
7040 poor organisation. For example pem_all.c contains every PEM function
7041 which has a knock on effect of linking in large amounts of (unused)
7042 ASN1 code. Grouping together similar functions and splitting unrelated
7043 functions prevents this.
7046 *) Cleanup of EVP macros.
7049 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7050 correct _ecb suffix.
7053 *) Add initial OCSP responder support to ocsp application. The
7054 revocation information is handled using the text based index
7055 use by the ca application. The responder can either handle
7056 requests generated internally, supplied in files (for example
7057 via a CGI script) or using an internal minimal server.
7060 *) Add configuration choices to get zlib compression for TLS.
7063 *) Changes to Kerberos SSL for RFC 2712 compliance:
7064 1. Implemented real KerberosWrapper, instead of just using
7065 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7066 2. Implemented optional authenticator field of KerberosWrapper.
7068 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7069 and authenticator structs; see crypto/krb5/.
7071 Generalized Kerberos calls to support multiple Kerberos libraries.
7072 [Vern Staats <staatsvr@asc.hpc.mil>,
7073 Jeffrey Altman <jaltman@columbia.edu>
7074 via Richard Levitte]
7076 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7077 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7078 values for each of the key sizes rather than having just
7079 parameters (and 'speed' generating keys each time).
7082 *) Speed up EVP routines.
7085 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7086 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7087 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7088 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7090 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7091 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7092 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7095 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7097 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7100 *) Added the OS2-EMX target.
7101 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7103 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7104 to support NCONF routines in extension code. New function CONF_set_nconf()
7105 to allow functions which take an NCONF to also handle the old LHASH
7106 structure: this means that the old CONF compatible routines can be
7107 retained (in particular wrt extensions) without having to duplicate the
7108 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7111 *) Enhance the general user interface with mechanisms for inner control
7112 and with possibilities to have yes/no kind of prompts.
7115 *) Change all calls to low level digest routines in the library and
7116 applications to use EVP. Add missing calls to HMAC_cleanup() and
7117 don't assume HMAC_CTX can be copied using memcpy().
7118 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7120 *) Add the possibility to control engines through control names but with
7121 arbitrary arguments instead of just a string.
7122 Change the key loaders to take a UI_METHOD instead of a callback
7123 function pointer. NOTE: this breaks binary compatibility with earlier
7124 versions of OpenSSL [engine].
7125 Adapt the nCipher code for these new conditions and add a card insertion
7129 *) Enhance the general user interface with mechanisms to better support
7130 dialog box interfaces, application-defined prompts, the possibility
7131 to use defaults (for example default passwords from somewhere else)
7132 and interrupts/cancellations.
7135 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7136 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7139 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7140 tidy up some unnecessarily weird code in 'sk_new()').
7141 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7143 *) Change the key loading routines for ENGINEs to use the same kind
7144 callback (pem_password_cb) as all other routines that need this
7148 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7149 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7150 than this minimum value is recommended.
7153 *) New random seeder for OpenVMS, using the system process statistics
7154 that are easily reachable.
7157 *) Windows apparently can't transparently handle global
7158 variables defined in DLLs. Initialisations such as:
7160 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7162 won't compile. This is used by the any applications that need to
7163 declare their own ASN1 modules. This was fixed by adding the option
7164 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7165 needed for static libraries under Win32.
7168 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7169 setting of purpose and trust fields. New X509_STORE trust and
7170 purpose functions and tidy up setting in other SSL functions.
7173 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7174 structure. These are inherited by X509_STORE_CTX when it is
7175 initialised. This allows various defaults to be set in the
7176 X509_STORE structure (such as flags for CRL checking and custom
7177 purpose or trust settings) for functions which only use X509_STORE_CTX
7178 internally such as S/MIME.
7180 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7181 trust settings if they are not set in X509_STORE. This allows X509_STORE
7182 purposes and trust (in S/MIME for example) to override any set by default.
7184 Add command line options for CRL checking to smime, s_client and s_server
7188 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7189 are set then the CRL is looked up in the X509_STORE structure and
7190 its validity and signature checked, then if the certificate is found
7191 in the CRL the verify fails with a revoked error.
7193 Various new CRL related callbacks added to X509_STORE_CTX structure.
7195 Command line options added to 'verify' application to support this.
7197 This needs some additional work, such as being able to handle multiple
7198 CRLs with different times, extension based lookup (rather than just
7199 by subject name) and ultimately more complete V2 CRL extension
7203 *) Add a general user interface API (crypto/ui/). This is designed
7204 to replace things like des_read_password and friends (backward
7205 compatibility functions using this new API are provided).
7206 The purpose is to remove prompting functions from the DES code
7207 section as well as provide for prompting through dialog boxes in
7208 a window system and the like.
7211 *) Add "ex_data" support to ENGINE so implementations can add state at a
7212 per-structure level rather than having to store it globally.
7215 *) Make it possible for ENGINE structures to be copied when retrieved by
7216 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7217 This causes the "original" ENGINE structure to act like a template,
7218 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7219 operational state can be localised to each ENGINE structure, despite the
7220 fact they all share the same "methods". New ENGINE structures returned in
7221 this case have no functional references and the return value is the single
7222 structural reference. This matches the single structural reference returned
7223 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7227 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7228 needs to match any other type at all we need to manually clear the
7232 *) Changes to the "openssl engine" utility to include;
7233 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7234 about an ENGINE's available control commands.
7235 - executing control commands from command line arguments using the
7236 '-pre' and '-post' switches. '-post' is only used if '-t' is
7237 specified and the ENGINE is successfully initialised. The syntax for
7238 the individual commands are colon-separated, for example;
7239 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7242 *) New dynamic control command support for ENGINEs. ENGINEs can now
7243 declare their own commands (numbers), names (strings), descriptions,
7244 and input types for run-time discovery by calling applications. A
7245 subset of these commands are implicitly classed as "executable"
7246 depending on their input type, and only these can be invoked through
7247 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7248 can be based on user input, config files, etc). The distinction is
7249 that "executable" commands cannot return anything other than a boolean
7250 result and can only support numeric or string input, whereas some
7251 discoverable commands may only be for direct use through
7252 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7253 pointers, or other custom uses. The "executable" commands are to
7254 support parameterisations of ENGINE behaviour that can be
7255 unambiguously defined by ENGINEs and used consistently across any
7256 OpenSSL-based application. Commands have been added to all the
7257 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7258 control over shared-library paths without source code alterations.
7261 *) Changed all ENGINE implementations to dynamically allocate their
7262 ENGINEs rather than declaring them statically. Apart from this being
7263 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7264 this also allows the implementations to compile without using the
7265 internal engine_int.h header.
7268 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7269 'const' value. Any code that should be able to modify a RAND_METHOD
7270 should already have non-const pointers to it (ie. they should only
7271 modify their own ones).
7274 *) Made a variety of little tweaks to the ENGINE code.
7275 - "atalla" and "ubsec" string definitions were moved from header files
7276 to C code. "nuron" string definitions were placed in variables
7277 rather than hard-coded - allowing parameterisation of these values
7278 later on via ctrl() commands.
7279 - Removed unused "#if 0"'d code.
7280 - Fixed engine list iteration code so it uses ENGINE_free() to release
7281 structural references.
7282 - Constified the RAND_METHOD element of ENGINE structures.
7283 - Constified various get/set functions as appropriate and added
7284 missing functions (including a catch-all ENGINE_cpy that duplicates
7285 all ENGINE values onto a new ENGINE except reference counts/state).
7286 - Removed NULL parameter checks in get/set functions. Setting a method
7287 or function to NULL is a way of cancelling out a previously set
7288 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7289 and doesn't justify the extra error symbols and code.
7290 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7291 flags from engine_int.h to engine.h.
7292 - Changed prototypes for ENGINE handler functions (init(), finish(),
7293 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7296 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7297 to the algorithm using long division. The binary algorithm can be
7298 used only if the modulus is odd. On 32-bit systems, it is faster
7299 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7300 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7301 up to 450 bits. In 64-bit environments, the binary algorithm
7302 appears to be advantageous for much longer moduli; here we use it
7303 for moduli up to 2048 bits.
7306 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7307 could not support the combine flag in choice fields.
7310 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7311 extensions from a certificate request to the certificate.
7314 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7315 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7316 file: this allows the display of the certificate about to be
7317 signed to be customised, to allow certain fields to be included
7318 or excluded and extension details. The old system didn't display
7319 multicharacter strings properly, omitted fields not in the policy
7320 and couldn't display additional details such as extensions.
7323 *) Function EC_POINTs_mul for multiple scalar multiplication
7324 of an arbitrary number of elliptic curve points
7325 \sum scalars[i]*points[i],
7326 optionally including the generator defined for the EC_GROUP:
7327 scalar*generator + \sum scalars[i]*points[i].
7329 EC_POINT_mul is a simple wrapper function for the typical case
7330 that the point list has just one item (besides the optional
7334 *) First EC_METHODs for curves over GF(p):
7336 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7337 operations and provides various method functions that can also
7338 operate with faster implementations of modular arithmetic.
7340 EC_GFp_mont_method() reuses most functions that are part of
7341 EC_GFp_simple_method, but uses Montgomery arithmetic.
7343 [Bodo Moeller; point addition and point doubling
7344 implementation directly derived from source code provided by
7345 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7347 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7348 crypto/ec/ec_lib.c):
7350 Curves are EC_GROUP objects (with an optional group generator)
7351 based on EC_METHODs that are built into the library.
7353 Points are EC_POINT objects based on EC_GROUP objects.
7355 Most of the framework would be able to handle curves over arbitrary
7356 finite fields, but as there are no obvious types for fields other
7357 than GF(p), some functions are limited to that for now.
7360 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7361 that the file contains a complete HTTP response.
7364 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7365 change the def and num file printf format specifier from "%-40sXXX"
7366 to "%-39s XXX". The latter will always guarantee a space after the
7367 field while the former will cause them to run together if the field
7368 is 40 of more characters long.
7371 *) Constify the cipher and digest 'method' functions and structures
7372 and modify related functions to take constant EVP_MD and EVP_CIPHER
7376 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7377 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7380 *) Modify EVP_Digest*() routines so they now return values. Although the
7381 internal software routines can never fail additional hardware versions
7385 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7387 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7388 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7391 ERR_R_NESTED_ASN1_ERROR
7393 ERR_R_MISSING_ASN1_EOS
7394 were 4 .. 9, conflicting with
7395 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7397 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7398 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7400 Add new error code 'ERR_R_INTERNAL_ERROR'.
7403 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7407 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7408 sets the subject name for a new request or supersedes the
7409 subject name in a given request. Formats that can be parsed are
7410 'CN=Some Name, OU=myOU, C=IT'
7412 'CN=Some Name/OU=myOU/C=IT'.
7414 Add options '-batch' and '-verbose' to 'openssl req'.
7415 [Massimiliano Pala <madwolf@hackmasters.net>]
7417 *) Introduce the possibility to access global variables through
7418 functions on platform were that's the best way to handle exporting
7419 global variables in shared libraries. To enable this functionality,
7420 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7421 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7422 is normally done by Configure or something similar).
7424 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7425 in the source file (foo.c) like this:
7427 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7428 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7430 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7431 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7433 OPENSSL_DECLARE_GLOBAL(int,foo);
7434 #define foo OPENSSL_GLOBAL_REF(foo)
7435 OPENSSL_DECLARE_GLOBAL(double,bar);
7436 #define bar OPENSSL_GLOBAL_REF(bar)
7438 The #defines are very important, and therefore so is including the
7439 header file everywhere where the defined globals are used.
7441 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7442 of ASN.1 items, but that structure is a bit different.
7444 The largest change is in util/mkdef.pl which has been enhanced with
7445 better and easier to understand logic to choose which symbols should
7446 go into the Windows .def files as well as a number of fixes and code
7447 cleanup (among others, algorithm keywords are now sorted
7448 lexicographically to avoid constant rewrites).
7451 *) In BN_div() keep a copy of the sign of 'num' before writing the
7452 result to 'rm' because if rm==num the value will be overwritten
7453 and produce the wrong result if 'num' is negative: this caused
7454 problems with BN_mod() and BN_nnmod().
7457 *) Function OCSP_request_verify(). This checks the signature on an
7458 OCSP request and verifies the signer certificate. The signer
7459 certificate is just checked for a generic purpose and OCSP request
7463 *) Add OCSP_check_validity() function to check the validity of OCSP
7464 responses. OCSP responses are prepared in real time and may only
7465 be a few seconds old. Simply checking that the current time lies
7466 between thisUpdate and nextUpdate max reject otherwise valid responses
7467 caused by either OCSP responder or client clock inaccuracy. Instead
7468 we allow thisUpdate and nextUpdate to fall within a certain period of
7469 the current time. The age of the response can also optionally be
7470 checked. Two new options -validity_period and -status_age added to
7474 *) If signature or public key algorithm is unrecognized print out its
7475 OID rather that just UNKNOWN.
7478 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7479 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7480 ID to be generated from the issuer certificate alone which can then be
7481 passed to OCSP_id_issuer_cmp().
7484 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7485 ASN1 modules to export functions returning ASN1_ITEM pointers
7486 instead of the ASN1_ITEM structures themselves. This adds several
7487 new macros which allow the underlying ASN1 function/structure to
7488 be accessed transparently. As a result code should not use ASN1_ITEM
7489 references directly (such as &X509_it) but instead use the relevant
7490 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7491 use of the new ASN1 code on platforms where exporting structures
7492 is problematical (for example in shared libraries) but exporting
7493 functions returning pointers to structures is not.
7496 *) Add support for overriding the generation of SSL/TLS session IDs.
7497 These callbacks can be registered either in an SSL_CTX or per SSL.
7498 The purpose of this is to allow applications to control, if they wish,
7499 the arbitrary values chosen for use as session IDs, particularly as it
7500 can be useful for session caching in multiple-server environments. A
7501 command-line switch for testing this (and any client code that wishes
7502 to use such a feature) has been added to "s_server".
7503 [Geoff Thorpe, Lutz Jaenicke]
7505 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7506 of the form '#if defined(...) || defined(...) || ...' and
7507 '#if !defined(...) && !defined(...) && ...'. This also avoids
7508 the growing number of special cases it was previously handling.
7511 *) Make all configuration macros available for application by making
7512 sure they are available in opensslconf.h, by giving them names starting
7513 with "OPENSSL_" to avoid conflicts with other packages and by making
7514 sure e_os2.h will cover all platform-specific cases together with
7516 Additionally, it is now possible to define configuration/platform-
7517 specific names (called "system identities"). In the C code, these
7518 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7519 macro with the name beginning with "OPENSSL_SYS_", which is determined
7520 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7524 *) New option -set_serial to 'req' and 'x509' this allows the serial
7525 number to use to be specified on the command line. Previously self
7526 signed certificates were hard coded with serial number 0 and the
7527 CA options of 'x509' had to use a serial number in a file which was
7531 *) New options to 'ca' utility to support V2 CRL entry extensions.
7532 Currently CRL reason, invalidity date and hold instruction are
7533 supported. Add new CRL extensions to V3 code and some new objects.
7536 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7537 disable standard block padding (aka PKCS#5 padding) in the EVP
7538 API, which was previously mandatory. This means that the data is
7539 not padded in any way and so the total length much be a multiple
7540 of the block size, otherwise an error occurs.
7543 *) Initial (incomplete) OCSP SSL support.
7546 *) New function OCSP_parse_url(). This splits up a URL into its host,
7547 port and path components: primarily to parse OCSP URLs. New -url
7548 option to ocsp utility.
7551 *) New nonce behavior. The return value of OCSP_check_nonce() now
7552 reflects the various checks performed. Applications can decide
7553 whether to tolerate certain situations such as an absent nonce
7554 in a response when one was present in a request: the ocsp application
7555 just prints out a warning. New function OCSP_add1_basic_nonce()
7556 this is to allow responders to include a nonce in a response even if
7557 the request is nonce-less.
7560 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7561 skipped when using openssl x509 multiple times on a single input file,
7562 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7565 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7566 set string type: to handle setting ASN1_TIME structures. Fix ca
7567 utility to correctly initialize revocation date of CRLs.
7570 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7571 the clients preferred ciphersuites and rather use its own preferences.
7572 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7573 Internet Explorer by ensuring unchanged hash method during stepup.
7574 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7577 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7578 to aes and add a new 'exist' option to print out symbols that don't
7582 *) Additional options to ocsp utility to allow flags to be set and
7583 additional certificates supplied.
7586 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7587 OCSP client a number of certificate to only verify the response
7591 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7592 handle the new API. Currently only ECB, CBC modes supported. Add new
7595 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7596 Encryption Standard (AES) Ciphersuites for Transport Layer
7597 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7598 not enabled by default and were not part of the "ALL" ciphersuite
7599 alias because they were not yet official; they could be
7600 explicitly requested by specifying the "AESdraft" ciphersuite
7601 group alias. In the final release of OpenSSL 0.9.7, the group
7602 alias is called "AES" and is part of "ALL".)
7603 [Ben Laurie, Steve Henson, Bodo Moeller]
7605 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7606 request to response.
7609 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7610 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7611 extract information from a certificate request. OCSP_response_create()
7612 creates a response and optionally adds a basic response structure.
7613 OCSP_basic_add1_status() adds a complete single response to a basic
7614 response and returns the OCSP_SINGLERESP structure just added (to allow
7615 extensions to be included for example). OCSP_basic_add1_cert() adds a
7616 certificate to a basic response and OCSP_basic_sign() signs a basic
7617 response with various flags. New helper functions ASN1_TIME_check()
7618 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7619 (converts ASN1_TIME to GeneralizedTime).
7622 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7623 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7624 structure from a certificate. X509_pubkey_digest() digests the public_key
7625 contents: this is used in various key identifiers.
7628 *) Make sk_sort() tolerate a NULL argument.
7629 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7631 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7632 passed by the function are trusted implicitly. If any of them signed the
7633 response then it is assumed to be valid and is not verified.
7636 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7637 to data. This was previously part of the PKCS7 ASN1 code. This
7638 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7639 [Steve Henson, reported by Kenneth R. Robinette
7640 <support@securenetterm.com>]
7642 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7643 routines: without these tracing memory leaks is very painful.
7644 Fix leaks in PKCS12 and PKCS7 routines.
7647 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7648 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7649 effectively meant GeneralizedTime would never be used. Now it
7650 is initialised to -1 but X509_time_adj() now has to check the value
7651 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7652 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7653 [Steve Henson, reported by Kenneth R. Robinette
7654 <support@securenetterm.com>]
7656 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7657 result in a zero length in the ASN1_INTEGER structure which was
7658 not consistent with the structure when d2i_ASN1_INTEGER() was used
7659 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7660 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7661 where it did not print out a minus for negative ASN1_INTEGER.
7664 *) Add summary printout to ocsp utility. The various functions which
7665 convert status values to strings have been renamed to:
7666 OCSP_response_status_str(), OCSP_cert_status_str() and
7667 OCSP_crl_reason_str() and are no longer static. New options
7668 to verify nonce values and to disable verification. OCSP response
7669 printout format cleaned up.
7672 *) Add additional OCSP certificate checks. These are those specified
7673 in RFC2560. This consists of two separate checks: the CA of the
7674 certificate being checked must either be the OCSP signer certificate
7675 or the issuer of the OCSP signer certificate. In the latter case the
7676 OCSP signer certificate must contain the OCSP signing extended key
7677 usage. This check is performed by attempting to match the OCSP
7678 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7679 in the OCSP_CERTID structures of the response.
7682 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7683 and related routines. This uses the standard OpenSSL certificate
7684 verify routines to perform initial checks (just CA validity) and
7685 to obtain the certificate chain. Then additional checks will be
7686 performed on the chain. Currently the root CA is checked to see
7687 if it is explicitly trusted for OCSP signing. This is used to set
7688 a root CA as a global signing root: that is any certificate that
7689 chains to that CA is an acceptable OCSP signing certificate.
7692 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7693 extensions from a separate configuration file.
7694 As when reading extensions from the main configuration file,
7695 the '-extensions ...' option may be used for specifying the
7697 [Massimiliano Pala <madwolf@comune.modena.it>]
7699 *) New OCSP utility. Allows OCSP requests to be generated or
7700 read. The request can be sent to a responder and the output
7701 parsed, outputed or printed in text form. Not complete yet:
7702 still needs to check the OCSP response validity.
7705 *) New subcommands for 'openssl ca':
7706 'openssl ca -status <serial>' prints the status of the cert with
7707 the given serial number (according to the index file).
7708 'openssl ca -updatedb' updates the expiry status of certificates
7710 [Massimiliano Pala <madwolf@comune.modena.it>]
7712 *) New '-newreq-nodes' command option to CA.pl. This is like
7713 '-newreq', but calls 'openssl req' with the '-nodes' option
7714 so that the resulting key is not encrypted.
7715 [Damien Miller <djm@mindrot.org>]
7717 *) New configuration for the GNU Hurd.
7718 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7720 *) Initial code to implement OCSP basic response verify. This
7721 is currently incomplete. Currently just finds the signer's
7722 certificate and verifies the signature on the response.
7725 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7726 value of OPENSSLDIR. This is available via the new '-d' option
7727 to 'openssl version', and is also included in 'openssl version -a'.
7730 *) Allowing defining memory allocation callbacks that will be given
7731 file name and line number information in additional arguments
7732 (a const char* and an int). The basic functionality remains, as
7733 well as the original possibility to just replace malloc(),
7734 realloc() and free() by functions that do not know about these
7735 additional arguments. To register and find out the current
7736 settings for extended allocation functions, the following
7737 functions are provided:
7739 CRYPTO_set_mem_ex_functions
7740 CRYPTO_set_locked_mem_ex_functions
7741 CRYPTO_get_mem_ex_functions
7742 CRYPTO_get_locked_mem_ex_functions
7744 These work the same way as CRYPTO_set_mem_functions and friends.
7745 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7746 extended allocation function is enabled.
7747 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7748 a conventional allocation function is enabled.
7749 [Richard Levitte, Bodo Moeller]
7751 *) Finish off removing the remaining LHASH function pointer casts.
7752 There should no longer be any prototype-casting required when using
7753 the LHASH abstraction, and any casts that remain are "bugs". See
7754 the callback types and macros at the head of lhash.h for details
7755 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7758 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7759 If /dev/[u]random devices are not available or do not return enough
7760 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7762 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7763 /etc/entropy will be queried once each in this sequence, querying stops
7764 when enough entropy was collected without querying more sockets.
7767 *) Change the Unix RAND_poll() variant to be able to poll several
7768 random devices, as specified by DEVRANDOM, until a sufficient amount
7769 of data has been collected. We spend at most 10 ms on each file
7770 (select timeout) and read in non-blocking mode. DEVRANDOM now
7771 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7772 (previously it was just the string "/dev/urandom"), so on typical
7773 platforms the 10 ms delay will never occur.
7774 Also separate out the Unix variant to its own file, rand_unix.c.
7775 For VMS, there's a currently-empty rand_vms.c.
7778 *) Move OCSP client related routines to ocsp_cl.c. These
7779 provide utility functions which an application needing
7780 to issue a request to an OCSP responder and analyse the
7781 response will typically need: as opposed to those which an
7782 OCSP responder itself would need which will be added later.
7784 OCSP_request_sign() signs an OCSP request with an API similar
7785 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7786 response. OCSP_response_get1_basic() extracts basic response
7787 from response. OCSP_resp_find_status(): finds and extracts status
7788 information from an OCSP_CERTID structure (which will be created
7789 when the request structure is built). These are built from lower
7790 level functions which work on OCSP_SINGLERESP structures but
7791 won't normally be used unless the application wishes to examine
7792 extensions in the OCSP response for example.
7794 Replace nonce routines with a pair of functions.
7795 OCSP_request_add1_nonce() adds a nonce value and optionally
7796 generates a random value. OCSP_check_nonce() checks the
7797 validity of the nonce in an OCSP response.
7800 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7801 This doesn't copy the supplied OCSP_CERTID and avoids the
7802 need to free up the newly created id. Change return type
7803 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7804 This can then be used to add extensions to the request.
7805 Deleted OCSP_request_new(), since most of its functionality
7806 is now in OCSP_REQUEST_new() (and the case insensitive name
7807 clash) apart from the ability to set the request name which
7808 will be added elsewhere.
7811 *) Update OCSP API. Remove obsolete extensions argument from
7812 various functions. Extensions are now handled using the new
7813 OCSP extension code. New simple OCSP HTTP function which
7814 can be used to send requests and parse the response.
7817 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7818 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7819 uses the special reorder version of SET OF to sort the attributes
7820 and reorder them to match the encoded order. This resolves a long
7821 standing problem: a verify on a PKCS7 structure just after signing
7822 it used to fail because the attribute order did not match the
7823 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7824 it uses the received order. This is necessary to tolerate some broken
7825 software that does not order SET OF. This is handled by encoding
7826 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7827 to produce the required SET OF.
7830 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7831 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7832 files to get correct declarations of the ASN.1 item variables.
7835 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7836 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7837 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7838 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7839 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7840 ASN1_ITEM and no wrapper functions.
7843 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7844 replace the old function pointer based I/O routines. Change most of
7845 the *_d2i_bio() and *_d2i_fp() functions to use these.
7848 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7849 lines, recognize more "algorithms" that can be deselected, and make
7850 it complain about algorithm deselection that isn't recognised.
7853 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7854 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7855 to use new functions. Add NO_ASN1_OLD which can be set to remove
7856 some old style ASN1 functions: this can be used to determine if old
7857 code will still work when these eventually go away.
7860 *) New extension functions for OCSP structures, these follow the
7861 same conventions as certificates and CRLs.
7864 *) New function X509V3_add1_i2d(). This automatically encodes and
7865 adds an extension. Its behaviour can be customised with various
7866 flags to append, replace or delete. Various wrappers added for
7867 certificates and CRLs.
7870 *) Fix to avoid calling the underlying ASN1 print routine when
7871 an extension cannot be parsed. Correct a typo in the
7872 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7875 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7876 entries for variables.
7879 *) Add functionality to apps/openssl.c for detecting locking
7880 problems: As the program is single-threaded, all we have
7881 to do is register a locking callback using an array for
7882 storing which locks are currently held by the program.
7885 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7886 SSL_get_ex_data_X509_STORE_idx(), which is used in
7887 ssl_verify_cert_chain() and thus can be called at any time
7888 during TLS/SSL handshakes so that thread-safety is essential.
7889 Unfortunately, the ex_data design is not at all suited
7890 for multi-threaded use, so it probably should be abolished.
7893 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7894 [Broadcom, tweaked and integrated by Geoff Thorpe]
7896 *) Move common extension printing code to new function
7897 X509V3_print_extensions(). Reorganise OCSP print routines and
7898 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7901 *) New function X509_signature_print() to remove duplication in some
7905 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7906 set (this was treated exactly the same as SET OF previously). This
7907 is used to reorder the STACK representing the structure to match the
7908 encoding. This will be used to get round a problem where a PKCS7
7909 structure which was signed could not be verified because the STACK
7910 order did not reflect the encoded order.
7913 *) Reimplement the OCSP ASN1 module using the new code.
7916 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7917 for its ASN1 operations. The old style function pointers still exist
7918 for now but they will eventually go away.
7921 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7922 completely replaces the old ASN1 functionality with a table driven
7923 encoder and decoder which interprets an ASN1_ITEM structure describing
7924 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7925 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7926 has also been converted to the new form.
7929 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7930 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7931 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7932 for negative moduli.
7935 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7936 of not touching the result's sign bit.
7939 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7943 *) Changed the LHASH code to use prototypes for callbacks, and created
7944 macros to declare and implement thin (optionally static) functions
7945 that provide type-safety and avoid function pointer casting for the
7946 type-specific callbacks.
7949 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7951 [Veers Staats <staatsvr@asc.hpc.mil>,
7952 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7954 *) Reformat the FAQ so the different questions and answers can be divided
7955 in sections depending on the subject.
7958 *) Have the zlib compression code load ZLIB.DLL dynamically under
7962 *) New function BN_mod_sqrt for computing square roots modulo a prime
7963 (using the probabilistic Tonelli-Shanks algorithm unless
7964 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7965 be handled deterministically).
7966 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7968 *) Make BN_mod_inverse faster by explicitly handling small quotients
7969 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7970 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7973 *) New function BN_kronecker.
7976 *) Fix BN_gcd so that it works on negative inputs; the result is
7977 positive unless both parameters are zero.
7978 Previously something reasonably close to an infinite loop was
7979 possible because numbers could be growing instead of shrinking
7980 in the implementation of Euclid's algorithm.
7983 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7984 sign of the number in question.
7986 Fix BN_is_word(a,w) to work correctly for w == 0.
7988 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7989 because its test if the absolute value of 'a' equals 'w'.
7990 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7991 it exists mostly for use in the implementations of BN_is_zero(),
7992 BN_is_one(), and BN_is_word().
7995 *) New function BN_swap.
7998 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7999 the exponentiation functions are more likely to produce reasonable
8000 results on negative inputs.
8003 *) Change BN_mod_mul so that the result is always non-negative.
8004 Previously, it could be negative if one of the factors was negative;
8005 I don't think anyone really wanted that behaviour.
8008 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8009 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8010 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8011 and add new functions:
8020 BN_mod_lshift1_quick
8024 These functions always generate non-negative results.
8026 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8027 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8029 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8030 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8031 be reduced modulo m.
8032 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8035 The following entry accidentally appeared in the CHANGES file
8036 distributed with OpenSSL 0.9.7. The modifications described in
8037 it do *not* apply to OpenSSL 0.9.7.
8039 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8040 was actually never needed) and in BN_mul(). The removal in BN_mul()
8041 required a small change in bn_mul_part_recursive() and the addition
8042 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8043 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8044 bn_sub_words() and bn_add_words() except they take arrays with
8049 *) In 'openssl passwd', verify passwords read from the terminal
8050 unless the '-salt' option is used (which usually means that
8051 verification would just waste user's time since the resulting
8052 hash is going to be compared with some given password hash)
8053 or the new '-noverify' option is used.
8055 This is an incompatible change, but it does not affect
8056 non-interactive use of 'openssl passwd' (passwords on the command
8057 line, '-stdin' option, '-in ...' option) and thus should not
8061 *) Remove all references to RSAref, since there's no more need for it.
8064 *) Make DSO load along a path given through an environment variable
8065 (SHLIB_PATH) with shl_load().
8068 *) Constify the ENGINE code as a result of BIGNUM constification.
8069 Also constify the RSA code and most things related to it. In a
8070 few places, most notable in the depth of the ASN.1 code, ugly
8071 casts back to non-const were required (to be solved at a later
8075 *) Make it so the openssl application has all engines loaded by default.
8078 *) Constify the BIGNUM routines a little more.
8081 *) Add the following functions:
8083 ENGINE_load_cswift()
8085 ENGINE_load_atalla()
8087 ENGINE_load_builtin_engines()
8089 That way, an application can itself choose if external engines that
8090 are built-in in OpenSSL shall ever be used or not. The benefit is
8091 that applications won't have to be linked with libdl or other dso
8092 libraries unless it's really needed.
8094 Changed 'openssl engine' to load all engines on demand.
8095 Changed the engine header files to avoid the duplication of some
8096 declarations (they differed!).
8099 *) 'openssl engine' can now list capabilities.
8102 *) Better error reporting in 'openssl engine'.
8105 *) Never call load_dh_param(NULL) in s_server.
8108 *) Add engine application. It can currently list engines by name and
8109 identity, and test if they are actually available.
8112 *) Improve RPM specification file by forcing symbolic linking and making
8113 sure the installed documentation is also owned by root.root.
8114 [Damien Miller <djm@mindrot.org>]
8116 *) Give the OpenSSL applications more possibilities to make use of
8117 keys (public as well as private) handled by engines.
8120 *) Add OCSP code that comes from CertCo.
8123 *) Add VMS support for the Rijndael code.
8126 *) Added untested support for Nuron crypto accelerator.
8129 *) Add support for external cryptographic devices. This code was
8130 previously distributed separately as the "engine" branch.
8131 [Geoff Thorpe, Richard Levitte]
8133 *) Rework the filename-translation in the DSO code. It is now possible to
8134 have far greater control over how a "name" is turned into a filename
8135 depending on the operating environment and any oddities about the
8136 different shared library filenames on each system.
8139 *) Support threads on FreeBSD-elf in Configure.
8142 *) Fix for SHA1 assembly problem with MASM: it produces
8143 warnings about corrupt line number information when assembling
8144 with debugging information. This is caused by the overlapping
8146 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8149 NCONF_get_number() has no error checking at all. As a replacement,
8150 NCONF_get_number_e() is defined (_e for "error checking") and is
8151 promoted strongly. The old NCONF_get_number is kept around for
8152 binary backward compatibility.
8153 Make it possible for methods to load from something other than a BIO,
8154 by providing a function pointer that is given a name instead of a BIO.
8155 For example, this could be used to load configuration data from an
8159 *) Fix for non blocking accept BIOs. Added new I/O special reason
8160 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8161 with non blocking I/O was not possible because no retry code was
8162 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8166 *) Added the beginnings of Rijndael support.
8169 *) Fix for bug in DirectoryString mask setting. Add support for
8170 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8171 to allow certificate printing to more controllable, additional
8172 'certopt' option to 'x509' to allow new printing options to be
8176 *) Clean old EAY MD5 hack from e_os.h.
8179 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8181 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8182 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8183 [Joe Orton, Steve Henson]
8185 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8187 *) Fix additional bug revealed by the NISCC test suite:
8189 Stop bug triggering large recursion when presented with
8190 certain ASN.1 tags (CVE-2003-0851)
8193 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8195 *) Fix various bugs revealed by running the NISCC test suite:
8197 Stop out of bounds reads in the ASN1 code when presented with
8198 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8200 If verify callback ignores invalid public key errors don't try to check
8201 certificate signature with the NULL public key.
8205 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8206 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8210 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8211 extra data after the compression methods not only for TLS 1.0
8212 but also for SSL 3.0 (as required by the specification).
8213 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8215 *) Change X509_certificate_type() to mark the key as exported/exportable
8216 when it's 512 *bits* long, not 512 bytes.
8219 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8221 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8222 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8223 a protocol version number mismatch like a decryption error
8224 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8227 *) Turn on RSA blinding by default in the default implementation
8228 to avoid a timing attack. Applications that don't want it can call
8229 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8230 They would be ill-advised to do so in most cases.
8231 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8233 *) Change RSA blinding code so that it works when the PRNG is not
8234 seeded (in this case, the secret RSA exponent is abused as
8235 an unpredictable seed -- if it is not unpredictable, there
8236 is no point in blinding anyway). Make RSA blinding thread-safe
8237 by remembering the creator's thread ID in rsa->blinding and
8238 having all other threads use local one-time blinding factors
8239 (this requires more computation than sharing rsa->blinding, but
8240 avoids excessive locking; and if an RSA object is not shared
8241 between threads, blinding will still be very fast).
8244 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8246 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8247 via timing by performing a MAC computation even if incorrect
8248 block cipher padding has been found. This is a countermeasure
8249 against active attacks where the attacker has to distinguish
8250 between bad padding and a MAC verification error. (CVE-2003-0078)
8252 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8253 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8254 Martin Vuagnoux (EPFL, Ilion)]
8256 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8258 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8259 memory from it's contents. This is done with a counter that will
8260 place alternating values in each byte. This can be used to solve
8261 two issues: 1) the removal of calls to memset() by highly optimizing
8262 compilers, and 2) cleansing with other values than 0, since those can
8263 be read through on certain media, for example a swap space on disk.
8266 *) Bugfix: client side session caching did not work with external caching,
8267 because the session->cipher setting was not restored when reloading
8268 from the external cache. This problem was masked, when
8269 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8270 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8273 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8274 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8275 [Zeev Lieber <zeev-l@yahoo.com>]
8277 *) Undo an undocumented change introduced in 0.9.6e which caused
8278 repeated calls to OpenSSL_add_all_ciphers() and
8279 OpenSSL_add_all_digests() to be ignored, even after calling
8283 *) Change the default configuration reader to deal with last line not
8284 being properly terminated.
8287 *) Change X509_NAME_cmp() so it applies the special rules on handling
8288 DN values that are of type PrintableString, as well as RDNs of type
8289 emailAddress where the value has the type ia5String.
8290 [stefank@valicert.com via Richard Levitte]
8292 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8293 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8294 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8295 the bitwise-OR of the two for use by the majority of applications
8296 wanting this behaviour, and update the docs. The documented
8297 behaviour and actual behaviour were inconsistent and had been
8298 changing anyway, so this is more a bug-fix than a behavioural
8300 [Geoff Thorpe, diagnosed by Nadav Har'El]
8302 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8303 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8306 *) Fix initialization code race conditions in
8307 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8308 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8309 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8310 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8311 ssl2_get_cipher_by_char(),
8312 ssl3_get_cipher_by_char().
8313 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8315 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8316 the cached sessions are flushed, as the remove_cb() might use ex_data
8317 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8318 (see [openssl.org #212]).
8319 [Geoff Thorpe, Lutz Jaenicke]
8321 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8322 length, instead of the encoding length to d2i_ASN1_OBJECT.
8325 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8327 *) [In 0.9.6g-engine release:]
8328 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8329 [Lynn Gazis <lgazis@rainbow.com>]
8331 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8333 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8334 and get fix the header length calculation.
8335 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8336 Alon Kantor <alonk@checkpoint.com> (and others),
8339 *) Use proper error handling instead of 'assertions' in buffer
8340 overflow checks added in 0.9.6e. This prevents DoS (the
8341 assertions could call abort()).
8342 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8344 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8346 *) Add various sanity checks to asn1_get_length() to reject
8347 the ASN1 length bytes if they exceed sizeof(long), will appear
8348 negative or the content length exceeds the length of the
8350 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8352 *) Fix cipher selection routines: ciphers without encryption had no flags
8353 for the cipher strength set and where therefore not handled correctly
8354 by the selection routines (PR #130).
8357 *) Fix EVP_dsa_sha macro.
8361 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8362 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8363 that was added in OpenSSL 0.9.6d.
8365 As the countermeasure turned out to be incompatible with some
8366 broken SSL implementations, the new option is part of SSL_OP_ALL.
8367 SSL_OP_ALL is usually employed when compatibility with weird SSL
8368 implementations is desired (e.g. '-bugs' option to 's_client' and
8369 's_server'), so the new option is automatically set in many
8373 *) Changes in security patch:
8375 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8376 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8377 Air Force Materiel Command, USAF, under agreement number
8380 *) Add various sanity checks to asn1_get_length() to reject
8381 the ASN1 length bytes if they exceed sizeof(long), will appear
8382 negative or the content length exceeds the length of the
8383 supplied buffer. (CVE-2002-0659)
8384 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8386 *) Assertions for various potential buffer overflows, not known to
8388 [Ben Laurie (CHATS)]
8390 *) Various temporary buffers to hold ASCII versions of integers were
8391 too small for 64 bit platforms. (CVE-2002-0655)
8392 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8394 *) Remote buffer overflow in SSL3 protocol - an attacker could
8395 supply an oversized session ID to a client. (CVE-2002-0656)
8396 [Ben Laurie (CHATS)]
8398 *) Remote buffer overflow in SSL2 protocol - an attacker could
8399 supply an oversized client master key. (CVE-2002-0656)
8400 [Ben Laurie (CHATS)]
8402 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8404 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8405 encoded as NULL) with id-dsa-with-sha1.
8406 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8408 *) Check various X509_...() return values in apps/req.c.
8409 [Nils Larsch <nla@trustcenter.de>]
8411 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8412 an end-of-file condition would erroneously be flagged, when the CRLF
8413 was just at the end of a processed block. The bug was discovered when
8414 processing data through a buffering memory BIO handing the data to a
8415 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8416 <ptsekov@syntrex.com> and Nedelcho Stanev.
8419 *) Implement a countermeasure against a vulnerability recently found
8420 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8421 before application data chunks to avoid the use of known IVs
8422 with data potentially chosen by the attacker.
8425 *) Fix length checks in ssl3_get_client_hello().
8428 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8429 to prevent ssl3_read_internal() from incorrectly assuming that
8430 ssl3_read_bytes() found application data while handshake
8431 processing was enabled when in fact s->s3->in_read_app_data was
8432 merely automatically cleared during the initial handshake.
8433 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8435 *) Fix object definitions for Private and Enterprise: they were not
8436 recognized in their shortname (=lowercase) representation. Extend
8437 obj_dat.pl to issue an error when using undefined keywords instead
8438 of silently ignoring the problem (Svenning Sorensen
8439 <sss@sss.dnsalias.net>).
8442 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8443 generators, i.e. generators other than 2 and 5. (Previously, the
8444 code did not properly initialise the 'add' and 'rem' values to
8445 BN_generate_prime().)
8447 In the new general case, we do not insist that 'generator' is
8448 actually a primitive root: This requirement is rather pointless;
8449 a generator of the order-q subgroup is just as good, if not
8453 *) Map new X509 verification errors to alerts. Discovered and submitted by
8454 Tom Wu <tom@arcot.com>.
8457 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8458 returning non-zero before the data has been completely received
8459 when using non-blocking I/O.
8460 [Bodo Moeller; problem pointed out by John Hughes]
8462 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8463 [Ben Laurie, Lutz Jaenicke]
8465 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8466 Yoram Zahavi <YoramZ@gilian.com>).
8469 *) Add information about CygWin 1.3 and on, and preserve proper
8470 configuration for the versions before that.
8471 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8473 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8474 check whether we deal with a copy of a session and do not delete from
8475 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8476 <izhar@checkpoint.com>.
8479 *) Do not store session data into the internal session cache, if it
8480 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8481 flag is set). Proposed by Aslam <aslam@funk.com>.
8484 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8488 *) [In 0.9.6d-engine release:]
8489 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8490 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8492 *) Add the configuration target linux-s390x.
8493 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8495 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8496 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8497 variable as an indication that a ClientHello message has been
8498 received. As the flag value will be lost between multiple
8499 invocations of ssl3_accept when using non-blocking I/O, the
8500 function may not be aware that a handshake has actually taken
8501 place, thus preventing a new session from being added to the
8504 To avoid this problem, we now set s->new_session to 2 instead of
8505 using a local variable.
8506 [Lutz Jaenicke, Bodo Moeller]
8508 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8509 if the SSL_R_LENGTH_MISMATCH error is detected.
8510 [Geoff Thorpe, Bodo Moeller]
8512 *) New 'shared_ldflag' column in Configure platform table.
8515 *) Fix EVP_CIPHER_mode macro.
8516 ["Dan S. Camper" <dan@bti.net>]
8518 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8519 type, we must throw them away by setting rr->length to 0.
8520 [D P Chang <dpc@qualys.com>]
8522 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8524 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8525 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8526 worked incorrectly for those cases where range = 10..._2 and
8527 3*range is two bits longer than range.)
8530 *) Only add signing time to PKCS7 structures if it is not already
8534 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8535 OBJ_ld_ce should be OBJ_id_ce.
8536 Also some ip-pda OIDs in crypto/objects/objects.txt were
8537 incorrect (cf. RFC 3039).
8538 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8540 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8541 returns early because it has nothing to do.
8542 [Andy Schneider <andy.schneider@bjss.co.uk>]
8544 *) [In 0.9.6c-engine release:]
8545 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8546 [Andy Schneider <andy.schneider@bjss.co.uk>]
8548 *) [In 0.9.6c-engine release:]
8549 Add support for Cryptographic Appliance's keyserver technology.
8550 (Use engine 'keyclient')
8551 [Cryptographic Appliances and Geoff Thorpe]
8553 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8554 is called via tools/c89.sh because arguments have to be
8555 rearranged (all '-L' options must appear before the first object
8557 [Richard Shapiro <rshapiro@abinitio.com>]
8559 *) [In 0.9.6c-engine release:]
8560 Add support for Broadcom crypto accelerator cards, backported
8562 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8564 *) [In 0.9.6c-engine release:]
8565 Add support for SureWare crypto accelerator cards from
8566 Baltimore Technologies. (Use engine 'sureware')
8567 [Baltimore Technologies and Mark Cox]
8569 *) [In 0.9.6c-engine release:]
8570 Add support for crypto accelerator cards from Accelerated
8571 Encryption Processing, www.aep.ie. (Use engine 'aep')
8572 [AEP Inc. and Mark Cox]
8574 *) Add a configuration entry for gcc on UnixWare.
8575 [Gary Benson <gbenson@redhat.com>]
8577 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8578 messages are stored in a single piece (fixed-length part and
8579 variable-length part combined) and fix various bugs found on the way.
8582 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8583 instead. BIO_gethostbyname() does not know what timeouts are
8584 appropriate, so entries would stay in cache even when they have
8586 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8588 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8589 faced with a pathologically small ClientHello fragment that does
8590 not contain client_version: Instead of aborting with an error,
8591 simply choose the highest available protocol version (i.e.,
8592 TLS 1.0 unless it is disabled). In practice, ClientHello
8593 messages are never sent like this, but this change gives us
8594 strictly correct behaviour at least for TLS.
8597 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8598 never resets s->method to s->ctx->method when called from within
8599 one of the SSL handshake functions.
8600 [Bodo Moeller; problem pointed out by Niko Baric]
8602 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8603 (sent using the client's version number) if client_version is
8604 smaller than the protocol version in use. Also change
8605 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8606 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8607 the client will at least see that alert.
8610 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8614 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8615 client receives HelloRequest while in a handshake.
8616 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8618 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8619 should end in 'break', not 'goto end' which circumvents various
8620 cleanups done in state SSL_ST_OK. But session related stuff
8621 must be disabled for SSL_ST_OK in the case that we just sent a
8624 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8625 before just sending a HelloRequest.
8626 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8628 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8629 reveal whether illegal block cipher padding was found or a MAC
8630 verification error occurred. (Neither SSLerr() codes nor alerts
8631 are directly visible to potential attackers, but the information
8632 may leak via logfiles.)
8634 Similar changes are not required for the SSL 2.0 implementation
8635 because the number of padding bytes is sent in clear for SSL 2.0,
8636 and the extra bytes are just ignored. However ssl/s2_pkt.c
8637 failed to verify that the purported number of padding bytes is in
8641 *) Add OpenUNIX-8 support including shared libraries
8642 (Boyd Lynn Gerber <gerberb@zenez.com>).
8645 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8646 'wristwatch attack' using huge encoding parameters (cf.
8647 James H. Manger's CRYPTO 2001 paper). Note that the
8648 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8649 encoding parameters and hence was not vulnerable.
8652 *) BN_sqr() bug fix.
8653 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8655 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8656 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8657 followed by modular reduction.
8658 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8660 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8661 equivalent based on BN_pseudo_rand() instead of BN_rand().
8664 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8665 This function was broken, as the check for a new client hello message
8666 to handle SGC did not allow these large messages.
8667 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8670 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8673 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8674 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8677 *) Rework the configuration and shared library support for Tru64 Unix.
8678 The configuration part makes use of modern compiler features and
8679 still retains old compiler behavior for those that run older versions
8680 of the OS. The shared library support part includes a variant that
8681 uses the RPATH feature, and is available through the special
8682 configuration target "alpha-cc-rpath", which will never be selected
8684 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8686 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8687 with the same message size as in ssl3_get_certificate_request().
8688 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8689 messages might inadvertently be reject as too long.
8690 [Petr Lampa <lampa@fee.vutbr.cz>]
8692 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8695 *) Modified SSL library such that the verify_callback that has been set
8696 specificly for an SSL object with SSL_set_verify() is actually being
8697 used. Before the change, a verify_callback set with this function was
8698 ignored and the verify_callback() set in the SSL_CTX at the time of
8699 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8700 to allow the necessary settings.
8703 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8704 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8705 done automatically (in contradiction to the requirements of the C
8706 standard). This made problems when used from OpenSSH.
8709 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8710 dh->length and always used
8712 BN_rand_range(priv_key, dh->p).
8714 BN_rand_range() is not necessary for Diffie-Hellman, and this
8715 specific range makes Diffie-Hellman unnecessarily inefficient if
8716 dh->length (recommended exponent length) is much smaller than the
8717 length of dh->p. We could use BN_rand_range() if the order of
8718 the subgroup was stored in the DH structure, but we only have
8723 BN_rand(priv_key, l, ...)
8725 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8731 RSA_eay_public_encrypt
8732 RSA_eay_private_decrypt
8733 RSA_eay_private_encrypt (signing)
8734 RSA_eay_public_decrypt (signature verification)
8736 (default implementations for RSA_public_encrypt,
8737 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8738 always reject numbers >= n.
8741 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8742 to synchronize access to 'locking_thread'. This is necessary on
8743 systems where access to 'locking_thread' (an 'unsigned long'
8744 variable) is not atomic.
8747 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8748 *before* setting the 'crypto_lock_rand' flag. The previous code had
8749 a race condition if 0 is a valid thread ID.
8750 [Travis Vitek <vitek@roguewave.com>]
8752 *) Add support for shared libraries under Irix.
8753 [Albert Chin-A-Young <china@thewrittenword.com>]
8755 *) Add configuration option to build on Linux on both big-endian and
8757 [Ralf Baechle <ralf@uni-koblenz.de>]
8759 *) Add the possibility to create shared libraries on HP-UX.
8762 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8764 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8765 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8766 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8767 PRNG state recovery was possible based on the output of
8768 one PRNG request appropriately sized to gain knowledge on
8769 'md' followed by enough consecutive 1-byte PRNG requests
8770 to traverse all of 'state'.
8772 1. When updating 'md_local' (the current thread's copy of 'md')
8773 during PRNG output generation, hash all of the previous
8774 'md_local' value, not just the half used for PRNG output.
8776 2. Make the number of bytes from 'state' included into the hash
8777 independent from the number of PRNG bytes requested.
8779 The first measure alone would be sufficient to avoid
8780 Markku-Juhani's attack. (Actually it had never occurred
8781 to me that the half of 'md_local' used for chaining was the
8782 half from which PRNG output bytes were taken -- I had always
8783 assumed that the secret half would be used.) The second
8784 measure makes sure that additional data from 'state' is never
8785 mixed into 'md_local' in small portions; this heuristically
8786 further strengthens the PRNG.
8789 *) Fix crypto/bn/asm/mips3.s.
8792 *) When only the key is given to "enc", the IV is undefined. Print out
8793 an error message in this case.
8796 *) Handle special case when X509_NAME is empty in X509 printing routines.
8799 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8800 positive and less than q.
8803 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8804 used: it isn't thread safe and the add_lock_callback should handle
8806 [Paul Rose <Paul.Rose@bridge.com>]
8808 *) Verify that incoming data obeys the block size in
8809 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8813 [Ulf Möller, Bodo Möller]
8815 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8816 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8817 when fixing the server behaviour for backwards-compatible 'client
8818 hello' messages. (Note that the attack is impractical against
8819 SSL 3.0 and TLS 1.0 anyway because length and version checking
8820 means that the probability of guessing a valid ciphertext is
8821 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8824 Before 0.9.5, the countermeasure (hide the error by generating a
8825 random 'decryption result') did not work properly because
8826 ERR_clear_error() was missing, meaning that SSL_get_error() would
8827 detect the supposedly ignored error.
8829 Both problems are now fixed.
8832 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8833 (previously it was 1024).
8836 *) Fix for compatibility mode trust settings: ignore trust settings
8837 unless some valid trust or reject settings are present.
8840 *) Fix for blowfish EVP: its a variable length cipher.
8843 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8844 parameters in DSA public key structures and return an error in the
8845 DSA routines if parameters are absent.
8848 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8849 in the current directory if neither $RANDFILE nor $HOME was set.
8850 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8851 caused some confusion to Windows users who haven't defined $HOME.
8852 Thus RAND_file_name() is changed again: e_os.h can define a
8853 DEFAULT_HOME, which will be used if $HOME is not set.
8854 For Windows, we use "C:"; on other platforms, we still require
8855 environment variables.
8857 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8858 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8859 having multiple threads call RAND_poll() concurrently.
8862 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8863 combination of a flag and a thread ID variable.
8864 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8865 flag), *other* threads can enter ssleay_add_bytes without obeying
8866 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8867 that they do not hold after the first thread unsets add_do_not_lock).
8870 *) Change bctest again: '-x' expressions are not available in all
8874 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8876 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8877 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8879 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8880 the default extension for executables, if any. Also, make the perl
8881 scripts that use symlink() to test if it really exists and use "cp"
8882 if it doesn't. All this made OpenSSL compilable and installable in
8886 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8887 If SEQUENCE is length is indefinite just set c->slen to the total
8888 amount of data available.
8889 [Steve Henson, reported by shige@FreeBSD.org]
8890 [This change does not apply to 0.9.7.]
8892 *) Change bctest to avoid here-documents inside command substitution
8893 (workaround for FreeBSD /bin/sh bug).
8894 For compatibility with Ultrix, avoid shell functions (introduced
8895 in the bctest version that searches along $PATH).
8898 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8899 with des_encrypt() defined on some operating systems, like Solaris
8903 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8904 On the Importance of Eliminating Errors in Cryptographic
8905 Computations, J. Cryptology 14 (2001) 2, 101-119,
8906 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8909 *) MIPS assembler BIGNUM division bug fix.
8912 *) Disabled incorrect Alpha assembler code.
8915 *) Fix PKCS#7 decode routines so they correctly update the length
8916 after reading an EOC for the EXPLICIT tag.
8918 [This change does not apply to 0.9.7.]
8920 *) Fix bug in PKCS#12 key generation routines. This was triggered
8921 if a 3DES key was generated with a 0 initial byte. Include
8922 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8923 (but broken) behaviour.
8926 *) Enhance bctest to search for a working bc along $PATH and print
8928 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8930 *) Fix memory leaks in err.c: free err_data string if necessary;
8931 don't write to the wrong index in ERR_set_error_data.
8934 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8938 *) Replace rdtsc with _emit statements for VC++ version 5.
8939 [Jeremy Cooper <jeremy@baymoo.org>]
8941 *) Make it possible to reuse SSLv2 sessions.
8944 *) In copy_email() check for >= 0 as a return value for
8945 X509_NAME_get_index_by_NID() since 0 is a valid index.
8946 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8948 *) Avoid coredump with unsupported or invalid public keys by checking if
8949 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8950 PKCS7_verify() fails with non detached data.
8953 *) Don't use getenv in library functions when run as setuid/setgid.
8954 New function OPENSSL_issetugid().
8957 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8958 due to incorrect handling of multi-threading:
8960 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8962 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8964 3. Count how many times MemCheck_off() has been called so that
8965 nested use can be treated correctly. This also avoids
8966 inband-signalling in the previous code (which relied on the
8967 assumption that thread ID 0 is impossible).
8970 *) Add "-rand" option also to s_client and s_server.
8973 *) Fix CPU detection on Irix 6.x.
8974 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8975 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8977 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8980 [This change does not apply to 0.9.7.]
8982 *) Use the cached encoding of an X509_NAME structure rather than
8983 copying it. This is apparently the reason for the libsafe "errors"
8984 but the code is actually correct.
8987 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8988 Bleichenbacher's DSA attack.
8989 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8990 to be set and top=0 forces the highest bit to be set; top=-1 is new
8991 and leaves the highest bit random.
8992 [Ulf Moeller, Bodo Moeller]
8994 *) In the NCONF_...-based implementations for CONF_... queries
8995 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8996 a temporary CONF structure with the data component set to NULL
8997 (which gives segmentation faults in lh_retrieve).
8998 Instead, use NULL for the CONF pointer in CONF_get_string and
8999 CONF_get_number (which may use environment variables) and directly
9000 return NULL from CONF_get_section.
9003 *) Fix potential buffer overrun for EBCDIC.
9006 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9007 keyUsage if basicConstraints absent for a CA.
9010 *) Make SMIME_write_PKCS7() write mail header values with a format that
9011 is more generally accepted (no spaces before the semicolon), since
9012 some programs can't parse those values properly otherwise. Also make
9013 sure BIO's that break lines after each write do not create invalid
9017 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9018 macros previously used would not encode an empty SEQUENCE OF
9019 and break the signature.
9021 [This change does not apply to 0.9.7.]
9023 *) Zero the premaster secret after deriving the master secret in
9027 *) Add some EVP_add_digest_alias registrations (as found in
9028 OpenSSL_add_all_digests()) to SSL_library_init()
9029 aka OpenSSL_add_ssl_algorithms(). This provides improved
9030 compatibility with peers using X.509 certificates
9031 with unconventional AlgorithmIdentifier OIDs.
9034 *) Fix for Irix with NO_ASM.
9035 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9037 *) ./config script fixes.
9038 [Ulf Moeller, Richard Levitte]
9040 *) Fix 'openssl passwd -1'.
9043 *) Change PKCS12_key_gen_asc() so it can cope with non null
9044 terminated strings whose length is passed in the passlen
9045 parameter, for example from PEM callbacks. This was done
9046 by adding an extra length parameter to asc2uni().
9047 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9049 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9050 call failed, free the DSA structure.
9053 *) Fix to uni2asc() to cope with zero length Unicode strings.
9054 These are present in some PKCS#12 files.
9057 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9058 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9059 when writing a 32767 byte record.
9060 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9062 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9063 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9065 (RSA objects have a reference count access to which is protected
9066 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9067 so they are meant to be shared between threads.)
9068 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9069 "Reddie, Steven" <Steven.Reddie@ca.com>]
9071 *) Fix a deadlock in CRYPTO_mem_leaks().
9074 *) Use better test patterns in bntest.
9077 *) rand_win.c fix for Borland C.
9080 *) BN_rshift bugfix for n == 0.
9083 *) Add a 'bctest' script that checks for some known 'bc' bugs
9084 so that 'make test' does not abort just because 'bc' is broken.
9087 *) Store verify_result within SSL_SESSION also for client side to
9088 avoid potential security hole. (Re-used sessions on the client side
9089 always resulted in verify_result==X509_V_OK, not using the original
9090 result of the server certificate verification.)
9093 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9094 SSL3_RT_APPLICATION_DATA, return 0.
9095 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9099 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9100 releases, have been re-implemented by renaming the previous
9101 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9102 and ssl3_read_internal, respectively, and adding 'peek' parameters
9103 to them. The new ssl[23]_{read,peek} functions are calls to
9104 ssl[23]_read_internal with the 'peek' flag set appropriately.
9105 A 'peek' parameter has also been added to ssl3_read_bytes, which
9106 does the actual work for ssl3_read_internal.
9109 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9110 the method-specific "init()" handler. Also clean up ex_data after
9111 calling the method-specific "finish()" handler. Previously, this was
9112 happening the other way round.
9115 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9116 The previous value, 12, was not always sufficient for BN_mod_exp().
9119 *) Make sure that shared libraries get the internal name engine with
9120 the full version number and not just 0. This should mark the
9121 shared libraries as not backward compatible. Of course, this should
9122 be changed again when we can guarantee backward binary compatibility.
9125 *) Fix typo in get_cert_by_subject() in by_dir.c
9126 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9128 *) Rework the system to generate shared libraries:
9130 - Make note of the expected extension for the shared libraries and
9131 if there is a need for symbolic links from for example libcrypto.so.0
9132 to libcrypto.so.0.9.7. There is extended info in Configure for
9135 - Make as few rebuilds of the shared libraries as possible.
9137 - Still avoid linking the OpenSSL programs with the shared libraries.
9139 - When installing, install the shared libraries separately from the
9143 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9145 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9146 and not in SSL_clear because the latter is also used by the
9147 accept/connect functions; previously, the settings made by
9148 SSL_set_read_ahead would be lost during the handshake.
9149 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9151 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9152 Previously, it would create entries for disabled algorithms no
9156 *) Added several new manual pages for SSL_* function.
9159 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9161 *) In ssl23_get_client_hello, generate an error message when faced
9162 with an initial SSL 3.0/TLS record that is too small to contain the
9163 first two bytes of the ClientHello message, i.e. client_version.
9164 (Note that this is a pathologic case that probably has never happened
9165 in real life.) The previous approach was to use the version number
9166 from the record header as a substitute; but our protocol choice
9167 should not depend on that one because it is not authenticated
9168 by the Finished messages.
9171 *) More robust randomness gathering functions for Windows.
9172 [Jeffrey Altman <jaltman@columbia.edu>]
9174 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9175 not set then we don't setup the error code for issuer check errors
9176 to avoid possibly overwriting other errors which the callback does
9177 handle. If an application does set the flag then we assume it knows
9178 what it is doing and can handle the new informational codes
9182 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9183 a general "ANY" type, as such it should be able to decode anything
9184 including tagged types. However it didn't check the class so it would
9185 wrongly interpret tagged types in the same way as their universal
9186 counterpart and unknown types were just rejected. Changed so that the
9187 tagged and unknown types are handled in the same way as a SEQUENCE:
9188 that is the encoding is stored intact. There is also a new type
9189 "V_ASN1_OTHER" which is used when the class is not universal, in this
9190 case we have no idea what the actual type is so we just lump them all
9194 *) On VMS, stdout may very well lead to a file that is written to
9195 in a record-oriented fashion. That means that every write() will
9196 write a separate record, which will be read separately by the
9197 programs trying to read from it. This can be very confusing.
9199 The solution is to put a BIO filter in the way that will buffer
9200 text until a linefeed is reached, and then write everything a
9201 line at a time, so every record written will be an actual line,
9202 not chunks of lines and not (usually doesn't happen, but I've
9203 seen it once) several lines in one record. BIO_f_linebuffer() is
9206 Currently, it's a VMS-only method, because that's where it has
9207 been tested well enough.
9210 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9211 it can return incorrect results.
9212 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9213 but it was in 0.9.6-beta[12].)
9216 *) Disable the check for content being present when verifying detached
9217 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9218 include zero length content when signing messages.
9221 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9222 BIO_ctrl (for BIO pairs).
9225 *) Add DSO method for VMS.
9228 *) Bug fix: Montgomery multiplication could produce results with the
9232 *) Add RPM specification openssl.spec and modify it to build three
9233 packages. The default package contains applications, application
9234 documentation and run-time libraries. The devel package contains
9235 include files, static libraries and function documentation. The
9236 doc package contains the contents of the doc directory. The original
9237 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9240 *) Add a large number of documentation files for many SSL routines.
9241 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9243 *) Add a configuration entry for Sony News 4.
9244 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9246 *) Don't set the two most significant bits to one when generating a
9247 random number < q in the DSA library.
9250 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9251 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9252 the underlying transport is blocking) if a handshake took place.
9253 (The default behaviour is needed by applications such as s_client
9254 and s_server that use select() to determine when to use SSL_read;
9255 but for applications that know in advance when to expect data, it
9256 just makes things more complicated.)
9259 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9263 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9264 work better on such systems.
9265 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9267 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9268 Update PKCS12_parse() so it copies the friendlyName and the
9269 keyid to the certificates aux info.
9272 *) Fix bug in PKCS7_verify() which caused an infinite loop
9273 if there was more than one signature.
9274 [Sven Uszpelkat <su@celocom.de>]
9276 *) Major change in util/mkdef.pl to include extra information
9277 about each symbol, as well as presenting variables as well
9278 as functions. This change means that there's n more need
9279 to rebuild the .num files when some algorithms are excluded.
9282 *) Allow the verify time to be set by an application,
9283 rather than always using the current time.
9286 *) Phase 2 verify code reorganisation. The certificate
9287 verify code now looks up an issuer certificate by a
9288 number of criteria: subject name, authority key id
9289 and key usage. It also verifies self signed certificates
9290 by the same criteria. The main comparison function is
9291 X509_check_issued() which performs these checks.
9293 Lot of changes were necessary in order to support this
9294 without completely rewriting the lookup code.
9296 Authority and subject key identifier are now cached.
9298 The LHASH 'certs' is X509_STORE has now been replaced
9299 by a STACK_OF(X509_OBJECT). This is mainly because an
9300 LHASH can't store or retrieve multiple objects with
9301 the same hash value.
9303 As a result various functions (which were all internal
9304 use only) have changed to handle the new X509_STORE
9305 structure. This will break anything that messed round
9306 with X509_STORE internally.
9308 The functions X509_STORE_add_cert() now checks for an
9309 exact match, rather than just subject name.
9311 The X509_STORE API doesn't directly support the retrieval
9312 of multiple certificates matching a given criteria, however
9313 this can be worked round by performing a lookup first
9314 (which will fill the cache with candidate certificates)
9315 and then examining the cache for matches. This is probably
9316 the best we can do without throwing out X509_LOOKUP
9317 entirely (maybe later...).
9319 The X509_VERIFY_CTX structure has been enhanced considerably.
9321 All certificate lookup operations now go via a get_issuer()
9322 callback. Although this currently uses an X509_STORE it
9323 can be replaced by custom lookups. This is a simple way
9324 to bypass the X509_STORE hackery necessary to make this
9325 work and makes it possible to use more efficient techniques
9326 in future. A very simple version which uses a simple
9327 STACK for its trusted certificate store is also provided
9328 using X509_STORE_CTX_trusted_stack().
9330 The verify_cb() and verify() callbacks now have equivalents
9331 in the X509_STORE_CTX structure.
9333 X509_STORE_CTX also has a 'flags' field which can be used
9334 to customise the verify behaviour.
9337 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9338 excludes S/MIME capabilities.
9341 *) When a certificate request is read in keep a copy of the
9342 original encoding of the signed data and use it when outputting
9343 again. Signatures then use the original encoding rather than
9344 a decoded, encoded version which may cause problems if the
9345 request is improperly encoded.
9348 *) For consistency with other BIO_puts implementations, call
9349 buffer_write(b, ...) directly in buffer_puts instead of calling
9352 In BIO_puts, increment b->num_write as in BIO_write.
9353 [Peter.Sylvester@EdelWeb.fr]
9355 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9356 BN_zero, we may not return a BIGNUM with an array consisting of
9360 *) Avoid calling abort() from within the library when problems are
9361 detected, except if preprocessor symbols have been defined
9362 (such as REF_CHECK, BN_DEBUG etc.).
9365 *) New openssl application 'rsautl'. This utility can be
9366 used for low level RSA operations. DER public key
9367 BIO/fp routines also added.
9370 *) New Configure entry and patches for compiling on QNX 4.
9371 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9373 *) A demo state-machine implementation was sponsored by
9374 Nuron (http://www.nuron.com/) and is now available in
9375 demos/state_machine.
9378 *) New options added to the 'dgst' utility for signature
9379 generation and verification.
9382 *) Unrecognized PKCS#7 content types are now handled via a
9383 catch all ASN1_TYPE structure. This allows unsupported
9384 types to be stored as a "blob" and an application can
9385 encode and decode it manually.
9388 *) Fix various signed/unsigned issues to make a_strex.c
9390 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9392 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9393 length if passed a buffer. ASN1_INTEGER_to_BN failed
9394 if passed a NULL BN and its argument was negative.
9395 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9397 *) Modification to PKCS#7 encoding routines to output definite
9398 length encoding. Since currently the whole structures are in
9399 memory there's not real point in using indefinite length
9400 constructed encoding. However if OpenSSL is compiled with
9401 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9404 *) Added BIO_vprintf() and BIO_vsnprintf().
9407 *) Added more prefixes to parse for in the the strings written
9408 through a logging bio, to cover all the levels that are available
9409 through syslog. The prefixes are now:
9411 PANIC, EMERG, EMR => LOG_EMERG
9412 ALERT, ALR => LOG_ALERT
9413 CRIT, CRI => LOG_CRIT
9414 ERROR, ERR => LOG_ERR
9415 WARNING, WARN, WAR => LOG_WARNING
9416 NOTICE, NOTE, NOT => LOG_NOTICE
9417 INFO, INF => LOG_INFO
9418 DEBUG, DBG => LOG_DEBUG
9420 and as before, if none of those prefixes are present at the
9421 beginning of the string, LOG_ERR is chosen.
9423 On Win32, the LOG_* levels are mapped according to this:
9425 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9426 LOG_WARNING => EVENTLOG_WARNING_TYPE
9427 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9431 *) Made it possible to reconfigure with just the configuration
9432 argument "reconf" or "reconfigure". The command line arguments
9433 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9434 and are retrieved from there when reconfiguring.
9438 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9440 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9443 *) The obj_dat.pl script was messing up the sorting of object
9444 names. The reason was that it compared the quoted version
9445 of strings as a result "OCSP" > "OCSP Signing" because
9446 " > SPACE. Changed script to store unquoted versions of
9447 names and add quotes on output. It was also omitting some
9448 names from the lookup table if they were given a default
9449 value (that is if SN is missing it is given the same
9450 value as LN and vice versa), these are now added on the
9451 grounds that if an object has a name we should be able to
9452 look it up. Finally added warning output when duplicate
9453 short or long names are found.
9456 *) Changes needed for Tandem NSK.
9457 [Scott Uroff <scott@xypro.com>]
9459 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9460 RSA_padding_check_SSLv23(), special padding was never detected
9461 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9462 version rollback attacks was not effective.
9464 In s23_clnt.c, don't use special rollback-attack detection padding
9465 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9466 client; similarly, in s23_srvr.c, don't do the rollback check if
9467 SSL 2.0 is the only protocol enabled in the server.
9470 *) Make it possible to get hexdumps of unprintable data with 'openssl
9471 asn1parse'. By implication, the functions ASN1_parse_dump() and
9472 BIO_dump_indent() are added.
9475 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9476 these print out strings and name structures based on various
9477 flags including RFC2253 support and proper handling of
9478 multibyte characters. Added options to the 'x509' utility
9479 to allow the various flags to be set.
9482 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9483 Also change the functions X509_cmp_current_time() and
9484 X509_gmtime_adj() work with an ASN1_TIME structure,
9485 this will enable certificates using GeneralizedTime in validity
9486 dates to be checked.
9489 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9490 negative public key encodings) on by default,
9491 NO_NEG_PUBKEY_BUG can be set to disable it.
9494 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9495 content octets. An i2c_ASN1_OBJECT is unnecessary because
9496 the encoding can be trivially obtained from the structure.
9499 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9500 not read locks (CRYPTO_r_[un]lock).
9503 *) A first attempt at creating official support for shared
9504 libraries through configuration. I've kept it so the
9505 default is static libraries only, and the OpenSSL programs
9506 are always statically linked for now, but there are
9507 preparations for dynamic linking in place.
9508 This has been tested on Linux and Tru64.
9511 *) Randomness polling function for Win9x, as described in:
9512 Peter Gutmann, Software Generation of Practically Strong
9516 *) Fix so PRNG is seeded in req if using an already existing
9520 *) New options to smime application. -inform and -outform
9521 allow alternative formats for the S/MIME message including
9522 PEM and DER. The -content option allows the content to be
9523 specified separately. This should allow things like Netscape
9524 form signing output easier to verify.
9527 *) Fix the ASN1 encoding of tags using the 'long form'.
9530 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9531 STRING types. These convert content octets to and from the
9532 underlying type. The actual tag and length octets are
9533 already assumed to have been read in and checked. These
9534 are needed because all other string types have virtually
9535 identical handling apart from the tag. By having versions
9536 of the ASN1 functions that just operate on content octets
9537 IMPLICIT tagging can be handled properly. It also allows
9538 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9539 and ASN1_INTEGER are identical apart from the tag.
9542 *) Change the handling of OID objects as follows:
9544 - New object identifiers are inserted in objects.txt, following
9545 the syntax given in objects.README.
9546 - objects.pl is used to process obj_mac.num and create a new
9548 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9551 This is currently kind of a hack, and the perl code in objects.pl
9552 isn't very elegant, but it works as I intended. The simplest way
9553 to check that it worked correctly is to look in obj_dat.h and
9554 check the array nid_objs and make sure the objects haven't moved
9555 around (this is important!). Additions are OK, as well as
9556 consistent name changes.
9559 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9562 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9563 The given file adds to whatever has already been seeded into the
9564 random pool through the RANDFILE configuration file option or
9565 environment variable, or the default random state file.
9568 *) mkstack.pl now sorts each macro group into lexical order.
9569 Previously the output order depended on the order the files
9570 appeared in the directory, resulting in needless rewriting
9574 *) Patches to make OpenSSL compile under Win32 again. Mostly
9575 work arounds for the VC++ problem that it treats func() as
9576 func(void). Also stripped out the parts of mkdef.pl that
9577 added extra typesafe functions: these no longer exist.
9580 *) Reorganisation of the stack code. The macros are now all
9581 collected in safestack.h . Each macro is defined in terms of
9582 a "stack macro" of the form SKM_<name>(type, a, b). The
9583 DEBUG_SAFESTACK is now handled in terms of function casts,
9584 this has the advantage of retaining type safety without the
9585 use of additional functions. If DEBUG_SAFESTACK is not defined
9586 then the non typesafe macros are used instead. Also modified the
9587 mkstack.pl script to handle the new form. Needs testing to see
9588 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9589 the default if no major problems. Similar behaviour for ASN1_SET_OF
9590 and PKCS12_STACK_OF.
9593 *) When some versions of IIS use the 'NET' form of private key the
9594 key derivation algorithm is different. Normally MD5(password) is
9595 used as a 128 bit RC4 key. In the modified case
9596 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9597 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9598 as the old Netscape_RSA functions except they have an additional
9599 'sgckey' parameter which uses the modified algorithm. Also added
9600 an -sgckey command line option to the rsa utility. Thanks to
9601 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9602 algorithm to openssl-dev.
9605 *) The evp_local.h macros were using 'c.##kname' which resulted in
9606 invalid expansion on some systems (SCO 5.0.5 for example).
9607 Corrected to 'c.kname'.
9608 [Phillip Porch <root@theporch.com>]
9610 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9611 a STACK of email addresses from a certificate or request, these look
9612 in the subject name and the subject alternative name extensions and
9613 omit any duplicate addresses.
9616 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9617 This makes DSA verification about 2 % faster.
9620 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9621 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9622 plus overhead for 1024 bit moduli).
9623 This makes exponentiations about 0.5 % faster for 1024 bit
9624 exponents (as measured by "openssl speed rsa2048").
9627 *) Rename memory handling macros to avoid conflicts with other
9629 Malloc => OPENSSL_malloc
9630 Malloc_locked => OPENSSL_malloc_locked
9631 Realloc => OPENSSL_realloc
9632 Free => OPENSSL_free
9635 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9636 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9639 *) CygWin32 support.
9640 [John Jarvie <jjarvie@newsguy.com>]
9642 *) The type-safe stack code has been rejigged. It is now only compiled
9643 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9644 by default all type-specific stack functions are "#define"d back to
9645 standard stack functions. This results in more streamlined output
9646 but retains the type-safety checking possibilities of the original
9650 *) The STACK code has been cleaned up, and certain type declarations
9651 that didn't make a lot of sense have been brought in line. This has
9652 also involved a cleanup of sorts in safestack.h to more correctly
9653 map type-safe stack functions onto their plain stack counterparts.
9654 This work has also resulted in a variety of "const"ifications of
9655 lots of the code, especially "_cmp" operations which should normally
9656 be prototyped with "const" parameters anyway.
9659 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9660 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9661 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9662 where all of 'md' is used each time the PRNG is used, but 'state'
9663 is used only indexed by a cyclic counter. As entropy may not be
9664 well distributed from the beginning, 'md' is important as a
9665 chaining variable. However, the output function chains only half
9666 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9667 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9668 in all of 'state' being rewritten, with the new values depending
9669 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9672 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9673 the handshake is continued after ssl_verify_cert_chain();
9674 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9675 can lead to 'unexplainable' connection aborts later.
9676 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9678 *) Major EVP API cipher revision.
9679 Add hooks for extra EVP features. This allows various cipher
9680 parameters to be set in the EVP interface. Support added for variable
9681 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9682 setting of RC2 and RC5 parameters.
9684 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9687 Remove lots of duplicated code from the EVP library. For example *every*
9688 cipher init() function handles the 'iv' in the same way according to the
9689 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9690 for CFB and OFB modes they zero ctx->num.
9692 New functionality allows removal of S/MIME code RC2 hack.
9694 Most of the routines have the same form and so can be declared in terms
9697 By shifting this to the top level EVP_CipherInit() it can be removed from
9698 all individual ciphers. If the cipher wants to handle IVs or keys
9699 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9702 Change lots of functions like EVP_EncryptUpdate() to now return a
9703 value: although software versions of the algorithms cannot fail
9704 any installed hardware versions can.
9707 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9708 this option is set, tolerate broken clients that send the negotiated
9709 protocol version number instead of the requested protocol version
9713 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9714 i.e. non-zero for export ciphersuites, zero otherwise.
9715 Previous versions had this flag inverted, inconsistent with
9716 rsa_tmp_cb (..._TMP_RSA_CB).
9717 [Bodo Moeller; problem reported by Amit Chopra]
9719 *) Add missing DSA library text string. Work around for some IIS
9720 key files with invalid SEQUENCE encoding.
9723 *) Add a document (doc/standards.txt) that list all kinds of standards
9724 and so on that are implemented in OpenSSL.
9727 *) Enhance c_rehash script. Old version would mishandle certificates
9728 with the same subject name hash and wouldn't handle CRLs at all.
9729 Added -fingerprint option to crl utility, to support new c_rehash
9733 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9736 *) Fix for SSL server purpose checking. Server checking was
9737 rejecting certificates which had extended key usage present
9738 but no ssl client purpose.
9739 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9741 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9742 is a little unclear about how a blank password is handled.
9743 Since the password in encoded as a BMPString with terminating
9744 double NULL a zero length password would end up as just the
9745 double NULL. However no password at all is different and is
9746 handled differently in the PKCS#12 key generation code. NS
9747 treats a blank password as zero length. MSIE treats it as no
9748 password on export: but it will try both on import. We now do
9749 the same: PKCS12_parse() tries zero length and no password if
9750 the password is set to "" or NULL (NULL is now a valid password:
9751 it wasn't before) as does the pkcs12 application.
9754 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9755 perror when PEM_read_bio_X509_REQ fails, the error message must
9756 be obtained from the error queue.
9759 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9760 it in ERR_remove_state if appropriate, and change ERR_get_state
9761 accordingly to avoid race conditions (this is necessary because
9762 thread_hash is no longer constant once set).
9765 *) Bugfix for linux-elf makefile.one.
9768 *) RSA_get_default_method() will now cause a default
9769 RSA_METHOD to be chosen if one doesn't exist already.
9770 Previously this was only set during a call to RSA_new()
9771 or RSA_new_method(NULL) meaning it was possible for
9772 RSA_get_default_method() to return NULL.
9775 *) Added native name translation to the existing DSO code
9776 that will convert (if the flag to do so is set) filenames
9777 that are sufficiently small and have no path information
9778 into a canonical native form. Eg. "blah" converted to
9779 "libblah.so" or "blah.dll" etc.
9782 *) New function ERR_error_string_n(e, buf, len) which is like
9783 ERR_error_string(e, buf), but writes at most 'len' bytes
9784 including the 0 terminator. For ERR_error_string_n, 'buf'
9786 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9788 *) CONF library reworked to become more general. A new CONF
9789 configuration file reader "class" is implemented as well as a
9790 new functions (NCONF_*, for "New CONF") to handle it. The now
9791 old CONF_* functions are still there, but are reimplemented to
9792 work in terms of the new functions. Also, a set of functions
9793 to handle the internal storage of the configuration data is
9794 provided to make it easier to write new configuration file
9795 reader "classes" (I can definitely see something reading a
9796 configuration file in XML format, for example), called _CONF_*,
9797 or "the configuration storage API"...
9799 The new configuration file reading functions are:
9801 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9802 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9804 NCONF_default, NCONF_WIN32
9806 NCONF_dump_fp, NCONF_dump_bio
9808 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9809 NCONF_new creates a new CONF object. This works in the same way
9810 as other interfaces in OpenSSL, like the BIO interface.
9811 NCONF_dump_* dump the internal storage of the configuration file,
9812 which is useful for debugging. All other functions take the same
9813 arguments as the old CONF_* functions wth the exception of the
9814 first that must be a `CONF *' instead of a `LHASH *'.
9816 To make it easer to use the new classes with the old CONF_* functions,
9817 the function CONF_set_default_method is provided.
9820 *) Add '-tls1' option to 'openssl ciphers', which was already
9821 mentioned in the documentation but had not been implemented.
9822 (This option is not yet really useful because even the additional
9823 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9826 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9827 OpenSSL-based applications) load shared libraries and bind to
9828 them in a portable way.
9829 [Geoff Thorpe, with contributions from Richard Levitte]
9831 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9833 *) Make sure _lrotl and _lrotr are only used with MSVC.
9835 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9836 (the default implementation of RAND_status).
9838 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9839 to '-clrext' (= clear extensions), as intended and documented.
9840 [Bodo Moeller; inconsistency pointed out by Michael Attili
9843 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9844 was larger than the MD block size.
9845 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9847 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9848 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9849 using the passed key: if the passed key was a private key the result
9850 of X509_print(), for example, would be to print out all the private key
9854 *) des_quad_cksum() byte order bug fix.
9855 [Ulf Möller, using the problem description in krb4-0.9.7, where
9856 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9858 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9860 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9862 *) For easily testing in shell scripts whether some command
9863 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9864 returns with exit code 0 iff no command of the given name is available.
9865 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9866 the output goes to stdout and nothing is printed to stderr.
9867 Additional arguments are always ignored.
9869 Since for each cipher there is a command of the same name,
9870 the 'no-cipher' compilation switches can be tested this way.
9872 ('openssl no-XXX' is not able to detect pseudo-commands such
9873 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9876 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9879 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9880 is set; it will be thrown away anyway because each handshake creates
9882 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9883 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9884 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9885 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
9888 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9889 'Q' and 'R' lose their special meanings (quit/renegotiate).
9890 This is part of what -quiet does; unlike -quiet, -ign_eof
9891 does not suppress any output.
9894 *) Add compatibility options to the purpose and trust code. The
9895 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9896 accepts a certificate or CA, this was the previous behaviour,
9897 with all the associated security issues.
9899 X509_TRUST_COMPAT is the old trust behaviour: only and
9900 automatically trust self signed roots in certificate store. A
9901 new trust setting X509_TRUST_DEFAULT is used to specify that
9902 a purpose has no associated trust setting and it should instead
9903 use the value in the default purpose.
9906 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9907 and fix a memory leak.
9910 *) In util/mkerr.pl (which implements 'make errors'), preserve
9911 reason strings from the previous version of the .c file, as
9912 the default to have only downcase letters (and digits) in
9913 automatically generated reasons codes is not always appropriate.
9916 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9917 using strerror. Previously, ERR_reason_error_string() returned
9918 library names as reason strings for SYSerr; but SYSerr is a special
9919 case where small numbers are errno values, not library numbers.
9922 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9923 converts DSA parameters into DH parameters. (When creating parameters,
9924 DSA_generate_parameters is used.)
9927 *) Include 'length' (recommended exponent length) in C code generated
9928 by 'openssl dhparam -C'.
9931 *) The second argument to set_label in perlasm was already being used
9932 so couldn't be used as a "file scope" flag. Moved to third argument
9936 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9937 instead of RAND_bytes for encryption IVs and salts.
9940 *) Include RAND_status() into RAND_METHOD instead of implementing
9941 it only for md_rand.c Otherwise replacing the PRNG by calling
9942 RAND_set_rand_method would be impossible.
9945 *) Don't let DSA_generate_key() enter an infinite loop if the random
9946 number generation fails.
9949 *) New 'rand' application for creating pseudo-random output.
9952 *) Added configuration support for Linux/IA64
9953 [Rolf Haberrecker <rolf@suse.de>]
9955 *) Assembler module support for Mingw32.
9958 *) Shared library support for HPUX (in shlib/).
9959 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9961 *) Shared library support for Solaris gcc.
9962 [Lutz Behnke <behnke@trustcenter.de>]
9964 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9966 *) PKCS7_encrypt() was adding text MIME headers twice because they
9967 were added manually and by SMIME_crlf_copy().
9970 *) In bntest.c don't call BN_rand with zero bits argument.
9971 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9973 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9974 case was implemented. This caused BN_div_recp() to fail occasionally.
9977 *) Add an optional second argument to the set_label() in the perl
9978 assembly language builder. If this argument exists and is set
9979 to 1 it signals that the assembler should use a symbol whose
9980 scope is the entire file, not just the current function. This
9981 is needed with MASM which uses the format label:: for this scope.
9982 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9984 *) Change the ASN1 types so they are typedefs by default. Before
9985 almost all types were #define'd to ASN1_STRING which was causing
9986 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9990 *) Change names of new functions to the new get1/get0 naming
9991 convention: After 'get1', the caller owns a reference count
9992 and has to call ..._free; 'get0' returns a pointer to some
9993 data structure without incrementing reference counters.
9994 (Some of the existing 'get' functions increment a reference
9995 counter, some don't.)
9996 Similarly, 'set1' and 'add1' functions increase reference
9997 counters or duplicate objects.
10000 *) Allow for the possibility of temp RSA key generation failure:
10001 the code used to assume it always worked and crashed on failure.
10004 *) Fix potential buffer overrun problem in BIO_printf().
10005 [Ulf Möller, using public domain code by Patrick Powell; problem
10006 pointed out by David Sacerdote <das33@cornell.edu>]
10008 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10009 RAND_egd() and RAND_status(). In the command line application,
10010 the EGD socket can be specified like a seed file using RANDFILE
10014 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10015 Some CAs (e.g. Verisign) distribute certificates in this form.
10018 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10019 list to exclude them. This means that no special compilation option
10020 is needed to use anonymous DH: it just needs to be included in the
10024 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10025 EVP_MD_type. The old functionality is available in a new macro called
10026 EVP_MD_md(). Change code that uses it and update docs.
10029 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10030 where the 'void *' argument is replaced by a function pointer argument.
10031 Previously 'void *' was abused to point to functions, which works on
10032 many platforms, but is not correct. As these functions are usually
10033 called by macros defined in OpenSSL header files, most source code
10034 should work without changes.
10037 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10038 sections with information on -D... compiler switches used for
10039 compiling the library so that applications can see them. To enable
10040 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10041 must be defined. E.g.,
10042 #define OPENSSL_ALGORITHM_DEFINES
10043 #include <openssl/opensslconf.h>
10044 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10045 [Richard Levitte, Ulf and Bodo Möller]
10047 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10051 *) Change the 'other' type in certificate aux info to a STACK_OF
10052 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10053 the required ASN1 format: arbitrary types determined by an OID.
10056 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10057 argument to 'req'. This is not because the function is newer or
10058 better than others it just uses the work 'NEW' in the certificate
10059 request header lines. Some software needs this.
10062 *) Reorganise password command line arguments: now passwords can be
10063 obtained from various sources. Delete the PEM_cb function and make
10064 it the default behaviour: i.e. if the callback is NULL and the
10065 usrdata argument is not NULL interpret it as a null terminated pass
10066 phrase. If usrdata and the callback are NULL then the pass phrase
10067 is prompted for as usual.
10070 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10071 the support is automatically enabled. The resulting binaries will
10072 autodetect the card and use it if present.
10073 [Ben Laurie and Compaq Inc.]
10075 *) Work around for Netscape hang bug. This sends certificate request
10076 and server done in one record. Since this is perfectly legal in the
10077 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10078 the bugs/SSLv3 entry for more info.
10081 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10084 *) Add -rand argument to smime and pkcs12 applications and read/write
10088 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10091 *) Add command line password options to the remaining applications.
10094 *) Bug fix for BN_div_recp() for numerators with an even number of
10098 *) More tests in bntest.c, and changed test_bn output.
10101 *) ./config recognizes MacOS X now.
10104 *) Bug fix for BN_div() when the first words of num and divisor are
10105 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10108 *) Add support for various broken PKCS#8 formats, and command line
10109 options to produce them.
10112 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10113 get temporary BIGNUMs from a BN_CTX.
10116 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10120 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10121 include a #define from the old name to the new. The original intent
10122 was that statically linked binaries could for example just call
10123 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10124 link with digests. This never worked because SSLeay_add_all_digests()
10125 and SSLeay_add_all_ciphers() were in the same source file so calling
10126 one would link with the other. They are now in separate source files.
10129 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10132 *) Use a less unusual form of the Miller-Rabin primality test (it used
10133 a binary algorithm for exponentiation integrated into the Miller-Rabin
10134 loop, our standard modexp algorithms are faster).
10137 *) Support for the EBCDIC character set completed.
10138 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10140 *) Source code cleanups: use const where appropriate, eliminate casts,
10141 use void * instead of char * in lhash.
10144 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10145 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10146 this the server could overwrite ephemeral keys that the client
10150 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10151 using 50 iterations of the Rabin-Miller test.
10153 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10154 iterations of the Rabin-Miller test as required by the appendix
10155 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10156 As BN_is_prime_fasttest includes trial division, DSA parameter
10157 generation becomes much faster.
10159 This implies a change for the callback functions in DSA_is_prime
10160 and DSA_generate_parameters: The callback function is called once
10161 for each positive witness in the Rabin-Miller test, not just
10162 occasionally in the inner loop; and the parameters to the
10163 callback function now provide an iteration count for the outer
10164 loop rather than for the current invocation of the inner loop.
10165 DSA_generate_parameters additionally can call the callback
10166 function with an 'iteration count' of -1, meaning that a
10167 candidate has passed the trial division test (when q is generated
10168 from an application-provided seed, trial division is skipped).
10171 *) New function BN_is_prime_fasttest that optionally does trial
10172 division before starting the Rabin-Miller test and has
10173 an additional BN_CTX * argument (whereas BN_is_prime always
10174 has to allocate at least one BN_CTX).
10175 'callback(1, -1, cb_arg)' is called when a number has passed the
10176 trial division stage.
10179 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10183 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10186 *) New function BN_pseudo_rand().
10189 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10190 bignum version of BN_from_montgomery() with the working code from
10191 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10195 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10196 made it impossible to use the same SSL_SESSION data structure in
10197 SSL2 clients in multiple threads.
10200 *) The return value of RAND_load_file() no longer counts bytes obtained
10201 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10202 to seed the PRNG (previously an explicit byte count was required).
10203 [Ulf Möller, Bodo Möller]
10205 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10206 used (char *) instead of (void *) and had casts all over the place.
10209 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10212 *) Retain source code compatibility for BN_prime_checks macro:
10213 BN_is_prime(..., BN_prime_checks, ...) now uses
10214 BN_prime_checks_for_size to determine the appropriate number of
10215 Rabin-Miller iterations.
10218 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10219 DH_CHECK_P_NOT_SAFE_PRIME.
10220 (Check if this is true? OpenPGP calls them "strong".)
10223 *) Merge the functionality of "dh" and "gendh" programs into a new program
10224 "dhparam". The old programs are retained for now but will handle DH keys
10225 (instead of parameters) in future.
10228 *) Make the ciphers, s_server and s_client programs check the return values
10229 when a new cipher list is set.
10232 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10233 ciphers. Before when the 56bit ciphers were enabled the sorting was
10236 The syntax for the cipher sorting has been extended to support sorting by
10237 cipher-strength (using the strength_bits hard coded in the tables).
10238 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10240 Fix a bug in the cipher-command parser: when supplying a cipher command
10241 string with an "undefined" symbol (neither command nor alphanumeric
10242 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10243 an error is flagged.
10245 Due to the strength-sorting extension, the code of the
10246 ssl_create_cipher_list() function was completely rearranged. I hope that
10247 the readability was also increased :-)
10248 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10250 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10251 for the first serial number and places 2 in the serial number file. This
10252 avoids problems when the root CA is created with serial number zero and
10253 the first user certificate has the same issuer name and serial number
10257 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10258 the new code. Add documentation for this stuff.
10261 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10262 X509_*() to X509at_*() on the grounds that they don't handle X509
10263 structures and behave in an analogous way to the X509v3 functions:
10264 they shouldn't be called directly but wrapper functions should be used
10267 So we also now have some wrapper functions that call the X509at functions
10268 when passed certificate requests. (TO DO: similar things can be done with
10269 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10270 things. Some of these need some d2i or i2d and print functionality
10271 because they handle more complex structures.)
10274 *) Add missing #ifndefs that caused missing symbols when building libssl
10275 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10276 NO_RSA in ssl/s2*.c.
10277 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10279 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10280 has a return value which indicates the quality of the random data
10281 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10282 error queue. New function RAND_pseudo_bytes() generates output that is
10283 guaranteed to be unique but not unpredictable. RAND_add is like
10284 RAND_seed, but takes an extra argument for an entropy estimate
10285 (RAND_seed always assumes full entropy).
10288 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10289 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10290 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10291 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10292 false-positive rate of at most 2^-80 for random input.
10295 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10298 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10299 in the 0.9.5 release), this returns the chain
10300 from an X509_CTX structure with a dup of the stack and all
10301 the X509 reference counts upped: so the stack will exist
10302 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10305 Also make SSL_SESSION_print() print out the verify return
10309 *) Add manpage for the pkcs12 command. Also change the default
10310 behaviour so MAC iteration counts are used unless the new
10311 -nomaciter option is used. This improves file security and
10312 only older versions of MSIE (4.0 for example) need it.
10315 *) Honor the no-xxx Configure options when creating .DEF files.
10318 *) Add PKCS#10 attributes to field table: challengePassword,
10319 unstructuredName and unstructuredAddress. These are taken from
10320 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10321 international characters are used.
10323 More changes to X509_ATTRIBUTE code: allow the setting of types
10324 based on strings. Remove the 'loc' parameter when adding
10325 attributes because these will be a SET OF encoding which is sorted
10329 *) Initial changes to the 'req' utility to allow request generation
10330 automation. This will allow an application to just generate a template
10331 file containing all the field values and have req construct the
10334 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10335 used all over the place including certificate requests and PKCS#7
10336 structures. They are currently handled manually where necessary with
10337 some primitive wrappers for PKCS#7. The new functions behave in a
10338 manner analogous to the X509 extension functions: they allow
10339 attributes to be looked up by NID and added.
10341 Later something similar to the X509V3 code would be desirable to
10342 automatically handle the encoding, decoding and printing of the
10343 more complex types. The string types like challengePassword can
10344 be handled by the string table functions.
10346 Also modified the multi byte string table handling. Now there is
10347 a 'global mask' which masks out certain types. The table itself
10348 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10349 is useful when for example there is only one permissible type
10350 (as in countryName) and using the mask might result in no valid
10354 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10355 SSL_get_peer_finished to allow applications to obtain the latest
10356 Finished messages sent to the peer or expected from the peer,
10357 respectively. (SSL_get_peer_finished is usually the Finished message
10358 actually received from the peer, otherwise the protocol will be aborted.)
10360 As the Finished message are message digests of the complete handshake
10361 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10362 be used for external authentication procedures when the authentication
10363 provided by SSL/TLS is not desired or is not enough.
10366 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10367 the host supports BWX extension and if Compaq C is present on the
10368 $PATH. Just exploiting of the BWX extension results in 20-30%
10369 performance kick for some algorithms, e.g. DES and RC4 to mention
10370 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10374 *) Add support for MS "fast SGC". This is arguably a violation of the
10375 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10376 weak crypto and after checking the certificate is SGC a second one
10377 with strong crypto. MS SGC stops the first handshake after receiving
10378 the server certificate message and sends a second client hello. Since
10379 a server will typically do all the time consuming operations before
10380 expecting any further messages from the client (server key exchange
10381 is the most expensive) there is little difference between the two.
10383 To get OpenSSL to support MS SGC we have to permit a second client
10384 hello message after we have sent server done. In addition we have to
10385 reset the MAC if we do get this second client hello.
10388 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10389 if a DER encoded private key is RSA or DSA traditional format. Changed
10390 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10391 format DER encoded private key. Newer code should use PKCS#8 format which
10392 has the key type encoded in the ASN1 structure. Added DER private key
10393 support to pkcs8 application.
10396 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10397 ciphersuites has been selected (as required by the SSL 3/TLS 1
10398 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10399 is set, we interpret this as a request to violate the specification
10400 (the worst that can happen is a handshake failure, and 'correct'
10401 behaviour would result in a handshake failure anyway).
10404 *) In SSL_CTX_add_session, take into account that there might be multiple
10405 SSL_SESSION structures with the same session ID (e.g. when two threads
10406 concurrently obtain them from an external cache).
10407 The internal cache can handle only one SSL_SESSION with a given ID,
10408 so if there's a conflict, we now throw out the old one to achieve
10412 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10413 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10414 some routines that use cipher OIDs: some ciphers do not have OIDs
10415 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10419 *) Simplify the trust setting structure and code. Now we just have
10420 two sequences of OIDs for trusted and rejected settings. These will
10421 typically have values the same as the extended key usage extension
10422 and any application specific purposes.
10424 The trust checking code now has a default behaviour: it will just
10425 check for an object with the same NID as the passed id. Functions can
10426 be provided to override either the default behaviour or the behaviour
10427 for a given id. SSL client, server and email already have functions
10428 in place for compatibility: they check the NID and also return "trusted"
10429 if the certificate is self signed.
10432 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10433 traditional format into an EVP_PKEY structure.
10436 *) Add a password callback function PEM_cb() which either prompts for
10437 a password if usr_data is NULL or otherwise assumes it is a null
10438 terminated password. Allow passwords to be passed on command line
10439 environment or config files in a few more utilities.
10442 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10443 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10444 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10445 Update documentation.
10448 *) Support for ASN1 "NULL" type. This could be handled before by using
10449 ASN1_TYPE but there wasn't any function that would try to read a NULL
10450 and produce an error if it couldn't. For compatibility we also have
10451 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10452 don't allocate anything because they don't need to.
10455 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10457 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10459 *) Rebuild of the memory allocation routines used by OpenSSL code and
10460 possibly others as well. The purpose is to make an interface that
10461 provide hooks so anyone can build a separate set of allocation and
10462 deallocation routines to be used by OpenSSL, for example memory
10463 pool implementations, or something else, which was previously hard
10464 since Malloc(), Realloc() and Free() were defined as macros having
10465 the values malloc, realloc and free, respectively (except for Win32
10466 compilations). The same is provided for memory debugging code.
10467 OpenSSL already comes with functionality to find memory leaks, but
10468 this gives people a chance to debug other memory problems.
10470 With these changes, a new set of functions and macros have appeared:
10472 CRYPTO_set_mem_debug_functions() [F]
10473 CRYPTO_get_mem_debug_functions() [F]
10474 CRYPTO_dbg_set_options() [F]
10475 CRYPTO_dbg_get_options() [F]
10476 CRYPTO_malloc_debug_init() [M]
10478 The memory debug functions are NULL by default, unless the library
10479 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10480 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10481 gives the standard debugging functions that come with OpenSSL) or
10482 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10483 provided by the library user) must be used. When the standard
10484 debugging functions are used, CRYPTO_dbg_set_options can be used to
10485 request additional information:
10486 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10487 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10489 Also, things like CRYPTO_set_mem_functions will always give the
10490 expected result (the new set of functions is used for allocation
10491 and deallocation) at all times, regardless of platform and compiler
10494 To finish it up, some functions that were never use in any other
10495 way than through macros have a new API and new semantic:
10497 CRYPTO_dbg_malloc()
10498 CRYPTO_dbg_realloc()
10501 All macros of value have retained their old syntax.
10502 [Richard Levitte and Bodo Moeller]
10504 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10505 ordering of SMIMECapabilities wasn't in "strength order" and there
10506 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10510 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10511 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10512 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10514 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10515 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10516 functionality to handle multipart/signed properly) and a utility
10517 called 'smime' to call all this stuff. This is based on code I
10518 originally wrote for Celo who have kindly allowed it to be
10519 included in OpenSSL.
10522 *) Add variants des_set_key_checked and des_set_key_unchecked of
10523 des_set_key (aka des_key_sched). Global variable des_check_key
10524 decides which of these is called by des_set_key; this way
10525 des_check_key behaves as it always did, but applications and
10526 the library itself, which was buggy for des_check_key == 1,
10527 have a cleaner way to pick the version they need.
10530 *) New function PKCS12_newpass() which changes the password of a
10534 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10535 dynamic mix. In both cases the ids can be used as an index into the
10536 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10537 functions so they accept a list of the field values and the
10538 application doesn't need to directly manipulate the X509_TRUST
10542 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10546 *) Modify the way the V3 extension code looks up extensions. This now
10547 works in a similar way to the object code: we have some "standard"
10548 extensions in a static table which is searched with OBJ_bsearch()
10549 and the application can add dynamic ones if needed. The file
10550 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10551 updated whenever a new extension is added to the core code and kept
10552 in ext_nid order. There is a simple program 'tabtest.c' which checks
10553 this. New extensions are not added too often so this file can readily
10554 be maintained manually.
10556 There are two big advantages in doing things this way. The extensions
10557 can be looked up immediately and no longer need to be "added" using
10558 X509V3_add_standard_extensions(): this function now does nothing.
10559 [Side note: I get *lots* of email saying the extension code doesn't
10560 work because people forget to call this function]
10561 Also no dynamic allocation is done unless new extensions are added:
10562 so if we don't add custom extensions there is no need to call
10563 X509V3_EXT_cleanup().
10566 *) Modify enc utility's salting as follows: make salting the default. Add a
10567 magic header, so unsalted files fail gracefully instead of just decrypting
10568 to garbage. This is because not salting is a big security hole, so people
10569 should be discouraged from doing it.
10572 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10573 digest to be passed on the command line but it only used this
10574 parameter when signing a certificate. Modified so all relevant
10575 operations are affected by the digest parameter including the
10576 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10577 DSA key was used because it didn't fix the digest.
10580 *) Initial certificate chain verify code. Currently tests the untrusted
10581 certificates for consistency with the verify purpose (which is set
10582 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10584 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10585 this is because it will reject chains with invalid extensions whereas
10586 every previous version of OpenSSL and SSLeay made no checks at all.
10588 Trust code: checks the root CA for the relevant trust settings. Trust
10589 settings have an initial value consistent with the verify purpose: e.g.
10590 if the verify purpose is for SSL client use it expects the CA to be
10591 trusted for SSL client use. However the default value can be changed to
10592 permit custom trust settings: one example of this would be to only trust
10593 certificates from a specific "secure" set of CAs.
10595 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10596 which should be used for version portability: especially since the
10597 verify structure is likely to change more often now.
10599 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10600 to set them. If not set then assume SSL clients will verify SSL servers
10603 Two new options to the verify program: -untrusted allows a set of
10604 untrusted certificates to be passed in and -purpose which sets the
10605 intended purpose of the certificate. If a purpose is set then the
10606 new chain verify code is used to check extension consistency.
10609 *) Support for the authority information access extension.
10612 *) Modify RSA and DSA PEM read routines to transparently handle
10613 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10614 public keys in a format compatible with certificate
10615 SubjectPublicKeyInfo structures. Unfortunately there were already
10616 functions called *_PublicKey_* which used various odd formats so
10617 these are retained for compatibility: however the DSA variants were
10618 never in a public release so they have been deleted. Changed dsa/rsa
10619 utilities to handle the new format: note no releases ever handled public
10620 keys so we should be OK.
10622 The primary motivation for this change is to avoid the same fiasco
10623 that dogs private keys: there are several incompatible private key
10624 formats some of which are standard and some OpenSSL specific and
10625 require various evil hacks to allow partial transparent handling and
10626 even then it doesn't work with DER formats. Given the option anything
10627 other than PKCS#8 should be dumped: but the other formats have to
10628 stay in the name of compatibility.
10630 With public keys and the benefit of hindsight one standard format
10631 is used which works with EVP_PKEY, RSA or DSA structures: though
10632 it clearly returns an error if you try to read the wrong kind of key.
10634 Added a -pubkey option to the 'x509' utility to output the public key.
10635 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10636 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10637 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10638 that do the same as the EVP_PKEY_assign_*() except they up the
10639 reference count of the added key (they don't "swallow" the
10643 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10644 CRLs would fail if the file contained no certificates or no CRLs:
10645 added a new function to read in both types and return the number
10646 read: this means that if none are read it will be an error. The
10647 DER versions of the certificate and CRL reader would always fail
10648 because it isn't possible to mix certificates and CRLs in DER format
10649 without choking one or the other routine. Changed this to just read
10650 a certificate: this is the best we can do. Also modified the code
10651 in apps/verify.c to take notice of return codes: it was previously
10652 attempting to read in certificates from NULL pointers and ignoring
10653 any errors: this is one reason why the cert and CRL reader seemed
10654 to work. It doesn't check return codes from the default certificate
10655 routines: these may well fail if the certificates aren't installed.
10658 *) Code to support otherName option in GeneralName.
10661 *) First update to verify code. Change the verify utility
10662 so it warns if it is passed a self signed certificate:
10663 for consistency with the normal behaviour. X509_verify
10664 has been modified to it will now verify a self signed
10665 certificate if *exactly* the same certificate appears
10666 in the store: it was previously impossible to trust a
10667 single self signed certificate. This means that:
10668 openssl verify ss.pem
10669 now gives a warning about a self signed certificate but
10670 openssl verify -CAfile ss.pem ss.pem
10674 *) For servers, store verify_result in SSL_SESSION data structure
10675 (and add it to external session representation).
10676 This is needed when client certificate verifications fails,
10677 but an application-provided verification callback (set by
10678 SSL_CTX_set_cert_verify_callback) allows accepting the session
10679 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10680 but returns 1): When the session is reused, we have to set
10681 ssl->verify_result to the appropriate error code to avoid
10683 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10685 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10686 case in PKCS7_dataInit() where the signed PKCS7 structure
10687 didn't contain any existing data because it was being created.
10688 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10690 *) Add a salt to the key derivation routines in enc.c. This
10691 forms the first 8 bytes of the encrypted file. Also add a
10692 -S option to allow a salt to be input on the command line.
10695 *) New function X509_cmp(). Oddly enough there wasn't a function
10696 to compare two certificates. We do this by working out the SHA1
10697 hash and comparing that. X509_cmp() will be needed by the trust
10701 *) SSL_get1_session() is like SSL_get_session(), but increments
10702 the reference count in the SSL_SESSION returned.
10703 [Geoff Thorpe <geoff@eu.c2.net>]
10705 *) Fix for 'req': it was adding a null to request attributes.
10706 Also change the X509_LOOKUP and X509_INFO code to handle
10707 certificate auxiliary information.
10710 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10714 *) Add the possibility to add extra information to the memory leak
10715 detecting output, to form tracebacks, showing from where each
10716 allocation was originated: CRYPTO_push_info("constant string") adds
10717 the string plus current file name and line number to a per-thread
10718 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10719 is like calling CYRPTO_pop_info() until the stack is empty.
10720 Also updated memory leak detection code to be multi-thread-safe.
10723 *) Add options -text and -noout to pkcs7 utility and delete the
10724 encryption options which never did anything. Update docs.
10727 *) Add options to some of the utilities to allow the pass phrase
10728 to be included on either the command line (not recommended on
10729 OSes like Unix) or read from the environment. Update the
10730 manpages and fix a few bugs.
10733 *) Add a few manpages for some of the openssl commands.
10736 *) Fix the -revoke option in ca. It was freeing up memory twice,
10737 leaking and not finding already revoked certificates.
10740 *) Extensive changes to support certificate auxiliary information.
10741 This involves the use of X509_CERT_AUX structure and X509_AUX
10742 functions. An X509_AUX function such as PEM_read_X509_AUX()
10743 can still read in a certificate file in the usual way but it
10744 will also read in any additional "auxiliary information". By
10745 doing things this way a fair degree of compatibility can be
10746 retained: existing certificates can have this information added
10747 using the new 'x509' options.
10749 Current auxiliary information includes an "alias" and some trust
10750 settings. The trust settings will ultimately be used in enhanced
10751 certificate chain verification routines: currently a certificate
10752 can only be trusted if it is self signed and then it is trusted
10756 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10757 The problem was that one of the replacement routines had not been working
10758 since SSLeay releases. For now the offending routine has been replaced
10759 with non-optimised assembler. Even so, this now gives around 95%
10760 performance improvement for 1024 bit RSA signs.
10763 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10764 handling. Most clients have the effective key size in bits equal to
10765 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10766 A few however don't do this and instead use the size of the decrypted key
10767 to determine the RC2 key length and the AlgorithmIdentifier to determine
10768 the effective key length. In this case the effective key length can still
10769 be 40 bits but the key length can be 168 bits for example. This is fixed
10770 by manually forcing an RC2 key into the EVP_PKEY structure because the
10771 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10772 the key length and effective key length are equal.
10775 *) Add a bunch of functions that should simplify the creation of
10776 X509_NAME structures. Now you should be able to do:
10777 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10778 and have it automatically work out the correct field type and fill in
10779 the structures. The more adventurous can try:
10780 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10781 and it will (hopefully) work out the correct multibyte encoding.
10784 *) Change the 'req' utility to use the new field handling and multibyte
10785 copy routines. Before the DN field creation was handled in an ad hoc
10786 way in req, ca, and x509 which was rather broken and didn't support
10787 BMPStrings or UTF8Strings. Since some software doesn't implement
10788 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10789 using the dirstring_type option. See the new comment in the default
10790 openssl.cnf for more info.
10793 *) Make crypto/rand/md_rand.c more robust:
10794 - Assure unique random numbers after fork().
10795 - Make sure that concurrent threads access the global counter and
10796 md serializably so that we never lose entropy in them
10797 or use exactly the same state in multiple threads.
10798 Access to the large state is not always serializable because
10799 the additional locking could be a performance killer, and
10800 md should be large enough anyway.
10803 *) New file apps/app_rand.c with commonly needed functionality
10804 for handling the random seed file.
10806 Use the random seed file in some applications that previously did not:
10808 dsaparam -genkey (which also ignored its '-rand' option),
10811 x509 (when signing).
10812 Except on systems with /dev/urandom, it is crucial to have a random
10813 seed file at least for key creation, DSA signing, and for DH exchanges;
10814 for RSA signatures we could do without one.
10816 gendh and gendsa (unlike genrsa) used to read only the first byte
10817 of each file listed in the '-rand' option. The function as previously
10818 found in genrsa is now in app_rand.c and is used by all programs
10819 that support '-rand'.
10822 *) In RAND_write_file, use mode 0600 for creating files;
10823 don't just chmod when it may be too late.
10826 *) Report an error from X509_STORE_load_locations
10827 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10830 *) New function ASN1_mbstring_copy() this copies a string in either
10831 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10832 into an ASN1_STRING type. A mask of permissible types is passed
10833 and it chooses the "minimal" type to use or an error if not type
10837 *) Add function equivalents to the various macros in asn1.h. The old
10838 macros are retained with an M_ prefix. Code inside the library can
10839 use the M_ macros. External code (including the openssl utility)
10840 should *NOT* in order to be "shared library friendly".
10843 *) Add various functions that can check a certificate's extensions
10844 to see if it usable for various purposes such as SSL client,
10845 server or S/MIME and CAs of these types. This is currently
10846 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10847 verification. Also added a -purpose flag to x509 utility to
10848 print out all the purposes.
10851 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10855 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10856 for, obtain and decode and extension and obtain its critical flag.
10857 This allows all the necessary extension code to be handled in a
10858 single function call.
10861 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10862 platforms. See crypto/rc4/rc4_enc.c for further details.
10865 *) New -noout option to asn1parse. This causes no output to be produced
10866 its main use is when combined with -strparse and -out to extract data
10867 from a file (which may not be in ASN.1 format).
10870 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10871 when producing the local key id.
10872 [Richard Levitte <levitte@stacken.kth.se>]
10874 *) New option -dhparam in s_server. This allows a DH parameter file to be
10875 stated explicitly. If it is not stated then it tries the first server
10876 certificate file. The previous behaviour hard coded the filename
10880 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10881 a public key to be input or output. For example:
10882 openssl rsa -in key.pem -pubout -out pubkey.pem
10883 Also added necessary DSA public key functions to handle this.
10886 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10887 in the message. This was handled by allowing
10888 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10889 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10891 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10892 to the end of the strings whereas this didn't. This would cause problems
10893 if strings read with d2i_ASN1_bytes() were later modified.
10894 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10896 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10897 data and it contains EOF it will end up returning an error. This is
10898 caused by input 46 bytes long. The cause is due to the way base64
10899 BIOs find the start of base64 encoded data. They do this by trying a
10900 trial decode on each line until they find one that works. When they
10901 do a flag is set and it starts again knowing it can pass all the
10902 data directly through the decoder. Unfortunately it doesn't reset
10903 the context it uses. This means that if EOF is reached an attempt
10904 is made to pass two EOFs through the context and this causes the
10905 resulting error. This can also cause other problems as well. As is
10906 usual with these problems it takes *ages* to find and the fix is
10907 trivial: move one line.
10908 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10910 *) Ugly workaround to get s_client and s_server working under Windows. The
10911 old code wouldn't work because it needed to select() on sockets and the
10912 tty (for keypresses and to see if data could be written). Win32 only
10913 supports select() on sockets so we select() with a 1s timeout on the
10914 sockets and then see if any characters are waiting to be read, if none
10915 are present then we retry, we also assume we can always write data to
10916 the tty. This isn't nice because the code then blocks until we've
10917 received a complete line of data and it is effectively polling the
10918 keyboard at 1s intervals: however it's quite a bit better than not
10919 working at all :-) A dedicated Windows application might handle this
10920 with an event loop for example.
10923 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10924 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10925 will be called when RSA_sign() and RSA_verify() are used. This is useful
10926 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10927 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10928 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10929 This necessitated the support of an extra signature type NID_md5_sha1
10930 for SSL signatures and modifications to the SSL library to use it instead
10931 of calling RSA_public_decrypt() and RSA_private_encrypt().
10934 *) Add new -verify -CAfile and -CApath options to the crl program, these
10935 will lookup a CRL issuers certificate and verify the signature in a
10936 similar way to the verify program. Tidy up the crl program so it
10937 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10938 less strict. It will now permit CRL extensions even if it is not
10939 a V2 CRL: this will allow it to tolerate some broken CRLs.
10942 *) Initialize all non-automatic variables each time one of the openssl
10943 sub-programs is started (this is necessary as they may be started
10944 multiple times from the "OpenSSL>" prompt).
10945 [Lennart Bang, Bodo Moeller]
10947 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10948 removing all other RSA functionality (this is what NO_RSA does). This
10949 is so (for example) those in the US can disable those operations covered
10950 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10954 *) Non-copying interface to BIO pairs.
10955 (still largely untested)
10958 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10959 ASCII string. This was handled independently in various places before.
10962 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10963 UTF8 strings a character at a time.
10966 *) Use client_version from client hello to select the protocol
10967 (s23_srvr.c) and for RSA client key exchange verification
10968 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10971 *) Add various utility functions to handle SPKACs, these were previously
10972 handled by poking round in the structure internals. Added new function
10973 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10974 print, verify and generate SPKACs. Based on an original idea from
10975 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10978 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10981 *) Allow the config file extension section to be overwritten on the
10982 command line. Based on an original idea from Massimiliano Pala
10983 <madwolf@comune.modena.it>. The new option is called -extensions
10984 and can be applied to ca, req and x509. Also -reqexts to override
10985 the request extensions in req and -crlexts to override the crl extensions
10989 *) Add new feature to the SPKAC handling in ca. Now you can include
10990 the same field multiple times by preceding it by "XXXX." for example:
10993 this is the same syntax as used in the req config file.
10996 *) Allow certificate extensions to be added to certificate requests. These
10997 are specified in a 'req_extensions' option of the req section of the
10998 config file. They can be printed out with the -text option to req but
10999 are otherwise ignored at present.
11002 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11003 data read consists of only the final block it would not decrypted because
11004 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11005 A misplaced 'break' also meant the decrypted final block might not be
11006 copied until the next read.
11009 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11010 a few extra parameters to the DH structure: these will be useful if
11011 for example we want the value of 'q' or implement X9.42 DH.
11014 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11015 provides hooks that allow the default DSA functions or functions on a
11016 "per key" basis to be replaced. This allows hardware acceleration and
11017 hardware key storage to be handled without major modification to the
11018 library. Also added low level modexp hooks and CRYPTO_EX structure and
11019 associated functions.
11022 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11023 as "read only": it can't be written to and the buffer it points to will
11024 not be freed. Reading from a read only BIO is much more efficient than
11025 a normal memory BIO. This was added because there are several times when
11026 an area of memory needs to be read from a BIO. The previous method was
11027 to create a memory BIO and write the data to it, this results in two
11028 copies of the data and an O(n^2) reading algorithm. There is a new
11029 function BIO_new_mem_buf() which creates a read only memory BIO from
11030 an area of memory. Also modified the PKCS#7 routines to use read only
11034 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11035 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11036 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11037 but a retry condition occurred while trying to read the rest.
11040 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11041 NID_pkcs7_encrypted by default: this was wrong since this should almost
11042 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11043 the encrypted data type: this is a more sensible place to put it and it
11044 allows the PKCS#12 code to be tidied up that duplicated this
11048 *) Changed obj_dat.pl script so it takes its input and output files on
11049 the command line. This should avoid shell escape redirection problems
11053 *) Initial support for certificate extension requests, these are included
11054 in things like Xenroll certificate requests. Included functions to allow
11055 extensions to be obtained and added.
11058 *) -crlf option to s_client and s_server for sending newlines as
11059 CRLF (as required by many protocols).
11062 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11064 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11065 [Ralf S. Engelschall]
11067 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11068 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11070 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11074 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11075 DH parameters/keys (q is lost during that conversion, but the resulting
11076 DH parameters contain its length).
11078 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11079 much faster than DH_generate_parameters (which creates parameters
11080 where p = 2*q + 1), and also the smaller q makes DH computations
11081 much more efficient (160-bit exponentiation instead of 1024-bit
11082 exponentiation); so this provides a convenient way to support DHE
11083 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11084 utter importance to use
11085 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11087 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11088 when such DH parameters are used, because otherwise small subgroup
11089 attacks may become possible!
11092 *) Avoid memory leak in i2d_DHparams.
11095 *) Allow the -k option to be used more than once in the enc program:
11096 this allows the same encrypted message to be read by multiple recipients.
11099 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11100 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11101 it will always use the numerical form of the OID, even if it has a short
11105 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11106 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11107 otherwise bn_mod_exp was called. In the case of hardware keys for example
11108 no private key components need be present and it might store extra data
11109 in the RSA structure, which cannot be accessed from bn_mod_exp.
11110 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11111 private key operations.
11114 *) Added support for SPARC Linux.
11117 *) pem_password_cb function type incompatibly changed from
11118 typedef int pem_password_cb(char *buf, int size, int rwflag);
11120 ....(char *buf, int size, int rwflag, void *userdata);
11121 so that applications can pass data to their callbacks:
11122 The PEM[_ASN1]_{read,write}... functions and macros now take an
11123 additional void * argument, which is just handed through whenever
11124 the password callback is called.
11125 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11127 New function SSL_CTX_set_default_passwd_cb_userdata.
11129 Compatibility note: As many C implementations push function arguments
11130 onto the stack in reverse order, the new library version is likely to
11131 interoperate with programs that have been compiled with the old
11132 pem_password_cb definition (PEM_whatever takes some data that
11133 happens to be on the stack as its last argument, and the callback
11134 just ignores this garbage); but there is no guarantee whatsoever that
11137 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11138 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11139 problems not only on Windows, but also on some Unix platforms.
11140 To avoid problematic command lines, these definitions are now in an
11141 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11142 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11145 *) MIPS III/IV assembler module is reimplemented.
11148 *) More DES library cleanups: remove references to srand/rand and
11149 delete an unused file.
11152 *) Add support for the the free Netwide assembler (NASM) under Win32,
11153 since not many people have MASM (ml) and it can be hard to obtain.
11154 This is currently experimental but it seems to work OK and pass all
11155 the tests. Check out INSTALL.W32 for info.
11158 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11159 without temporary keys kept an extra copy of the server key,
11160 and connections with temporary keys did not free everything in case
11164 *) New function RSA_check_key and new openssl rsa option -check
11165 for verifying the consistency of RSA keys.
11166 [Ulf Moeller, Bodo Moeller]
11168 *) Various changes to make Win32 compile work:
11169 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11170 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11171 comparison" warnings.
11172 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11175 *) Add a debugging option to PKCS#5 v2 key generation function: when
11176 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11177 derived keys are printed to stderr.
11180 *) Copy the flags in ASN1_STRING_dup().
11181 [Roman E. Pavlov <pre@mo.msk.ru>]
11183 *) The x509 application mishandled signing requests containing DSA
11184 keys when the signing key was also DSA and the parameters didn't match.
11186 It was supposed to omit the parameters when they matched the signing key:
11187 the verifying software was then supposed to automatically use the CA's
11188 parameters if they were absent from the end user certificate.
11190 Omitting parameters is no longer recommended. The test was also
11191 the wrong way round! This was probably due to unusual behaviour in
11192 EVP_cmp_parameters() which returns 1 if the parameters match.
11193 This meant that parameters were omitted when they *didn't* match and
11194 the certificate was useless. Certificates signed with 'ca' didn't have
11196 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11198 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11199 The interface is as follows:
11200 Applications can use
11201 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11202 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11203 "off" is now the default.
11204 The library internally uses
11205 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11206 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11207 to disable memory-checking temporarily.
11209 Some inconsistent states that previously were possible (and were
11210 even the default) are now avoided.
11212 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11213 with each memory chunk allocated; this is occasionally more helpful
11214 than just having a counter.
11216 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11218 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11222 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11223 which largely parallels "options", but is for changing API behaviour,
11224 whereas "options" are about protocol behaviour.
11225 Initial "mode" flags are:
11227 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11228 a single record has been written.
11229 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11230 retries use the same buffer location.
11231 (But all of the contents must be
11235 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11238 *) Fix problems with no-hmac etc.
11239 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11241 *) New functions RSA_get_default_method(), RSA_set_method() and
11242 RSA_get_method(). These allows replacement of RSA_METHODs without having
11243 to mess around with the internals of an RSA structure.
11246 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11247 Also really enable memory leak checks in openssl.c and in some
11249 [Chad C. Mulligan, Bodo Moeller]
11251 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11252 up the length of negative integers. This has now been simplified to just
11253 store the length when it is first determined and use it later, rather
11254 than trying to keep track of where data is copied and updating it to
11256 [Steve Henson, reported by Brien Wheeler
11257 <bwheeler@authentica-security.com>]
11259 *) Add a new function PKCS7_signatureVerify. This allows the verification
11260 of a PKCS#7 signature but with the signing certificate passed to the
11261 function itself. This contrasts with PKCS7_dataVerify which assumes the
11262 certificate is present in the PKCS#7 structure. This isn't always the
11263 case: certificates can be omitted from a PKCS#7 structure and be
11264 distributed by "out of band" means (such as a certificate database).
11267 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11268 function prototypes in pem.h, also change util/mkdef.pl to add the
11269 necessary function names.
11272 *) mk1mf.pl (used by Windows builds) did not properly read the
11273 options set by Configure in the top level Makefile, and Configure
11274 was not even able to write more than one option correctly.
11275 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11278 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11279 file to be loaded from a BIO or FILE pointer. The BIO version will
11280 for example allow memory BIOs to contain config info.
11283 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11284 Whoever hopes to achieve shared-library compatibility across versions
11285 must use this, not the compile-time macro.
11286 (Exercise 0.9.4: Which is the minimum library version required by
11288 Note: All this applies only to multi-threaded programs, others don't
11292 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11293 through a BIO pair triggered the default case, i.e.
11294 SSLerr(...,SSL_R_UNKNOWN_STATE).
11297 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11298 can use the SSL library even if none of the specific BIOs is
11302 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11303 for the encoded length.
11304 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11306 *) Add initial documentation of the X509V3 functions.
11309 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11310 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11311 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11312 secure PKCS#8 private key format with a high iteration count.
11315 *) Fix determination of Perl interpreter: A perl or perl5
11316 _directory_ in $PATH was also accepted as the interpreter.
11317 [Ralf S. Engelschall]
11319 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11320 wrong with it but it was very old and did things like calling
11321 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11322 unusual formatting.
11325 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11326 to use the new extension code.
11329 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11330 with macros. This should make it easier to change their form, add extra
11331 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11335 *) Add to configuration table a new entry that can specify an alternative
11336 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11337 according to Mark Crispin <MRC@Panda.COM>.
11341 *) DES CBC did not update the IV. Weird.
11344 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11345 Changing the behaviour of the former might break existing programs --
11346 where IV updating is needed, des_ncbc_encrypt can be used.
11349 *) When bntest is run from "make test" it drives bc to check its
11350 calculations, as well as internally checking them. If an internal check
11351 fails, it needs to cause bc to give a non-zero result or make test carries
11352 on without noticing the failure. Fixed.
11355 *) DES library cleanups.
11358 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11359 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11360 ciphers. NOTE: although the key derivation function has been verified
11361 against some published test vectors it has not been extensively tested
11362 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11366 *) Instead of "mkdir -p", which is not fully portable, use new
11367 Perl script "util/mkdir-p.pl".
11370 *) Rewrite the way password based encryption (PBE) is handled. It used to
11371 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11372 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11373 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11374 the 'parameter' field of the AlgorithmIdentifier is passed to the
11375 underlying key generation function so it must do its own ASN1 parsing.
11376 This has also changed the EVP_PBE_CipherInit() function which now has a
11377 'parameter' argument instead of literal salt and iteration count values
11378 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11381 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11382 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11383 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11384 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11385 value was just used as a "magic string" and not used directly its
11386 value doesn't matter.
11389 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11393 *) "linux-sparc64" configuration (ultrapenguin).
11394 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11395 "linux-sparc" configuration.
11396 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11398 *) config now generates no-xxx options for missing ciphers.
11401 *) Support the EBCDIC character set (work in progress).
11402 File ebcdic.c not yet included because it has a different license.
11403 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11405 *) Support BS2000/OSD-POSIX.
11406 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11408 *) Make callbacks for key generation use void * instead of char *.
11411 *) Make S/MIME samples compile (not yet tested).
11414 *) Additional typesafe stacks.
11417 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11421 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11423 *) New configuration variant "sco5-gcc".
11425 *) Updated some demos.
11426 [Sean O Riordain, Wade Scholine]
11428 *) Add missing BIO_free at exit of pkcs12 application.
11431 *) Fix memory leak in conf.c.
11434 *) Updates for Win32 to assembler version of MD5.
11437 *) Set #! path to perl in apps/der_chop to where we found it
11438 instead of using a fixed path.
11441 *) SHA library changes for irix64-mips4-cc.
11444 *) Improvements for VMS support.
11448 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11450 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11451 This also avoids the problems with SC4.2 and unpatched SC5.
11452 [Andy Polyakov <appro@fy.chalmers.se>]
11454 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11455 These are required because of the typesafe stack would otherwise break
11456 existing code. If old code used a structure member which used to be STACK
11457 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11458 sk_num or sk_value it would produce an error because the num, data members
11459 are not present in STACK_OF. Now it just produces a warning. sk_set
11460 replaces the old method of assigning a value to sk_value
11461 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11462 that does this will no longer work (and should use sk_set instead) but
11463 this could be regarded as a "questionable" behaviour anyway.
11466 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11467 correctly handle encrypted S/MIME data.
11470 *) Change type of various DES function arguments from des_cblock
11471 (which means, in function argument declarations, pointer to char)
11472 to des_cblock * (meaning pointer to array with 8 char elements),
11473 which allows the compiler to do more typechecking; it was like
11474 that back in SSLeay, but with lots of ugly casts.
11476 Introduce new type const_des_cblock.
11479 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11480 problems: find RecipientInfo structure that matches recipient certificate
11481 and initialise the ASN1 structures properly based on passed cipher.
11484 *) Belatedly make the BN tests actually check the results.
11487 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11488 to and from BNs: it was completely broken. New compilation option
11489 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11490 key elements as negative integers.
11493 *) Reorganize and speed up MD5.
11494 [Andy Polyakov <appro@fy.chalmers.se>]
11497 [Richard Levitte <richard@levitte.org>]
11499 *) New option -out to asn1parse to allow the parsed structure to be
11500 output to a file. This is most useful when combined with the -strparse
11501 option to examine the output of things like OCTET STRINGS.
11504 *) Make SSL library a little more fool-proof by not requiring any longer
11505 that SSL_set_{accept,connect}_state be called before
11506 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11507 in many applications because usually everything *appeared* to work as
11508 intended anyway -- now it really works as intended).
11511 *) Move openssl.cnf out of lib/.
11514 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11515 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11516 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11517 [Ralf S. Engelschall]
11519 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11520 handle PKCS#7 enveloped data properly.
11521 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11523 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11524 copying pointers. The cert_st handling is changed by this in
11525 various ways (and thus what used to be known as ctx->default_cert
11526 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11527 any longer when s->cert does not give us what we need).
11528 ssl_cert_instantiate becomes obsolete by this change.
11529 As soon as we've got the new code right (possibly it already is?),
11530 we have solved a couple of bugs of the earlier code where s->cert
11531 was used as if it could not have been shared with other SSL structures.
11533 Note that using the SSL API in certain dirty ways now will result
11534 in different behaviour than observed with earlier library versions:
11535 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11536 does not influence s as it used to.
11538 In order to clean up things more thoroughly, inside SSL_SESSION
11539 we don't use CERT any longer, but a new structure SESS_CERT
11540 that holds per-session data (if available); currently, this is
11541 the peer's certificate chain and, for clients, the server's certificate
11542 and temporary key. CERT holds only those values that can have
11543 meaningful defaults in an SSL_CTX.
11546 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11547 from the internal representation. Various PKCS#7 fixes: remove some
11548 evil casts and set the enc_dig_alg field properly based on the signing
11552 *) Allow PKCS#12 password to be set from the command line or the
11553 environment. Let 'ca' get its config file name from the environment
11554 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11558 *) Allow certificate policies extension to use an IA5STRING for the
11559 organization field. This is contrary to the PKIX definition but
11560 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11564 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11565 without disallowing inline assembler and the like for non-pedantic builds.
11568 *) Support Borland C++ builder.
11569 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11571 *) Support Mingw32.
11574 *) SHA-1 cleanups and performance enhancements.
11575 [Andy Polyakov <appro@fy.chalmers.se>]
11577 *) Sparc v8plus assembler for the bignum library.
11578 [Andy Polyakov <appro@fy.chalmers.se>]
11580 *) Accept any -xxx and +xxx compiler options in Configure.
11583 *) Update HPUX configuration.
11586 *) Add missing sk_<type>_unshift() function to safestack.h
11587 [Ralf S. Engelschall]
11589 *) New function SSL_CTX_use_certificate_chain_file that sets the
11590 "extra_cert"s in addition to the certificate. (This makes sense
11591 only for "PEM" format files, as chains as a whole are not
11595 *) Support verify_depth from the SSL API.
11596 x509_vfy.c had what can be considered an off-by-one-error:
11597 Its depth (which was not part of the external interface)
11598 was actually counting the number of certificates in a chain;
11599 now it really counts the depth.
11602 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11603 instead of X509err, which often resulted in confusing error
11604 messages since the error codes are not globally unique
11605 (e.g. an alleged error in ssl3_accept when a certificate
11606 didn't match the private key).
11608 *) New function SSL_CTX_set_session_id_context that allows to set a default
11609 value (so that you don't need SSL_set_session_id_context for each
11610 connection using the SSL_CTX).
11613 *) OAEP decoding bug fix.
11616 *) Support INSTALL_PREFIX for package builders, as proposed by
11620 *) New Configure options "threads" and "no-threads". For systems
11621 where the proper compiler options are known (currently Solaris
11622 and Linux), "threads" is the default.
11625 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11628 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11629 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11630 such as /usr/local/bin.
11633 *) "make linux-shared" to build shared libraries.
11634 [Niels Poppe <niels@netbox.org>]
11636 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11639 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11640 extension adding in x509 utility.
11643 *) Remove NOPROTO sections and error code comments.
11646 *) Partial rewrite of the DEF file generator to now parse the ANSI
11650 *) New Configure options --prefix=DIR and --openssldir=DIR.
11653 *) Complete rewrite of the error code script(s). It is all now handled
11654 by one script at the top level which handles error code gathering,
11655 header rewriting and C source file generation. It should be much better
11656 than the old method: it now uses a modified version of Ulf's parser to
11657 read the ANSI prototypes in all header files (thus the old K&R definitions
11658 aren't needed for error creation any more) and do a better job of
11659 translating function codes into names. The old 'ASN1 error code imbedded
11660 in a comment' is no longer necessary and it doesn't use .err files which
11661 have now been deleted. Also the error code call doesn't have to appear all
11662 on one line (which resulted in some large lines...).
11665 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11668 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11669 0 (which usually indicates a closed connection), but continue reading.
11672 *) Fix some race conditions.
11675 *) Add support for CRL distribution points extension. Add Certificate
11676 Policies and CRL distribution points documentation.
11679 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11682 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11683 8 of keying material. Merlin has also confirmed interop with this fix
11684 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11685 [Merlin Hughes <merlin@baltimore.ie>]
11687 *) Fix lots of warnings.
11688 [Richard Levitte <levitte@stacken.kth.se>]
11690 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11691 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11692 [Richard Levitte <levitte@stacken.kth.se>]
11694 *) Fix problems with sizeof(long) == 8.
11695 [Andy Polyakov <appro@fy.chalmers.se>]
11697 *) Change functions to ANSI C.
11700 *) Fix typos in error codes.
11701 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11703 *) Remove defunct assembler files from Configure.
11706 *) SPARC v8 assembler BIGNUM implementation.
11707 [Andy Polyakov <appro@fy.chalmers.se>]
11709 *) Support for Certificate Policies extension: both print and set.
11710 Various additions to support the r2i method this uses.
11713 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11714 return a const string when you are expecting an allocated buffer.
11717 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11718 types DirectoryString and DisplayText.
11721 *) Add code to allow r2i extensions to access the configuration database,
11722 add an LHASH database driver and add several ctx helper functions.
11725 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11726 fail when they extended the size of a BIGNUM.
11729 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11730 support typesafe stack.
11733 *) Fix typo in SSL_[gs]et_options().
11734 [Nils Frostberg <nils@medcom.se>]
11736 *) Delete various functions and files that belonged to the (now obsolete)
11737 old X509V3 handling code.
11740 *) New Configure option "rsaref".
11743 *) Don't auto-generate pem.h.
11746 *) Introduce type-safe ASN.1 SETs.
11749 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11750 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11752 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11753 that links with OpenSSL (well at least cause lots of warnings), but fear
11754 not: the conversion is trivial, and it eliminates loads of evil casts. A
11755 few STACKed things have been converted already. Feel free to convert more.
11756 In the fullness of time, I'll do away with the STACK type altogether.
11759 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11760 specified in <certfile> by updating the entry in the index.txt file.
11761 This way one no longer has to edit the index.txt file manually for
11762 revoking a certificate. The -revoke option does the gory details now.
11763 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11765 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11766 `-text' option at all and this way the `-noout -text' combination was
11767 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11768 [Ralf S. Engelschall]
11770 *) Make sure a corresponding plain text error message exists for the
11771 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11772 verify callback function determined that a certificate was revoked.
11773 [Ralf S. Engelschall]
11775 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11776 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11777 all available ciphers including rc5, which was forgotten until now.
11778 In order to let the testing shell script know which algorithms
11779 are available, a new (up to now undocumented) command
11780 "openssl list-cipher-commands" is used.
11783 *) Bugfix: s_client occasionally would sleep in select() when
11784 it should have checked SSL_pending() first.
11787 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11788 the raw DSA values prior to ASN.1 encoding.
11791 *) Tweaks to Configure
11792 [Niels Poppe <niels@netbox.org>]
11794 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11798 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11801 *) New config option to avoid instructions that are illegal on the 80386.
11802 The default code is faster, but requires at least a 486.
11805 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11806 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11807 same as SSL2_VERSION anyway.
11810 *) New "-showcerts" option for s_client.
11813 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11814 application. Various cleanups and fixes.
11817 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11818 modify error routines to work internally. Add error codes and PBE init
11819 to library startup routines.
11822 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11823 packing functions to asn1 and evp. Changed function names and error
11824 codes along the way.
11827 *) PKCS12 integration: and so it begins... First of several patches to
11828 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11829 objects to objects.h
11832 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11833 and display support for Thawte strong extranet extension.
11836 *) Add LinuxPPC support.
11837 [Jeff Dubrule <igor@pobox.org>]
11839 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11840 bn_div_words in alpha.s.
11841 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11843 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11844 OAEP isn't supported when OpenSSL is built with RSAref.
11845 [Ulf Moeller <ulf@fitug.de>]
11847 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11848 so they no longer are missing under -DNOPROTO.
11849 [Soren S. Jorvang <soren@t.dk>]
11852 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11854 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11855 doesn't work when the session is reused. Coming soon!
11858 *) Fix a security hole, that allows sessions to be reused in the wrong
11859 context thus bypassing client cert protection! All software that uses
11860 client certs and session caches in multiple contexts NEEDS PATCHING to
11861 allow session reuse! A fuller solution is in the works.
11862 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11864 *) Some more source tree cleanups (removed obsolete files
11865 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11866 permission on "config" script to be executable) and a fix for the INSTALL
11868 [Ulf Moeller <ulf@fitug.de>]
11870 *) Remove some legacy and erroneous uses of malloc, free instead of
11872 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11874 *) Make rsa_oaep_test return non-zero on error.
11875 [Ulf Moeller <ulf@fitug.de>]
11877 *) Add support for native Solaris shared libraries. Configure
11878 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11879 if someone would make that last step automatic.
11880 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11882 *) ctx_size was not built with the right compiler during "make links". Fixed.
11885 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11886 except NULL ciphers". This means the default cipher list will no longer
11887 enable NULL ciphers. They need to be specifically enabled e.g. with
11888 the string "DEFAULT:eNULL".
11891 *) Fix to RSA private encryption routines: if p < q then it would
11892 occasionally produce an invalid result. This will only happen with
11893 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11896 *) Be less restrictive and allow also `perl util/perlpath.pl
11897 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11898 because this way one can also use an interpreter named `perl5' (which is
11899 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11900 installed as `perl').
11901 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11903 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11904 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11906 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11907 advapi32.lib to Win32 build and change the pem test comparison
11908 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11909 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11910 and crypto/des/ede_cbcm_enc.c.
11913 *) DES quad checksum was broken on big-endian architectures. Fixed.
11916 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11917 Win32 test batch file so it (might) work again. The Win32 test batch file
11918 is horrible: I feel ill....
11921 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11922 in e_os.h. Audit of header files to check ANSI and non ANSI
11923 sections: 10 functions were absent from non ANSI section and not exported
11924 from Windows DLLs. Fixed up libeay.num for new functions.
11927 *) Make `openssl version' output lines consistent.
11928 [Ralf S. Engelschall]
11930 *) Fix Win32 symbol export lists for BIO functions: Added
11931 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11932 to ms/libeay{16,32}.def.
11933 [Ralf S. Engelschall]
11935 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11936 fine under Unix and passes some trivial tests I've now added. But the
11937 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11938 added to make sure no one expects that this stuff really works in the
11939 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11940 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11942 [Ralf S. Engelschall]
11944 *) Fix the generation of two part addresses in perl.
11945 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11947 *) Add config entry for Linux on MIPS.
11948 [John Tobey <jtobey@channel1.com>]
11950 *) Make links whenever Configure is run, unless we are on Windoze.
11953 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11954 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11958 *) Add a useful kludge to allow package maintainers to specify compiler and
11959 other platforms details on the command line without having to patch the
11960 Configure script everytime: One now can use ``perl Configure
11961 <id>:<details>'', i.e. platform ids are allowed to have details appended
11962 to them (separated by colons). This is treated as there would be a static
11963 pre-configured entry in Configure's %table under key <id> with value
11964 <details> and ``perl Configure <id>'' is called. So, when you want to
11965 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11966 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11967 now, which overrides the FreeBSD-elf entry on-the-fly.
11968 [Ralf S. Engelschall]
11970 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11973 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11974 on the `perl Configure ...' command line. This way one can compile
11975 OpenSSL libraries with Position Independent Code (PIC) which is needed
11976 for linking it into DSOs.
11977 [Ralf S. Engelschall]
11979 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11983 *) Cleaned up the LICENSE document: The official contact for any license
11984 questions now is the OpenSSL core team under openssl-core@openssl.org.
11985 And add a paragraph about the dual-license situation to make sure people
11986 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11987 to the OpenSSL toolkit.
11988 [Ralf S. Engelschall]
11990 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11991 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11992 Additionally cleaned up the `make links' target: Remove unnecessary
11993 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11994 to speed processing and no longer clutter the display with confusing
11995 stuff. Instead only the actually done links are displayed.
11996 [Ralf S. Engelschall]
11998 *) Permit null encryption ciphersuites, used for authentication only. It used
11999 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12000 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12004 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12005 signed attributes when verifying signatures (this would break them),
12006 the detached data encoding was wrong and public keys obtained using
12007 X509_get_pubkey() weren't freed.
12010 *) Add text documentation for the BUFFER functions. Also added a work around
12011 to a Win95 console bug. This was triggered by the password read stuff: the
12012 last character typed gets carried over to the next fread(). If you were
12013 generating a new cert request using 'req' for example then the last
12014 character of the passphrase would be CR which would then enter the first
12018 *) Added the new `Includes OpenSSL Cryptography Software' button as
12019 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12020 button and can be used by applications based on OpenSSL to show the
12021 relationship to the OpenSSL project.
12022 [Ralf S. Engelschall]
12024 *) Remove confusing variables in function signatures in files
12025 ssl/ssl_lib.c and ssl/ssl.h.
12026 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12028 *) Don't install bss_file.c under PREFIX/include/
12029 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12031 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12032 functions that return function pointers and has support for NT specific
12033 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12034 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12035 unsigned to signed types: this was killing the Win32 compile.
12038 *) Add new certificate file to stack functions,
12039 SSL_add_dir_cert_subjects_to_stack() and
12040 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12041 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12042 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12043 This means that Apache-SSL and similar packages don't have to mess around
12044 to add as many CAs as they want to the preferred list.
12047 *) Experiment with doxygen documentation. Currently only partially applied to
12049 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12050 openssl.doxy as the configuration file.
12053 *) Get rid of remaining C++-style comments which strict C compilers hate.
12054 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12056 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12057 compiled in by default: it has problems with large keys.
12060 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12061 DH private keys and/or callback functions which directly correspond to
12062 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12063 is needed for applications which have to configure certificates on a
12064 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12066 For the RSA certificate situation is makes no difference, but
12067 for the DSA certificate situation this fixes the "no shared cipher"
12068 problem where the OpenSSL cipher selection procedure failed because the
12069 temporary keys were not overtaken from the context and the API provided
12070 no way to reconfigure them.
12071 The new functions now let applications reconfigure the stuff and they
12072 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12073 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12074 non-public-API function ssl_cert_instantiate() is used as a helper
12075 function and also to reduce code redundancy inside ssl_rsa.c.
12076 [Ralf S. Engelschall]
12078 *) Move s_server -dcert and -dkey options out of the undocumented feature
12079 area because they are useful for the DSA situation and should be
12080 recognized by the users.
12081 [Ralf S. Engelschall]
12083 *) Fix the cipher decision scheme for export ciphers: the export bits are
12084 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12085 SSL_EXP_MASK. So, the original variable has to be used instead of the
12086 already masked variable.
12087 [Richard Levitte <levitte@stacken.kth.se>]
12089 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12090 [Richard Levitte <levitte@stacken.kth.se>]
12092 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12093 from `int' to `unsigned int' because it's a length and initialized by
12094 EVP_DigestFinal() which expects an `unsigned int *'.
12095 [Richard Levitte <levitte@stacken.kth.se>]
12097 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12098 script. Instead use the usual Shell->Perl transition trick.
12099 [Ralf S. Engelschall]
12101 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12102 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12103 -noout -modulus' as it's already the case for `openssl rsa -noout
12104 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12105 currently the public key is printed (a decision which was already done by
12106 `openssl dsa -modulus' in the past) which serves a similar purpose.
12107 Additionally the NO_RSA no longer completely removes the whole -modulus
12108 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12110 [Ralf S. Engelschall]
12112 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12113 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12114 [Arne Ansper <arne@ats.cyber.ee>]
12116 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12117 to be added. Now both 'req' and 'ca' can use new objects defined in the
12121 *) Add cool BIO that does syslog (or event log on NT).
12122 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12124 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12125 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12126 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12127 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12130 *) Add preliminary config info for new extension code.
12133 *) Make RSA_NO_PADDING really use no padding.
12134 [Ulf Moeller <ulf@fitug.de>]
12136 *) Generate errors when private/public key check is done.
12139 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12140 for some CRL extensions and new objects added.
12143 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12144 key usage extension and fuller support for authority key id.
12147 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12148 padding method for RSA, which is recommended for new applications in PKCS
12149 #1 v2.0 (RFC 2437, October 1998).
12150 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12151 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12152 against Bleichbacher's attack on RSA.
12153 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12156 *) Updates to the new SSL compression code
12157 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12159 *) Fix so that the version number in the master secret, when passed
12160 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12161 (because the server will not accept higher), that the version number
12162 is 0x03,0x01, not 0x03,0x00
12163 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12165 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12166 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12167 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12170 *) Support for RAW extensions where an arbitrary extension can be
12171 created by including its DER encoding. See apps/openssl.cnf for
12175 *) Make sure latest Perl versions don't interpret some generated C array
12176 code as Perl array code in the crypto/err/err_genc.pl script.
12177 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12179 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12180 not many people have the assembler. Various Win32 compilation fixes and
12181 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12182 build instructions.
12185 *) Modify configure script 'Configure' to automatically create crypto/date.h
12186 file under Win32 and also build pem.h from pem.org. New script
12187 util/mkfiles.pl to create the MINFO file on environments that can't do a
12188 'make files': perl util/mkfiles.pl >MINFO should work.
12191 *) Major rework of DES function declarations, in the pursuit of correctness
12192 and purity. As a result, many evil casts evaporated, and some weirdness,
12193 too. You may find this causes warnings in your code. Zapping your evil
12194 casts will probably fix them. Mostly.
12197 *) Fix for a typo in asn1.h. Bug fix to object creation script
12198 obj_dat.pl. It considered a zero in an object definition to mean
12199 "end of object": none of the objects in objects.h have any zeros
12200 so it wasn't spotted.
12201 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12203 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12204 Masking (CBCM). In the absence of test vectors, the best I have been able
12205 to do is check that the decrypt undoes the encrypt, so far. Send me test
12206 vectors if you have them.
12209 *) Correct calculation of key length for export ciphers (too much space was
12210 allocated for null ciphers). This has not been tested!
12213 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12214 message is now correct (it understands "crypto" and "ssl" on its
12215 command line). There is also now an "update" option. This will update
12216 the util/ssleay.num and util/libeay.num files with any new functions.
12218 perl util/mkdef.pl crypto ssl update
12219 it will update them.
12222 *) Overhauled the Perl interface (perl/*):
12223 - ported BN stuff to OpenSSL's different BN library
12224 - made the perl/ source tree CVS-aware
12225 - renamed the package from SSLeay to OpenSSL (the files still contain
12226 their history because I've copied them in the repository)
12227 - removed obsolete files (the test scripts will be replaced
12228 by better Test::Harness variants in the future)
12229 [Ralf S. Engelschall]
12231 *) First cut for a very conservative source tree cleanup:
12232 1. merge various obsolete readme texts into doc/ssleay.txt
12233 where we collect the old documents and readme texts.
12234 2. remove the first part of files where I'm already sure that we no
12235 longer need them because of three reasons: either they are just temporary
12236 files which were left by Eric or they are preserved original files where
12237 I've verified that the diff is also available in the CVS via "cvs diff
12238 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12239 the crypto/md/ stuff).
12240 [Ralf S. Engelschall]
12242 *) More extension code. Incomplete support for subject and issuer alt
12243 name, issuer and authority key id. Change the i2v function parameters
12244 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12245 what that's for :-) Fix to ASN1 macro which messed up
12246 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12249 *) Preliminary support for ENUMERATED type. This is largely copied from the
12253 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12254 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12256 *) Make sure `make rehash' target really finds the `openssl' program.
12257 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12259 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12260 like to hear about it if this slows down other processors.
12263 *) Add CygWin32 platform information to Configure script.
12264 [Alan Batie <batie@aahz.jf.intel.com>]
12266 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12267 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12269 *) New program nseq to manipulate netscape certificate sequences
12272 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12276 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12277 but the BN code had some problems that would cause failures when
12278 doing certificate verification and some other functions.
12279 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12281 *) Add ASN1 and PEM code to support netscape certificate sequences.
12284 *) Add ASN1 and PEM code to support netscape certificate sequences.
12287 *) Add several PKIX and private extended key usage OIDs.
12290 *) Modify the 'ca' program to handle the new extension code. Modify
12291 openssl.cnf for new extension format, add comments.
12294 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12295 and add a sample to openssl.cnf so req -x509 now adds appropriate
12299 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12300 error code, add initial support to X509_print() and x509 application.
12303 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12304 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12305 stuff is currently isolated and isn't even compiled yet.
12308 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12309 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12310 Removed the versions check from X509 routines when loading extensions:
12311 this allows certain broken certificates that don't set the version
12312 properly to be processed.
12315 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12316 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12317 can still be regenerated with "make depend".
12320 *) Spelling mistake in C version of CAST-128.
12321 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12323 *) Changes to the error generation code. The perl script err-code.pl
12324 now reads in the old error codes and retains the old numbers, only
12325 adding new ones if necessary. It also only changes the .err files if new
12326 codes are added. The makefiles have been modified to only insert errors
12327 when needed (to avoid needlessly modifying header files). This is done
12328 by only inserting errors if the .err file is newer than the auto generated
12329 C file. To rebuild all the error codes from scratch (the old behaviour)
12330 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12331 or delete all the .err files.
12334 *) CAST-128 was incorrectly implemented for short keys. The C version has
12335 been fixed, but is untested. The assembler versions are also fixed, but
12336 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12337 to regenerate it if needed.
12338 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12339 Hagino <itojun@kame.net>]
12341 *) File was opened incorrectly in randfile.c.
12342 [Ulf Möller <ulf@fitug.de>]
12344 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12345 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12346 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12347 al: it's just almost always a UTCTime. Note this patch adds new error
12348 codes so do a "make errors" if there are problems.
12351 *) Correct Linux 1 recognition in config.
12352 [Ulf Möller <ulf@fitug.de>]
12354 *) Remove pointless MD5 hash when using DSA keys in ca.
12355 [Anonymous <nobody@replay.com>]
12357 *) Generate an error if given an empty string as a cert directory. Also
12358 generate an error if handed NULL (previously returned 0 to indicate an
12359 error, but didn't set one).
12360 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12362 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12365 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12366 parameters. This was causing a warning which killed off the Win32 compile.
12369 *) Remove C++ style comments from crypto/bn/bn_local.h.
12370 [Neil Costigan <neil.costigan@celocom.com>]
12372 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12373 based on a text string, looking up short and long names and finally
12374 "dot" format. The "dot" format stuff didn't work. Added new function
12375 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12376 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12377 OID is not part of the table.
12380 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12381 X509_LOOKUP_by_alias().
12384 *) Sort openssl functions by name.
12387 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12388 encryption from sample DSA keys (in case anyone is interested the password
12392 *) Make _all_ *_free functions accept a NULL pointer.
12393 [Frans Heymans <fheymans@isaserver.be>]
12395 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12397 [Anonymous <nobody@replay.com>]
12399 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12400 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12402 *) Don't blow it for numeric -newkey arguments to apps/req.
12403 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12405 *) Temp key "for export" tests were wrong in s3_srvr.c.
12406 [Anonymous <nobody@replay.com>]
12408 *) Add prototype for temp key callback functions
12409 SSL_CTX_set_tmp_{rsa,dh}_callback().
12412 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12413 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12416 *) X509_name_add_entry() freed the wrong thing after an error.
12417 [Arne Ansper <arne@ats.cyber.ee>]
12419 *) rsa_eay.c would attempt to free a NULL context.
12420 [Arne Ansper <arne@ats.cyber.ee>]
12422 *) BIO_s_socket() had a broken should_retry() on Windoze.
12423 [Arne Ansper <arne@ats.cyber.ee>]
12425 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12426 [Arne Ansper <arne@ats.cyber.ee>]
12428 *) Make sure the already existing X509_STORE->depth variable is initialized
12429 in X509_STORE_new(), but document the fact that this variable is still
12430 unused in the certificate verification process.
12431 [Ralf S. Engelschall]
12433 *) Fix the various library and apps files to free up pkeys obtained from
12434 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12437 *) Fix reference counting in X509_PUBKEY_get(). This makes
12438 demos/maurice/example2.c work, amongst others, probably.
12439 [Steve Henson and Ben Laurie]
12441 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12442 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12443 are no longer created. This way we have a single and consistent command
12444 line interface `openssl <command>', similar to `cvs <command>'.
12445 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12447 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12448 BIT STRING wrapper always have zero unused bits.
12451 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12454 *) Make the top-level INSTALL documentation easier to understand.
12457 *) Makefiles updated to exit if an error occurs in a sub-directory
12458 make (including if user presses ^C) [Paul Sutton]
12460 *) Make Montgomery context stuff explicit in RSA data structure.
12463 *) Fix build order of pem and err to allow for generated pem.h.
12466 *) Fix renumbering bug in X509_NAME_delete_entry().
12469 *) Enhanced the err-ins.pl script so it makes the error library number
12470 global and can add a library name. This is needed for external ASN1 and
12471 other error libraries.
12474 *) Fixed sk_insert which never worked properly.
12477 *) Fix ASN1 macros so they can handle indefinite length constructed
12478 EXPLICIT tags. Some non standard certificates use these: they can now
12482 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12483 into a single doc/ssleay.txt bundle. This way the information is still
12484 preserved but no longer messes up this directory. Now it's new room for
12485 the new set of documentation files.
12486 [Ralf S. Engelschall]
12488 *) SETs were incorrectly DER encoded. This was a major pain, because they
12489 shared code with SEQUENCEs, which aren't coded the same. This means that
12490 almost everything to do with SETs or SEQUENCEs has either changed name or
12491 number of arguments.
12492 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12494 *) Fix test data to work with the above.
12497 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12498 was already fixed by Eric for 0.9.1 it seems.
12499 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12501 *) Autodetect FreeBSD3.
12504 *) Fix various bugs in Configure. This affects the following platforms:
12508 unixware-2.0-pentium
12512 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12513 before they are needed.
12516 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12520 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12522 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12523 changed SSLeay to OpenSSL in version strings.
12524 [Ralf S. Engelschall]
12526 *) Some fixups to the top-level documents.
12529 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12530 because the symlink to include/ was missing.
12531 [Ralf S. Engelschall]
12533 *) Incorporated the popular no-RSA/DSA-only patches
12534 which allow to compile a RSA-free SSLeay.
12535 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12537 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12538 when "ssleay" is still not found.
12539 [Ralf S. Engelschall]
12541 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12542 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12544 *) Updated the README file.
12545 [Ralf S. Engelschall]
12547 *) Added various .cvsignore files in the CVS repository subdirs
12548 to make a "cvs update" really silent.
12549 [Ralf S. Engelschall]
12551 *) Recompiled the error-definition header files and added
12552 missing symbols to the Win32 linker tables.
12553 [Ralf S. Engelschall]
12555 *) Cleaned up the top-level documents;
12556 o new files: CHANGES and LICENSE
12557 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12558 o merged COPYRIGHT into LICENSE
12559 o removed obsolete TODO file
12560 o renamed MICROSOFT to INSTALL.W32
12561 [Ralf S. Engelschall]
12563 *) Removed dummy files from the 0.9.1b source tree:
12564 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12565 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12566 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12567 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12568 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12569 [Ralf S. Engelschall]
12571 *) Added various platform portability fixes.
12574 *) The Genesis of the OpenSSL rpject:
12575 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12576 Young and Tim J. Hudson created while they were working for C2Net until
12578 [The OpenSSL Project]
12581 Changes between 0.9.0b and 0.9.1b [not released]
12583 *) Updated a few CA certificates under certs/
12586 *) Changed some BIGNUM api stuff.
12589 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12590 DGUX x86, Linux Alpha, etc.
12593 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12594 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12598 *) Add -strparse option to asn1pars program which parses nested
12600 [Dr Stephen Henson <shenson@bigfoot.com>]
12602 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12605 *) DSA fix for "ca" program.
12608 *) Added "-genkey" option to "dsaparam" program.
12611 *) Added RIPE MD160 (rmd160) message digest.
12614 *) Added -a (all) option to "ssleay version" command.
12617 *) Added PLATFORM define which is the id given to Configure.
12620 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12623 *) Extended the ASN.1 parser routines.
12626 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12629 *) Added a BN_CTX to the BN library.
12632 *) Fixed the weak key values in DES library
12635 *) Changed API in EVP library for cipher aliases.
12638 *) Added support for RC2/64bit cipher.
12641 *) Converted the lhash library to the crypto/mem.c functions.
12644 *) Added more recognized ASN.1 object ids.
12647 *) Added more RSA padding checks for SSL/TLS.
12650 *) Added BIO proxy/filter functionality.
12653 *) Added extra_certs to SSL_CTX which can be used
12654 send extra CA certificates to the client in the CA cert chain sending
12655 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12658 *) Now Fortezza is denied in the authentication phase because
12659 this is key exchange mechanism is not supported by SSLeay at all.
12662 *) Additional PKCS1 checks.
12665 *) Support the string "TLSv1" for all TLS v1 ciphers.
12668 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12669 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12672 *) Fixed a few memory leaks.
12675 *) Fixed various code and comment typos.
12678 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12679 bytes sent in the client random.
12680 [Edward Bishop <ebishop@spyglass.com>]