5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the git commit log; for example,
7 https://github.com/openssl/openssl/commits/ and pick the appropriate
10 Changes between 1.1.0g and 1.1.1 [xx XXX xxxx]
12 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
13 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
14 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
15 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
16 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
17 configuraton has been separated out. See the ciphers man page or the
18 SSL_CTX_set_ciphersuites() man page for more information.
21 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
22 in responder mode now supports the new "-multi" option, which
23 spawns the specified number of child processes to handle OCSP
24 requests. The "-timeout" option now also limits the OCSP
25 responder's patience to wait to receive the full client request
26 on a newly accepted connection. Child processes are respawned
27 as needed, and the CA index file is automatically reloaded
28 when changed. This makes it possible to run the "ocsp" responder
29 as a long-running service, making the OpenSSL CA somewhat more
30 feature-complete. In this mode, most diagnostic messages logged
31 after entering the event loop are logged via syslog(3) rather than
35 *) Added support for X448 and Ed448. Heavily based on original work by
39 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
40 objects loaded. This adds the functions OSSL_STORE_expect() and
41 OSSL_STORE_find() as well as needed tools to construct searches and
42 get the search data out of them.
45 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
46 version of OpenSSL should review their configuration settings to ensure
47 that they are still appropriate for TLSv1.3. For further information see:
48 https://www.openssl.org/blog/blog/2018/02/08/tlsv1.3/
50 NOTE: In this pre-release of OpenSSL a draft version of the
51 TLSv1.3 standard has been implemented. Implementations of different draft
52 versions of the standard do not inter-operate, and this version will not
53 inter-operate with an implementation of the final standard when it is
54 eventually published. Different pre-release versions may implement
55 different versions of the draft. The final version of OpenSSL 1.1.1 will
56 implement the final version of the standard.
57 TODO(TLS1.3): Remove the above note before final release
60 *) Grand redesign of the OpenSSL random generator
62 The default RAND method now utilizes an AES-CTR DRBG according to
63 NIST standard SP 800-90Ar1. The new random generator is essentially
64 a port of the default random generator from the OpenSSL FIPS 2.0
65 object module. It is a hybrid deterministic random bit generator
66 using an AES-CTR bit stream and which seeds and reseeds itself
67 automatically using trusted system entropy sources.
69 Some of its new features are:
70 o Support for multiple DRBG instances with seed chaining.
71 o Add a public DRBG instance for the default RAND method.
72 o Add a dedicated DRBG instance for generating long term private keys.
73 o Make the DRBG instances fork-safe.
74 o Keep all global DRBG instances on the secure heap if it is enabled.
75 o Add a DRBG instance to every SSL instance for lock free operation
76 and to increase unpredictability.
77 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
79 *) Changed Configure so it only says what it does and doesn't dump
80 so much data. Instead, ./configdata.pm should be used as a script
81 to display all sorts of configuration data.
84 *) Added processing of "make variables" to Configure.
87 *) Added SHA512/224 and SHA512/256 algorithm support.
90 *) The last traces of Netware support, first removed in 1.1.0, have
94 *) Get rid of Makefile.shared, and in the process, make the processing
95 of certain files (rc.obj, or the .def/.map/.opt files produced from
96 the ordinal files) more visible and hopefully easier to trace and
97 debug (or make silent).
100 *) Make it possible to have environment variable assignments as
101 arguments to config / Configure.
104 *) Add multi-prime RSA (RFC 8017) support.
107 *) Add SM3 implemented according to GB/T 32905-2016
108 [ Jack Lloyd <jack.lloyd@ribose.com>,
109 Ronald Tse <ronald.tse@ribose.com>,
110 Erick Borsboom <erick.borsboom@ribose.com> ]
112 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
113 as documented in RFC6066.
114 Based on a patch from Tomasz Moń
115 [Filipe Raimundo da Silva]
117 *) Add SM4 implemented according to GB/T 32907-2016.
118 [ Jack Lloyd <jack.lloyd@ribose.com>,
119 Ronald Tse <ronald.tse@ribose.com>,
120 Erick Borsboom <erick.borsboom@ribose.com> ]
122 *) Reimplement -newreq-nodes and ERR_error_string_n; the
123 original author does not agree with the license change.
126 *) Add ARIA AEAD TLS support.
129 *) Some macro definitions to support VS6 have been removed. Visual
130 Studio 6 has not worked since 1.1.0
133 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
134 without clearing the errors.
137 *) Add "atfork" functions. If building on a system that without
138 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
139 requirements. The RAND facility now uses/requires this.
145 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
146 not possible to disable entirely. However, it's still possible to
147 disable the console reading UI method, UI_OpenSSL() (use UI_null()
150 To disable, configure with 'no-ui-console'. 'no-ui' is still
151 possible to use as an alias. Check at compile time with the
152 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
153 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
156 *) Add a STORE module, which implements a uniform and URI based reader of
157 stores that can contain keys, certificates, CRLs and numerous other
158 objects. The main API is loosely based on a few stdio functions,
159 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
160 OSSL_STORE_error and OSSL_STORE_close.
161 The implementation uses backends called "loaders" to implement arbitrary
162 URI schemes. There is one built in "loader" for the 'file' scheme.
165 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
166 then adjusted to work on FreeBSD 8.4 as well.
167 Enable by configuring with 'enable-devcryptoeng'. This is done by default
168 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
171 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
172 util/mkerr.pl, which is adapted to allow those prefixes, leading to
173 error code calls like this:
175 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
177 With this change, we claim the namespaces OSSL and OPENSSL in a manner
178 that can be encoded in C. For the foreseeable future, this will only
180 [Richard Levitte and Tim Hudson]
182 *) Removed BSD cryptodev engine.
185 *) Add a build target 'build_all_generated', to build all generated files
186 and only that. This can be used to prepare everything that requires
187 things like perl for a system that lacks perl and then move everything
188 to that system and do the rest of the build there.
191 *) In the UI interface, make it possible to duplicate the user data. This
192 can be used by engines that need to retain the data for a longer time
193 than just the call where this user data is passed.
196 *) Ignore the '-named_curve auto' value for compatibility of applications
198 [Tomas Mraz <tmraz@fedoraproject.org>]
200 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
201 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
202 alerts across multiple records (some of which could be empty). In practice
203 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
204 prohibts this altogether and other libraries (BoringSSL, NSS) do not
205 support this at all. Supporting it adds significant complexity to the
206 record layer, and its removal is unlikely to cause inter-operability
210 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
211 with Z. These are meant to replace LONG and ZLONG and to be size safe.
212 The use of LONG and ZLONG is discouraged and scheduled for deprecation
216 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
217 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
218 [Richard Levitte, Andy Polyakov]
220 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
224 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
225 platform rather than 'mingw'.
228 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
229 success if they are asked to add an object which already exists
230 in the store. This change cascades to other functions which load
231 certificates and CRLs.
234 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
235 facilitate stack unwinding even from assembly subroutines.
238 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
239 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
242 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
243 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
244 which is the minimum version we support.
247 *) Certificate time validation (X509_cmp_time) enforces stricter
248 compliance with RFC 5280. Fractional seconds and timezone offsets
249 are no longer allowed.
252 *) Add support for ARIA
255 *) s_client will now send the Server Name Indication (SNI) extension by
256 default unless the new "-noservername" option is used. The server name is
257 based on the host provided to the "-connect" option unless overridden by
261 *) Add support for SipHash
264 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
265 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
266 prevent issues where no progress is being made and the peer continually
267 sends unrecognised record types, using up resources processing them.
270 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
271 using the algorithm defined in
272 https://www.akkadia.org/drepper/SHA-crypt.txt
275 *) Heartbeat support has been removed; the ABI is changed for now.
276 [Richard Levitte, Rich Salz]
278 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
281 *) The RSA "null" method, which was partially supported to avoid patent
282 issues, has been replaced to always returns NULL.
285 Changes between 1.1.0g and 1.1.0h [xx XXX xxxx]
287 *) Removed the OS390-Unix config target. It relied on a script that doesn't
291 *) rsaz_1024_mul_avx2 overflow bug on x86_64
293 There is an overflow bug in the AVX2 Montgomery multiplication procedure
294 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
295 Analysis suggests that attacks against RSA and DSA as a result of this
296 defect would be very difficult to perform and are not believed likely.
297 Attacks against DH1024 are considered just feasible, because most of the
298 work necessary to deduce information about a private key may be performed
299 offline. The amount of resources required for such an attack would be
300 significant. However, for an attack on TLS to be meaningful, the server
301 would have to share the DH1024 private key among multiple clients, which is
302 no longer an option since CVE-2016-0701.
304 This only affects processors that support the AVX2 but not ADX extensions
305 like Intel Haswell (4th generation).
307 This issue was reported to OpenSSL by David Benjamin (Google). The issue
308 was originally found via the OSS-Fuzz project.
312 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
314 *) bn_sqrx8x_internal carry bug on x86_64
316 There is a carry propagating bug in the x86_64 Montgomery squaring
317 procedure. No EC algorithms are affected. Analysis suggests that attacks
318 against RSA and DSA as a result of this defect would be very difficult to
319 perform and are not believed likely. Attacks against DH are considered just
320 feasible (although very difficult) because most of the work necessary to
321 deduce information about a private key may be performed offline. The amount
322 of resources required for such an attack would be very significant and
323 likely only accessible to a limited number of attackers. An attacker would
324 additionally need online access to an unpatched system using the target
325 private key in a scenario with persistent DH parameters and a private
326 key that is shared between multiple clients.
328 This only affects processors that support the BMI1, BMI2 and ADX extensions
329 like Intel Broadwell (5th generation) and later or AMD Ryzen.
331 This issue was reported to OpenSSL by the OSS-Fuzz project.
335 *) Malformed X.509 IPAddressFamily could cause OOB read
337 If an X.509 certificate has a malformed IPAddressFamily extension,
338 OpenSSL could do a one-byte buffer overread. The most likely result
339 would be an erroneous display of the certificate in text format.
341 This issue was reported to OpenSSL by the OSS-Fuzz project.
345 Changes between 1.1.0e and 1.1.0f [25 May 2017]
347 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
348 platform rather than 'mingw'.
351 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
352 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
353 which is the minimum version we support.
356 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
358 *) Encrypt-Then-Mac renegotiation crash
360 During a renegotiation handshake if the Encrypt-Then-Mac extension is
361 negotiated where it was not in the original handshake (or vice-versa) then
362 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
363 and servers are affected.
365 This issue was reported to OpenSSL by Joe Orton (Red Hat).
369 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
371 *) Truncated packet could crash via OOB read
373 If one side of an SSL/TLS path is running on a 32-bit host and a specific
374 cipher is being used, then a truncated packet can cause that host to
375 perform an out-of-bounds read, usually resulting in a crash.
377 This issue was reported to OpenSSL by Robert Święcki of Google.
381 *) Bad (EC)DHE parameters cause a client crash
383 If a malicious server supplies bad parameters for a DHE or ECDHE key
384 exchange then this can result in the client attempting to dereference a
385 NULL pointer leading to a client crash. This could be exploited in a Denial
388 This issue was reported to OpenSSL by Guido Vranken.
392 *) BN_mod_exp may produce incorrect results on x86_64
394 There is a carry propagating bug in the x86_64 Montgomery squaring
395 procedure. No EC algorithms are affected. Analysis suggests that attacks
396 against RSA and DSA as a result of this defect would be very difficult to
397 perform and are not believed likely. Attacks against DH are considered just
398 feasible (although very difficult) because most of the work necessary to
399 deduce information about a private key may be performed offline. The amount
400 of resources required for such an attack would be very significant and
401 likely only accessible to a limited number of attackers. An attacker would
402 additionally need online access to an unpatched system using the target
403 private key in a scenario with persistent DH parameters and a private
404 key that is shared between multiple clients. For example this can occur by
405 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
406 similar to CVE-2015-3193 but must be treated as a separate problem.
408 This issue was reported to OpenSSL by the OSS-Fuzz project.
412 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
414 *) ChaCha20/Poly1305 heap-buffer-overflow
416 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
417 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
418 crash. This issue is not considered to be exploitable beyond a DoS.
420 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
424 *) CMS Null dereference
426 Applications parsing invalid CMS structures can crash with a NULL pointer
427 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
428 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
429 structure callback if an attempt is made to free certain invalid encodings.
430 Only CHOICE structures using a callback which do not handle NULL value are
433 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
437 *) Montgomery multiplication may produce incorrect results
439 There is a carry propagating bug in the Broadwell-specific Montgomery
440 multiplication procedure that handles input lengths divisible by, but
441 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
442 and DH private keys are impossible. This is because the subroutine in
443 question is not used in operations with the private key itself and an input
444 of the attacker's direct choice. Otherwise the bug can manifest itself as
445 transient authentication and key negotiation failures or reproducible
446 erroneous outcome of public-key operations with specially crafted input.
447 Among EC algorithms only Brainpool P-512 curves are affected and one
448 presumably can attack ECDH key negotiation. Impact was not analyzed in
449 detail, because pre-requisites for attack are considered unlikely. Namely
450 multiple clients have to choose the curve in question and the server has to
451 share the private key among them, neither of which is default behaviour.
452 Even then only clients that chose the curve will be affected.
454 This issue was publicly reported as transient failures and was not
455 initially recognized as a security issue. Thanks to Richard Morgan for
456 providing reproducible case.
460 *) Removed automatic addition of RPATH in shared libraries and executables,
461 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
464 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
466 *) Fix Use After Free for large message sizes
468 The patch applied to address CVE-2016-6307 resulted in an issue where if a
469 message larger than approx 16k is received then the underlying buffer to
470 store the incoming message is reallocated and moved. Unfortunately a
471 dangling pointer to the old location is left which results in an attempt to
472 write to the previously freed location. This is likely to result in a
473 crash, however it could potentially lead to execution of arbitrary code.
475 This issue only affects OpenSSL 1.1.0a.
477 This issue was reported to OpenSSL by Robert Święcki.
481 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
483 *) OCSP Status Request extension unbounded memory growth
485 A malicious client can send an excessively large OCSP Status Request
486 extension. If that client continually requests renegotiation, sending a
487 large OCSP Status Request extension each time, then there will be unbounded
488 memory growth on the server. This will eventually lead to a Denial Of
489 Service attack through memory exhaustion. Servers with a default
490 configuration are vulnerable even if they do not support OCSP. Builds using
491 the "no-ocsp" build time option are not affected.
493 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
497 *) SSL_peek() hang on empty record
499 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
500 sends an empty record. This could be exploited by a malicious peer in a
501 Denial Of Service attack.
503 This issue was reported to OpenSSL by Alex Gaynor.
507 *) Excessive allocation of memory in tls_get_message_header() and
508 dtls1_preprocess_fragment()
510 A (D)TLS message includes 3 bytes for its length in the header for the
511 message. This would allow for messages up to 16Mb in length. Messages of
512 this length are excessive and OpenSSL includes a check to ensure that a
513 peer is sending reasonably sized messages in order to avoid too much memory
514 being consumed to service a connection. A flaw in the logic of version
515 1.1.0 means that memory for the message is allocated too early, prior to
516 the excessive message length check. Due to way memory is allocated in
517 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
518 to service a connection. This could lead to a Denial of Service through
519 memory exhaustion. However, the excessive message length check still takes
520 place, and this would cause the connection to immediately fail. Assuming
521 that the application calls SSL_free() on the failed connection in a timely
522 manner then the 21Mb of allocated memory will then be immediately freed
523 again. Therefore the excessive memory allocation will be transitory in
524 nature. This then means that there is only a security impact if:
526 1) The application does not call SSL_free() in a timely manner in the event
527 that the connection fails
529 2) The application is working in a constrained environment where there is
530 very little free memory
532 3) The attacker initiates multiple connection attempts such that there are
533 multiple connections in a state where memory has been allocated for the
534 connection; SSL_free() has not yet been called; and there is insufficient
535 memory to service the multiple requests.
537 Except in the instance of (1) above any Denial Of Service is likely to be
538 transitory because as soon as the connection fails the memory is
539 subsequently freed again in the SSL_free() call. However there is an
540 increased risk during this period of application crashes due to the lack of
541 memory - which would then mean a more serious Denial of Service.
543 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
544 (CVE-2016-6307 and CVE-2016-6308)
547 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
548 had to be removed. Primary reason is that vendor assembler can't
549 assemble our modules with -KPIC flag. As result it, assembly
550 support, was not even available as option. But its lack means
551 lack of side-channel resistant code, which is incompatible with
552 security by todays standards. Fortunately gcc is readily available
553 prepackaged option, which we firmly point at...
556 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
558 *) Windows command-line tool supports UTF-8 opt-in option for arguments
559 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
560 (to any value) allows Windows user to access PKCS#12 file generated
561 with Windows CryptoAPI and protected with non-ASCII password, as well
562 as files generated under UTF-8 locale on Linux also protected with
566 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
567 have been disabled by default and removed from DEFAULT, just like RC4.
568 See the RC4 item below to re-enable both.
571 *) The method for finding the storage location for the Windows RAND seed file
572 has changed. First we check %RANDFILE%. If that is not set then we check
573 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
574 all else fails we fall back to C:\.
577 *) The EVP_EncryptUpdate() function has had its return type changed from void
578 to int. A return of 0 indicates and error while a return of 1 indicates
582 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
583 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
584 off the constant time implementation for RSA, DSA and DH have been made
585 no-ops and deprecated.
588 *) Windows RAND implementation was simplified to only get entropy by
589 calling CryptGenRandom(). Various other RAND-related tickets
591 [Joseph Wylie Yandle, Rich Salz]
593 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
594 and OPENSSL_LH_, respectively. The old names are available
595 with API compatibility. They new names are now completely documented.
598 *) Unify TYPE_up_ref(obj) methods signature.
599 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
600 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
601 int (instead of void) like all others TYPE_up_ref() methods.
602 So now these methods also check the return value of CRYPTO_atomic_add(),
603 and the validity of object reference counter.
604 [fdasilvayy@gmail.com]
606 *) With Windows Visual Studio builds, the .pdb files are installed
607 alongside the installed libraries and executables. For a static
608 library installation, ossl_static.pdb is the associate compiler
609 generated .pdb file to be used when linking programs.
612 *) Remove openssl.spec. Packaging files belong with the packagers.
615 *) Automatic Darwin/OSX configuration has had a refresh, it will now
616 recognise x86_64 architectures automatically. You can still decide
617 to build for a different bitness with the environment variable
618 KERNEL_BITS (can be 32 or 64), for example:
620 KERNEL_BITS=32 ./config
624 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
625 256 bit AES and HMAC with SHA256.
628 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
631 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
634 *) To enable users to have their own config files and build file templates,
635 Configure looks in the directory indicated by the environment variable
636 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
637 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
638 name and is used as is.
641 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
642 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
643 X509_CERT_FILE_CTX was removed.
646 *) "shared" builds are now the default. To create only static libraries use
647 the "no-shared" Configure option.
650 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
651 All of these option have not worked for some while and are fundamental
655 *) Make various cleanup routines no-ops and mark them as deprecated. Most
656 global cleanup functions are no longer required because they are handled
657 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
658 Explicitly de-initing can cause problems (e.g. where a library that uses
659 OpenSSL de-inits, but an application is still using it). The affected
660 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
661 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
662 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
666 *) --strict-warnings no longer enables runtime debugging options
667 such as REF_DEBUG. Instead, debug options are automatically
668 enabled with '--debug' builds.
669 [Andy Polyakov, Emilia Käsper]
671 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
672 have been moved out of the public header files. New functions for managing
673 these have been added.
676 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
677 objects have been moved out of the public header files. New
678 functions for managing these have been added.
681 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
682 have been moved out of the public header files. New functions for managing
683 these have been added.
686 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
687 moved out of the public header files. New functions for managing these
691 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
694 *) Removed the mk1mf build scripts.
697 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
698 it is always safe to #include a header now.
701 *) Removed the aged BC-32 config and all its supporting scripts
704 *) Removed support for Ultrix, Netware, and OS/2.
707 *) Add support for HKDF.
710 *) Add support for blake2b and blake2s
713 *) Added support for "pipelining". Ciphers that have the
714 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
715 encryptions/decryptions simultaneously. There are currently no built-in
716 ciphers with this property but the expectation is that engines will be able
717 to offer it to significantly improve throughput. Support has been extended
718 into libssl so that multiple records for a single connection can be
719 processed in one go (for >=TLS 1.1).
722 *) Added the AFALG engine. This is an async capable engine which is able to
723 offload work to the Linux kernel. In this initial version it only supports
724 AES128-CBC. The kernel must be version 4.1.0 or greater.
727 *) OpenSSL now uses a new threading API. It is no longer necessary to
728 set locking callbacks to use OpenSSL in a multi-threaded environment. There
729 are two supported threading models: pthreads and windows threads. It is
730 also possible to configure OpenSSL at compile time for "no-threads". The
731 old threading API should no longer be used. The functions have been
732 replaced with "no-op" compatibility macros.
733 [Alessandro Ghedini, Matt Caswell]
735 *) Modify behavior of ALPN to invoke callback after SNI/servername
736 callback, such that updates to the SSL_CTX affect ALPN.
739 *) Add SSL_CIPHER queries for authentication and key-exchange.
742 *) Changes to the DEFAULT cipherlist:
743 - Prefer (EC)DHE handshakes over plain RSA.
744 - Prefer AEAD ciphers over legacy ciphers.
745 - Prefer ECDSA over RSA when both certificates are available.
746 - Prefer TLSv1.2 ciphers/PRF.
747 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
751 *) Change the ECC default curve list to be this, in order: x25519,
752 secp256r1, secp521r1, secp384r1.
755 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
756 disabled by default. They can be re-enabled using the
757 enable-weak-ssl-ciphers option to Configure.
760 *) If the server has ALPN configured, but supports no protocols that the
761 client advertises, send a fatal "no_application_protocol" alert.
762 This behaviour is SHALL in RFC 7301, though it isn't universally
763 implemented by other servers.
766 *) Add X25519 support.
767 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
768 for public and private key encoding using the format documented in
769 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
770 key generation and key derivation.
772 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
776 *) Deprecate SRP_VBASE_get_by_user.
777 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
778 In order to fix an unavoidable memory leak (CVE-2016-0798),
779 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
780 seed, even if the seed is configured.
782 Users should use SRP_VBASE_get1_by_user instead. Note that in
783 SRP_VBASE_get1_by_user, caller must free the returned value. Note
784 also that even though configuring the SRP seed attempts to hide
785 invalid usernames by continuing the handshake with fake
786 credentials, this behaviour is not constant time and no strong
787 guarantees are made that the handshake is indistinguishable from
788 that of a valid user.
791 *) Configuration change; it's now possible to build dynamic engines
792 without having to build shared libraries and vice versa. This
793 only applies to the engines in engines/, those in crypto/engine/
794 will always be built into libcrypto (i.e. "static").
796 Building dynamic engines is enabled by default; to disable, use
797 the configuration option "disable-dynamic-engine".
799 The only requirements for building dynamic engines are the
800 presence of the DSO module and building with position independent
801 code, so they will also automatically be disabled if configuring
802 with "disable-dso" or "disable-pic".
804 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
805 are also taken away from openssl/opensslconf.h, as they are
809 *) Configuration change; if there is a known flag to compile
810 position independent code, it will always be applied on the
811 libcrypto and libssl object files, and never on the application
812 object files. This means other libraries that use routines from
813 libcrypto / libssl can be made into shared libraries regardless
814 of how OpenSSL was configured.
816 If this isn't desirable, the configuration options "disable-pic"
817 or "no-pic" can be used to disable the use of PIC. This will
818 also disable building shared libraries and dynamic engines.
821 *) Removed JPAKE code. It was experimental and has no wide use.
824 *) The INSTALL_PREFIX Makefile variable has been renamed to
825 DESTDIR. That makes for less confusion on what this variable
826 is for. Also, the configuration option --install_prefix is
830 *) Heartbeat for TLS has been removed and is disabled by default
831 for DTLS; configure with enable-heartbeats. Code that uses the
832 old #define's might need to be updated.
833 [Emilia Käsper, Rich Salz]
835 *) Rename REF_CHECK to REF_DEBUG.
838 *) New "unified" build system
840 The "unified" build system is aimed to be a common system for all
841 platforms we support. With it comes new support for VMS.
843 This system builds supports building in a different directory tree
844 than the source tree. It produces one Makefile (for unix family
845 or lookalikes), or one descrip.mms (for VMS).
847 The source of information to make the Makefile / descrip.mms is
848 small files called 'build.info', holding the necessary
849 information for each directory with source to compile, and a
850 template in Configurations, like unix-Makefile.tmpl or
853 With this change, the library names were also renamed on Windows
854 and on VMS. They now have names that are closer to the standard
855 on Unix, and include the major version number, and in certain
856 cases, the architecture they are built for. See "Notes on shared
857 libraries" in INSTALL.
859 We rely heavily on the perl module Text::Template.
862 *) Added support for auto-initialisation and de-initialisation of the library.
863 OpenSSL no longer requires explicit init or deinit routines to be called,
864 except in certain circumstances. See the OPENSSL_init_crypto() and
865 OPENSSL_init_ssl() man pages for further information.
868 *) The arguments to the DTLSv1_listen function have changed. Specifically the
869 "peer" argument is now expected to be a BIO_ADDR object.
871 *) Rewrite of BIO networking library. The BIO library lacked consistent
872 support of IPv6, and adding it required some more extensive
873 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
874 which hold all types of addresses and chains of address information.
875 It also introduces a new API, with functions like BIO_socket,
876 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
877 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
878 have been adapted accordingly.
881 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
885 *) CRIME protection: disable compression by default, even if OpenSSL is
886 compiled with zlib enabled. Applications can still enable compression
887 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
888 using the SSL_CONF library to configure compression.
891 *) The signature of the session callback configured with
892 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
893 was explicitly marked as 'const unsigned char*' instead of
897 *) Always DPURIFY. Remove the use of uninitialized memory in the
898 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
901 *) Removed many obsolete configuration items, including
902 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
903 MD2_CHAR, MD2_INT, MD2_LONG
905 IDEA_SHORT, IDEA_LONG
906 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
907 [Rich Salz, with advice from Andy Polyakov]
909 *) Many BN internals have been moved to an internal header file.
910 [Rich Salz with help from Andy Polyakov]
912 *) Configuration and writing out the results from it has changed.
913 Files such as Makefile include/openssl/opensslconf.h and are now
914 produced through general templates, such as Makefile.in and
915 crypto/opensslconf.h.in and some help from the perl module
918 Also, the center of configuration information is no longer
919 Makefile. Instead, Configure produces a perl module in
920 configdata.pm which holds most of the config data (in the hash
921 table %config), the target data that comes from the target
922 configuration in one of the Configurations/*.conf files (in
926 *) To clarify their intended purposes, the Configure options
927 --prefix and --openssldir change their semantics, and become more
928 straightforward and less interdependent.
930 --prefix shall be used exclusively to give the location INSTALLTOP
931 where programs, scripts, libraries, include files and manuals are
932 going to be installed. The default is now /usr/local.
934 --openssldir shall be used exclusively to give the default
935 location OPENSSLDIR where certificates, private keys, CRLs are
936 managed. This is also where the default openssl.cnf gets
938 If the directory given with this option is a relative path, the
939 values of both the --prefix value and the --openssldir value will
940 be combined to become OPENSSLDIR.
941 The default for --openssldir is INSTALLTOP/ssl.
943 Anyone who uses --openssldir to specify where OpenSSL is to be
944 installed MUST change to use --prefix instead.
947 *) The GOST engine was out of date and therefore it has been removed. An up
948 to date GOST engine is now being maintained in an external repository.
949 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
950 support for GOST ciphersuites (these are only activated if a GOST engine
954 *) EGD is no longer supported by default; use enable-egd when
956 [Ben Kaduk and Rich Salz]
958 *) The distribution now has Makefile.in files, which are used to
959 create Makefile's when Configure is run. *Configure must be run
960 before trying to build now.*
963 *) The return value for SSL_CIPHER_description() for error conditions
967 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
969 Obtaining and performing DNSSEC validation of TLSA records is
970 the application's responsibility. The application provides
971 the TLSA records of its choice to OpenSSL, and these are then
972 used to authenticate the peer.
974 The TLSA records need not even come from DNS. They can, for
975 example, be used to implement local end-entity certificate or
976 trust-anchor "pinning", where the "pin" data takes the form
977 of TLSA records, which can augment or replace verification
978 based on the usual WebPKI public certification authorities.
981 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
982 continues to support deprecated interfaces in default builds.
983 However, applications are strongly advised to compile their
984 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
985 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
986 or the 1.1.0 releases.
988 In environments in which all applications have been ported to
989 not use any deprecated interfaces OpenSSL's Configure script
990 should be used with the --api=1.1.0 option to entirely remove
991 support for the deprecated features from the library and
992 unconditionally disable them in the installed headers.
993 Essentially the same effect can be achieved with the "no-deprecated"
994 argument to Configure, except that this will always restrict
995 the build to just the latest API, rather than a fixed API
998 As applications are ported to future revisions of the API,
999 they should update their compile-time OPENSSL_API_COMPAT define
1000 accordingly, but in most cases should be able to continue to
1001 compile with later releases.
1003 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1004 0x10000000L and 0x00908000L, respectively. However those
1005 versions did not support the OPENSSL_API_COMPAT feature, and
1006 so applications are not typically tested for explicit support
1007 of just the undeprecated features of either release.
1010 *) Add support for setting the minimum and maximum supported protocol.
1011 It can bet set via the SSL_set_min_proto_version() and
1012 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1013 MaxProtocol. It's recommended to use the new APIs to disable
1014 protocols instead of disabling individual protocols using
1015 SSL_set_options() or SSL_CONF's Protocol. This change also
1016 removes support for disabling TLS 1.2 in the OpenSSL TLS
1017 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1020 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1023 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1024 and integrates ECDSA and ECDH functionality into EC. Implementations can
1025 now redirect key generation and no longer need to convert to or from
1028 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1029 include the ec.h header file instead.
1032 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1033 ciphers who are no longer supported and drops support the ephemeral RSA key
1034 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1037 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1038 opaque. For HMAC_CTX, the following constructors and destructors
1041 HMAC_CTX *HMAC_CTX_new(void);
1042 void HMAC_CTX_free(HMAC_CTX *ctx);
1044 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1045 destroy such methods has been added. See EVP_MD_meth_new(3) and
1046 EVP_CIPHER_meth_new(3) for documentation.
1049 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1050 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1051 EVP_MD_CTX_reset() should be called instead to reinitialise
1052 an already created structure.
1053 2) For consistency with the majority of our object creators and
1054 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1055 EVP_MD_CTX_(new|free). The old names are retained as macros
1056 for deprecated builds.
1059 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1060 cryptographic operations to be performed asynchronously as long as an
1061 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1062 further details. Libssl has also had this capability integrated with the
1063 introduction of the new mode SSL_MODE_ASYNC and associated error
1064 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1065 pages. This work was developed in partnership with Intel Corp.
1068 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1069 always enabled now. If you want to disable the support you should
1070 exclude it using the list of supported ciphers. This also means that the
1071 "-no_ecdhe" option has been removed from s_server.
1074 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1075 SSL_{CTX_}set1_curves() which can set a list.
1078 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1079 curve you want to support using SSL_{CTX_}set1_curves().
1082 *) State machine rewrite. The state machine code has been significantly
1083 refactored in order to remove much duplication of code and solve issues
1084 with the old code (see ssl/statem/README for further details). This change
1085 does have some associated API changes. Notably the SSL_state() function
1086 has been removed and replaced by SSL_get_state which now returns an
1087 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1088 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1092 *) All instances of the string "ssleay" in the public API were replaced
1093 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1094 Some error codes related to internal RSA_eay API's were renamed.
1097 *) The demo files in crypto/threads were moved to demo/threads.
1100 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1102 [Matt Caswell, Rich Salz]
1104 *) New ASN.1 embed macro.
1106 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1107 structure is not allocated: it is part of the parent. That is instead of
1115 This reduces memory fragmentation and make it impossible to accidentally
1116 set a mandatory field to NULL.
1118 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1119 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1120 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1124 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1127 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1128 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1129 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1130 DES and RC4 ciphersuites.
1133 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1134 This changes the decoding behaviour for some invalid messages,
1135 though the change is mostly in the more lenient direction, and
1136 legacy behaviour is preserved as much as possible.
1139 *) Fix no-stdio build.
1140 [ David Woodhouse <David.Woodhouse@intel.com> and also
1141 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1143 *) New testing framework
1144 The testing framework has been largely rewritten and is now using
1145 perl and the perl modules Test::Harness and an extended variant of
1146 Test::More called OpenSSL::Test to do its work. All test scripts in
1147 test/ have been rewritten into test recipes, and all direct calls to
1148 executables in test/Makefile have become individual recipes using the
1149 simplified testing OpenSSL::Test::Simple.
1151 For documentation on our testing modules, do:
1153 perldoc test/testlib/OpenSSL/Test/Simple.pm
1154 perldoc test/testlib/OpenSSL/Test.pm
1158 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1159 are used; the latter aborts on memory leaks (usually checked on exit).
1160 Some undocumented "set malloc, etc., hooks" functions were removed
1161 and others were changed. All are now documented.
1164 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1166 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1168 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1169 from RFC4279, RFC4785, RFC5487, RFC5489.
1171 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1172 original RSA_PSK patch.
1175 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1176 era flag was never set throughout the codebase (only read). Also removed
1177 SSL3_FLAGS_POP_BUFFER which was only used if
1178 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1181 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1182 to be "oneline" instead of "compat".
1185 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1186 not aware of clients that still exhibit this bug, and the workaround
1187 hasn't been working properly for a while.
1190 *) The return type of BIO_number_read() and BIO_number_written() as well as
1191 the corresponding num_read and num_write members in the BIO structure has
1192 changed from unsigned long to uint64_t. On platforms where an unsigned
1193 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1197 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1198 OpenSSL without support for them. It also means that maintaining
1199 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1200 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1203 *) Removed support for the two export grade static DH ciphersuites
1204 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1205 were newly added (along with a number of other static DH ciphersuites) to
1206 1.0.2. However the two export ones have *never* worked since they were
1207 introduced. It seems strange in any case to be adding new export
1208 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1211 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1212 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1213 and turned into macros which simply call the new preferred function names
1214 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1215 should use the new names instead. Also as part of this change the ssl23.h
1216 header file has been removed.
1219 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1220 code and the associated standard is no longer considered fit-for-purpose.
1223 *) RT2547 was closed. When generating a private key, try to make the
1224 output file readable only by the owner. This behavior change might
1225 be noticeable when interacting with other software.
1227 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1231 *) Added HTTP GET support to the ocsp command.
1234 *) Changed default digest for the dgst and enc commands from MD5 to
1238 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1241 *) Added support for TLS extended master secret from
1242 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1243 initial patch which was a great help during development.
1246 *) All libssl internal structures have been removed from the public header
1247 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1248 now redundant). Users should not attempt to access internal structures
1249 directly. Instead they should use the provided API functions.
1252 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1253 Access to deprecated functions can be re-enabled by running config with
1254 "enable-deprecated". In addition applications wishing to use deprecated
1255 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1256 will, by default, disable some transitive includes that previously existed
1257 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1260 *) Added support for OCB mode. OpenSSL has been granted a patent license
1261 compatible with the OpenSSL license for use of OCB. Details are available
1262 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1263 for OCB can be removed by calling config with no-ocb.
1266 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1267 compatible client hello.
1270 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1271 done while fixing the error code for the key-too-small case.
1272 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1274 *) CA.sh has been removed; use CA.pl instead.
1277 *) Removed old DES API.
1280 *) Remove various unsupported platforms:
1286 Sinix/ReliantUNIX RM400
1291 16-bit platforms such as WIN16
1294 *) Clean up OPENSSL_NO_xxx #define's
1295 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1296 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1297 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1298 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1299 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1300 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1301 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1302 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1303 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1304 Remove MS_STATIC; it's a relic from platforms <32 bits.
1307 *) Cleaned up dead code
1308 Remove all but one '#ifdef undef' which is to be looked at.
1311 *) Clean up calling of xxx_free routines.
1312 Just like free(), fix most of the xxx_free routines to accept
1313 NULL. Remove the non-null checks from callers. Save much code.
1316 *) Add secure heap for storage of private keys (when possible).
1317 Add BIO_s_secmem(), CBIGNUM, etc.
1318 Contributed by Akamai Technologies under our Corporate CLA.
1321 *) Experimental support for a new, fast, unbiased prime candidate generator,
1322 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1323 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1325 *) New output format NSS in the sess_id command line tool. This allows
1326 exporting the session id and the master key in NSS keylog format.
1327 [Martin Kaiser <martin@kaiser.cx>]
1329 *) Harmonize version and its documentation. -f flag is used to display
1331 [mancha <mancha1@zoho.com>]
1333 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1334 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1335 [mancha <mancha1@zoho.com>]
1337 *) Fix some double frees. These are not thought to be exploitable.
1338 [mancha <mancha1@zoho.com>]
1340 *) A missing bounds check in the handling of the TLS heartbeat extension
1341 can be used to reveal up to 64k of memory to a connected client or
1344 Thanks for Neel Mehta of Google Security for discovering this bug and to
1345 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1346 preparing the fix (CVE-2014-0160)
1347 [Adam Langley, Bodo Moeller]
1349 *) Fix for the attack described in the paper "Recovering OpenSSL
1350 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1351 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1352 http://eprint.iacr.org/2014/140
1354 Thanks to Yuval Yarom and Naomi Benger for discovering this
1355 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1356 [Yuval Yarom and Naomi Benger]
1358 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1359 this fixes a limitation in previous versions of OpenSSL.
1362 *) Experimental encrypt-then-mac support.
1364 Experimental support for encrypt then mac from
1365 draft-gutmann-tls-encrypt-then-mac-02.txt
1367 To enable it set the appropriate extension number (0x42 for the test
1368 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1370 For non-compliant peers (i.e. just about everything) this should have no
1373 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1377 *) Add EVP support for key wrapping algorithms, to avoid problems with
1378 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1379 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1380 algorithms and include tests cases.
1383 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1387 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1388 MGF1 digest and OAEP label.
1391 *) Make openssl verify return errors.
1392 [Chris Palmer <palmer@google.com> and Ben Laurie]
1394 *) New function ASN1_TIME_diff to calculate the difference between two
1395 ASN1_TIME structures or one structure and the current time.
1398 *) Update fips_test_suite to support multiple command line options. New
1399 test to induce all self test errors in sequence and check expected
1403 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1404 sign or verify all in one operation.
1407 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1408 test programs and fips_test_suite. Includes functionality to parse
1409 the minimal script output of fipsalgest.pl directly.
1412 *) Add authorisation parameter to FIPS_module_mode_set().
1415 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1418 *) Use separate DRBG fields for internal and external flags. New function
1419 FIPS_drbg_health_check() to perform on demand health checking. Add
1420 generation tests to fips_test_suite with reduced health check interval to
1421 demonstrate periodic health checking. Add "nodh" option to
1422 fips_test_suite to skip very slow DH test.
1425 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1429 *) More extensive health check for DRBG checking many more failure modes.
1430 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1431 combination: call this in fips_test_suite.
1434 *) Add support for canonical generation of DSA parameter 'g'. See
1437 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1438 POST to handle HMAC cases.
1441 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1442 to return numerical and string versions of the FIPS module number.
1445 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1446 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1447 outside the validated module in the FIPS capable OpenSSL.
1450 *) Minor change to DRBG entropy callback semantics. In some cases
1451 there is no multiple of the block length between min_len and
1452 max_len. Allow the callback to return more than max_len bytes
1453 of entropy but discard any extra: it is the callback's responsibility
1454 to ensure that the extra data discarded does not impact the
1455 requested amount of entropy.
1458 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1459 information in FIPS186-3, SP800-57 and SP800-131A.
1462 *) CCM support via EVP. Interface is very similar to GCM case except we
1463 must supply all data in one chunk (i.e. no update, final) and the
1464 message length must be supplied if AAD is used. Add algorithm test
1468 *) Initial version of POST overhaul. Add POST callback to allow the status
1469 of POST to be monitored and/or failures induced. Modify fips_test_suite
1470 to use callback. Always run all selftests even if one fails.
1473 *) XTS support including algorithm test driver in the fips_gcmtest program.
1474 Note: this does increase the maximum key length from 32 to 64 bytes but
1475 there should be no binary compatibility issues as existing applications
1476 will never use XTS mode.
1479 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1480 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1481 performs algorithm blocking for unapproved PRNG types. Also do not
1482 set PRNG type in FIPS_mode_set(): leave this to the application.
1483 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1484 the standard OpenSSL PRNG: set additional data to a date time vector.
1487 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1488 This shouldn't present any incompatibility problems because applications
1489 shouldn't be using these directly and any that are will need to rethink
1490 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1493 *) Extensive self tests and health checking required by SP800-90 DRBG.
1494 Remove strength parameter from FIPS_drbg_instantiate and always
1495 instantiate at maximum supported strength.
1498 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1501 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1504 *) New function DH_compute_key_padded() to compute a DH key and pad with
1505 leading zeroes if needed: this complies with SP800-56A et al.
1508 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1509 anything, incomplete, subject to change and largely untested at present.
1512 *) Modify fipscanisteronly build option to only build the necessary object
1513 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1516 *) Add experimental option FIPSSYMS to give all symbols in
1517 fipscanister.o and FIPS or fips prefix. This will avoid
1518 conflicts with future versions of OpenSSL. Add perl script
1519 util/fipsas.pl to preprocess assembly language source files
1520 and rename any affected symbols.
1523 *) Add selftest checks and algorithm block of non-fips algorithms in
1524 FIPS mode. Remove DES2 from selftests.
1527 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1528 return internal method without any ENGINE dependencies. Add new
1529 tiny fips sign and verify functions.
1532 *) New build option no-ec2m to disable characteristic 2 code.
1535 *) New build option "fipscanisteronly". This only builds fipscanister.o
1536 and (currently) associated fips utilities. Uses the file Makefile.fips
1537 instead of Makefile.org as the prototype.
1540 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1541 Update fips_gcmtest to use IV generator.
1544 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1545 setting output buffer to NULL. The *Final function must be
1546 called although it will not retrieve any additional data. The tag
1547 can be set or retrieved with a ctrl. The IV length is by default 12
1548 bytes (96 bits) but can be set to an alternative value. If the IV
1549 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1553 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1554 underlying do_cipher function handles all cipher semantics itself
1555 including padding and finalisation. This is useful if (for example)
1556 an ENGINE cipher handles block padding itself. The behaviour of
1557 do_cipher is subtly changed if this flag is set: the return value
1558 is the number of characters written to the output buffer (zero is
1559 no longer an error code) or a negative error code. Also if the
1560 input buffer is NULL and length 0 finalisation should be performed.
1563 *) If a candidate issuer certificate is already part of the constructed
1564 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1567 *) Improve forward-security support: add functions
1569 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1570 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1572 for use by SSL/TLS servers; the callback function will be called whenever a
1573 new session is created, and gets to decide whether the session may be
1574 cached to make it resumable (return 0) or not (return 1). (As by the
1575 SSL/TLS protocol specifications, the session_id sent by the server will be
1576 empty to indicate that the session is not resumable; also, the server will
1577 not generate RFC 4507 (RFC 5077) session tickets.)
1579 A simple reasonable callback implementation is to return is_forward_secure.
1580 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1581 by the SSL/TLS server library, indicating whether it can provide forward
1583 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1585 *) New -verify_name option in command line utilities to set verification
1589 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1590 Add CMAC pkey methods.
1593 *) Experimental renegotiation in s_server -www mode. If the client
1594 browses /reneg connection is renegotiated. If /renegcert it is
1595 renegotiated requesting a certificate.
1598 *) Add an "external" session cache for debugging purposes to s_server. This
1599 should help trace issues which normally are only apparent in deployed
1600 multi-process servers.
1603 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1604 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1605 BIO_set_cipher() and some obscure PEM functions were changed so they
1606 can now return an error. The RAND changes required a change to the
1607 RAND_METHOD structure.
1610 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1611 a gcc attribute to warn if the result of a function is ignored. This
1612 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1613 whose return value is often ignored.
1616 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1617 These allow SCTs (signed certificate timestamps) to be requested and
1618 validated when establishing a connection.
1619 [Rob Percival <robpercival@google.com>]
1621 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1623 *) Prevent padding oracle in AES-NI CBC MAC check
1625 A MITM attacker can use a padding oracle attack to decrypt traffic
1626 when the connection uses an AES CBC cipher and the server support
1629 This issue was introduced as part of the fix for Lucky 13 padding
1630 attack (CVE-2013-0169). The padding check was rewritten to be in
1631 constant time by making sure that always the same bytes are read and
1632 compared against either the MAC or padding bytes. But it no longer
1633 checked that there was enough data to have both the MAC and padding
1636 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1640 *) Fix EVP_EncodeUpdate overflow
1642 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1643 Base64 encoding of binary data. If an attacker is able to supply very large
1644 amounts of input data then a length check can overflow resulting in a heap
1647 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1648 the PEM_write_bio* family of functions. These are mainly used within the
1649 OpenSSL command line applications, so any application which processes data
1650 from an untrusted source and outputs it as a PEM file should be considered
1651 vulnerable to this issue. User applications that call these APIs directly
1652 with large amounts of untrusted data may also be vulnerable.
1654 This issue was reported by Guido Vranken.
1658 *) Fix EVP_EncryptUpdate overflow
1660 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1661 is able to supply very large amounts of input data after a previous call to
1662 EVP_EncryptUpdate() with a partial block then a length check can overflow
1663 resulting in a heap corruption. Following an analysis of all OpenSSL
1664 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1665 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1666 the first called function after an EVP_EncryptInit(), and therefore that
1667 specific call must be safe. The second form is where the length passed to
1668 EVP_EncryptUpdate() can be seen from the code to be some small value and
1669 therefore there is no possibility of an overflow. Since all instances are
1670 one of these two forms, it is believed that there can be no overflows in
1671 internal code due to this problem. It should be noted that
1672 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1673 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1674 of these calls have also been analysed too and it is believed there are no
1675 instances in internal usage where an overflow could occur.
1677 This issue was reported by Guido Vranken.
1681 *) Prevent ASN.1 BIO excessive memory allocation
1683 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1684 a short invalid encoding can cause allocation of large amounts of memory
1685 potentially consuming excessive resources or exhausting memory.
1687 Any application parsing untrusted data through d2i BIO functions is
1688 affected. The memory based functions such as d2i_X509() are *not* affected.
1689 Since the memory based functions are used by the TLS library, TLS
1690 applications are not affected.
1692 This issue was reported by Brian Carpenter.
1698 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1699 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1700 in arbitrary stack data being returned in the buffer.
1702 This issue was reported by Guido Vranken.
1706 *) Modify behavior of ALPN to invoke callback after SNI/servername
1707 callback, such that updates to the SSL_CTX affect ALPN.
1710 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1714 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1715 methods are enabled and ssl2 is disabled the methods return NULL.
1718 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1720 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1721 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1722 provide any "EXPORT" or "LOW" strength ciphers.
1725 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1726 is by default disabled at build-time. Builds that are not configured with
1727 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1728 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1729 will need to explicitly call either of:
1731 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1733 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1735 as appropriate. Even if either of those is used, or the application
1736 explicitly uses the version-specific SSLv2_method() or its client and
1737 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1738 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1739 ciphers, and SSLv2 56-bit DES are no longer available.
1743 *) Fix a double-free in DSA code
1745 A double free bug was discovered when OpenSSL parses malformed DSA private
1746 keys and could lead to a DoS attack or memory corruption for applications
1747 that receive DSA private keys from untrusted sources. This scenario is
1750 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1755 *) Disable SRP fake user seed to address a server memory leak.
1757 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1759 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1760 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1761 was changed to ignore the "fake user" SRP seed, even if the seed
1764 Users should use SRP_VBASE_get1_by_user instead. Note that in
1765 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1766 also that even though configuring the SRP seed attempts to hide
1767 invalid usernames by continuing the handshake with fake
1768 credentials, this behaviour is not constant time and no strong
1769 guarantees are made that the handshake is indistinguishable from
1770 that of a valid user.
1774 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1776 In the BN_hex2bn function the number of hex digits is calculated using an
1777 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1778 large values of |i| this can result in |bn_expand| not allocating any
1779 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1780 field as NULL leading to a subsequent NULL ptr deref. For very large values
1781 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1782 In this case memory is allocated to the internal BIGNUM data field, but it
1783 is insufficiently sized leading to heap corruption. A similar issue exists
1784 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1785 is ever called by user applications with very large untrusted hex/dec data.
1786 This is anticipated to be a rare occurrence.
1788 All OpenSSL internal usage of these functions use data that is not expected
1789 to be untrusted, e.g. config file data or application command line
1790 arguments. If user developed applications generate config file data based
1791 on untrusted data then it is possible that this could also lead to security
1792 consequences. This is also anticipated to be rare.
1794 This issue was reported to OpenSSL by Guido Vranken.
1798 *) Fix memory issues in BIO_*printf functions
1800 The internal |fmtstr| function used in processing a "%s" format string in
1801 the BIO_*printf functions could overflow while calculating the length of a
1802 string and cause an OOB read when printing very long strings.
1804 Additionally the internal |doapr_outch| function can attempt to write to an
1805 OOB memory location (at an offset from the NULL pointer) in the event of a
1806 memory allocation failure. In 1.0.2 and below this could be caused where
1807 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1808 could be in processing a very long "%s" format string. Memory leaks can
1811 The first issue may mask the second issue dependent on compiler behaviour.
1812 These problems could enable attacks where large amounts of untrusted data
1813 is passed to the BIO_*printf functions. If applications use these functions
1814 in this way then they could be vulnerable. OpenSSL itself uses these
1815 functions when printing out human-readable dumps of ASN.1 data. Therefore
1816 applications that print this data could be vulnerable if the data is from
1817 untrusted sources. OpenSSL command line applications could also be
1818 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1819 as command line arguments.
1821 Libssl is not considered directly vulnerable. Additionally certificates etc
1822 received via remote connections via libssl are also unlikely to be able to
1823 trigger these issues because of message size limits enforced within libssl.
1825 This issue was reported to OpenSSL Guido Vranken.
1829 *) Side channel attack on modular exponentiation
1831 A side-channel attack was found which makes use of cache-bank conflicts on
1832 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1833 of RSA keys. The ability to exploit this issue is limited as it relies on
1834 an attacker who has control of code in a thread running on the same
1835 hyper-threaded core as the victim thread which is performing decryptions.
1837 This issue was reported to OpenSSL by Yuval Yarom, The University of
1838 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1839 Nadia Heninger, University of Pennsylvania with more information at
1840 http://cachebleed.info.
1844 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1845 if no keysize is specified with default_bits. This fixes an
1846 omission in an earlier change that changed all RSA/DSA key generation
1847 apps to use 2048 bits by default.
1850 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1851 *) DH small subgroups
1853 Historically OpenSSL only ever generated DH parameters based on "safe"
1854 primes. More recently (in version 1.0.2) support was provided for
1855 generating X9.42 style parameter files such as those required for RFC 5114
1856 support. The primes used in such files may not be "safe". Where an
1857 application is using DH configured with parameters based on primes that are
1858 not "safe" then an attacker could use this fact to find a peer's private
1859 DH exponent. This attack requires that the attacker complete multiple
1860 handshakes in which the peer uses the same private DH exponent. For example
1861 this could be used to discover a TLS server's private DH exponent if it's
1862 reusing the private DH exponent or it's using a static DH ciphersuite.
1864 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1865 TLS. It is not on by default. If the option is not set then the server
1866 reuses the same private DH exponent for the life of the server process and
1867 would be vulnerable to this attack. It is believed that many popular
1868 applications do set this option and would therefore not be at risk.
1870 The fix for this issue adds an additional check where a "q" parameter is
1871 available (as is the case in X9.42 based parameters). This detects the
1872 only known attack, and is the only possible defense for static DH
1873 ciphersuites. This could have some performance impact.
1875 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1876 default and cannot be disabled. This could have some performance impact.
1878 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1882 *) SSLv2 doesn't block disabled ciphers
1884 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1885 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1886 been disabled, provided that the SSLv2 protocol was not also disabled via
1889 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1890 and Sebastian Schinzel.
1894 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1896 *) BN_mod_exp may produce incorrect results on x86_64
1898 There is a carry propagating bug in the x86_64 Montgomery squaring
1899 procedure. No EC algorithms are affected. Analysis suggests that attacks
1900 against RSA and DSA as a result of this defect would be very difficult to
1901 perform and are not believed likely. Attacks against DH are considered just
1902 feasible (although very difficult) because most of the work necessary to
1903 deduce information about a private key may be performed offline. The amount
1904 of resources required for such an attack would be very significant and
1905 likely only accessible to a limited number of attackers. An attacker would
1906 additionally need online access to an unpatched system using the target
1907 private key in a scenario with persistent DH parameters and a private
1908 key that is shared between multiple clients. For example this can occur by
1909 default in OpenSSL DHE based SSL/TLS ciphersuites.
1911 This issue was reported to OpenSSL by Hanno Böck.
1915 *) Certificate verify crash with missing PSS parameter
1917 The signature verification routines will crash with a NULL pointer
1918 dereference if presented with an ASN.1 signature using the RSA PSS
1919 algorithm and absent mask generation function parameter. Since these
1920 routines are used to verify certificate signature algorithms this can be
1921 used to crash any certificate verification operation and exploited in a
1922 DoS attack. Any application which performs certificate verification is
1923 vulnerable including OpenSSL clients and servers which enable client
1926 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1930 *) X509_ATTRIBUTE memory leak
1932 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1933 memory. This structure is used by the PKCS#7 and CMS routines so any
1934 application which reads PKCS#7 or CMS data from untrusted sources is
1935 affected. SSL/TLS is not affected.
1937 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1942 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1943 This changes the decoding behaviour for some invalid messages,
1944 though the change is mostly in the more lenient direction, and
1945 legacy behaviour is preserved as much as possible.
1948 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1950 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1952 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1954 *) Alternate chains certificate forgery
1956 During certificate verification, OpenSSL will attempt to find an
1957 alternative certificate chain if the first attempt to build such a chain
1958 fails. An error in the implementation of this logic can mean that an
1959 attacker could cause certain checks on untrusted certificates to be
1960 bypassed, such as the CA flag, enabling them to use a valid leaf
1961 certificate to act as a CA and "issue" an invalid certificate.
1963 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1967 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1969 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1970 incompatibility in the handling of HMAC. The previous ABI has now been
1974 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1976 *) Malformed ECParameters causes infinite loop
1978 When processing an ECParameters structure OpenSSL enters an infinite loop
1979 if the curve specified is over a specially malformed binary polynomial
1982 This can be used to perform denial of service against any
1983 system which processes public keys, certificate requests or
1984 certificates. This includes TLS clients and TLS servers with
1985 client authentication enabled.
1987 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1991 *) Exploitable out-of-bounds read in X509_cmp_time
1993 X509_cmp_time does not properly check the length of the ASN1_TIME
1994 string and can read a few bytes out of bounds. In addition,
1995 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1998 An attacker can use this to craft malformed certificates and CRLs of
1999 various sizes and potentially cause a segmentation fault, resulting in
2000 a DoS on applications that verify certificates or CRLs. TLS clients
2001 that verify CRLs are affected. TLS clients and servers with client
2002 authentication enabled may be affected if they use custom verification
2005 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2006 independently by Hanno Böck.
2010 *) PKCS7 crash with missing EnvelopedContent
2012 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2013 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2014 with missing content and trigger a NULL pointer dereference on parsing.
2016 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2017 structures from untrusted sources are affected. OpenSSL clients and
2018 servers are not affected.
2020 This issue was reported to OpenSSL by Michal Zalewski (Google).
2024 *) CMS verify infinite loop with unknown hash function
2026 When verifying a signedData message the CMS code can enter an infinite loop
2027 if presented with an unknown hash function OID. This can be used to perform
2028 denial of service against any system which verifies signedData messages using
2030 This issue was reported to OpenSSL by Johannes Bauer.
2034 *) Race condition handling NewSessionTicket
2036 If a NewSessionTicket is received by a multi-threaded client when attempting to
2037 reuse a previous ticket then a race condition can occur potentially leading to
2038 a double free of the ticket data.
2042 *) Only support 256-bit or stronger elliptic curves with the
2043 'ecdh_auto' setting (server) or by default (client). Of supported
2044 curves, prefer P-256 (both).
2047 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2049 *) ClientHello sigalgs DoS fix
2051 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2052 invalid signature algorithms extension a NULL pointer dereference will
2053 occur. This can be exploited in a DoS attack against the server.
2055 This issue was was reported to OpenSSL by David Ramos of Stanford
2058 [Stephen Henson and Matt Caswell]
2060 *) Multiblock corrupted pointer fix
2062 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2063 feature only applies on 64 bit x86 architecture platforms that support AES
2064 NI instructions. A defect in the implementation of "multiblock" can cause
2065 OpenSSL's internal write buffer to become incorrectly set to NULL when
2066 using non-blocking IO. Typically, when the user application is using a
2067 socket BIO for writing, this will only result in a failed connection.
2068 However if some other BIO is used then it is likely that a segmentation
2069 fault will be triggered, thus enabling a potential DoS attack.
2071 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2075 *) Segmentation fault in DTLSv1_listen fix
2077 The DTLSv1_listen function is intended to be stateless and processes the
2078 initial ClientHello from many peers. It is common for user code to loop
2079 over the call to DTLSv1_listen until a valid ClientHello is received with
2080 an associated cookie. A defect in the implementation of DTLSv1_listen means
2081 that state is preserved in the SSL object from one invocation to the next
2082 that can lead to a segmentation fault. Errors processing the initial
2083 ClientHello can trigger this scenario. An example of such an error could be
2084 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2087 This issue was reported to OpenSSL by Per Allansson.
2091 *) Segmentation fault in ASN1_TYPE_cmp fix
2093 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2094 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2095 certificate signature algorithm consistency this can be used to crash any
2096 certificate verification operation and exploited in a DoS attack. Any
2097 application which performs certificate verification is vulnerable including
2098 OpenSSL clients and servers which enable client authentication.
2102 *) Segmentation fault for invalid PSS parameters fix
2104 The signature verification routines will crash with a NULL pointer
2105 dereference if presented with an ASN.1 signature using the RSA PSS
2106 algorithm and invalid parameters. Since these routines are used to verify
2107 certificate signature algorithms this can be used to crash any
2108 certificate verification operation and exploited in a DoS attack. Any
2109 application which performs certificate verification is vulnerable including
2110 OpenSSL clients and servers which enable client authentication.
2112 This issue was was reported to OpenSSL by Brian Carpenter.
2116 *) ASN.1 structure reuse memory corruption fix
2118 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2119 memory corruption via an invalid write. Such reuse is and has been
2120 strongly discouraged and is believed to be rare.
2122 Applications that parse structures containing CHOICE or ANY DEFINED BY
2123 components may be affected. Certificate parsing (d2i_X509 and related
2124 functions) are however not affected. OpenSSL clients and servers are
2129 *) PKCS7 NULL pointer dereferences fix
2131 The PKCS#7 parsing code does not handle missing outer ContentInfo
2132 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2133 missing content and trigger a NULL pointer dereference on parsing.
2135 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2136 otherwise parse PKCS#7 structures from untrusted sources are
2137 affected. OpenSSL clients and servers are not affected.
2139 This issue was reported to OpenSSL by Michal Zalewski (Google).
2143 *) DoS via reachable assert in SSLv2 servers fix
2145 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2146 servers that both support SSLv2 and enable export cipher suites by sending
2147 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2149 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2150 (OpenSSL development team).
2154 *) Empty CKE with client auth and DHE fix
2156 If client auth is used then a server can seg fault in the event of a DHE
2157 ciphersuite being selected and a zero length ClientKeyExchange message
2158 being sent by the client. This could be exploited in a DoS attack.
2162 *) Handshake with unseeded PRNG fix
2164 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2165 with an unseeded PRNG. The conditions are:
2166 - The client is on a platform where the PRNG has not been seeded
2167 automatically, and the user has not seeded manually
2168 - A protocol specific client method version has been used (i.e. not
2169 SSL_client_methodv23)
2170 - A ciphersuite is used that does not require additional random data from
2171 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2173 If the handshake succeeds then the client random that has been used will
2174 have been generated from a PRNG with insufficient entropy and therefore the
2175 output may be predictable.
2177 For example using the following command with an unseeded openssl will
2178 succeed on an unpatched platform:
2180 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2184 *) Use After Free following d2i_ECPrivatekey error fix
2186 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2187 could cause a use after free condition. This, in turn, could cause a double
2188 free in several private key parsing functions (such as d2i_PrivateKey
2189 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2190 for applications that receive EC private keys from untrusted
2191 sources. This scenario is considered rare.
2193 This issue was discovered by the BoringSSL project and fixed in their
2198 *) X509_to_X509_REQ NULL pointer deref fix
2200 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2201 the certificate key is invalid. This function is rarely used in practice.
2203 This issue was discovered by Brian Carpenter.
2207 *) Removed the export ciphers from the DEFAULT ciphers
2210 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2212 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2213 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2214 So far those who have to target multiple platforms would compromise
2215 and argue that binary targeting say ARMv5 would still execute on
2216 ARMv8. "Universal" build resolves this compromise by providing
2217 near-optimal performance even on newer platforms.
2220 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2221 (other platforms pending).
2222 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2224 *) Add support for the SignedCertificateTimestampList certificate and
2225 OCSP response extensions from RFC6962.
2228 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2229 for corner cases. (Certain input points at infinity could lead to
2230 bogus results, with non-infinity inputs mapped to infinity too.)
2233 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2234 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2235 common cases are optimized and there still is room for further
2236 improvements. Vector Permutation AES for Altivec is also added.
2239 *) Add support for little-endian ppc64 Linux target.
2240 [Marcelo Cerri (IBM)]
2242 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2243 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2244 are optimized and there still is room for further improvements.
2245 Both 32- and 64-bit modes are supported.
2246 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2248 *) Improved ARMv7 NEON support.
2251 *) Support for SPARC Architecture 2011 crypto extensions, first
2252 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2253 SHA256/512, MD5, GHASH and modular exponentiation.
2254 [Andy Polyakov, David Miller]
2256 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2258 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2260 *) Support for new and upcoming Intel processors, including AVX2,
2261 BMI and SHA ISA extensions. This includes additional "stitched"
2262 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2265 This work was sponsored by Intel Corp.
2268 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2269 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2270 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2273 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2274 this fixes a limitation in previous versions of OpenSSL.
2277 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2278 MGF1 digest and OAEP label.
2281 *) Add EVP support for key wrapping algorithms, to avoid problems with
2282 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2283 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2284 algorithms and include tests cases.
2287 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2289 [Douglas E. Engert, Steve Henson]
2291 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2292 difference in days and seconds between two tm or ASN1_TIME structures.
2295 *) Add -rev test option to s_server to just reverse order of characters
2296 received by client and send back to server. Also prints an abbreviated
2297 summary of the connection parameters.
2300 *) New option -brief for s_client and s_server to print out a brief summary
2301 of connection parameters.
2304 *) Add callbacks for arbitrary TLS extensions.
2305 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2307 *) New option -crl_download in several openssl utilities to download CRLs
2308 from CRLDP extension in certificates.
2311 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2314 *) New function X509_CRL_diff to generate a delta CRL from the difference
2315 of two full CRLs. Add support to "crl" utility.
2318 *) New functions to set lookup_crls function and to retrieve
2319 X509_STORE from X509_STORE_CTX.
2322 *) Print out deprecated issuer and subject unique ID fields in
2326 *) Extend OCSP I/O functions so they can be used for simple general purpose
2327 HTTP as well as OCSP. New wrapper function which can be used to download
2328 CRLs using the OCSP API.
2331 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2334 *) SSL_CONF* functions. These provide a common framework for application
2335 configuration using configuration files or command lines.
2338 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2339 message callback and prints the results. Needs compile time option
2340 "enable-ssl-trace". New options to s_client and s_server to enable
2344 *) New ctrl and macro to retrieve supported points extensions.
2345 Print out extension in s_server and s_client.
2348 *) New functions to retrieve certificate signature and signature
2352 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2356 *) New Suite B modes for TLS code. These use and enforce the requirements
2357 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2358 only use Suite B curves. The Suite B modes can be set by using the
2359 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2362 *) New chain verification flags for Suite B levels of security. Check
2363 algorithms are acceptable when flags are set in X509_verify_cert.
2366 *) Make tls1_check_chain return a set of flags indicating checks passed
2367 by a certificate chain. Add additional tests to handle client
2368 certificates: checks for matching certificate type and issuer name
2372 *) If an attempt is made to use a signature algorithm not in the peer
2373 preference list abort the handshake. If client has no suitable
2374 signature algorithms in response to a certificate request do not
2375 use the certificate.
2378 *) If server EC tmp key is not in client preference list abort handshake.
2381 *) Add support for certificate stores in CERT structure. This makes it
2382 possible to have different stores per SSL structure or one store in
2383 the parent SSL_CTX. Include distinct stores for certificate chain
2384 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2385 to build and store a certificate chain in CERT structure: returning
2386 an error if the chain cannot be built: this will allow applications
2387 to test if a chain is correctly configured.
2389 Note: if the CERT based stores are not set then the parent SSL_CTX
2390 store is used to retain compatibility with existing behaviour.
2394 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2395 mask based on the current session, check mask when sending client
2396 hello and checking the requested ciphersuite.
2399 *) New ctrls to retrieve and set certificate types in a certificate
2400 request message. Print out received values in s_client. If certificate
2401 types is not set with custom values set sensible values based on
2402 supported signature algorithms.
2405 *) Support for distinct client and server supported signature algorithms.
2408 *) Add certificate callback. If set this is called whenever a certificate
2409 is required by client or server. An application can decide which
2410 certificate chain to present based on arbitrary criteria: for example
2411 supported signature algorithms. Add very simple example to s_server.
2412 This fixes many of the problems and restrictions of the existing client
2413 certificate callback: for example you can now clear an existing
2414 certificate and specify the whole chain.
2417 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2418 the certificate can be used for (if anything). Set valid_flags field
2419 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2420 to have similar checks in it.
2422 Add new "cert_flags" field to CERT structure and include a "strict mode".
2423 This enforces some TLS certificate requirements (such as only permitting
2424 certificate signature algorithms contained in the supported algorithms
2425 extension) which some implementations ignore: this option should be used
2426 with caution as it could cause interoperability issues.
2429 *) Update and tidy signature algorithm extension processing. Work out
2430 shared signature algorithms based on preferences and peer algorithms
2431 and print them out in s_client and s_server. Abort handshake if no
2432 shared signature algorithms.
2435 *) Add new functions to allow customised supported signature algorithms
2436 for SSL and SSL_CTX structures. Add options to s_client and s_server
2440 *) New function SSL_certs_clear() to delete all references to certificates
2441 from an SSL structure. Before this once a certificate had been added
2442 it couldn't be removed.
2445 *) Integrate hostname, email address and IP address checking with certificate
2446 verification. New verify options supporting checking in openssl utility.
2449 *) Fixes and wildcard matching support to hostname and email checking
2450 functions. Add manual page.
2451 [Florian Weimer (Red Hat Product Security Team)]
2453 *) New functions to check a hostname email or IP address against a
2454 certificate. Add options x509 utility to print results of checks against
2458 *) Fix OCSP checking.
2459 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2461 *) Initial experimental support for explicitly trusted non-root CAs.
2462 OpenSSL still tries to build a complete chain to a root but if an
2463 intermediate CA has a trust setting included that is used. The first
2464 setting is used: whether to trust (e.g., -addtrust option to the x509
2468 *) Add -trusted_first option which attempts to find certificates in the
2469 trusted store even if an untrusted chain is also supplied.
2472 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2473 platform support for Linux and Android.
2476 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2479 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2480 When in FIPS mode the approved implementations are used as normal,
2481 when not in FIPS mode the internal unapproved versions are used instead.
2482 This means that the FIPS capable OpenSSL isn't forced to use the
2483 (often lower performance) FIPS implementations outside FIPS mode.
2486 *) Transparently support X9.42 DH parameters when calling
2487 PEM_read_bio_DHparameters. This means existing applications can handle
2488 the new parameter format automatically.
2491 *) Initial experimental support for X9.42 DH parameter format: mainly
2492 to support use of 'q' parameter for RFC5114 parameters.
2495 *) Add DH parameters from RFC5114 including test data to dhtest.
2498 *) Support for automatic EC temporary key parameter selection. If enabled
2499 the most preferred EC parameters are automatically used instead of
2500 hardcoded fixed parameters. Now a server just has to call:
2501 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2502 support ECDH and use the most appropriate parameters.
2505 *) Enhance and tidy EC curve and point format TLS extension code. Use
2506 static structures instead of allocation if default values are used.
2507 New ctrls to set curves we wish to support and to retrieve shared curves.
2508 Print out shared curves in s_server. New options to s_server and s_client
2509 to set list of supported curves.
2512 *) New ctrls to retrieve supported signature algorithms and
2513 supported curve values as an array of NIDs. Extend openssl utility
2514 to print out received values.
2517 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2518 between NIDs and the more common NIST names such as "P-256". Enhance
2519 ecparam utility and ECC method to recognise the NIST names for curves.
2522 *) Enhance SSL/TLS certificate chain handling to support different
2523 chains for each certificate instead of one chain in the parent SSL_CTX.
2526 *) Support for fixed DH ciphersuite client authentication: where both
2527 server and client use DH certificates with common parameters.
2530 *) Support for fixed DH ciphersuites: those requiring DH server
2534 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2536 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2537 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2538 X509_CINF_get_signature were reverted post internal team review.
2540 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2542 *) Build fixes for the Windows and OpenVMS platforms
2543 [Matt Caswell and Richard Levitte]
2545 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2547 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2548 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2549 dereference. This could lead to a Denial Of Service attack. Thanks to
2550 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2554 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2555 dtls1_buffer_record function under certain conditions. In particular this
2556 could occur if an attacker sent repeated DTLS records with the same
2557 sequence number but for the next epoch. The memory leak could be exploited
2558 by an attacker in a Denial of Service attack through memory exhaustion.
2559 Thanks to Chris Mueller for reporting this issue.
2563 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2564 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2565 method would be set to NULL which could later result in a NULL pointer
2566 dereference. Thanks to Frank Schmirler for reporting this issue.
2570 *) Abort handshake if server key exchange message is omitted for ephemeral
2573 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2574 reporting this issue.
2578 *) Remove non-export ephemeral RSA code on client and server. This code
2579 violated the TLS standard by allowing the use of temporary RSA keys in
2580 non-export ciphersuites and could be used by a server to effectively
2581 downgrade the RSA key length used to a value smaller than the server
2582 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2583 INRIA or reporting this issue.
2587 *) Fixed issue where DH client certificates are accepted without verification.
2588 An OpenSSL server will accept a DH certificate for client authentication
2589 without the certificate verify message. This effectively allows a client to
2590 authenticate without the use of a private key. This only affects servers
2591 which trust a client certificate authority which issues certificates
2592 containing DH keys: these are extremely rare and hardly ever encountered.
2593 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2598 *) Ensure that the session ID context of an SSL is updated when its
2599 SSL_CTX is updated via SSL_set_SSL_CTX.
2601 The session ID context is typically set from the parent SSL_CTX,
2602 and can vary with the CTX.
2605 *) Fix various certificate fingerprint issues.
2607 By using non-DER or invalid encodings outside the signed portion of a
2608 certificate the fingerprint can be changed without breaking the signature.
2609 Although no details of the signed portion of the certificate can be changed
2610 this can cause problems with some applications: e.g. those using the
2611 certificate fingerprint for blacklists.
2613 1. Reject signatures with non zero unused bits.
2615 If the BIT STRING containing the signature has non zero unused bits reject
2616 the signature. All current signature algorithms require zero unused bits.
2618 2. Check certificate algorithm consistency.
2620 Check the AlgorithmIdentifier inside TBS matches the one in the
2621 certificate signature. NB: this will result in signature failure
2622 errors for some broken certificates.
2624 Thanks to Konrad Kraszewski from Google for reporting this issue.
2626 3. Check DSA/ECDSA signatures use DER.
2628 Re-encode DSA/ECDSA signatures and compare with the original received
2629 signature. Return an error if there is a mismatch.
2631 This will reject various cases including garbage after signature
2632 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2633 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2634 (negative or with leading zeroes).
2636 Further analysis was conducted and fixes were developed by Stephen Henson
2637 of the OpenSSL core team.
2642 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2643 results on some platforms, including x86_64. This bug occurs at random
2644 with a very low probability, and is not known to be exploitable in any
2645 way, though its exact impact is difficult to determine. Thanks to Pieter
2646 Wuille (Blockstream) who reported this issue and also suggested an initial
2647 fix. Further analysis was conducted by the OpenSSL development team and
2648 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2649 the OpenSSL core team.
2653 *) Do not resume sessions on the server if the negotiated protocol
2654 version does not match the session's version. Resuming with a different
2655 version, while not strictly forbidden by the RFC, is of questionable
2656 sanity and breaks all known clients.
2657 [David Benjamin, Emilia Käsper]
2659 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2660 early CCS messages during renegotiation. (Note that because
2661 renegotiation is encrypted, this early CCS was not exploitable.)
2664 *) Tighten client-side session ticket handling during renegotiation:
2665 ensure that the client only accepts a session ticket if the server sends
2666 the extension anew in the ServerHello. Previously, a TLS client would
2667 reuse the old extension state and thus accept a session ticket if one was
2668 announced in the initial ServerHello.
2670 Similarly, ensure that the client requires a session ticket if one
2671 was advertised in the ServerHello. Previously, a TLS client would
2672 ignore a missing NewSessionTicket message.
2675 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2677 *) SRTP Memory Leak.
2679 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2680 sends a carefully crafted handshake message, to cause OpenSSL to fail
2681 to free up to 64k of memory causing a memory leak. This could be
2682 exploited in a Denial Of Service attack. This issue affects OpenSSL
2683 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2684 whether SRTP is used or configured. Implementations of OpenSSL that
2685 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2687 The fix was developed by the OpenSSL team.
2691 *) Session Ticket Memory Leak.
2693 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2694 integrity of that ticket is first verified. In the event of a session
2695 ticket integrity check failing, OpenSSL will fail to free memory
2696 causing a memory leak. By sending a large number of invalid session
2697 tickets an attacker could exploit this issue in a Denial Of Service
2702 *) Build option no-ssl3 is incomplete.
2704 When OpenSSL is configured with "no-ssl3" as a build option, servers
2705 could accept and complete a SSL 3.0 handshake, and clients could be
2706 configured to send them.
2708 [Akamai and the OpenSSL team]
2710 *) Add support for TLS_FALLBACK_SCSV.
2711 Client applications doing fallback retries should call
2712 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2714 [Adam Langley, Bodo Moeller]
2716 *) Add additional DigestInfo checks.
2718 Re-encode DigestInto in DER and check against the original when
2719 verifying RSA signature: this will reject any improperly encoded
2720 DigestInfo structures.
2722 Note: this is a precautionary measure and no attacks are currently known.
2726 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2728 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2729 SRP code can be overrun an internal buffer. Add sanity check that
2730 g, A, B < N to SRP code.
2732 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2733 Group for discovering this issue.
2737 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2738 TLS 1.0 instead of higher protocol versions when the ClientHello message
2739 is badly fragmented. This allows a man-in-the-middle attacker to force a
2740 downgrade to TLS 1.0 even if both the server and the client support a
2741 higher protocol version, by modifying the client's TLS records.
2743 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2744 researching this issue.
2748 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2749 to a denial of service attack. A malicious server can crash the client
2750 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2751 ciphersuite and sending carefully crafted handshake messages.
2753 Thanks to Felix Gröbert (Google) for discovering and researching this
2758 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2759 to leak memory. This can be exploited through a Denial of Service attack.
2760 Thanks to Adam Langley for discovering and researching this issue.
2764 *) An attacker can force openssl to consume large amounts of memory whilst
2765 processing DTLS handshake messages. This can be exploited through a
2766 Denial of Service attack.
2767 Thanks to Adam Langley for discovering and researching this issue.
2771 *) An attacker can force an error condition which causes openssl to crash
2772 whilst processing DTLS packets due to memory being freed twice. This
2773 can be exploited through a Denial of Service attack.
2774 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2779 *) If a multithreaded client connects to a malicious server using a resumed
2780 session and the server sends an ec point format extension it could write
2781 up to 255 bytes to freed memory.
2783 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2788 *) A malicious server can crash an OpenSSL client with a null pointer
2789 dereference (read) by specifying an SRP ciphersuite even though it was not
2790 properly negotiated with the client. This can be exploited through a
2791 Denial of Service attack.
2793 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2794 discovering and researching this issue.
2798 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2799 X509_name_oneline, X509_name_print_ex et al. to leak some information
2800 from the stack. Applications may be affected if they echo pretty printing
2801 output to the attacker.
2803 Thanks to Ivan Fratric (Google) for discovering this issue.
2805 [Emilia Käsper, and Steve Henson]
2807 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2808 for corner cases. (Certain input points at infinity could lead to
2809 bogus results, with non-infinity inputs mapped to infinity too.)
2812 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2814 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2815 handshake can force the use of weak keying material in OpenSSL
2816 SSL/TLS clients and servers.
2818 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2819 researching this issue. (CVE-2014-0224)
2820 [KIKUCHI Masashi, Steve Henson]
2822 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2823 OpenSSL DTLS client the code can be made to recurse eventually crashing
2826 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2828 [Imre Rad, Steve Henson]
2830 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2831 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2832 client or server. This is potentially exploitable to run arbitrary
2833 code on a vulnerable client or server.
2835 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2836 [Jüri Aedla, Steve Henson]
2838 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2839 are subject to a denial of service attack.
2841 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2842 this issue. (CVE-2014-3470)
2843 [Felix Gröbert, Ivan Fratric, Steve Henson]
2845 *) Harmonize version and its documentation. -f flag is used to display
2847 [mancha <mancha1@zoho.com>]
2849 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2850 in i2d_ECPrivateKey.
2851 [mancha <mancha1@zoho.com>]
2853 *) Fix some double frees. These are not thought to be exploitable.
2854 [mancha <mancha1@zoho.com>]
2856 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2858 *) A missing bounds check in the handling of the TLS heartbeat extension
2859 can be used to reveal up to 64k of memory to a connected client or
2862 Thanks for Neel Mehta of Google Security for discovering this bug and to
2863 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2864 preparing the fix (CVE-2014-0160)
2865 [Adam Langley, Bodo Moeller]
2867 *) Fix for the attack described in the paper "Recovering OpenSSL
2868 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2869 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2870 http://eprint.iacr.org/2014/140
2872 Thanks to Yuval Yarom and Naomi Benger for discovering this
2873 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2874 [Yuval Yarom and Naomi Benger]
2876 *) TLS pad extension: draft-agl-tls-padding-03
2878 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2879 TLS client Hello record length value would otherwise be > 255 and
2880 less that 512 pad with a dummy extension containing zeroes so it
2881 is at least 512 bytes long.
2883 [Adam Langley, Steve Henson]
2885 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2887 *) Fix for TLS record tampering bug. A carefully crafted invalid
2888 handshake could crash OpenSSL with a NULL pointer exception.
2889 Thanks to Anton Johansson for reporting this issues.
2892 *) Keep original DTLS digest and encryption contexts in retransmission
2893 structures so we can use the previous session parameters if they need
2894 to be resent. (CVE-2013-6450)
2897 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2898 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2899 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2900 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2901 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2902 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2903 [Rob Stradling, Adam Langley]
2905 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2907 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2908 supporting platforms or when small records were transferred.
2909 [Andy Polyakov, Steve Henson]
2911 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2913 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2915 This addresses the flaw in CBC record processing discovered by
2916 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2917 at: http://www.isg.rhul.ac.uk/tls/
2919 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2920 Security Group at Royal Holloway, University of London
2921 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2922 Emilia Käsper for the initial patch.
2924 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2926 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2927 ciphersuites which can be exploited in a denial of service attack.
2928 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2929 and detecting this bug and to Wolfgang Ettlinger
2930 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2934 *) Return an error when checking OCSP signatures when key is NULL.
2935 This fixes a DoS attack. (CVE-2013-0166)
2938 *) Make openssl verify return errors.
2939 [Chris Palmer <palmer@google.com> and Ben Laurie]
2941 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2942 the right response is stapled. Also change SSL_get_certificate()
2943 so it returns the certificate actually sent.
2944 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2945 [Rob Stradling <rob.stradling@comodo.com>]
2947 *) Fix possible deadlock when decoding public keys.
2950 *) Don't use TLS 1.0 record version number in initial client hello
2954 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2956 *) Sanity check record length before skipping explicit IV in TLS
2957 1.2, 1.1 and DTLS to fix DoS attack.
2959 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2960 fuzzing as a service testing platform.
2964 *) Initialise tkeylen properly when encrypting CMS messages.
2965 Thanks to Solar Designer of Openwall for reporting this issue.
2968 *) In FIPS mode don't try to use composite ciphers as they are not
2972 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2974 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2975 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2976 mean any application compiled against OpenSSL 1.0.0 headers setting
2977 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
2978 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2979 0x10000000L Any application which was previously compiled against
2980 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2981 will need to be recompiled as a result. Letting be results in
2982 inability to disable specifically TLS 1.1 and in client context,
2983 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2986 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
2987 disable just protocol X, but all protocols above X *if* there are
2988 protocols *below* X still enabled. In more practical terms it means
2989 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2990 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2991 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2995 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2997 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2998 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2999 in CRYPTO_realloc_clean.
3001 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3002 issue and to Adam Langley <agl@chromium.org> for fixing it.
3004 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3006 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3009 *) Workarounds for some broken servers that "hang" if a client hello
3010 record length exceeds 255 bytes.
3012 1. Do not use record version number > TLS 1.0 in initial client
3013 hello: some (but not all) hanging servers will now work.
3014 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3015 the number of ciphers sent in the client hello. This should be
3016 set to an even number, such as 50, for example by passing:
3017 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3018 Most broken servers should now work.
3019 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3020 TLS 1.2 client support entirely.
3023 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3026 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3028 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3029 STRING form instead of a DigestInfo.
3032 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3033 and the RSA_sign/RSA_verify functions. This was made more apparent when
3034 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3035 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3036 the correct format in RSA_verify so both forms transparently work.
3039 *) Some servers which support TLS 1.0 can choke if we initially indicate
3040 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3041 encrypted premaster secret. As a workaround use the maximum permitted
3042 client version in client hello, this should keep such servers happy
3043 and still work with previous versions of OpenSSL.
3046 *) Add support for TLS/DTLS heartbeats.
3047 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3049 *) Add support for SCTP.
3050 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3052 *) Improved PRNG seeding for VOS.
3053 [Paul Green <Paul.Green@stratus.com>]
3055 *) Extensive assembler packs updates, most notably:
3057 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3058 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3059 - x86_64: bit-sliced AES implementation;
3060 - ARM: NEON support, contemporary platforms optimizations;
3061 - s390x: z196 support;
3062 - *: GHASH and GF(2^m) multiplication implementations;
3066 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3067 (removal of unnecessary code)
3068 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3070 *) Add TLS key material exporter from RFC 5705.
3073 *) Add DTLS-SRTP negotiation from RFC 5764.
3076 *) Add Next Protocol Negotiation,
3077 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3078 disabled with a no-npn flag to config or Configure. Code donated
3080 [Adam Langley <agl@google.com> and Ben Laurie]
3082 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3083 NIST-P256, NIST-P521, with constant-time single point multiplication on
3084 typical inputs. Compiler support for the nonstandard type __uint128_t is
3085 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3086 Code made available under Apache License version 2.0.
3088 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3089 line to include this in your build of OpenSSL, and run "make depend" (or
3090 "make update"). This enables the following EC_METHODs:
3092 EC_GFp_nistp224_method()
3093 EC_GFp_nistp256_method()
3094 EC_GFp_nistp521_method()
3096 EC_GROUP_new_by_curve_name() will automatically use these (while
3097 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3099 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3101 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3102 all platforms. Move ssize_t definition from e_os.h to the public
3103 header file e_os2.h as it now appears in public header file cms.h
3106 *) New -sigopt option to the ca, req and x509 utilities. Additional
3107 signature parameters can be passed using this option and in
3111 *) Add RSA PSS signing function. This will generate and set the
3112 appropriate AlgorithmIdentifiers for PSS based on those in the
3113 corresponding EVP_MD_CTX structure. No application support yet.
3116 *) Support for companion algorithm specific ASN1 signing routines.
3117 New function ASN1_item_sign_ctx() signs a pre-initialised
3118 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3119 the appropriate parameters.
3122 *) Add new algorithm specific ASN1 verification initialisation function
3123 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3124 handling will be the same no matter what EVP_PKEY_METHOD is used.
3125 Add a PSS handler to support verification of PSS signatures: checked
3126 against a number of sample certificates.
3129 *) Add signature printing for PSS. Add PSS OIDs.
3130 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3132 *) Add algorithm specific signature printing. An individual ASN1 method
3133 can now print out signatures instead of the standard hex dump.
3135 More complex signatures (e.g. PSS) can print out more meaningful
3136 information. Include DSA version that prints out the signature
3140 *) Password based recipient info support for CMS library: implementing
3144 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3145 neatly separates the code into cipher and PBE sections and is required
3146 for some algorithms that split PBES2 into separate pieces (such as
3147 password based CMS).
3150 *) Session-handling fixes:
3151 - Fix handling of connections that are resuming with a session ID,
3152 but also support Session Tickets.
3153 - Fix a bug that suppressed issuing of a new ticket if the client
3154 presented a ticket with an expired session.
3155 - Try to set the ticket lifetime hint to something reasonable.
3156 - Make tickets shorter by excluding irrelevant information.
3157 - On the client side, don't ignore renewed tickets.
3158 [Adam Langley, Bodo Moeller (Google)]
3160 *) Fix PSK session representation.
3163 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3165 This work was sponsored by Intel.
3168 *) Add GCM support to TLS library. Some custom code is needed to split
3169 the IV between the fixed (from PRF) and explicit (from TLS record)
3170 portions. This adds all GCM ciphersuites supported by RFC5288 and
3171 RFC5289. Generalise some AES* cipherstrings to include GCM and
3172 add a special AESGCM string for GCM only.
3175 *) Expand range of ctrls for AES GCM. Permit setting invocation
3176 field on decrypt and retrieval of invocation field only on encrypt.
3179 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3180 As required by RFC5289 these ciphersuites cannot be used if for
3181 versions of TLS earlier than 1.2.
3184 *) For FIPS capable OpenSSL interpret a NULL default public key method
3185 as unset and return the appropriate default but do *not* set the default.
3186 This means we can return the appropriate method in applications that
3187 switch between FIPS and non-FIPS modes.
3190 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3191 ENGINE is used then we cannot handle that in the FIPS module so we
3192 keep original code iff non-FIPS operations are allowed.
3195 *) Add -attime option to openssl utilities.
3196 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3198 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3201 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3202 FIPS EC methods unconditionally for now.
3205 *) New build option no-ec2m to disable characteristic 2 code.
3208 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3209 all cases can be covered as some introduce binary incompatibilities.
3212 *) Redirect RSA operations to FIPS module including keygen,
3213 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3216 *) Add similar low level API blocking to ciphers.
3219 *) Low level digest APIs are not approved in FIPS mode: any attempt
3220 to use these will cause a fatal error. Applications that *really* want
3221 to use them can use the private_* version instead.
3224 *) Redirect cipher operations to FIPS module for FIPS builds.
3227 *) Redirect digest operations to FIPS module for FIPS builds.
3230 *) Update build system to add "fips" flag which will link in fipscanister.o
3231 for static and shared library builds embedding a signature if needed.
3234 *) Output TLS supported curves in preference order instead of numerical
3235 order. This is currently hardcoded for the highest order curves first.
3236 This should be configurable so applications can judge speed vs strength.
3239 *) Add TLS v1.2 server support for client authentication.
3242 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3246 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3247 FIPS modules versions.
3250 *) Add TLS v1.2 client side support for client authentication. Keep cache
3251 of handshake records longer as we don't know the hash algorithm to use
3252 until after the certificate request message is received.
3255 *) Initial TLS v1.2 client support. Add a default signature algorithms
3256 extension including all the algorithms we support. Parse new signature
3257 format in client key exchange. Relax some ECC signing restrictions for
3258 TLS v1.2 as indicated in RFC5246.
3261 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3262 to new signature format when needed using client digest preference.
3263 All server ciphersuites should now work correctly in TLS v1.2. No client
3264 support yet and no support for client certificates.
3267 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3268 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3269 ciphersuites. At present only RSA key exchange ciphersuites work with
3270 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3271 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3272 and version checking.
3275 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3276 with this defined it will not be affected by any changes to ssl internal
3277 structures. Add several utility functions to allow openssl application
3278 to work with OPENSSL_NO_SSL_INTERN defined.
3281 *) A long standing patch to add support for SRP from EdelWeb (Peter
3282 Sylvester and Christophe Renou) was integrated.
3283 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3284 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3287 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3290 *) Permit abbreviated handshakes when renegotiating using the function
3291 SSL_renegotiate_abbreviated().
3292 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3294 *) Add call to ENGINE_register_all_complete() to
3295 ENGINE_load_builtin_engines(), so some implementations get used
3296 automatically instead of needing explicit application support.
3299 *) Add support for TLS key exporter as described in RFC5705.
3300 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3302 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3303 a few changes are required:
3305 Add SSL_OP_NO_TLSv1_1 flag.
3306 Add TLSv1_1 methods.
3307 Update version checking logic to handle version 1.1.
3308 Add explicit IV handling (ported from DTLS code).
3309 Add command line options to s_client/s_server.
3312 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3314 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3315 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3316 content decryption and always return the same error. Note: this attack
3317 needs on average 2^20 messages so it only affects automated senders. The
3318 old behaviour can be re-enabled in the CMS code by setting the
3319 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3320 an MMA defence is not necessary.
3321 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3322 this issue. (CVE-2012-0884)
3325 *) Fix CVE-2011-4619: make sure we really are receiving a
3326 client hello before rejecting multiple SGC restarts. Thanks to
3327 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3330 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3332 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3333 Thanks to Antonio Martin, Enterprise Secure Access Research and
3334 Development, Cisco Systems, Inc. for discovering this bug and
3335 preparing a fix. (CVE-2012-0050)
3338 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3340 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3341 of the Vaudenay padding oracle attack on CBC mode encryption
3342 which enables an efficient plaintext recovery attack against
3343 the OpenSSL implementation of DTLS. Their attack exploits timing
3344 differences arising during decryption processing. A research
3345 paper describing this attack can be found at:
3346 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3347 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3348 Security Group at Royal Holloway, University of London
3349 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3350 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3351 for preparing the fix. (CVE-2011-4108)
3352 [Robin Seggelmann, Michael Tuexen]
3354 *) Clear bytes used for block padding of SSL 3.0 records.
3356 [Adam Langley (Google)]
3358 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3359 Kadianakis <desnacked@gmail.com> for discovering this issue and
3360 Adam Langley for preparing the fix. (CVE-2011-4619)
3361 [Adam Langley (Google)]
3363 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3364 [Andrey Kulikov <amdeich@gmail.com>]
3366 *) Prevent malformed RFC3779 data triggering an assertion failure.
3367 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3368 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3369 [Rob Austein <sra@hactrn.net>]
3371 *) Improved PRNG seeding for VOS.
3372 [Paul Green <Paul.Green@stratus.com>]
3374 *) Fix ssl_ciph.c set-up race.
3375 [Adam Langley (Google)]
3377 *) Fix spurious failures in ecdsatest.c.
3378 [Emilia Käsper (Google)]
3380 *) Fix the BIO_f_buffer() implementation (which was mixing different
3381 interpretations of the '..._len' fields).
3382 [Adam Langley (Google)]
3384 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3385 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3386 threads won't reuse the same blinding coefficients.
3388 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3389 lock to call BN_BLINDING_invert_ex, and avoids one use of
3390 BN_BLINDING_update for each BN_BLINDING structure (previously,
3391 the last update always remained unused).
3392 [Emilia Käsper (Google)]
3394 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3395 [Bob Buckholz (Google)]
3397 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3399 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3400 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3401 [Kaspar Brand <ossl@velox.ch>]
3403 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3404 for multi-threaded use of ECDH. (CVE-2011-3210)
3405 [Adam Langley (Google)]
3407 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3410 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3411 signature public key algorithm by using OID xref utilities instead.
3412 Before this you could only use some ECC ciphersuites with SHA1 only.
3415 *) Add protection against ECDSA timing attacks as mentioned in the paper
3416 by Billy Bob Brumley and Nicola Tuveri, see:
3418 http://eprint.iacr.org/2011/232.pdf
3420 [Billy Bob Brumley and Nicola Tuveri]
3422 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3424 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3425 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3427 *) Fix bug in string printing code: if *any* escaping is enabled we must
3428 escape the escape character (backslash) or the resulting string is
3432 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3434 *) Disable code workaround for ancient and obsolete Netscape browsers
3435 and servers: an attacker can use it in a ciphersuite downgrade attack.
3436 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3439 *) Fixed J-PAKE implementation error, originally discovered by
3440 Sebastien Martini, further info and confirmation from Stefan
3441 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3444 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3446 *) Fix extension code to avoid race conditions which can result in a buffer
3447 overrun vulnerability: resumed sessions must not be modified as they can
3448 be shared by multiple threads. CVE-2010-3864
3451 *) Fix WIN32 build system to correctly link an ENGINE directory into
3455 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3457 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3459 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3461 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3463 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3464 context. The operation can be customised via the ctrl mechanism in
3465 case ENGINEs want to include additional functionality.
3468 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3471 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3472 output hashes compatible with older versions of OpenSSL.
3473 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3475 *) Fix compression algorithm handling: if resuming a session use the
3476 compression algorithm of the resumed session instead of determining
3477 it from client hello again. Don't allow server to change algorithm.
3480 *) Add load_crls() function to apps tidying load_certs() too. Add option
3481 to verify utility to allow additional CRLs to be included.
3484 *) Update OCSP request code to permit adding custom headers to the request:
3485 some responders need this.
3488 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3490 [Julia Lawall <julia@diku.dk>]
3492 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3493 needlessly dereferenced structures, used obsolete functions and
3494 didn't handle all updated verify codes correctly.
3497 *) Disable MD2 in the default configuration.
3500 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3501 indicate the initial BIO being pushed or popped. This makes it possible
3502 to determine whether the BIO is the one explicitly called or as a result
3503 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3504 it handles reference counts correctly and doesn't zero out the I/O bio
3505 when it is not being explicitly popped. WARNING: applications which
3506 included workarounds for the old buggy behaviour will need to be modified
3507 or they could free up already freed BIOs.
3510 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3511 renaming to all platforms (within the 0.9.8 branch, this was
3512 done conditionally on Netware platforms to avoid a name clash).
3513 [Guenter <lists@gknw.net>]
3515 *) Add ECDHE and PSK support to DTLS.
3516 [Michael Tuexen <tuexen@fh-muenster.de>]
3518 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3522 *) Add "missing" function EVP_MD_flags() (without this the only way to
3523 retrieve a digest flags is by accessing the structure directly. Update
3524 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3525 or cipher is registered as in the "from" argument. Print out all
3526 registered digests in the dgst usage message instead of manually
3527 attempting to work them out.
3530 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3531 this allows the use of compression and extensions. Change default cipher
3532 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3533 by default unless an application cipher string requests it.
3536 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3537 key ids to find matching certificates and keys but some PKCS#12 files
3538 don't follow the (somewhat unwritten) rules and this strategy fails.
3539 Now just gather all certificates together and the first private key
3540 then look for the first certificate that matches the key.
3543 *) Support use of registered digest and cipher names for dgst and cipher
3544 commands instead of having to add each one as a special case. So now
3551 openssl dgst -sha256 foo
3553 and this works for ENGINE based algorithms too.
3557 *) Update Gost ENGINE to support parameter files.
3558 [Victor B. Wagner <vitus@cryptocom.ru>]
3560 *) Support GeneralizedTime in ca utility.
3561 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3563 *) Enhance the hash format used for certificate directory links. The new
3564 form uses the canonical encoding (meaning equivalent names will work
3565 even if they aren't identical) and uses SHA1 instead of MD5. This form
3566 is incompatible with the older format and as a result c_rehash should
3567 be used to rebuild symbolic links.
3570 *) Make PKCS#8 the default write format for private keys, replacing the
3571 traditional format. This form is standardised, more secure and doesn't
3572 include an implicit MD5 dependency.
3575 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3576 committed to OpenSSL should pass this lot as a minimum.
3579 *) Add session ticket override functionality for use by EAP-FAST.
3580 [Jouni Malinen <j@w1.fi>]
3582 *) Modify HMAC functions to return a value. Since these can be implemented
3583 in an ENGINE errors can occur.
3586 *) Type-checked OBJ_bsearch_ex.
3589 *) Type-checked OBJ_bsearch. Also some constification necessitated
3590 by type-checking. Still to come: TXT_DB, bsearch(?),
3591 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3595 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3596 seconds to a tm structure directly, instead of going through OS
3597 specific date routines. This avoids any issues with OS routines such
3598 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3599 and X509_time_adj_ex() to cover the extended range. The existing
3600 X509_time_adj() is still usable and will no longer have any date issues.
3603 *) Delta CRL support. New use deltas option which will attempt to locate
3604 and search any appropriate delta CRLs available.
3606 This work was sponsored by Google.
3609 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3610 code and add additional score elements. Validate alternate CRL paths
3611 as part of the CRL checking and indicate a new error "CRL path validation
3612 error" in this case. Applications wanting additional details can use
3613 the verify callback and check the new "parent" field. If this is not
3614 NULL CRL path validation is taking place. Existing applications won't
3615 see this because it requires extended CRL support which is off by
3618 This work was sponsored by Google.
3621 *) Support for freshest CRL extension.
3623 This work was sponsored by Google.
3626 *) Initial indirect CRL support. Currently only supported in the CRLs
3627 passed directly and not via lookup. Process certificate issuer
3628 CRL entry extension and lookup CRL entries by bother issuer name
3629 and serial number. Check and process CRL issuer entry in IDP extension.
3631 This work was sponsored by Google.
3634 *) Add support for distinct certificate and CRL paths. The CRL issuer
3635 certificate is validated separately in this case. Only enabled if
3636 an extended CRL support flag is set: this flag will enable additional
3637 CRL functionality in future.
3639 This work was sponsored by Google.
3642 *) Add support for policy mappings extension.
3644 This work was sponsored by Google.
3647 *) Fixes to pathlength constraint, self issued certificate handling,
3648 policy processing to align with RFC3280 and PKITS tests.
3650 This work was sponsored by Google.
3653 *) Support for name constraints certificate extension. DN, email, DNS
3654 and URI types are currently supported.
3656 This work was sponsored by Google.
3659 *) To cater for systems that provide a pointer-based thread ID rather
3660 than numeric, deprecate the current numeric thread ID mechanism and
3661 replace it with a structure and associated callback type. This
3662 mechanism allows a numeric "hash" to be extracted from a thread ID in
3663 either case, and on platforms where pointers are larger than 'long',
3664 mixing is done to help ensure the numeric 'hash' is usable even if it
3665 can't be guaranteed unique. The default mechanism is to use "&errno"
3666 as a pointer-based thread ID to distinguish between threads.
3668 Applications that want to provide their own thread IDs should now use
3669 CRYPTO_THREADID_set_callback() to register a callback that will call
3670 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3672 Note that ERR_remove_state() is now deprecated, because it is tied
3673 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3674 to free the current thread's error state should be replaced by
3675 ERR_remove_thread_state(NULL).
3677 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3678 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3679 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3680 application was previously providing a numeric thread callback that
3681 was inappropriate for distinguishing threads, then uniqueness might
3682 have been obtained with &errno that happened immediately in the
3683 intermediate development versions of OpenSSL; this is no longer the
3684 case, the numeric thread callback will now override the automatic use
3686 [Geoff Thorpe, with help from Bodo Moeller]
3688 *) Initial support for different CRL issuing certificates. This covers a
3689 simple case where the self issued certificates in the chain exist and
3690 the real CRL issuer is higher in the existing chain.
3692 This work was sponsored by Google.
3695 *) Removed effectively defunct crypto/store from the build.
3698 *) Revamp of STACK to provide stronger type-checking. Still to come:
3699 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3700 ASN1_STRING, CONF_VALUE.
3703 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3704 RAM on SSL connections. This option can save about 34k per idle SSL.
3707 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3708 STACK, TXT_DB, bsearch, qsort.
3711 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3712 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3713 support for data, signedData, compressedData, digestedData and
3714 encryptedData, envelopedData types included. Scripts to check against
3715 RFC4134 examples draft and interop and consistency checks of many
3716 content types and variants.
3719 *) Add options to enc utility to support use of zlib compression BIO.
3722 *) Extend mk1mf to support importing of options and assembly language
3723 files from Configure script, currently only included in VC-WIN32.
3724 The assembly language rules can now optionally generate the source
3725 files from the associated perl scripts.
3728 *) Implement remaining functionality needed to support GOST ciphersuites.
3729 Interop testing has been performed using CryptoPro implementations.
3730 [Victor B. Wagner <vitus@cryptocom.ru>]
3732 *) s390x assembler pack.
3735 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3739 *) Implement Opaque PRF Input TLS extension as specified in
3740 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3741 official specification yet and no extension type assignment by
3742 IANA exists, this extension (for now) will have to be explicitly
3743 enabled when building OpenSSL by providing the extension number
3744 to use. For example, specify an option
3746 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3748 to the "config" or "Configure" script to enable the extension,
3749 assuming extension number 0x9527 (which is a completely arbitrary
3750 and unofficial assignment based on the MD5 hash of the Internet
3751 Draft). Note that by doing so, you potentially lose
3752 interoperability with other TLS implementations since these might
3753 be using the same extension number for other purposes.
3755 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3756 opaque PRF input value to use in the handshake. This will create
3757 an internal copy of the length-'len' string at 'src', and will
3758 return non-zero for success.
3760 To get more control and flexibility, provide a callback function
3763 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3764 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3768 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3771 Callback function 'cb' will be called in handshakes, and is
3772 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3773 Argument 'arg' is for application purposes (the value as given to
3774 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3775 be provided to the callback function). The callback function
3776 has to return non-zero to report success: usually 1 to use opaque
3777 PRF input just if possible, or 2 to enforce use of the opaque PRF
3778 input. In the latter case, the library will abort the handshake
3779 if opaque PRF input is not successfully negotiated.
3781 Arguments 'peerinput' and 'len' given to the callback function
3782 will always be NULL and 0 in the case of a client. A server will
3783 see the client's opaque PRF input through these variables if
3784 available (NULL and 0 otherwise). Note that if the server
3785 provides an opaque PRF input, the length must be the same as the
3786 length of the client's opaque PRF input.
3788 Note that the callback function will only be called when creating
3789 a new session (session resumption can resume whatever was
3790 previously negotiated), and will not be called in SSL 2.0
3791 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3792 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3793 for applications that need to enforce opaque PRF input.
3797 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3800 [Victor B. Wagner <vitus@cryptocom.ru>]
3802 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3803 RFC4507bis. The encrypted ticket format is an encrypted encoded
3804 SSL_SESSION structure, that way new session features are automatically
3807 If a client application caches session in an SSL_SESSION structure
3808 support is transparent because tickets are now stored in the encoded
3811 The SSL_CTX structure automatically generates keys for ticket
3812 protection in servers so again support should be possible
3813 with no application modification.
3815 If a client or server wishes to disable RFC4507 support then the option
3816 SSL_OP_NO_TICKET can be set.
3818 Add a TLS extension debugging callback to allow the contents of any client
3819 or server extensions to be examined.
3821 This work was sponsored by Google.
3824 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3825 OpenSSL should now compile cleanly on gcc 4.2
3826 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3828 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3829 support including streaming MAC support: this is required for GOST
3830 ciphersuite support.
3831 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3833 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3834 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3835 to output in BER and PEM format.
3838 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3839 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3840 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3841 ENGINE support for HMAC keys which are unextractable. New -mac and
3842 -macopt options to dgst utility.
3845 *) New option -sigopt to dgst utility. Update dgst to use
3846 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3847 alternative signing parameters such as X9.31 or PSS in the dgst
3851 *) Change ssl_cipher_apply_rule(), the internal function that does
3852 the work each time a ciphersuite string requests enabling
3853 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3854 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3855 the order of disabled ciphersuites such that those ciphersuites
3856 that most recently went from enabled to disabled not only stay
3857 in order with respect to each other, but also have higher priority
3858 than other disabled ciphersuites the next time ciphersuites are
3861 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3862 the same ciphersuites as with "HIGH" alone, but in a specific
3863 order where the PSK ciphersuites come first (since they are the
3864 most recently disabled ciphersuites when "HIGH" is parsed).
3866 Also, change ssl_create_cipher_list() (using this new
3867 functionality) such that between otherwise identical
3868 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
3872 *) Change ssl_create_cipher_list() so that it automatically
3873 arranges the ciphersuites in reasonable order before starting
3874 to process the rule string. Thus, the definition for "DEFAULT"
3875 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3876 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3877 This makes it much easier to arrive at a reasonable default order
3878 in applications for which anonymous ciphers are OK (meaning
3879 that you can't actually use DEFAULT).
3880 [Bodo Moeller; suggested by Victor Duchovni]
3882 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3883 processing) into multiple integers instead of setting
3884 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3885 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3886 (These masks as well as the individual bit definitions are hidden
3887 away into the non-exported interface ssl/ssl_locl.h, so this
3888 change to the definition of the SSL_CIPHER structure shouldn't
3889 affect applications.) This give us more bits for each of these
3890 categories, so there is no longer a need to coagulate AES128 and
3891 AES256 into a single algorithm bit, and to coagulate Camellia128
3892 and Camellia256 into a single algorithm bit, which has led to all
3895 Thus, among other things, the kludge introduced in 0.9.7m and
3896 0.9.8e for masking out AES256 independently of AES128 or masking
3897 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3899 With the change, we also introduce new ciphersuite aliases that
3900 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3904 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3905 Use the leftmost N bytes of the signature input if the input is
3906 larger than the prime q (with N being the size in bytes of q).
3909 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3910 it yet and it is largely untested.
3913 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3916 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3917 some compilers (gcc 4.2 and later) reject their use. Safestack is
3918 reimplemented. Update ASN1 to avoid use of legacy functions.
3921 *) Win32/64 targets are linked with Winsock2.
3924 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3925 to external functions. This can be used to increase CRL handling
3926 efficiency especially when CRLs are very large by (for example) storing
3927 the CRL revoked certificates in a database.
3930 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3931 new CRLs added to a directory can be used. New command line option
3932 -verify_return_error to s_client and s_server. This causes real errors
3933 to be returned by the verify callback instead of carrying on no matter
3934 what. This reflects the way a "real world" verify callback would behave.
3937 *) GOST engine, supporting several GOST algorithms and public key formats.
3938 Kindly donated by Cryptocom.
3941 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3942 partitioned by DP are handled but no indirect CRL or reason partitioning
3943 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3944 selected via a scoring technique which handles IDP and AKID in CRLs.
3947 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3948 will ultimately be used for all verify operations: this will remove the
3949 X509_STORE dependency on certificate verification and allow alternative
3950 lookup methods. X509_STORE based implementations of these two callbacks.
3953 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3954 Modify get_crl() to find a valid (unexpired) CRL if possible.
3957 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3958 this would be called X509_CRL_cmp() but that name is already used by
3959 a function that just compares CRL issuer names. Cache several CRL
3960 extensions in X509_CRL structure and cache CRLDP in X509.
3963 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3964 this maps equivalent X509_NAME structures into a consistent structure.
3965 Name comparison can then be performed rapidly using memcmp().
3968 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3972 *) Allow digests to supply their own micalg string for S/MIME type using
3973 the ctrl EVP_MD_CTRL_MICALG.
3976 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3977 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3978 ctrl. It can then customise the structure before and/or after signing
3982 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3983 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3984 to free up any added signature OIDs.
3987 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3988 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3989 digest and cipher tables. New options added to openssl utility:
3990 list-message-digest-algorithms and list-cipher-algorithms.
3993 *) Change the array representation of binary polynomials: the list
3994 of degrees of non-zero coefficients is now terminated with -1.
3995 Previously it was terminated with 0, which was also part of the
3996 value; thus, the array representation was not applicable to
3997 polynomials where t^0 has coefficient zero. This change makes
3998 the array representation useful in a more general context.
4001 *) Various modifications and fixes to SSL/TLS cipher string
4002 handling. For ECC, the code now distinguishes between fixed ECDH
4003 with RSA certificates on the one hand and with ECDSA certificates
4004 on the other hand, since these are separate ciphersuites. The
4005 unused code for Fortezza ciphersuites has been removed.
4007 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4008 (not "ECDHE"). For consistency with the code for DH
4009 certificates, use of ECDH certificates is now considered ECDH
4010 authentication, not RSA or ECDSA authentication (the latter is
4011 merely the CA's signing algorithm and not actively used in the
4014 The temporary ciphersuite alias "ECCdraft" is no longer
4015 available, and ECC ciphersuites are no longer excluded from "ALL"
4016 and "DEFAULT". The following aliases now exist for RFC 4492
4017 ciphersuites, most of these by analogy with the DH case:
4019 kECDHr - ECDH cert, signed with RSA
4020 kECDHe - ECDH cert, signed with ECDSA
4021 kECDH - ECDH cert (signed with either RSA or ECDSA)
4022 kEECDH - ephemeral ECDH
4023 ECDH - ECDH cert or ephemeral ECDH
4029 AECDH - anonymous ECDH
4030 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4034 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4035 Use correct micalg parameters depending on digest(s) in signed message.
4038 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4039 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4042 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4043 an engine to register a method. Add ENGINE lookups for methods and
4044 functional reference processing.
4047 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4048 EVP_{Sign,Verify}* which allow an application to customise the signature
4052 *) New -resign option to smime utility. This adds one or more signers
4053 to an existing PKCS#7 signedData structure. Also -md option to use an
4054 alternative message digest algorithm for signing.
4057 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4058 create PKCS7 structures containing multiple signers. Update smime
4059 application to support multiple signers.
4062 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4066 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4067 Reorganize PBE internals to lookup from a static table using NIDs,
4068 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4069 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4070 PRF which will be automatically used with PBES2.
4073 *) Replace the algorithm specific calls to generate keys in "req" with the
4077 *) Update PKCS#7 enveloped data routines to use new API. This is now
4078 supported by any public key method supporting the encrypt operation. A
4079 ctrl is added to allow the public key algorithm to examine or modify
4080 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4084 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4085 a default digest type to use. In most cases this will be SHA1 but some
4086 algorithms (such as GOST) need to specify an alternative digest. The
4087 return value indicates how strong the preference is 1 means optional and
4088 2 is mandatory (that is it is the only supported type). Modify
4089 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4090 use the default md. Update openssl utilities to use the default digest
4091 type for signing if it is not explicitly indicated.
4094 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4095 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4096 signing method from the key type. This effectively removes the link
4097 between digests and public key types.
4100 *) Add an OID cross reference table and utility functions. Its purpose is to
4101 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4102 rsaEncryption. This will allow some of the algorithm specific hackery
4103 needed to use the correct OID to be removed.
4106 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4107 structures for PKCS7_sign(). They are now set up by the relevant public
4111 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4114 *) Add support for key derivation (agreement) in the API, DH method and
4118 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4119 public and private key formats. As a side effect these add additional
4120 command line functionality not previously available: DSA signatures can be
4121 generated and verified using pkeyutl and DH key support and generation in
4126 [Oliver Tappe <zooey@hirschkaefer.de>]
4128 *) New make target "install_html_docs" installs HTML renditions of the
4130 [Oliver Tappe <zooey@hirschkaefer.de>]
4132 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4133 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4134 support key and parameter generation and add initial key generation
4135 functionality for RSA.
4138 *) Add functions for main EVP_PKEY_method operations. The undocumented
4139 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4140 EVP_PKEY_{encrypt,decrypt}_old.
4143 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4144 key API, doesn't do much yet.
4147 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4148 public key algorithms. New option to openssl utility:
4149 "list-public-key-algorithms" to print out info.
4152 *) Implement the Supported Elliptic Curves Extension for
4153 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4156 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4157 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4160 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4161 utilities such as rsa, dsa, dsaparam etc except they process any key
4165 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4166 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4167 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4171 *) Initial support for pluggable public key ASN1.
4172 De-spaghettify the public key ASN1 handling. Move public and private
4173 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4174 algorithm specific handling to a single module within the relevant
4175 algorithm directory. Add functions to allow (near) opaque processing
4176 of public and private key structures.
4179 *) Implement the Supported Point Formats Extension for
4180 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4183 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4184 for the psk identity [hint] and the psk callback functions to the
4185 SSL_SESSION, SSL and SSL_CTX structure.
4188 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4192 SSL_CTX_use_psk_identity_hint
4193 SSL_get_psk_identity_hint
4194 SSL_get_psk_identity
4195 SSL_use_psk_identity_hint
4197 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4199 *) Add RFC 3161 compliant time stamp request creation, response generation
4200 and response verification functionality.
4201 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4203 *) Add initial support for TLS extensions, specifically for the server_name
4204 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4205 have new members for a host name. The SSL data structure has an
4206 additional member SSL_CTX *initial_ctx so that new sessions can be
4207 stored in that context to allow for session resumption, even after the
4208 SSL has been switched to a new SSL_CTX in reaction to a client's
4209 server_name extension.
4211 New functions (subject to change):
4213 SSL_get_servername()
4214 SSL_get_servername_type()
4217 New CTRL codes and macros (subject to change):
4219 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4220 - SSL_CTX_set_tlsext_servername_callback()
4221 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4222 - SSL_CTX_set_tlsext_servername_arg()
4223 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4225 openssl s_client has a new '-servername ...' option.
4227 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4228 '-key2 ...', '-servername_fatal' (subject to change). This allows
4229 testing the HostName extension for a specific single host name ('-cert'
4230 and '-key' remain fallbacks for handshakes without HostName
4231 negotiation). If the unrecognized_name alert has to be sent, this by
4232 default is a warning; it becomes fatal with the '-servername_fatal'
4235 [Peter Sylvester, Remy Allais, Christophe Renou]
4237 *) Whirlpool hash implementation is added.
4240 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4241 bn(64,32). Because of instruction set limitations it doesn't have
4242 any negative impact on performance. This was done mostly in order
4243 to make it possible to share assembler modules, such as bn_mul_mont
4244 implementations, between 32- and 64-bit builds without hassle.
4247 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4248 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4252 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4253 dedicated Montgomery multiplication procedure, is introduced.
4254 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4255 "64-bit" performance on certain 32-bit targets.
4258 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4259 in SSL structures. New SSL ctrl to set maximum send fragment size.
4260 Save memory by setting the I/O buffer sizes dynamically instead of
4261 using the maximum available value.
4264 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4265 in addition to the text details.
4268 *) Very, very preliminary EXPERIMENTAL support for printing of general
4269 ASN1 structures. This currently produces rather ugly output and doesn't
4270 handle several customised structures at all.
4273 *) Integrated support for PVK file format and some related formats such
4274 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4275 these in the 'rsa' and 'dsa' utilities.
4278 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4281 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4282 place for the (very old) "NETSCAPE" format certificates which are now
4283 handled using new ASN1 code equivalents.
4286 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4287 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4288 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4291 *) Modify CRL distribution points extension code to print out previously
4292 unsupported fields. Enhance extension setting code to allow setting of
4296 *) Add print and set support for Issuing Distribution Point CRL extension.
4299 *) Change 'Configure' script to enable Camellia by default.
4302 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4304 *) When rejecting SSL/TLS records due to an incorrect version number, never
4305 update s->server with a new major version number. As of
4306 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4307 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4308 the previous behavior could result in a read attempt at NULL when
4309 receiving specific incorrect SSL/TLS records once record payload
4310 protection is active. (CVE-2010-0740)
4311 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4313 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4314 could be crashed if the relevant tables were not present (e.g. chrooted).
4315 [Tomas Hoger <thoger@redhat.com>]
4317 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4319 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4320 [Martin Olsson, Neel Mehta]
4322 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4323 accommodate for stack sorting, always a write lock!).
4326 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4327 excessive delays in the RAND_poll(): over a minute. As a workaround
4328 include a time check in the inner Heap32Next loop too.
4331 *) The code that handled flushing of data in SSL/TLS originally used the
4332 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4333 the problem outlined in PR#1949. The fix suggested there however can
4334 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4335 of Apache). So instead simplify the code to flush unconditionally.
4336 This should be fine since flushing with no data to flush is a no op.
4339 *) Handle TLS versions 2.0 and later properly and correctly use the
4340 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4341 off ancient servers have a habit of sticking around for a while...
4344 *) Modify compression code so it frees up structures without using the
4345 ex_data callbacks. This works around a problem where some applications
4346 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4347 restarting) then use compression (e.g. SSL with compression) later.
4348 This results in significant per-connection memory leaks and
4349 has caused some security issues including CVE-2008-1678 and
4353 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4354 change when encrypting or decrypting.
4357 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4358 connect and renegotiate with servers which do not support RI.
4359 Until RI is more widely deployed this option is enabled by default.
4362 *) Add "missing" ssl ctrls to clear options and mode.
4365 *) If client attempts to renegotiate and doesn't support RI respond with
4366 a no_renegotiation alert as required by RFC5746. Some renegotiating
4367 TLS clients will continue a connection gracefully when they receive
4368 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4369 waiting for a server hello which it will never receive. Now we treat a
4370 received no_renegotiation alert as a fatal error. This is because
4371 applications requesting a renegotiation might well expect it to succeed
4372 and would have no code in place to handle the server denying it so the
4373 only safe thing to do is to terminate the connection.
4376 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4377 peer supports secure renegotiation and 0 otherwise. Print out peer
4378 renegotiation support in s_client/s_server.
4381 *) Replace the highly broken and deprecated SPKAC certification method with
4382 the updated NID creation version. This should correctly handle UTF8.
4385 *) Implement RFC5746. Re-enable renegotiation but require the extension
4386 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4387 turns out to be a bad idea. It has been replaced by
4388 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4389 SSL_CTX_set_options(). This is really not recommended unless you
4390 know what you are doing.
4391 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4393 *) Fixes to stateless session resumption handling. Use initial_ctx when
4394 issuing and attempting to decrypt tickets in case it has changed during
4395 servername handling. Use a non-zero length session ID when attempting
4396 stateless session resumption: this makes it possible to determine if
4397 a resumption has occurred immediately after receiving server hello
4398 (several places in OpenSSL subtly assume this) instead of later in
4402 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4403 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4404 fixes for a few places where the return code is not checked
4406 [Julia Lawall <julia@diku.dk>]
4408 *) Add --strict-warnings option to Configure script to include devteam
4409 warnings in other configurations.
4412 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4413 makes it possible to install openssl libraries in locations which
4414 have names other than "lib", for example "/usr/lib64" which some
4416 [Steve Henson, based on patch from Jeremy Utley]
4418 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4419 X690 8.9.12 and can produce some misleading textual output of OIDs.
4420 [Steve Henson, reported by Dan Kaminsky]
4422 *) Delete MD2 from algorithm tables. This follows the recommendation in
4423 several standards that it is not used in new applications due to
4424 several cryptographic weaknesses. For binary compatibility reasons
4425 the MD2 API is still compiled in by default.
4428 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4432 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4433 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4435 [Guenter <lists@gknw.net>]
4437 *) Fix the server certificate chain building code to use X509_verify_cert(),
4438 it used to have an ad-hoc builder which was unable to cope with anything
4439 other than a simple chain.
4440 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4442 *) Don't check self signed certificate signatures in X509_verify_cert()
4443 by default (a flag can override this): it just wastes time without
4444 adding any security. As a useful side effect self signed root CAs
4445 with non-FIPS digests are now usable in FIPS mode.
4448 *) In dtls1_process_out_of_seq_message() the check if the current message
4449 is already buffered was missing. For every new message was memory
4450 allocated, allowing an attacker to perform an denial of service attack
4451 with sending out of seq handshake messages until there is no memory
4452 left. Additionally every future message was buffered, even if the
4453 sequence number made no sense and would be part of another handshake.
4454 So only messages with sequence numbers less than 10 in advance will be
4455 buffered. (CVE-2009-1378)
4456 [Robin Seggelmann, discovered by Daniel Mentz]
4458 *) Records are buffered if they arrive with a future epoch to be
4459 processed after finishing the corresponding handshake. There is
4460 currently no limitation to this buffer allowing an attacker to perform
4461 a DOS attack with sending records with future epochs until there is no
4462 memory left. This patch adds the pqueue_size() function to determine
4463 the size of a buffer and limits the record buffer to 100 entries.
4465 [Robin Seggelmann, discovered by Daniel Mentz]
4467 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4468 parent structure is freed. (CVE-2009-1379)
4471 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4472 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4475 [Ilya O. <vrghost@gmail.com>]
4477 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4479 *) Disable renegotiation completely - this fixes a severe security
4480 problem (CVE-2009-3555) at the cost of breaking all
4481 renegotiation. Renegotiation can be re-enabled by setting
4482 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4483 run-time. This is really not recommended unless you know what
4487 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4489 *) Don't set val to NULL when freeing up structures, it is freed up by
4490 underlying code. If sizeof(void *) > sizeof(long) this can result in
4491 zeroing past the valid field. (CVE-2009-0789)
4492 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4494 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4495 checked correctly. This would allow some invalid signed attributes to
4496 appear to verify correctly. (CVE-2009-0591)
4497 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4499 *) Reject UniversalString and BMPString types with invalid lengths. This
4500 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4501 a legal length. (CVE-2009-0590)
4504 *) Set S/MIME signing as the default purpose rather than setting it
4505 unconditionally. This allows applications to override it at the store
4509 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4510 to handle some structures.
4513 *) Improve efficiency of mem_gets: don't search whole buffer each time
4515 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4517 *) New -hex option for openssl rand.
4520 *) Print out UTF8String and NumericString when parsing ASN1.
4523 *) Support NumericString type for name components.
4526 *) Allow CC in the environment to override the automatically chosen
4527 compiler. Note that nothing is done to ensure flags work with the
4531 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4533 *) Properly check EVP_VerifyFinal() and similar return values
4535 [Ben Laurie, Bodo Moeller, Google Security Team]
4537 *) Enable TLS extensions by default.
4540 *) Allow the CHIL engine to be loaded, whether the application is
4541 multithreaded or not. (This does not release the developer from the
4542 obligation to set up the dynamic locking callbacks.)
4543 [Sander Temme <sander@temme.net>]
4545 *) Use correct exit code if there is an error in dgst command.
4546 [Steve Henson; problem pointed out by Roland Dirlewanger]
4548 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4549 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4552 *) Add experimental JPAKE support, including demo authentication in
4553 s_client and s_server.
4556 *) Set the comparison function in v3_addr_canonize().
4557 [Rob Austein <sra@hactrn.net>]
4559 *) Add support for XMPP STARTTLS in s_client.
4560 [Philip Paeps <philip@freebsd.org>]
4562 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4563 to ensure that even with this option, only ciphersuites in the
4564 server's preference list will be accepted. (Note that the option
4565 applies only when resuming a session, so the earlier behavior was
4566 just about the algorithm choice for symmetric cryptography.)
4569 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4571 *) Fix NULL pointer dereference if a DTLS server received
4572 ChangeCipherSpec as first record (CVE-2009-1386).
4575 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4576 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4579 *) The fix in 0.9.8c that supposedly got rid of unsafe
4580 double-checked locking was incomplete for RSA blinding,
4581 addressing just one layer of what turns out to have been
4582 doubly unsafe triple-checked locking.
4584 So now fix this for real by retiring the MONT_HELPER macro
4585 in crypto/rsa/rsa_eay.c.
4587 [Bodo Moeller; problem pointed out by Marius Schilder]
4589 *) Various precautionary measures:
4591 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4593 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4594 (NB: This would require knowledge of the secret session ticket key
4595 to exploit, in which case you'd be SOL either way.)
4597 - Change bn_nist.c so that it will properly handle input BIGNUMs
4598 outside the expected range.
4600 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4603 [Neel Mehta, Bodo Moeller]
4605 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4606 the load fails. Useful for distros.
4607 [Ben Laurie and the FreeBSD team]
4609 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4612 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4615 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4617 This work was sponsored by Logica.
4620 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4621 keystores. Support for SSL/TLS client authentication too.
4622 Not compiled unless enable-capieng specified to Configure.
4624 This work was sponsored by Logica.
4627 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4628 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4629 attribute creation routines such as certificate requests and PKCS#12
4633 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4635 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4636 handshake which could lead to a client crash as found using the
4637 Codenomicon TLS test suite (CVE-2008-1672)
4638 [Steve Henson, Mark Cox]
4640 *) Fix double free in TLS server name extensions which could lead to
4641 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4644 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4646 Clear the error queue to ensure that error entries left from
4647 older function calls do not interfere with the correct operation.
4648 [Lutz Jaenicke, Erik de Castro Lopo]
4650 *) Remove root CA certificates of commercial CAs:
4652 The OpenSSL project does not recommend any specific CA and does not
4653 have any policy with respect to including or excluding any CA.
4654 Therefore it does not make any sense to ship an arbitrary selection
4655 of root CA certificates with the OpenSSL software.
4658 *) RSA OAEP patches to fix two separate invalid memory reads.
4659 The first one involves inputs when 'lzero' is greater than
4660 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4661 before the beginning of from). The second one involves inputs where
4662 the 'db' section contains nothing but zeroes (there is a one-byte
4663 invalid read after the end of 'db').
4664 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4666 *) Partial backport from 0.9.9-dev:
4668 Introduce bn_mul_mont (dedicated Montgomery multiplication
4669 procedure) as a candidate for BIGNUM assembler implementation.
4670 While 0.9.9-dev uses assembler for various architectures, only
4671 x86_64 is available by default here in the 0.9.8 branch, and
4672 32-bit x86 is available through a compile-time setting.
4674 To try the 32-bit x86 assembler implementation, use Configure
4675 option "enable-montasm" (which exists only for this backport).
4677 As "enable-montasm" for 32-bit x86 disclaims code stability
4678 anyway, in this constellation we activate additional code
4679 backported from 0.9.9-dev for further performance improvements,
4680 namely BN_from_montgomery_word. (To enable this otherwise,
4681 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4683 [Andy Polyakov (backport partially by Bodo Moeller)]
4685 *) Add TLS session ticket callback. This allows an application to set
4686 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4687 values. This is useful for key rollover for example where several key
4688 sets may exist with different names.
4691 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4692 This was broken until now in 0.9.8 releases, such that the only way
4693 a registered ENGINE could be used (assuming it initialises
4694 successfully on the host) was to explicitly set it as the default
4695 for the relevant algorithms. This is in contradiction with 0.9.7
4696 behaviour and the documentation. With this fix, when an ENGINE is
4697 registered into a given algorithm's table of implementations, the
4698 'uptodate' flag is reset so that auto-discovery will be used next
4699 time a new context for that algorithm attempts to select an
4701 [Ian Lister (tweaked by Geoff Thorpe)]
4703 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4704 implementation in the following ways:
4706 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4709 Lack of BER streaming support means one pass streaming processing is
4710 only supported if data is detached: setting the streaming flag is
4711 ignored for embedded content.
4713 CMS support is disabled by default and must be explicitly enabled
4714 with the enable-cms configuration option.
4717 *) Update the GMP engine glue to do direct copies between BIGNUM and
4718 mpz_t when openssl and GMP use the same limb size. Otherwise the
4719 existing "conversion via a text string export" trick is still used.
4720 [Paul Sheer <paulsheer@gmail.com>]
4722 *) Zlib compression BIO. This is a filter BIO which compressed and
4723 uncompresses any data passed through it.
4726 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4727 RFC3394 compatible AES key wrapping.
4730 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4731 sets string data without copying. X509_ALGOR_set0() and
4732 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4733 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4734 from an X509_ATTRIBUTE structure optionally checking it occurs only
4735 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4739 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4740 to get the expected BN_FLG_CONSTTIME behavior.
4741 [Bodo Moeller (Google)]
4745 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4746 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4747 - added some more tests to do_tests.pl
4748 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4749 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4750 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4751 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4752 - various changes to netware.pl to enable gcc-cross builds on Win32
4754 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4755 - various changes to fix missing prototype warnings
4756 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4757 - added AES, WHIRLPOOL and CPUID assembler code to build files
4758 - added missing AES assembler make rules to mk1mf.pl
4759 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4760 [Guenter Knauf <eflash@gmx.net>]
4762 *) Implement certificate status request TLS extension defined in RFC3546.
4763 A client can set the appropriate parameters and receive the encoded
4764 OCSP response via a callback. A server can query the supplied parameters
4765 and set the encoded OCSP response in the callback. Add simplified examples
4766 to s_client and s_server.
4769 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4771 *) Fix various bugs:
4772 + Binary incompatibility of ssl_ctx_st structure
4773 + DTLS interoperation with non-compliant servers
4774 + Don't call get_session_cb() without proposed session
4775 + Fix ia64 assembler code
4776 [Andy Polyakov, Steve Henson]
4778 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4780 *) DTLS Handshake overhaul. There were longstanding issues with
4781 OpenSSL DTLS implementation, which were making it impossible for
4782 RFC 4347 compliant client to communicate with OpenSSL server.
4783 Unfortunately just fixing these incompatibilities would "cut off"
4784 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4785 server keeps tolerating non RFC compliant syntax. The opposite is
4786 not true, 0.9.8f client can not communicate with earlier server.
4787 This update even addresses CVE-2007-4995.
4790 *) Changes to avoid need for function casts in OpenSSL: some compilers
4791 (gcc 4.2 and later) reject their use.
4792 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4795 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4796 RFC4507bis. The encrypted ticket format is an encrypted encoded
4797 SSL_SESSION structure, that way new session features are automatically
4800 If a client application caches session in an SSL_SESSION structure
4801 support is transparent because tickets are now stored in the encoded
4804 The SSL_CTX structure automatically generates keys for ticket
4805 protection in servers so again support should be possible
4806 with no application modification.
4808 If a client or server wishes to disable RFC4507 support then the option
4809 SSL_OP_NO_TICKET can be set.
4811 Add a TLS extension debugging callback to allow the contents of any client
4812 or server extensions to be examined.
4814 This work was sponsored by Google.
4817 *) Add initial support for TLS extensions, specifically for the server_name
4818 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4819 have new members for a host name. The SSL data structure has an
4820 additional member SSL_CTX *initial_ctx so that new sessions can be
4821 stored in that context to allow for session resumption, even after the
4822 SSL has been switched to a new SSL_CTX in reaction to a client's
4823 server_name extension.
4825 New functions (subject to change):
4827 SSL_get_servername()
4828 SSL_get_servername_type()
4831 New CTRL codes and macros (subject to change):
4833 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4834 - SSL_CTX_set_tlsext_servername_callback()
4835 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4836 - SSL_CTX_set_tlsext_servername_arg()
4837 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4839 openssl s_client has a new '-servername ...' option.
4841 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4842 '-key2 ...', '-servername_fatal' (subject to change). This allows
4843 testing the HostName extension for a specific single host name ('-cert'
4844 and '-key' remain fallbacks for handshakes without HostName
4845 negotiation). If the unrecognized_name alert has to be sent, this by
4846 default is a warning; it becomes fatal with the '-servername_fatal'
4849 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4851 *) Add AES and SSE2 assembly language support to VC++ build.
4854 *) Mitigate attack on final subtraction in Montgomery reduction.
4857 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4858 (which previously caused an internal error).
4861 *) Squeeze another 10% out of IGE mode when in != out.
4864 *) AES IGE mode speedup.
4865 [Dean Gaudet (Google)]
4867 *) Add the Korean symmetric 128-bit cipher SEED (see
4868 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4869 add SEED ciphersuites from RFC 4162:
4871 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4872 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4873 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4874 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4876 To minimize changes between patchlevels in the OpenSSL 0.9.8
4877 series, SEED remains excluded from compilation unless OpenSSL
4878 is configured with 'enable-seed'.
4879 [KISA, Bodo Moeller]
4881 *) Mitigate branch prediction attacks, which can be practical if a
4882 single processor is shared, allowing a spy process to extract
4883 information. For detailed background information, see
4884 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4885 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4886 and Necessary Software Countermeasures"). The core of the change
4887 are new versions BN_div_no_branch() and
4888 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4889 respectively, which are slower, but avoid the security-relevant
4890 conditional branches. These are automatically called by BN_div()
4891 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4892 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4893 remove a conditional branch.
4895 BN_FLG_CONSTTIME is the new name for the previous
4896 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4897 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4898 in the exponent causes BN_mod_exp_mont() to use the alternative
4899 implementation in BN_mod_exp_mont_consttime().) The old name
4900 remains as a deprecated alias.
4902 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4903 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4904 constant-time implementations for more than just exponentiation.
4905 Here too the old name is kept as a deprecated alias.
4907 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4908 the BN_BLINDING structure gets an independent copy of the
4909 modulus. This means that the previous "BIGNUM *m" argument to
4910 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4911 essentially becomes "const BIGNUM *m", although we can't actually
4912 change this in the header file before 0.9.9. It allows
4913 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4914 enable BN_FLG_CONSTTIME.
4916 [Matthew D Wood (Intel Corp)]
4918 *) In the SSL/TLS server implementation, be strict about session ID
4919 context matching (which matters if an application uses a single
4920 external cache for different purposes). Previously,
4921 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4922 set. This did ensure strict client verification, but meant that,
4923 with applications using a single external cache for quite
4924 different requirements, clients could circumvent ciphersuite
4925 restrictions for a given session ID context by starting a session
4926 in a different context.
4929 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4930 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4931 authentication-only ciphersuites.
4934 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4935 not complete and could lead to a possible single byte overflow
4936 (CVE-2007-5135) [Ben Laurie]
4938 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4940 *) Since AES128 and AES256 (and similarly Camellia128 and
4941 Camellia256) share a single mask bit in the logic of
4942 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4943 kludge to work properly if AES128 is available and AES256 isn't
4944 (or if Camellia128 is available and Camellia256 isn't).
4947 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4948 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4949 When a point or a seed is encoded in a BIT STRING, we need to
4950 prevent the removal of trailing zero bits to get the proper DER
4951 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4952 of a NamedBitList, for which trailing 0 bits need to be removed.)
4955 *) Have SSL/TLS server implementation tolerate "mismatched" record
4956 protocol version while receiving ClientHello even if the
4957 ClientHello is fragmented. (The server can't insist on the
4958 particular protocol version it has chosen before the ServerHello
4959 message has informed the client about his choice.)
4962 *) Add RFC 3779 support.
4963 [Rob Austein for ARIN, Ben Laurie]
4965 *) Load error codes if they are not already present instead of using a
4966 static variable. This allows them to be cleanly unloaded and reloaded.
4967 Improve header file function name parsing.
4970 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4971 or CAPABILITY handshake as required by RFCs.
4974 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4976 *) Introduce limits to prevent malicious keys being able to
4977 cause a denial of service. (CVE-2006-2940)
4978 [Steve Henson, Bodo Moeller]
4980 *) Fix ASN.1 parsing of certain invalid structures that can result
4981 in a denial of service. (CVE-2006-2937) [Steve Henson]
4983 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4984 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4986 *) Fix SSL client code which could crash if connecting to a
4987 malicious SSLv2 server. (CVE-2006-4343)
4988 [Tavis Ormandy and Will Drewry, Google Security Team]
4990 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4991 match only those. Before that, "AES256-SHA" would be interpreted
4992 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4993 the same strength classification in 0.9.7h) as we currently only
4994 have a single AES bit in the ciphersuite description bitmap.
4995 That change, however, also applied to ciphersuite strings such as
4996 "RC4-MD5" that intentionally matched multiple ciphersuites --
4997 namely, SSL 2.0 ciphersuites in addition to the more common ones
4998 from SSL 3.0/TLS 1.0.
5000 So we change the selection algorithm again: Naming an explicit
5001 ciphersuite selects this one ciphersuite, and any other similar
5002 ciphersuite (same bitmap) from *other* protocol versions.
5003 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5004 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5006 Since SSL 2.0 does not have any ciphersuites for which the
5007 128/256 bit distinction would be relevant, this works for now.
5008 The proper fix will be to use different bits for AES128 and
5009 AES256, which would have avoided the problems from the beginning;
5010 however, bits are scarce, so we can only do this in a new release
5011 (not just a patchlevel) when we can change the SSL_CIPHER
5012 definition to split the single 'unsigned long mask' bitmap into
5013 multiple values to extend the available space.
5017 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5019 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5020 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5022 *) Add AES IGE and biIGE modes.
5025 *) Change the Unix randomness entropy gathering to use poll() when
5026 possible instead of select(), since the latter has some
5027 undesirable limitations.
5028 [Darryl Miles via Richard Levitte and Bodo Moeller]
5030 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5031 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5032 cannot be implicitly activated as part of, e.g., the "AES" alias.
5033 However, please upgrade to OpenSSL 0.9.9[-dev] for
5034 non-experimental use of the ECC ciphersuites to get TLS extension
5035 support, which is required for curve and point format negotiation
5036 to avoid potential handshake problems.
5039 *) Disable rogue ciphersuites:
5041 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5042 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5043 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5045 The latter two were purportedly from
5046 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5049 Also deactivate the remaining ciphersuites from
5050 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5051 unofficial, and the ID has long expired.
5054 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5055 dual-core machines) and other potential thread-safety issues.
5058 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5059 versions), which is now available for royalty-free use
5060 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5061 Also, add Camellia TLS ciphersuites from RFC 4132.
5063 To minimize changes between patchlevels in the OpenSSL 0.9.8
5064 series, Camellia remains excluded from compilation unless OpenSSL
5065 is configured with 'enable-camellia'.
5068 *) Disable the padding bug check when compression is in use. The padding
5069 bug check assumes the first packet is of even length, this is not
5070 necessarily true if compression is enabled and can result in false
5071 positives causing handshake failure. The actual bug test is ancient
5072 code so it is hoped that implementations will either have fixed it by
5073 now or any which still have the bug do not support compression.
5076 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5078 *) When applying a cipher rule check to see if string match is an explicit
5079 cipher suite and only match that one cipher suite if it is.
5082 *) Link in manifests for VC++ if needed.
5083 [Austin Ziegler <halostatue@gmail.com>]
5085 *) Update support for ECC-based TLS ciphersuites according to
5086 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5087 TLS extensions, which are supported starting with the 0.9.9
5088 branch, not in the OpenSSL 0.9.8 branch).
5091 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5092 opaque EVP_CIPHER_CTX handling.
5095 *) Fixes and enhancements to zlib compression code. We now only use
5096 "zlib1.dll" and use the default __cdecl calling convention on Win32
5097 to conform with the standards mentioned here:
5098 http://www.zlib.net/DLL_FAQ.txt
5099 Static zlib linking now works on Windows and the new --with-zlib-include
5100 --with-zlib-lib options to Configure can be used to supply the location
5101 of the headers and library. Gracefully handle case where zlib library
5105 *) Several fixes and enhancements to the OID generation code. The old code
5106 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5107 handle numbers larger than ULONG_MAX, truncated printing and had a
5108 non standard OBJ_obj2txt() behaviour.
5111 *) Add support for building of engines under engine/ as shared libraries
5112 under VC++ build system.
5115 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5116 Hopefully, we will not see any false combination of paths any more.
5119 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5121 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5122 (part of SSL_OP_ALL). This option used to disable the
5123 countermeasure against man-in-the-middle protocol-version
5124 rollback in the SSL 2.0 server implementation, which is a bad
5125 idea. (CVE-2005-2969)
5127 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5128 for Information Security, National Institute of Advanced Industrial
5129 Science and Technology [AIST], Japan)]
5131 *) Add two function to clear and return the verify parameter flags.
5134 *) Keep cipherlists sorted in the source instead of sorting them at
5135 runtime, thus removing the need for a lock.
5138 *) Avoid some small subgroup attacks in Diffie-Hellman.
5139 [Nick Mathewson and Ben Laurie]
5141 *) Add functions for well-known primes.
5144 *) Extended Windows CE support.
5145 [Satoshi Nakamura and Andy Polyakov]
5147 *) Initialize SSL_METHOD structures at compile time instead of during
5148 runtime, thus removing the need for a lock.
5151 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5152 attempting to decrypt each encrypted key in turn. Add support to
5156 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5158 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5161 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5164 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5165 key into the same file any more.
5168 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5171 *) Add -utf8 command line and config file option to 'ca'.
5172 [Stefan <stf@udoma.org]
5174 *) Removed the macro des_crypt(), as it seems to conflict with some
5175 libraries. Use DES_crypt().
5178 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5179 involves renaming the source and generated shared-libs for
5180 both. The engines will accept the corrected or legacy ids
5181 ('ncipher' and '4758_cca' respectively) when binding. NB,
5182 this only applies when building 'shared'.
5183 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5185 *) Add attribute functions to EVP_PKEY structure. Modify
5186 PKCS12_create() to recognize a CSP name attribute and
5187 use it. Make -CSP option work again in pkcs12 utility.
5190 *) Add new functionality to the bn blinding code:
5191 - automatic re-creation of the BN_BLINDING parameters after
5192 a fixed number of uses (currently 32)
5193 - add new function for parameter creation
5194 - introduce flags to control the update behaviour of the
5195 BN_BLINDING parameters
5196 - hide BN_BLINDING structure
5197 Add a second BN_BLINDING slot to the RSA structure to improve
5198 performance when a single RSA object is shared among several
5202 *) Add support for DTLS.
5203 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5205 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5206 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5209 *) Remove buggy and incomplete DH cert support from
5210 ssl/ssl_rsa.c and ssl/s3_both.c
5213 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5214 the apps/openssl applications.
5217 *) Compile clean with "-Wall -Wmissing-prototypes
5218 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5219 DEBUG_SAFESTACK must also be set.
5222 *) Change ./Configure so that certain algorithms can be disabled by default.
5223 The new counterpiece to "no-xxx" is "enable-xxx".
5225 The patented RC5 and MDC2 algorithms will now be disabled unless
5226 "enable-rc5" and "enable-mdc2", respectively, are specified.
5228 (IDEA remains enabled despite being patented. This is because IDEA
5229 is frequently required for interoperability, and there is no license
5230 fee for non-commercial use. As before, "no-idea" can be used to
5231 avoid this algorithm.)
5235 *) Add processing of proxy certificates (see RFC 3820). This work was
5236 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5237 EGEE (Enabling Grids for E-science in Europe).
5240 *) RC4 performance overhaul on modern architectures/implementations, such
5241 as Intel P4, IA-64 and AMD64.
5244 *) New utility extract-section.pl. This can be used specify an alternative
5245 section number in a pod file instead of having to treat each file as
5246 a separate case in Makefile. This can be done by adding two lines to the
5249 =for comment openssl_section:XXX
5251 The blank line is mandatory.
5255 *) New arguments -certform, -keyform and -pass for s_client and s_server
5256 to allow alternative format key and certificate files and passphrase
5260 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5261 update associated structures and add various utility functions.
5263 Add new policy related verify parameters, include policy checking in
5264 standard verify code. Enhance 'smime' application with extra parameters
5265 to support policy checking and print out.
5268 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5269 Nehemiah processors. These extensions support AES encryption in hardware
5270 as well as RNG (though RNG support is currently disabled).
5271 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5273 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5276 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5277 [Andy Polyakov and a number of other people]
5279 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5280 implementation contributed by IBM.
5281 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5283 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5284 exponent rather than 'unsigned long'. There is a corresponding change to
5285 the new 'rsa_keygen' element of the RSA_METHOD structure.
5286 [Jelte Jansen, Geoff Thorpe]
5288 *) Functionality for creating the initial serial number file is now
5289 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5291 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5292 number file to 1, which is bound to cause problems. To avoid
5293 the problems while respecting compatibility between different 0.9.7
5294 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5295 CA.pl for serial number initialization. With the new release 0.9.8,
5296 we can fix the problem directly in the 'ca' utility.)
5299 *) Reduced header interdependencies by declaring more opaque objects in
5300 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5301 give fewer recursive includes, which could break lazy source code - so
5302 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5303 developers should define this symbol when building and using openssl to
5304 ensure they track the recommended behaviour, interfaces, [etc], but
5305 backwards-compatible behaviour prevails when this isn't defined.
5308 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5311 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5312 This will generate a random key of the appropriate length based on the
5313 cipher context. The EVP_CIPHER can provide its own random key generation
5314 routine to support keys of a specific form. This is used in the des and
5315 3des routines to generate a key of the correct parity. Update S/MIME
5316 code to use new functions and hence generate correct parity DES keys.
5317 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5318 valid (weak or incorrect parity).
5321 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5322 as looking them up. This is useful when the verified structure may contain
5323 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5324 present unless the new PKCS7_NO_CRL flag is asserted.
5327 *) Extend ASN1 oid configuration module. It now additionally accepts the
5330 shortName = some long name, 1.2.3.4
5333 *) Reimplemented the BN_CTX implementation. There is now no more static
5334 limitation on the number of variables it can handle nor the depth of the
5335 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5336 information can now expand as required, and rather than having a single
5337 static array of bignums, BN_CTX now uses a linked-list of such arrays
5338 allowing it to expand on demand whilst maintaining the usefulness of
5339 BN_CTX's "bundling".
5342 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5343 to allow all RSA operations to function using a single BN_CTX.
5346 *) Preliminary support for certificate policy evaluation and checking. This
5347 is initially intended to pass the tests outlined in "Conformance Testing
5348 of Relying Party Client Certificate Path Processing Logic" v1.07.
5351 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5352 remained unused and not that useful. A variety of other little bignum
5353 tweaks and fixes have also been made continuing on from the audit (see
5357 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5358 associated ASN1, EVP and SSL functions and old ASN1 macros.
5361 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5362 and this should never fail. So the return value from the use of
5363 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5364 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5367 *) BN_CTX_get() should return zero-valued bignums, providing the same
5368 initialised value as BN_new().
5369 [Geoff Thorpe, suggested by Ulf Möller]
5371 *) Support for inhibitAnyPolicy certificate extension.
5374 *) An audit of the BIGNUM code is underway, for which debugging code is
5375 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5376 is considered valid when processing BIGNUMs, and causes execution to
5377 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5378 further steps are taken to deliberately pollute unused data in BIGNUM
5379 structures to try and expose faulty code further on. For now, openssl will
5380 (in its default mode of operation) continue to tolerate the inconsistent
5381 forms that it has tolerated in the past, but authors and packagers should
5382 consider trying openssl and their own applications when compiled with
5383 these debugging symbols defined. It will help highlight potential bugs in
5384 their own code, and will improve the test coverage for OpenSSL itself. At
5385 some point, these tighter rules will become openssl's default to improve
5386 maintainability, though the assert()s and other overheads will remain only
5387 in debugging configurations. See bn.h for more details.
5388 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5390 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5391 that can only be obtained through BN_CTX_new() (which implicitly
5392 initialises it). The presence of this function only made it possible
5393 to overwrite an existing structure (and cause memory leaks).
5396 *) Because of the callback-based approach for implementing LHASH as a
5397 template type, lh_insert() adds opaque objects to hash-tables and
5398 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5399 to clean up those corresponding objects before destroying the hash table
5400 (and losing the object pointers). So some over-zealous constifications in
5401 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5402 objects as "const" and the lh_doall[_arg] callback wrappers are not
5403 prototyped to have "const" restrictions on the object pointers they are
5404 given (and so aren't required to cast them away any more).
5407 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5408 (speed) prefers to use its own implementation. The two implementations
5409 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5410 its object type properly exposed (MS_TM) instead of casting to/from "char
5411 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5412 aren't necessarily the greatest nomenclatures - but this is what was used
5413 internally to the implementation so I've used that for now.
5416 *) Ensure that deprecated functions do not get compiled when
5417 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5418 the self-tests were still using deprecated key-generation functions so
5419 these have been updated also.
5422 *) Reorganise PKCS#7 code to separate the digest location functionality
5423 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5424 New function PKCS7_set_digest() to set the digest type for PKCS#7
5425 digestedData type. Add additional code to correctly generate the
5426 digestedData type and add support for this type in PKCS7 initialization
5430 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5431 structure of type "other".
5434 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5435 sure the loop does correctly stop and breaking ("division by zero")
5436 modulus operations are not performed. The (pre-generated) prime
5437 table crypto/bn/bn_prime.h was already correct, but it could not be
5438 re-generated on some platforms because of the "division by zero"
5439 situation in the script.
5440 [Ralf S. Engelschall]
5442 *) Update support for ECC-based TLS ciphersuites according to
5443 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5444 SHA-1 now is only used for "small" curves (where the
5445 representation of a field element takes up to 24 bytes); for
5446 larger curves, the field element resulting from ECDH is directly
5447 used as premaster secret.
5448 [Douglas Stebila (Sun Microsystems Laboratories)]
5450 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5451 curve secp160r1 to the tests.
5452 [Douglas Stebila (Sun Microsystems Laboratories)]
5454 *) Add the possibility to load symbols globally with DSO.
5455 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5457 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5458 control of the error stack.
5461 *) Add support for STORE in ENGINE.
5464 *) Add the STORE type. The intention is to provide a common interface
5465 to certificate and key stores, be they simple file-based stores, or
5466 HSM-type store, or LDAP stores, or...
5467 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5470 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5471 pass a list of arguments to any function as well as provide a way
5472 for a function to pass data back to the caller.
5475 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5476 works like BUF_strdup() but can be used to duplicate a portion of
5477 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5481 *) Add the function sk_find_ex() which works like sk_find(), but will
5482 return an index to an element even if an exact match couldn't be
5483 found. The index is guaranteed to point at the element where the
5484 searched-for key would be inserted to preserve sorting order.
5487 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5488 takes an extra flags argument for optional functionality. Currently,
5489 the following flags are defined:
5491 OBJ_BSEARCH_VALUE_ON_NOMATCH
5492 This one gets OBJ_bsearch_ex() to return a pointer to the first
5493 element where the comparing function returns a negative or zero
5496 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5497 This one gets OBJ_bsearch_ex() to return a pointer to the first
5498 element where the comparing function returns zero. This is useful
5499 if there are more than one element where the comparing function
5503 *) Make it possible to create self-signed certificates with 'openssl ca'
5504 in such a way that the self-signed certificate becomes part of the
5505 CA database and uses the same mechanisms for serial number generation
5506 as all other certificate signing. The new flag '-selfsign' enables
5507 this functionality. Adapt CA.sh and CA.pl.in.
5510 *) Add functionality to check the public key of a certificate request
5511 against a given private. This is useful to check that a certificate
5512 request can be signed by that key (self-signing).
5515 *) Make it possible to have multiple active certificates with the same
5516 subject in the CA index file. This is done only if the keyword
5517 'unique_subject' is set to 'no' in the main CA section (default
5518 if 'CA_default') of the configuration file. The value is saved
5519 with the database itself in a separate index attribute file,
5520 named like the index file with '.attr' appended to the name.
5523 *) Generate multi-valued AVAs using '+' notation in config files for
5527 *) Support for nameConstraints certificate extension.
5530 *) Support for policyConstraints certificate extension.
5533 *) Support for policyMappings certificate extension.
5536 *) Make sure the default DSA_METHOD implementation only uses its
5537 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5538 and change its own handlers to be NULL so as to remove unnecessary
5539 indirection. This lets alternative implementations fallback to the
5540 default implementation more easily.
5543 *) Support for directoryName in GeneralName related extensions
5547 *) Make it possible to link applications using Makefile.shared.
5548 Make that possible even when linking against static libraries!
5551 *) Support for single pass processing for S/MIME signing. This now
5552 means that S/MIME signing can be done from a pipe, in addition
5553 cleartext signing (multipart/signed type) is effectively streaming
5554 and the signed data does not need to be all held in memory.
5556 This is done with a new flag PKCS7_STREAM. When this flag is set
5557 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5558 is done after the data is output (and digests calculated) in
5559 SMIME_write_PKCS7().
5562 *) Add full support for -rpath/-R, both in shared libraries and
5563 applications, at least on the platforms where it's known how
5567 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5568 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5569 will now compute a table of multiples of the generator that
5570 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5571 faster (notably in the case of a single point multiplication,
5572 scalar * generator).
5573 [Nils Larsch, Bodo Moeller]
5575 *) IPv6 support for certificate extensions. The various extensions
5576 which use the IP:a.b.c.d can now take IPv6 addresses using the
5577 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5581 *) Added an ENGINE that implements RSA by performing private key
5582 exponentiations with the GMP library. The conversions to and from
5583 GMP's mpz_t format aren't optimised nor are any montgomery forms
5584 cached, and on x86 it appears OpenSSL's own performance has caught up.
5585 However there are likely to be other architectures where GMP could
5586 provide a boost. This ENGINE is not built in by default, but it can be
5587 specified at Configure time and should be accompanied by the necessary
5588 linker additions, eg;
5589 ./config -DOPENSSL_USE_GMP -lgmp
5592 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5593 testing availability of engines with "-t" - the old behaviour is
5594 produced by increasing the feature's verbosity with "-tt".
5597 *) ECDSA routines: under certain error conditions uninitialized BN objects
5598 could be freed. Solution: make sure initialization is performed early
5599 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5603 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5604 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5605 software implementations. For DSA and DH, parameter generation can
5606 also be overridden by providing the appropriate method callbacks.
5609 *) Change the "progress" mechanism used in key-generation and
5610 primality testing to functions that take a new BN_GENCB pointer in
5611 place of callback/argument pairs. The new API functions have "_ex"
5612 postfixes and the older functions are reimplemented as wrappers for
5613 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5614 declarations of the old functions to help (graceful) attempts to
5615 migrate to the new functions. Also, the new key-generation API
5616 functions operate on a caller-supplied key-structure and return
5617 success/failure rather than returning a key or NULL - this is to
5618 help make "keygen" another member function of RSA_METHOD etc.
5620 Example for using the new callback interface:
5622 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5626 BN_GENCB_set(&my_cb, my_callback, my_arg);
5628 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5629 /* For the meaning of a, b in calls to my_callback(), see the
5630 * documentation of the function that calls the callback.
5631 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5632 * my_callback should return 1 if it wants BN_is_prime_ex()
5633 * to continue, or 0 to stop.
5638 *) Change the ZLIB compression method to be stateful, and make it
5639 available to TLS with the number defined in
5640 draft-ietf-tls-compression-04.txt.
5643 *) Add the ASN.1 structures and functions for CertificatePair, which
5644 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5646 CertificatePair ::= SEQUENCE {
5647 forward [0] Certificate OPTIONAL,
5648 reverse [1] Certificate OPTIONAL,
5649 -- at least one of the pair shall be present -- }
5651 Also implement the PEM functions to read and write certificate
5652 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5654 This needed to be defined, mostly for the sake of the LDAP
5655 attribute crossCertificatePair, but may prove useful elsewhere as
5659 *) Make it possible to inhibit symlinking of shared libraries in
5660 Makefile.shared, for Cygwin's sake.
5663 *) Extend the BIGNUM API by creating a function
5664 void BN_set_negative(BIGNUM *a, int neg);
5665 and a macro that behave like
5666 int BN_is_negative(const BIGNUM *a);
5668 to avoid the need to access 'a->neg' directly in applications.
5671 *) Implement fast modular reduction for pseudo-Mersenne primes
5672 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5673 EC_GROUP_new_curve_GFp() will now automatically use this
5675 [Nils Larsch <nla@trustcenter.de>]
5677 *) Add new lock type (CRYPTO_LOCK_BN).
5680 *) Change the ENGINE framework to automatically load engines
5681 dynamically from specific directories unless they could be
5682 found to already be built in or loaded. Move all the
5683 current engines except for the cryptodev one to a new
5685 The engines in engines/ are built as shared libraries if
5686 the "shared" options was given to ./Configure or ./config.
5687 Otherwise, they are inserted in libcrypto.a.
5688 /usr/local/ssl/engines is the default directory for dynamic
5689 engines, but that can be overridden at configure time through
5690 the usual use of --prefix and/or --openssldir, and at run
5691 time with the environment variable OPENSSL_ENGINES.
5692 [Geoff Thorpe and Richard Levitte]
5694 *) Add Makefile.shared, a helper makefile to build shared
5695 libraries. Adapt Makefile.org.
5698 *) Add version info to Win32 DLLs.
5699 [Peter 'Luna' Runestig" <peter@runestig.com>]
5701 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5702 can be added using this API to created arbitrary PKCS#12
5703 files while avoiding the low level API.
5705 New options to PKCS12_create(), key or cert can be NULL and
5706 will then be omitted from the output file. The encryption
5707 algorithm NIDs can be set to -1 for no encryption, the mac
5708 iteration count can be set to 0 to omit the mac.
5710 Enhance pkcs12 utility by making the -nokeys and -nocerts
5711 options work when creating a PKCS#12 file. New option -nomac
5712 to omit the mac, NONE can be set for an encryption algorithm.
5713 New code is modified to use the enhanced PKCS12_create()
5714 instead of the low level API.
5717 *) Extend ASN1 encoder to support indefinite length constructed
5718 encoding. This can output sequences tags and octet strings in
5719 this form. Modify pk7_asn1.c to support indefinite length
5720 encoding. This is experimental and needs additional code to
5721 be useful, such as an ASN1 bio and some enhanced streaming
5724 Extend template encode functionality so that tagging is passed
5725 down to the template encoder.
5728 *) Let 'openssl req' fail if an argument to '-newkey' is not
5729 recognized instead of using RSA as a default.
5732 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5733 As these are not official, they are not included in "ALL";
5734 the "ECCdraft" ciphersuite group alias can be used to select them.
5735 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5737 *) Add ECDH engine support.
5738 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5740 *) Add ECDH in new directory crypto/ecdh/.
5741 [Douglas Stebila (Sun Microsystems Laboratories)]
5743 *) Let BN_rand_range() abort with an error after 100 iterations
5744 without success (which indicates a broken PRNG).
5747 *) Change BN_mod_sqrt() so that it verifies that the input value
5748 is really the square of the return value. (Previously,
5749 BN_mod_sqrt would show GIGO behaviour.)
5752 *) Add named elliptic curves over binary fields from X9.62, SECG,
5753 and WAP/WTLS; add OIDs that were still missing.
5755 [Sheueling Chang Shantz and Douglas Stebila
5756 (Sun Microsystems Laboratories)]
5758 *) Extend the EC library for elliptic curves over binary fields
5759 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5762 EC_GF2m_simple_method
5766 EC_GROUP_new_curve_GF2m
5767 EC_GROUP_set_curve_GF2m
5768 EC_GROUP_get_curve_GF2m
5769 EC_POINT_set_affine_coordinates_GF2m
5770 EC_POINT_get_affine_coordinates_GF2m
5771 EC_POINT_set_compressed_coordinates_GF2m
5773 Point compression for binary fields is disabled by default for
5774 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5777 As binary polynomials are represented as BIGNUMs, various members
5778 of the EC_GROUP and EC_POINT data structures can be shared
5779 between the implementations for prime fields and binary fields;
5780 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5781 are essentially identical to their ..._GFp counterparts.
5782 (For simplicity, the '..._GFp' prefix has been dropped from
5783 various internal method names.)
5785 An internal 'field_div' method (similar to 'field_mul' and
5786 'field_sqr') has been added; this is used only for binary fields.
5788 [Sheueling Chang Shantz and Douglas Stebila
5789 (Sun Microsystems Laboratories)]
5791 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5792 through methods ('mul', 'precompute_mult').
5794 The generic implementations (now internally called 'ec_wNAF_mul'
5795 and 'ec_wNAF_precomputed_mult') remain the default if these
5796 methods are undefined.
5798 [Sheueling Chang Shantz and Douglas Stebila
5799 (Sun Microsystems Laboratories)]
5801 *) New function EC_GROUP_get_degree, which is defined through
5802 EC_METHOD. For curves over prime fields, this returns the bit
5803 length of the modulus.
5805 [Sheueling Chang Shantz and Douglas Stebila
5806 (Sun Microsystems Laboratories)]
5808 *) New functions EC_GROUP_dup, EC_POINT_dup.
5809 (These simply call ..._new and ..._copy).
5811 [Sheueling Chang Shantz and Douglas Stebila
5812 (Sun Microsystems Laboratories)]
5814 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5815 Polynomials are represented as BIGNUMs (where the sign bit is not
5816 used) in the following functions [macros]:
5819 BN_GF2m_sub [= BN_GF2m_add]
5820 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5821 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5822 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5824 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5825 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5826 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5827 BN_GF2m_cmp [= BN_ucmp]
5829 (Note that only the 'mod' functions are actually for fields GF(2^m).
5830 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5832 For some functions, an the irreducible polynomial defining a
5833 field can be given as an 'unsigned int[]' with strictly
5834 decreasing elements giving the indices of those bits that are set;
5835 i.e., p[] represents the polynomial
5836 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5838 p[0] > p[1] > ... > p[k] = 0.
5839 This applies to the following functions:
5844 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5845 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5847 BN_GF2m_mod_sqrt_arr
5848 BN_GF2m_mod_solve_quad_arr
5852 Conversion can be performed by the following functions:
5857 bntest.c has additional tests for binary polynomial arithmetic.
5859 Two implementations for BN_GF2m_mod_div() are available.
5860 The default algorithm simply uses BN_GF2m_mod_inv() and
5861 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5862 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5863 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5865 [Sheueling Chang Shantz and Douglas Stebila
5866 (Sun Microsystems Laboratories)]
5868 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5869 functionality is disabled at compile-time.
5870 [Douglas Stebila <douglas.stebila@sun.com>]
5872 *) Change default behaviour of 'openssl asn1parse' so that more
5873 information is visible when viewing, e.g., a certificate:
5875 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5876 mode the content of non-printable OCTET STRINGs is output in a
5877 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5878 avoid the appearance of a printable string.
5879 [Nils Larsch <nla@trustcenter.de>]
5881 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5883 EC_GROUP_set_asn1_flag()
5884 EC_GROUP_get_asn1_flag()
5885 EC_GROUP_set_point_conversion_form()
5886 EC_GROUP_get_point_conversion_form()
5887 These control ASN1 encoding details:
5888 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5889 has been set to OPENSSL_EC_NAMED_CURVE.
5890 - Points are encoded in uncompressed form by default; options for
5891 asn1_for are as for point2oct, namely
5892 POINT_CONVERSION_COMPRESSED
5893 POINT_CONVERSION_UNCOMPRESSED
5894 POINT_CONVERSION_HYBRID
5896 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5899 EC_GROUP_get0_seed()
5900 EC_GROUP_get_seed_len()
5901 This is used only for ASN1 purposes (so far).
5902 [Nils Larsch <nla@trustcenter.de>]
5904 *) Add 'field_type' member to EC_METHOD, which holds the NID
5905 of the appropriate field type OID. The new function
5906 EC_METHOD_get_field_type() returns this value.
5907 [Nils Larsch <nla@trustcenter.de>]
5912 EC_POINT_point2hex()
5913 EC_POINT_hex2point()
5914 providing useful interfaces to EC_POINT_point2oct() and
5915 EC_POINT_oct2point().
5916 [Nils Larsch <nla@trustcenter.de>]
5918 *) Change internals of the EC library so that the functions
5919 EC_GROUP_set_generator()
5920 EC_GROUP_get_generator()
5921 EC_GROUP_get_order()
5922 EC_GROUP_get_cofactor()
5923 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5924 to methods, which would lead to unnecessary code duplication when
5925 adding different types of curves.
5926 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5928 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5929 arithmetic, and such that modified wNAFs are generated
5930 (which avoid length expansion in many cases).
5933 *) Add a function EC_GROUP_check_discriminant() (defined via
5934 EC_METHOD) that verifies that the curve discriminant is non-zero.
5936 Add a function EC_GROUP_check() that makes some sanity tests
5937 on a EC_GROUP, its generator and order. This includes
5938 EC_GROUP_check_discriminant().
5939 [Nils Larsch <nla@trustcenter.de>]
5941 *) Add ECDSA in new directory crypto/ecdsa/.
5943 Add applications 'openssl ecparam' and 'openssl ecdsa'
5944 (these are based on 'openssl dsaparam' and 'openssl dsa').
5946 ECDSA support is also included in various other files across the
5947 library. Most notably,
5948 - 'openssl req' now has a '-newkey ecdsa:file' option;
5949 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5950 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5951 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5952 them suitable for ECDSA where domain parameters must be
5953 extracted before the specific public key;
5954 - ECDSA engine support has been added.
5955 [Nils Larsch <nla@trustcenter.de>]
5957 *) Include some named elliptic curves, and add OIDs from X9.62,
5958 SECG, and WAP/WTLS. Each curve can be obtained from the new
5960 EC_GROUP_new_by_curve_name(),
5961 and the list of available named curves can be obtained with
5962 EC_get_builtin_curves().
5963 Also add a 'curve_name' member to EC_GROUP objects, which can be
5965 EC_GROUP_set_curve_name()
5966 EC_GROUP_get_curve_name()
5967 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5969 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5970 was actually never needed) and in BN_mul(). The removal in BN_mul()
5971 required a small change in bn_mul_part_recursive() and the addition
5972 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5973 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5974 bn_sub_words() and bn_add_words() except they take arrays with
5978 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5980 *) Cleanse PEM buffers before freeing them since they may contain
5982 [Benjamin Bennett <ben@psc.edu>]
5984 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5985 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5986 authentication-only ciphersuites.
5989 *) Since AES128 and AES256 share a single mask bit in the logic of
5990 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5991 kludge to work properly if AES128 is available and AES256 isn't.
5994 *) Expand security boundary to match 1.1.1 module.
5997 *) Remove redundant features: hash file source, editing of test vectors
5998 modify fipsld to use external fips_premain.c signature.
6001 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6002 run algorithm test programs.
6005 *) Make algorithm test programs more tolerant of whitespace.
6008 *) Have SSL/TLS server implementation tolerate "mismatched" record
6009 protocol version while receiving ClientHello even if the
6010 ClientHello is fragmented. (The server can't insist on the
6011 particular protocol version it has chosen before the ServerHello
6012 message has informed the client about his choice.)
6015 *) Load error codes if they are not already present instead of using a
6016 static variable. This allows them to be cleanly unloaded and reloaded.
6019 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6021 *) Introduce limits to prevent malicious keys being able to
6022 cause a denial of service. (CVE-2006-2940)
6023 [Steve Henson, Bodo Moeller]
6025 *) Fix ASN.1 parsing of certain invalid structures that can result
6026 in a denial of service. (CVE-2006-2937) [Steve Henson]
6028 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6029 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6031 *) Fix SSL client code which could crash if connecting to a
6032 malicious SSLv2 server. (CVE-2006-4343)
6033 [Tavis Ormandy and Will Drewry, Google Security Team]
6035 *) Change ciphersuite string processing so that an explicit
6036 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6037 will no longer include "AES128-SHA"), and any other similar
6038 ciphersuite (same bitmap) from *other* protocol versions (so that
6039 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6040 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6041 changes from 0.9.8b and 0.9.8d.
6044 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6046 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6047 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6049 *) Change the Unix randomness entropy gathering to use poll() when
6050 possible instead of select(), since the latter has some
6051 undesirable limitations.
6052 [Darryl Miles via Richard Levitte and Bodo Moeller]
6054 *) Disable rogue ciphersuites:
6056 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6057 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6058 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6060 The latter two were purportedly from
6061 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6064 Also deactivate the remaining ciphersuites from
6065 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6066 unofficial, and the ID has long expired.
6069 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6070 dual-core machines) and other potential thread-safety issues.
6073 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6075 *) Adapt fipsld and the build system to link against the validated FIPS
6076 module in FIPS mode.
6079 *) Fixes for VC++ 2005 build under Windows.
6082 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6083 from a Windows bash shell such as MSYS. It is autodetected from the
6084 "config" script when run from a VC++ environment. Modify standard VC++
6085 build to use fipscanister.o from the GNU make build.
6088 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6090 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6091 The value now differs depending on if you build for FIPS or not.
6092 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6093 safely run with a non-FIPSed libcrypto, as it may crash because of
6094 the difference induced by this change.
6097 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6099 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6100 (part of SSL_OP_ALL). This option used to disable the
6101 countermeasure against man-in-the-middle protocol-version
6102 rollback in the SSL 2.0 server implementation, which is a bad
6103 idea. (CVE-2005-2969)
6105 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6106 for Information Security, National Institute of Advanced Industrial
6107 Science and Technology [AIST], Japan)]
6109 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6110 mainly for FIPS compliance and not fully integrated at this stage.
6113 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6114 the exponentiation using a fixed-length exponent. (Otherwise,
6115 the information leaked through timing could expose the secret key
6116 after many signatures; cf. Bleichenbacher's attack on DSA with
6120 *) Make a new fixed-window mod_exp implementation the default for
6121 RSA, DSA, and DH private-key operations so that the sequence of
6122 squares and multiplies and the memory access pattern are
6123 independent of the particular secret key. This will mitigate
6124 cache-timing and potential related attacks.
6126 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6127 and this is automatically used by BN_mod_exp_mont() if the new flag
6128 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6129 will use this BN flag for private exponents unless the flag
6130 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6131 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6133 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6135 *) Change the client implementation for SSLv23_method() and
6136 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6137 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6138 (Previously, the SSL 2.0 backwards compatible Client Hello
6139 message format would be used even with SSL_OP_NO_SSLv2.)
6142 *) Add support for smime-type MIME parameter in S/MIME messages which some
6146 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6147 a threadsafe manner. Modify rsa code to use new function and add calls
6148 to dsa and dh code (which had race conditions before).
6151 *) Include the fixed error library code in the C error file definitions
6152 instead of fixing them up at runtime. This keeps the error code
6153 structures constant.
6156 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6158 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6161 *) Fixes for newer kerberos headers. NB: the casts are needed because
6162 the 'length' field is signed on one version and unsigned on another
6163 with no (?) obvious way to tell the difference, without these VC++
6164 complains. Also the "definition" of FAR (blank) is no longer included
6165 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6166 some needed definitions.
6169 *) Undo Cygwin change.
6172 *) Added support for proxy certificates according to RFC 3820.
6173 Because they may be a security thread to unaware applications,
6174 they must be explicitly allowed in run-time. See
6175 docs/HOWTO/proxy_certificates.txt for further information.
6178 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6180 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6181 server and client random values. Previously
6182 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6183 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6185 This change has negligible security impact because:
6187 1. Server and client random values still have 24 bytes of pseudo random
6190 2. Server and client random values are sent in the clear in the initial
6193 3. The master secret is derived using the premaster secret (48 bytes in
6194 size for static RSA ciphersuites) as well as client server and random
6197 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6200 [Stephen Henson, reported by UK NISCC]
6202 *) Use Windows randomness collection on Cygwin.
6205 *) Fix hang in EGD/PRNGD query when communication socket is closed
6206 prematurely by EGD/PRNGD.
6207 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6209 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6212 *) Back-port of selected performance improvements from development
6213 branch, as well as improved support for PowerPC platforms.
6216 *) Add lots of checks for memory allocation failure, error codes to indicate
6217 failure and freeing up memory if a failure occurs.
6218 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6220 *) Add new -passin argument to dgst.
6223 *) Perform some character comparisons of different types in X509_NAME_cmp:
6224 this is needed for some certificates that re-encode DNs into UTF8Strings
6225 (in violation of RFC3280) and can't or won't issue name rollover
6229 *) Make an explicit check during certificate validation to see that
6230 the CA setting in each certificate on the chain is correct. As a
6231 side effect always do the following basic checks on extensions,
6232 not just when there's an associated purpose to the check:
6234 - if there is an unhandled critical extension (unless the user
6235 has chosen to ignore this fault)
6236 - if the path length has been exceeded (if one is set at all)
6237 - that certain extensions fit the associated purpose (if one has
6241 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6243 *) Avoid a race condition when CRLs are checked in a multi threaded
6244 environment. This would happen due to the reordering of the revoked
6245 entries during signature checking and serial number lookup. Now the
6246 encoding is cached and the serial number sort performed under a lock.
6247 Add new STACK function sk_is_sorted().
6250 *) Add Delta CRL to the extension code.
6253 *) Various fixes to s3_pkt.c so alerts are sent properly.
6254 [David Holmes <d.holmes@f5.com>]
6256 *) Reduce the chances of duplicate issuer name and serial numbers (in
6257 violation of RFC3280) using the OpenSSL certificate creation utilities.
6258 This is done by creating a random 64 bit value for the initial serial
6259 number when a serial number file is created or when a self signed
6260 certificate is created using 'openssl req -x509'. The initial serial
6261 number file is created using 'openssl x509 -next_serial' in CA.pl
6262 rather than being initialized to 1.
6265 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6267 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6268 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6269 [Joe Orton, Steve Henson]
6271 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6273 [Joe Orton, Steve Henson]
6275 *) Make it possible to have multiple active certificates with the same
6276 subject in the CA index file. This is done only if the keyword
6277 'unique_subject' is set to 'no' in the main CA section (default
6278 if 'CA_default') of the configuration file. The value is saved
6279 with the database itself in a separate index attribute file,
6280 named like the index file with '.attr' appended to the name.
6283 *) X509 verify fixes. Disable broken certificate workarounds when
6284 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6285 keyUsage extension present. Don't accept CRLs with unhandled critical
6286 extensions: since verify currently doesn't process CRL extensions this
6287 rejects a CRL with *any* critical extensions. Add new verify error codes
6291 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6292 A clarification of RFC2560 will require the use of OCTET STRINGs and
6293 some implementations cannot handle the current raw format. Since OpenSSL
6294 copies and compares OCSP nonces as opaque blobs without any attempt at
6295 parsing them this should not create any compatibility issues.
6298 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6299 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6300 this HMAC (and other) operations are several times slower than OpenSSL
6304 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6305 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6307 *) Use the correct content when signing type "other".
6310 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6312 *) Fix various bugs revealed by running the NISCC test suite:
6314 Stop out of bounds reads in the ASN1 code when presented with
6315 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6317 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6319 If verify callback ignores invalid public key errors don't try to check
6320 certificate signature with the NULL public key.
6324 *) New -ignore_err option in ocsp application to stop the server
6325 exiting on the first error in a request.
6328 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6329 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6333 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6334 extra data after the compression methods not only for TLS 1.0
6335 but also for SSL 3.0 (as required by the specification).
6336 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6338 *) Change X509_certificate_type() to mark the key as exported/exportable
6339 when it's 512 *bits* long, not 512 bytes.
6342 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6343 blocks during encryption.
6346 *) Various fixes to base64 BIO and non blocking I/O. On write
6347 flushes were not handled properly if the BIO retried. On read
6348 data was not being buffered properly and had various logic bugs.
6349 This also affects blocking I/O when the data being decoded is a
6353 *) Various S/MIME bugfixes and compatibility changes:
6354 output correct application/pkcs7 MIME type if
6355 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6356 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6357 of files as .eml work). Correctly handle very long lines in MIME
6361 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6363 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6364 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6365 a protocol version number mismatch like a decryption error
6366 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6369 *) Turn on RSA blinding by default in the default implementation
6370 to avoid a timing attack. Applications that don't want it can call
6371 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6372 They would be ill-advised to do so in most cases.
6373 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6375 *) Change RSA blinding code so that it works when the PRNG is not
6376 seeded (in this case, the secret RSA exponent is abused as
6377 an unpredictable seed -- if it is not unpredictable, there
6378 is no point in blinding anyway). Make RSA blinding thread-safe
6379 by remembering the creator's thread ID in rsa->blinding and
6380 having all other threads use local one-time blinding factors
6381 (this requires more computation than sharing rsa->blinding, but
6382 avoids excessive locking; and if an RSA object is not shared
6383 between threads, blinding will still be very fast).
6386 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6387 ENGINE as defaults for all supported algorithms irrespective of
6388 the 'flags' parameter. 'flags' is now honoured, so applications
6389 should make sure they are passing it correctly.
6392 *) Target "mingw" now allows native Windows code to be generated in
6393 the Cygwin environment as well as with the MinGW compiler.
6396 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6398 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6399 via timing by performing a MAC computation even if incorrect
6400 block cipher padding has been found. This is a countermeasure
6401 against active attacks where the attacker has to distinguish
6402 between bad padding and a MAC verification error. (CVE-2003-0078)
6404 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6405 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6406 Martin Vuagnoux (EPFL, Ilion)]
6408 *) Make the no-err option work as intended. The intention with no-err
6409 is not to have the whole error stack handling routines removed from
6410 libcrypto, it's only intended to remove all the function name and
6411 reason texts, thereby removing some of the footprint that may not
6412 be interesting if those errors aren't displayed anyway.
6414 NOTE: it's still possible for any application or module to have it's
6415 own set of error texts inserted. The routines are there, just not
6416 used by default when no-err is given.
6419 *) Add support for FreeBSD on IA64.
6420 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6422 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6423 Kerberos function mit_des_cbc_cksum(). Before this change,
6424 the value returned by DES_cbc_cksum() was like the one from
6425 mit_des_cbc_cksum(), except the bytes were swapped.
6426 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6428 *) Allow an application to disable the automatic SSL chain building.
6429 Before this a rather primitive chain build was always performed in
6430 ssl3_output_cert_chain(): an application had no way to send the
6431 correct chain if the automatic operation produced an incorrect result.
6433 Now the chain builder is disabled if either:
6435 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6437 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6439 The reasoning behind this is that an application would not want the
6440 auto chain building to take place if extra chain certificates are
6441 present and it might also want a means of sending no additional
6442 certificates (for example the chain has two certificates and the
6446 *) Add the possibility to build without the ENGINE framework.
6447 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6449 *) Under Win32 gmtime() can return NULL: check return value in
6450 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6453 *) DSA routines: under certain error conditions uninitialized BN objects
6454 could be freed. Solution: make sure initialization is performed early
6455 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6456 Nils Larsch <nla@trustcenter.de> via PR#459)
6459 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6460 checked on reconnect on the client side, therefore session resumption
6461 could still fail with a "ssl session id is different" error. This
6462 behaviour is masked when SSL_OP_ALL is used due to
6463 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6464 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6465 followup to PR #377.
6468 *) IA-32 assembler support enhancements: unified ELF targets, support
6469 for SCO/Caldera platforms, fix for Cygwin shared build.
6472 *) Add support for FreeBSD on sparc64. As a consequence, support for
6473 FreeBSD on non-x86 processors is separate from x86 processors on
6474 the config script, much like the NetBSD support.
6475 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6477 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6479 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6482 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6483 code (06) was taken as the first octet of the session ID and the last
6484 octet was ignored consequently. As a result SSLv2 client side session
6485 caching could not have worked due to the session ID mismatch between
6487 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6491 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6492 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6496 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6497 seems that in spite of existing for more than a year, many application
6498 author have done nothing to provide the necessary callbacks, which
6499 means that this particular engine will not work properly anywhere.
6500 This is a very unfortunate situation which forces us, in the name
6501 of usability, to give the hw_ncipher.c a static lock, which is part
6503 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6504 appear in 0.9.8 or later. We EXPECT application authors to have
6505 dealt properly with this when 0.9.8 is released (unless we actually
6506 make such changes in the libcrypto locking code that changes will
6507 have to be made anyway).
6510 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6511 octets have been read, EOF or an error occurs. Without this change
6512 some truncated ASN1 structures will not produce an error.
6515 *) Disable Heimdal support, since it hasn't been fully implemented.
6516 Still give the possibility to force the use of Heimdal, but with
6517 warnings and a request that patches get sent to openssl-dev.
6520 *) Add the VC-CE target, introduce the WINCE sysname, and add
6521 INSTALL.WCE and appropriate conditionals to make it build.
6522 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6524 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6525 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6526 edit numbers of the version.
6527 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6529 *) Introduce safe string copy and catenation functions
6530 (BUF_strlcpy() and BUF_strlcat()).
6531 [Ben Laurie (CHATS) and Richard Levitte]
6533 *) Avoid using fixed-size buffers for one-line DNs.
6534 [Ben Laurie (CHATS)]
6536 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6537 resizing buffers containing secrets, and use where appropriate.
6538 [Ben Laurie (CHATS)]
6540 *) Avoid using fixed size buffers for configuration file location.
6541 [Ben Laurie (CHATS)]
6543 *) Avoid filename truncation for various CA files.
6544 [Ben Laurie (CHATS)]
6546 *) Use sizeof in preference to magic numbers.
6547 [Ben Laurie (CHATS)]
6549 *) Avoid filename truncation in cert requests.
6550 [Ben Laurie (CHATS)]
6552 *) Add assertions to check for (supposedly impossible) buffer
6554 [Ben Laurie (CHATS)]
6556 *) Don't cache truncated DNS entries in the local cache (this could
6557 potentially lead to a spoofing attack).
6558 [Ben Laurie (CHATS)]
6560 *) Fix various buffers to be large enough for hex/decimal
6561 representations in a platform independent manner.
6562 [Ben Laurie (CHATS)]
6564 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6565 resizing buffers containing secrets, and use where appropriate.
6566 [Ben Laurie (CHATS)]
6568 *) Add BIO_indent() to avoid much slightly worrying code to do
6570 [Ben Laurie (CHATS)]
6572 *) Convert sprintf()/BIO_puts() to BIO_printf().
6573 [Ben Laurie (CHATS)]
6575 *) buffer_gets() could terminate with the buffer only half
6577 [Ben Laurie (CHATS)]
6579 *) Add assertions to prevent user-supplied crypto functions from
6580 overflowing internal buffers by having large block sizes, etc.
6581 [Ben Laurie (CHATS)]
6583 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6585 [Ben Laurie (CHATS)]
6587 *) Eliminate unused copy of key in RC4.
6588 [Ben Laurie (CHATS)]
6590 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6591 [Ben Laurie (CHATS)]
6593 *) Fix off-by-one error in EGD path.
6594 [Ben Laurie (CHATS)]
6596 *) If RANDFILE path is too long, ignore instead of truncating.
6597 [Ben Laurie (CHATS)]
6599 *) Eliminate unused and incorrectly sized X.509 structure
6601 [Ben Laurie (CHATS)]
6603 *) Eliminate unused and dangerous function knumber().
6604 [Ben Laurie (CHATS)]
6606 *) Eliminate unused and dangerous structure, KSSL_ERR.
6607 [Ben Laurie (CHATS)]
6609 *) Protect against overlong session ID context length in an encoded
6610 session object. Since these are local, this does not appear to be
6612 [Ben Laurie (CHATS)]
6614 *) Change from security patch (see 0.9.6e below) that did not affect
6615 the 0.9.6 release series:
6617 Remote buffer overflow in SSL3 protocol - an attacker could
6618 supply an oversized master key in Kerberos-enabled versions.
6620 [Ben Laurie (CHATS)]
6622 *) Change the SSL kerb5 codes to match RFC 2712.
6625 *) Make -nameopt work fully for req and add -reqopt switch.
6626 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6628 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6629 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6631 *) Make sure tests can be performed even if the corresponding algorithms
6632 have been removed entirely. This was also the last step to make
6633 OpenSSL compilable with DJGPP under all reasonable conditions.
6634 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6636 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6637 to allow version independent disabling of normally unselected ciphers,
6638 which may be activated as a side-effect of selecting a single cipher.
6640 (E.g., cipher list string "RSA" enables ciphersuites that are left
6641 out of "ALL" because they do not provide symmetric encryption.
6642 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6643 [Lutz Jaenicke, Bodo Moeller]
6645 *) Add appropriate support for separate platform-dependent build
6646 directories. The recommended way to make a platform-dependent
6647 build directory is the following (tested on Linux), maybe with
6650 # Place yourself outside of the OpenSSL source tree. In
6651 # this example, the environment variable OPENSSL_SOURCE
6652 # is assumed to contain the absolute OpenSSL source directory.
6653 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6654 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6655 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6656 mkdir -p `dirname $F`
6657 ln -s $OPENSSL_SOURCE/$F $F
6660 To be absolutely sure not to disturb the source tree, a "make clean"
6661 is a good thing. If it isn't successful, don't worry about it,
6662 it probably means the source directory is very clean.
6665 *) Make sure any ENGINE control commands make local copies of string
6666 pointers passed to them whenever necessary. Otherwise it is possible
6667 the caller may have overwritten (or deallocated) the original string
6668 data when a later ENGINE operation tries to use the stored values.
6669 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6671 *) Improve diagnostics in file reading and command-line digests.
6672 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6674 *) Add AES modes CFB and OFB to the object database. Correct an
6675 error in AES-CFB decryption.
6678 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6679 allows existing EVP_CIPHER_CTX structures to be reused after
6680 calling EVP_*Final(). This behaviour is used by encryption
6681 BIOs and some applications. This has the side effect that
6682 applications must explicitly clean up cipher contexts with
6683 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6686 *) Check the values of dna and dnb in bn_mul_recursive before calling
6687 bn_mul_comba (a non zero value means the a or b arrays do not contain
6688 n2 elements) and fallback to bn_mul_normal if either is not zero.
6691 *) Fix escaping of non-ASCII characters when using the -subj option
6692 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6695 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6696 form for "surname", serialNumber has no short form.
6697 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6698 therefore remove "mail" short name for "internet 7".
6699 The OID for unique identifiers in X509 certificates is
6700 x500UniqueIdentifier, not uniqueIdentifier.
6701 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6704 *) Add an "init" command to the ENGINE config module and auto initialize
6705 ENGINEs. Without any "init" command the ENGINE will be initialized
6706 after all ctrl commands have been executed on it. If init=1 the
6707 ENGINE is initialized at that point (ctrls before that point are run
6708 on the uninitialized ENGINE and after on the initialized one). If
6709 init=0 then the ENGINE will not be initialized at all.
6712 *) Fix the 'app_verify_callback' interface so that the user-defined
6713 argument is actually passed to the callback: In the
6714 SSL_CTX_set_cert_verify_callback() prototype, the callback
6715 declaration has been changed from
6718 int (*cb)(X509_STORE_CTX *,void *);
6719 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6720 i=s->ctx->app_verify_callback(&ctx)
6721 has been changed into
6722 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6724 To update applications using SSL_CTX_set_cert_verify_callback(),
6725 a dummy argument can be added to their callback functions.
6726 [D. K. Smetters <smetters@parc.xerox.com>]
6728 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6729 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6731 *) Add and OPENSSL_LOAD_CONF define which will cause
6732 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6733 This allows older applications to transparently support certain
6734 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6735 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6736 load the config file and OPENSSL_add_all_algorithms_conf() which will
6737 always load it have also been added.
6740 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6741 Adjust NIDs and EVP layer.
6742 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6744 *) Config modules support in openssl utility.
6746 Most commands now load modules from the config file,
6747 though in a few (such as version) this isn't done
6748 because it couldn't be used for anything.
6750 In the case of ca and req the config file used is
6751 the same as the utility itself: that is the -config
6752 command line option can be used to specify an
6756 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6757 use "openssl_conf" if filename is NULL use default openssl config file.
6760 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6761 config section name. Add a new flag to tolerate a missing config file
6762 and move code to CONF_modules_load_file().
6765 *) Support for crypto accelerator cards from Accelerated Encryption
6766 Processing, www.aep.ie. (Use engine 'aep')
6767 The support was copied from 0.9.6c [engine] and adapted/corrected
6768 to work with the new engine framework.
6769 [AEP Inc. and Richard Levitte]
6771 *) Support for SureWare crypto accelerator cards from Baltimore
6772 Technologies. (Use engine 'sureware')
6773 The support was copied from 0.9.6c [engine] and adapted
6774 to work with the new engine framework.
6777 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6778 make the newer ENGINE framework commands for the CHIL engine work.
6779 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6781 *) Make it possible to produce shared libraries on ReliantUNIX.
6782 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6784 *) Add the configuration target debug-linux-ppro.
6785 Make 'openssl rsa' use the general key loading routines
6786 implemented in apps.c, and make those routines able to
6787 handle the key format FORMAT_NETSCAPE and the variant
6789 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6791 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6792 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6794 *) Add -keyform to rsautl, and document -engine.
6795 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6797 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6798 BIO_R_NO_SUCH_FILE error code rather than the generic
6799 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6802 *) Add new functions
6804 ERR_peek_last_error_line
6805 ERR_peek_last_error_line_data.
6806 These are similar to
6809 ERR_peek_error_line_data,
6810 but report on the latest error recorded rather than the first one
6811 still in the error queue.
6812 [Ben Laurie, Bodo Moeller]
6814 *) default_algorithms option in ENGINE config module. This allows things
6816 default_algorithms = ALL
6817 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6820 *) Preliminary ENGINE config module.
6823 *) New experimental application configuration code.
6826 *) Change the AES code to follow the same name structure as all other
6827 symmetric ciphers, and behave the same way. Move everything to
6828 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6829 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6831 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6832 [Ben Laurie and Theo de Raadt]
6834 *) Add option to output public keys in req command.
6835 [Massimiliano Pala madwolf@openca.org]
6837 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6838 (up to about 10% better than before for P-192 and P-224).
6841 *) New functions/macros
6843 SSL_CTX_set_msg_callback(ctx, cb)
6844 SSL_CTX_set_msg_callback_arg(ctx, arg)
6845 SSL_set_msg_callback(ssl, cb)
6846 SSL_set_msg_callback_arg(ssl, arg)
6848 to request calling a callback function
6850 void cb(int write_p, int version, int content_type,
6851 const void *buf, size_t len, SSL *ssl, void *arg)
6853 whenever a protocol message has been completely received
6854 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6855 protocol version according to which the SSL library interprets
6856 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6857 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6858 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6859 specification (change_cipher_spec(20), alert(21), handshake(22)).
6860 'buf' and 'len' point to the actual message, 'ssl' to the
6861 SSL object, and 'arg' is the application-defined value set by
6862 SSL[_CTX]_set_msg_callback_arg().
6864 'openssl s_client' and 'openssl s_server' have new '-msg' options
6865 to enable a callback that displays all protocol messages.
6868 *) Change the shared library support so shared libraries are built as
6869 soon as the corresponding static library is finished, and thereby get
6870 openssl and the test programs linked against the shared library.
6871 This still only happens when the keyword "shard" has been given to
6872 the configuration scripts.
6874 NOTE: shared library support is still an experimental thing, and
6875 backward binary compatibility is still not guaranteed.
6876 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6878 *) Add support for Subject Information Access extension.
6879 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6881 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6882 additional bytes when new memory had to be allocated, not just
6883 when reusing an existing buffer.
6886 *) New command line and configuration option 'utf8' for the req command.
6887 This allows field values to be specified as UTF8 strings.
6890 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6891 runs for the former and machine-readable output for the latter.
6894 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6895 of the e-mail address in the DN (i.e., it will go into a certificate
6896 extension only). The new configuration file option 'email_in_dn = no'
6897 has the same effect.
6898 [Massimiliano Pala madwolf@openca.org]
6900 *) Change all functions with names starting with des_ to be starting
6901 with DES_ instead. Add wrappers that are compatible with libdes,
6902 but are named _ossl_old_des_*. Finally, add macros that map the
6903 des_* symbols to the corresponding _ossl_old_des_* if libdes
6904 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6905 desired, the des_* symbols will be mapped to DES_*, with one
6908 Since we provide two compatibility mappings, the user needs to
6909 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6910 compatibility is desired. The default (i.e., when that macro
6911 isn't defined) is OpenSSL 0.9.6c compatibility.
6913 There are also macros that enable and disable the support of old
6914 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6915 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6916 are defined, the default will apply: to support the old des routines.
6918 In either case, one must include openssl/des.h to get the correct
6919 definitions. Do not try to just include openssl/des_old.h, that
6922 NOTE: This is a major break of an old API into a new one. Software
6923 authors are encouraged to switch to the DES_ style functions. Some
6924 time in the future, des_old.h and the libdes compatibility functions
6925 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6926 default), and then completely removed.
6929 *) Test for certificates which contain unsupported critical extensions.
6930 If such a certificate is found during a verify operation it is
6931 rejected by default: this behaviour can be overridden by either
6932 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6933 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6934 X509_supported_extension() has also been added which returns 1 if a
6935 particular extension is supported.
6938 *) Modify the behaviour of EVP cipher functions in similar way to digests
6939 to retain compatibility with existing code.
6942 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6943 compatibility with existing code. In particular the 'ctx' parameter does
6944 not have to be to be initialized before the call to EVP_DigestInit() and
6945 it is tidied up after a call to EVP_DigestFinal(). New function
6946 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6947 EVP_MD_CTX_copy() changed to not require the destination to be
6948 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6949 requires the destination to be valid.
6951 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6952 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6955 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6956 so that complete 'Handshake' protocol structures are kept in memory
6957 instead of overwriting 'msg_type' and 'length' with 'body' data.
6960 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6961 [Massimo Santin via Richard Levitte]
6963 *) Major restructuring to the underlying ENGINE code. This includes
6964 reduction of linker bloat, separation of pure "ENGINE" manipulation
6965 (initialisation, etc) from functionality dealing with implementations
6966 of specific crypto interfaces. This change also introduces integrated
6967 support for symmetric ciphers and digest implementations - so ENGINEs
6968 can now accelerate these by providing EVP_CIPHER and EVP_MD
6969 implementations of their own. This is detailed in crypto/engine/README
6970 as it couldn't be adequately described here. However, there are a few
6971 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6972 were changed in the original introduction of ENGINE code have now
6973 reverted back - the hooking from this code to ENGINE is now a good
6974 deal more passive and at run-time, operations deal directly with
6975 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6976 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6977 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6978 they were not being used by the framework as there is no concept of a
6979 BIGNUM_METHOD and they could not be generalised to the new
6980 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6981 ENGINE_cpy() has been removed as it cannot be consistently defined in
6985 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6988 *) Change mkdef.pl to sort symbols that get the same entry number,
6989 and make sure the automatically generated functions ERR_load_*
6990 become part of libeay.num as well.
6993 *) New function SSL_renegotiate_pending(). This returns true once
6994 renegotiation has been requested (either SSL_renegotiate() call
6995 or HelloRequest/ClientHello received from the peer) and becomes
6996 false once a handshake has been completed.
6997 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6998 sends a HelloRequest, but does not ensure that a handshake takes
6999 place. SSL_renegotiate_pending() is useful for checking if the
7000 client has followed the request.)
7003 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7004 By default, clients may request session resumption even during
7005 renegotiation (if session ID contexts permit); with this option,
7006 session resumption is possible only in the first handshake.
7008 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7009 more bits available for options that should not be part of
7010 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7013 *) Add some demos for certificate and certificate request creation.
7016 *) Make maximum certificate chain size accepted from the peer application
7017 settable (SSL*_get/set_max_cert_list()), as proposed by
7018 "Douglas E. Engert" <deengert@anl.gov>.
7021 *) Add support for shared libraries for Unixware-7
7022 (Boyd Lynn Gerber <gerberb@zenez.com>).
7025 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7026 be done prior to destruction. Use this to unload error strings from
7027 ENGINEs that load their own error strings. NB: This adds two new API
7028 functions to "get" and "set" this destroy handler in an ENGINE.
7031 *) Alter all existing ENGINE implementations (except "openssl" and
7032 "openbsd") to dynamically instantiate their own error strings. This
7033 makes them more flexible to be built both as statically-linked ENGINEs
7034 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7035 Also, add stub code to each that makes building them as self-contained
7036 shared-libraries easier (see README.ENGINE).
7039 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7040 implementations into applications that are completely implemented in
7041 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7042 commands that can be used to configure what shared-library to load and
7043 to control aspects of the way it is handled. Also, made an update to
7044 the README.ENGINE file that brings its information up-to-date and
7045 provides some information and instructions on the "dynamic" ENGINE
7046 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7049 *) Make it possible to unload ranges of ERR strings with a new
7050 "ERR_unload_strings" function.
7053 *) Add a copy() function to EVP_MD.
7056 *) Make EVP_MD routines take a context pointer instead of just the
7057 md_data void pointer.
7060 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7061 that the digest can only process a single chunk of data
7062 (typically because it is provided by a piece of
7063 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7064 is only going to provide a single chunk of data, and hence the
7065 framework needn't accumulate the data for oneshot drivers.
7068 *) As with "ERR", make it possible to replace the underlying "ex_data"
7069 functions. This change also alters the storage and management of global
7070 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7071 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7072 index counters. The API functions that use this state have been changed
7073 to take a "class_index" rather than pointers to the class's local STACK
7074 and counter, and there is now an API function to dynamically create new
7075 classes. This centralisation allows us to (a) plug a lot of the
7076 thread-safety problems that existed, and (b) makes it possible to clean
7077 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7078 such data would previously have always leaked in application code and
7079 workarounds were in place to make the memory debugging turn a blind eye
7080 to it. Application code that doesn't use this new function will still
7081 leak as before, but their memory debugging output will announce it now
7082 rather than letting it slide.
7084 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7085 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7086 has a return value to indicate success or failure.
7089 *) Make it possible to replace the underlying "ERR" functions such that the
7090 global state (2 LHASH tables and 2 locks) is only used by the "default"
7091 implementation. This change also adds two functions to "get" and "set"
7092 the implementation prior to it being automatically set the first time
7093 any other ERR function takes place. Ie. an application can call "get",
7094 pass the return value to a module it has just loaded, and that module
7095 can call its own "set" function using that value. This means the
7096 module's "ERR" operations will use (and modify) the error state in the
7097 application and not in its own statically linked copy of OpenSSL code.
7100 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7101 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7102 the operation, and provides a more encapsulated way for external code
7103 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7104 to use these functions rather than manually incrementing the counts.
7106 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7109 *) Add EVP test program.
7112 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7115 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7116 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7117 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7118 These allow a CRL to be built without having to access X509_CRL fields
7119 directly. Modify 'ca' application to use new functions.
7122 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7123 bug workarounds. Rollback attack detection is a security feature.
7124 The problem will only arise on OpenSSL servers when TLSv1 is not
7125 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7126 Software authors not wanting to support TLSv1 will have special reasons
7127 for their choice and can explicitly enable this option.
7128 [Bodo Moeller, Lutz Jaenicke]
7130 *) Rationalise EVP so it can be extended: don't include a union of
7131 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7132 (similar to those existing for EVP_CIPHER_CTX).
7137 EVP_MD_CTX_init(&md); /* new function call */
7138 EVP_DigestInit(&md, EVP_sha1());
7139 EVP_DigestUpdate(&md, in, len);
7140 EVP_DigestFinal(&md, out, NULL);
7141 EVP_MD_CTX_cleanup(&md); /* new function call */
7145 *) Make DES key schedule conform to the usual scheme, as well as
7146 correcting its structure. This means that calls to DES functions
7147 now have to pass a pointer to a des_key_schedule instead of a
7148 plain des_key_schedule (which was actually always a pointer
7151 des_key_schedule ks;
7153 des_set_key_checked(..., &ks);
7154 des_ncbc_encrypt(..., &ks, ...);
7156 (Note that a later change renames 'des_...' into 'DES_...'.)
7159 *) Initial reduction of linker bloat: the use of some functions, such as
7160 PEM causes large amounts of unused functions to be linked in due to
7161 poor organisation. For example pem_all.c contains every PEM function
7162 which has a knock on effect of linking in large amounts of (unused)
7163 ASN1 code. Grouping together similar functions and splitting unrelated
7164 functions prevents this.
7167 *) Cleanup of EVP macros.
7170 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7171 correct _ecb suffix.
7174 *) Add initial OCSP responder support to ocsp application. The
7175 revocation information is handled using the text based index
7176 use by the ca application. The responder can either handle
7177 requests generated internally, supplied in files (for example
7178 via a CGI script) or using an internal minimal server.
7181 *) Add configuration choices to get zlib compression for TLS.
7184 *) Changes to Kerberos SSL for RFC 2712 compliance:
7185 1. Implemented real KerberosWrapper, instead of just using
7186 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7187 2. Implemented optional authenticator field of KerberosWrapper.
7189 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7190 and authenticator structs; see crypto/krb5/.
7192 Generalized Kerberos calls to support multiple Kerberos libraries.
7193 [Vern Staats <staatsvr@asc.hpc.mil>,
7194 Jeffrey Altman <jaltman@columbia.edu>
7195 via Richard Levitte]
7197 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7198 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7199 values for each of the key sizes rather than having just
7200 parameters (and 'speed' generating keys each time).
7203 *) Speed up EVP routines.
7206 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7207 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7208 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7209 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7211 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7212 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7213 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7216 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7218 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7221 *) Added the OS2-EMX target.
7222 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7224 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7225 to support NCONF routines in extension code. New function CONF_set_nconf()
7226 to allow functions which take an NCONF to also handle the old LHASH
7227 structure: this means that the old CONF compatible routines can be
7228 retained (in particular wrt extensions) without having to duplicate the
7229 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7232 *) Enhance the general user interface with mechanisms for inner control
7233 and with possibilities to have yes/no kind of prompts.
7236 *) Change all calls to low level digest routines in the library and
7237 applications to use EVP. Add missing calls to HMAC_cleanup() and
7238 don't assume HMAC_CTX can be copied using memcpy().
7239 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7241 *) Add the possibility to control engines through control names but with
7242 arbitrary arguments instead of just a string.
7243 Change the key loaders to take a UI_METHOD instead of a callback
7244 function pointer. NOTE: this breaks binary compatibility with earlier
7245 versions of OpenSSL [engine].
7246 Adapt the nCipher code for these new conditions and add a card insertion
7250 *) Enhance the general user interface with mechanisms to better support
7251 dialog box interfaces, application-defined prompts, the possibility
7252 to use defaults (for example default passwords from somewhere else)
7253 and interrupts/cancellations.
7256 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7257 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7260 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7261 tidy up some unnecessarily weird code in 'sk_new()').
7262 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7264 *) Change the key loading routines for ENGINEs to use the same kind
7265 callback (pem_password_cb) as all other routines that need this
7269 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7270 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7271 than this minimum value is recommended.
7274 *) New random seeder for OpenVMS, using the system process statistics
7275 that are easily reachable.
7278 *) Windows apparently can't transparently handle global
7279 variables defined in DLLs. Initialisations such as:
7281 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7283 won't compile. This is used by the any applications that need to
7284 declare their own ASN1 modules. This was fixed by adding the option
7285 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7286 needed for static libraries under Win32.
7289 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7290 setting of purpose and trust fields. New X509_STORE trust and
7291 purpose functions and tidy up setting in other SSL functions.
7294 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7295 structure. These are inherited by X509_STORE_CTX when it is
7296 initialised. This allows various defaults to be set in the
7297 X509_STORE structure (such as flags for CRL checking and custom
7298 purpose or trust settings) for functions which only use X509_STORE_CTX
7299 internally such as S/MIME.
7301 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7302 trust settings if they are not set in X509_STORE. This allows X509_STORE
7303 purposes and trust (in S/MIME for example) to override any set by default.
7305 Add command line options for CRL checking to smime, s_client and s_server
7309 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7310 are set then the CRL is looked up in the X509_STORE structure and
7311 its validity and signature checked, then if the certificate is found
7312 in the CRL the verify fails with a revoked error.
7314 Various new CRL related callbacks added to X509_STORE_CTX structure.
7316 Command line options added to 'verify' application to support this.
7318 This needs some additional work, such as being able to handle multiple
7319 CRLs with different times, extension based lookup (rather than just
7320 by subject name) and ultimately more complete V2 CRL extension
7324 *) Add a general user interface API (crypto/ui/). This is designed
7325 to replace things like des_read_password and friends (backward
7326 compatibility functions using this new API are provided).
7327 The purpose is to remove prompting functions from the DES code
7328 section as well as provide for prompting through dialog boxes in
7329 a window system and the like.
7332 *) Add "ex_data" support to ENGINE so implementations can add state at a
7333 per-structure level rather than having to store it globally.
7336 *) Make it possible for ENGINE structures to be copied when retrieved by
7337 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7338 This causes the "original" ENGINE structure to act like a template,
7339 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7340 operational state can be localised to each ENGINE structure, despite the
7341 fact they all share the same "methods". New ENGINE structures returned in
7342 this case have no functional references and the return value is the single
7343 structural reference. This matches the single structural reference returned
7344 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7348 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7349 needs to match any other type at all we need to manually clear the
7353 *) Changes to the "openssl engine" utility to include;
7354 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7355 about an ENGINE's available control commands.
7356 - executing control commands from command line arguments using the
7357 '-pre' and '-post' switches. '-post' is only used if '-t' is
7358 specified and the ENGINE is successfully initialised. The syntax for
7359 the individual commands are colon-separated, for example;
7360 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7363 *) New dynamic control command support for ENGINEs. ENGINEs can now
7364 declare their own commands (numbers), names (strings), descriptions,
7365 and input types for run-time discovery by calling applications. A
7366 subset of these commands are implicitly classed as "executable"
7367 depending on their input type, and only these can be invoked through
7368 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7369 can be based on user input, config files, etc). The distinction is
7370 that "executable" commands cannot return anything other than a boolean
7371 result and can only support numeric or string input, whereas some
7372 discoverable commands may only be for direct use through
7373 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7374 pointers, or other custom uses. The "executable" commands are to
7375 support parameterisations of ENGINE behaviour that can be
7376 unambiguously defined by ENGINEs and used consistently across any
7377 OpenSSL-based application. Commands have been added to all the
7378 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7379 control over shared-library paths without source code alterations.
7382 *) Changed all ENGINE implementations to dynamically allocate their
7383 ENGINEs rather than declaring them statically. Apart from this being
7384 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7385 this also allows the implementations to compile without using the
7386 internal engine_int.h header.
7389 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7390 'const' value. Any code that should be able to modify a RAND_METHOD
7391 should already have non-const pointers to it (ie. they should only
7392 modify their own ones).
7395 *) Made a variety of little tweaks to the ENGINE code.
7396 - "atalla" and "ubsec" string definitions were moved from header files
7397 to C code. "nuron" string definitions were placed in variables
7398 rather than hard-coded - allowing parameterisation of these values
7399 later on via ctrl() commands.
7400 - Removed unused "#if 0"'d code.
7401 - Fixed engine list iteration code so it uses ENGINE_free() to release
7402 structural references.
7403 - Constified the RAND_METHOD element of ENGINE structures.
7404 - Constified various get/set functions as appropriate and added
7405 missing functions (including a catch-all ENGINE_cpy that duplicates
7406 all ENGINE values onto a new ENGINE except reference counts/state).
7407 - Removed NULL parameter checks in get/set functions. Setting a method
7408 or function to NULL is a way of cancelling out a previously set
7409 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7410 and doesn't justify the extra error symbols and code.
7411 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7412 flags from engine_int.h to engine.h.
7413 - Changed prototypes for ENGINE handler functions (init(), finish(),
7414 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7417 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7418 to the algorithm using long division. The binary algorithm can be
7419 used only if the modulus is odd. On 32-bit systems, it is faster
7420 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7421 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7422 up to 450 bits. In 64-bit environments, the binary algorithm
7423 appears to be advantageous for much longer moduli; here we use it
7424 for moduli up to 2048 bits.
7427 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7428 could not support the combine flag in choice fields.
7431 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7432 extensions from a certificate request to the certificate.
7435 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7436 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7437 file: this allows the display of the certificate about to be
7438 signed to be customised, to allow certain fields to be included
7439 or excluded and extension details. The old system didn't display
7440 multicharacter strings properly, omitted fields not in the policy
7441 and couldn't display additional details such as extensions.
7444 *) Function EC_POINTs_mul for multiple scalar multiplication
7445 of an arbitrary number of elliptic curve points
7446 \sum scalars[i]*points[i],
7447 optionally including the generator defined for the EC_GROUP:
7448 scalar*generator + \sum scalars[i]*points[i].
7450 EC_POINT_mul is a simple wrapper function for the typical case
7451 that the point list has just one item (besides the optional
7455 *) First EC_METHODs for curves over GF(p):
7457 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7458 operations and provides various method functions that can also
7459 operate with faster implementations of modular arithmetic.
7461 EC_GFp_mont_method() reuses most functions that are part of
7462 EC_GFp_simple_method, but uses Montgomery arithmetic.
7464 [Bodo Moeller; point addition and point doubling
7465 implementation directly derived from source code provided by
7466 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7468 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7469 crypto/ec/ec_lib.c):
7471 Curves are EC_GROUP objects (with an optional group generator)
7472 based on EC_METHODs that are built into the library.
7474 Points are EC_POINT objects based on EC_GROUP objects.
7476 Most of the framework would be able to handle curves over arbitrary
7477 finite fields, but as there are no obvious types for fields other
7478 than GF(p), some functions are limited to that for now.
7481 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7482 that the file contains a complete HTTP response.
7485 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7486 change the def and num file printf format specifier from "%-40sXXX"
7487 to "%-39s XXX". The latter will always guarantee a space after the
7488 field while the former will cause them to run together if the field
7489 is 40 of more characters long.
7492 *) Constify the cipher and digest 'method' functions and structures
7493 and modify related functions to take constant EVP_MD and EVP_CIPHER
7497 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7498 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7501 *) Modify EVP_Digest*() routines so they now return values. Although the
7502 internal software routines can never fail additional hardware versions
7506 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7508 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7509 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7512 ERR_R_NESTED_ASN1_ERROR
7514 ERR_R_MISSING_ASN1_EOS
7515 were 4 .. 9, conflicting with
7516 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7518 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7519 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7521 Add new error code 'ERR_R_INTERNAL_ERROR'.
7524 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7528 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7529 sets the subject name for a new request or supersedes the
7530 subject name in a given request. Formats that can be parsed are
7531 'CN=Some Name, OU=myOU, C=IT'
7533 'CN=Some Name/OU=myOU/C=IT'.
7535 Add options '-batch' and '-verbose' to 'openssl req'.
7536 [Massimiliano Pala <madwolf@hackmasters.net>]
7538 *) Introduce the possibility to access global variables through
7539 functions on platform were that's the best way to handle exporting
7540 global variables in shared libraries. To enable this functionality,
7541 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7542 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7543 is normally done by Configure or something similar).
7545 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7546 in the source file (foo.c) like this:
7548 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7549 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7551 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7552 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7554 OPENSSL_DECLARE_GLOBAL(int,foo);
7555 #define foo OPENSSL_GLOBAL_REF(foo)
7556 OPENSSL_DECLARE_GLOBAL(double,bar);
7557 #define bar OPENSSL_GLOBAL_REF(bar)
7559 The #defines are very important, and therefore so is including the
7560 header file everywhere where the defined globals are used.
7562 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7563 of ASN.1 items, but that structure is a bit different.
7565 The largest change is in util/mkdef.pl which has been enhanced with
7566 better and easier to understand logic to choose which symbols should
7567 go into the Windows .def files as well as a number of fixes and code
7568 cleanup (among others, algorithm keywords are now sorted
7569 lexicographically to avoid constant rewrites).
7572 *) In BN_div() keep a copy of the sign of 'num' before writing the
7573 result to 'rm' because if rm==num the value will be overwritten
7574 and produce the wrong result if 'num' is negative: this caused
7575 problems with BN_mod() and BN_nnmod().
7578 *) Function OCSP_request_verify(). This checks the signature on an
7579 OCSP request and verifies the signer certificate. The signer
7580 certificate is just checked for a generic purpose and OCSP request
7584 *) Add OCSP_check_validity() function to check the validity of OCSP
7585 responses. OCSP responses are prepared in real time and may only
7586 be a few seconds old. Simply checking that the current time lies
7587 between thisUpdate and nextUpdate max reject otherwise valid responses
7588 caused by either OCSP responder or client clock inaccuracy. Instead
7589 we allow thisUpdate and nextUpdate to fall within a certain period of
7590 the current time. The age of the response can also optionally be
7591 checked. Two new options -validity_period and -status_age added to
7595 *) If signature or public key algorithm is unrecognized print out its
7596 OID rather that just UNKNOWN.
7599 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7600 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7601 ID to be generated from the issuer certificate alone which can then be
7602 passed to OCSP_id_issuer_cmp().
7605 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7606 ASN1 modules to export functions returning ASN1_ITEM pointers
7607 instead of the ASN1_ITEM structures themselves. This adds several
7608 new macros which allow the underlying ASN1 function/structure to
7609 be accessed transparently. As a result code should not use ASN1_ITEM
7610 references directly (such as &X509_it) but instead use the relevant
7611 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7612 use of the new ASN1 code on platforms where exporting structures
7613 is problematical (for example in shared libraries) but exporting
7614 functions returning pointers to structures is not.
7617 *) Add support for overriding the generation of SSL/TLS session IDs.
7618 These callbacks can be registered either in an SSL_CTX or per SSL.
7619 The purpose of this is to allow applications to control, if they wish,
7620 the arbitrary values chosen for use as session IDs, particularly as it
7621 can be useful for session caching in multiple-server environments. A
7622 command-line switch for testing this (and any client code that wishes
7623 to use such a feature) has been added to "s_server".
7624 [Geoff Thorpe, Lutz Jaenicke]
7626 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7627 of the form '#if defined(...) || defined(...) || ...' and
7628 '#if !defined(...) && !defined(...) && ...'. This also avoids
7629 the growing number of special cases it was previously handling.
7632 *) Make all configuration macros available for application by making
7633 sure they are available in opensslconf.h, by giving them names starting
7634 with "OPENSSL_" to avoid conflicts with other packages and by making
7635 sure e_os2.h will cover all platform-specific cases together with
7637 Additionally, it is now possible to define configuration/platform-
7638 specific names (called "system identities"). In the C code, these
7639 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7640 macro with the name beginning with "OPENSSL_SYS_", which is determined
7641 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7645 *) New option -set_serial to 'req' and 'x509' this allows the serial
7646 number to use to be specified on the command line. Previously self
7647 signed certificates were hard coded with serial number 0 and the
7648 CA options of 'x509' had to use a serial number in a file which was
7652 *) New options to 'ca' utility to support V2 CRL entry extensions.
7653 Currently CRL reason, invalidity date and hold instruction are
7654 supported. Add new CRL extensions to V3 code and some new objects.
7657 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7658 disable standard block padding (aka PKCS#5 padding) in the EVP
7659 API, which was previously mandatory. This means that the data is
7660 not padded in any way and so the total length much be a multiple
7661 of the block size, otherwise an error occurs.
7664 *) Initial (incomplete) OCSP SSL support.
7667 *) New function OCSP_parse_url(). This splits up a URL into its host,
7668 port and path components: primarily to parse OCSP URLs. New -url
7669 option to ocsp utility.
7672 *) New nonce behavior. The return value of OCSP_check_nonce() now
7673 reflects the various checks performed. Applications can decide
7674 whether to tolerate certain situations such as an absent nonce
7675 in a response when one was present in a request: the ocsp application
7676 just prints out a warning. New function OCSP_add1_basic_nonce()
7677 this is to allow responders to include a nonce in a response even if
7678 the request is nonce-less.
7681 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7682 skipped when using openssl x509 multiple times on a single input file,
7683 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7686 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7687 set string type: to handle setting ASN1_TIME structures. Fix ca
7688 utility to correctly initialize revocation date of CRLs.
7691 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7692 the clients preferred ciphersuites and rather use its own preferences.
7693 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7694 Internet Explorer by ensuring unchanged hash method during stepup.
7695 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7698 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7699 to aes and add a new 'exist' option to print out symbols that don't
7703 *) Additional options to ocsp utility to allow flags to be set and
7704 additional certificates supplied.
7707 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7708 OCSP client a number of certificate to only verify the response
7712 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7713 handle the new API. Currently only ECB, CBC modes supported. Add new
7716 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7717 Encryption Standard (AES) Ciphersuites for Transport Layer
7718 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7719 not enabled by default and were not part of the "ALL" ciphersuite
7720 alias because they were not yet official; they could be
7721 explicitly requested by specifying the "AESdraft" ciphersuite
7722 group alias. In the final release of OpenSSL 0.9.7, the group
7723 alias is called "AES" and is part of "ALL".)
7724 [Ben Laurie, Steve Henson, Bodo Moeller]
7726 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7727 request to response.
7730 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7731 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7732 extract information from a certificate request. OCSP_response_create()
7733 creates a response and optionally adds a basic response structure.
7734 OCSP_basic_add1_status() adds a complete single response to a basic
7735 response and returns the OCSP_SINGLERESP structure just added (to allow
7736 extensions to be included for example). OCSP_basic_add1_cert() adds a
7737 certificate to a basic response and OCSP_basic_sign() signs a basic
7738 response with various flags. New helper functions ASN1_TIME_check()
7739 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7740 (converts ASN1_TIME to GeneralizedTime).
7743 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7744 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7745 structure from a certificate. X509_pubkey_digest() digests the public_key
7746 contents: this is used in various key identifiers.
7749 *) Make sk_sort() tolerate a NULL argument.
7750 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7752 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7753 passed by the function are trusted implicitly. If any of them signed the
7754 response then it is assumed to be valid and is not verified.
7757 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7758 to data. This was previously part of the PKCS7 ASN1 code. This
7759 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7760 [Steve Henson, reported by Kenneth R. Robinette
7761 <support@securenetterm.com>]
7763 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7764 routines: without these tracing memory leaks is very painful.
7765 Fix leaks in PKCS12 and PKCS7 routines.
7768 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7769 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7770 effectively meant GeneralizedTime would never be used. Now it
7771 is initialised to -1 but X509_time_adj() now has to check the value
7772 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7773 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7774 [Steve Henson, reported by Kenneth R. Robinette
7775 <support@securenetterm.com>]
7777 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7778 result in a zero length in the ASN1_INTEGER structure which was
7779 not consistent with the structure when d2i_ASN1_INTEGER() was used
7780 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7781 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7782 where it did not print out a minus for negative ASN1_INTEGER.
7785 *) Add summary printout to ocsp utility. The various functions which
7786 convert status values to strings have been renamed to:
7787 OCSP_response_status_str(), OCSP_cert_status_str() and
7788 OCSP_crl_reason_str() and are no longer static. New options
7789 to verify nonce values and to disable verification. OCSP response
7790 printout format cleaned up.
7793 *) Add additional OCSP certificate checks. These are those specified
7794 in RFC2560. This consists of two separate checks: the CA of the
7795 certificate being checked must either be the OCSP signer certificate
7796 or the issuer of the OCSP signer certificate. In the latter case the
7797 OCSP signer certificate must contain the OCSP signing extended key
7798 usage. This check is performed by attempting to match the OCSP
7799 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7800 in the OCSP_CERTID structures of the response.
7803 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7804 and related routines. This uses the standard OpenSSL certificate
7805 verify routines to perform initial checks (just CA validity) and
7806 to obtain the certificate chain. Then additional checks will be
7807 performed on the chain. Currently the root CA is checked to see
7808 if it is explicitly trusted for OCSP signing. This is used to set
7809 a root CA as a global signing root: that is any certificate that
7810 chains to that CA is an acceptable OCSP signing certificate.
7813 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7814 extensions from a separate configuration file.
7815 As when reading extensions from the main configuration file,
7816 the '-extensions ...' option may be used for specifying the
7818 [Massimiliano Pala <madwolf@comune.modena.it>]
7820 *) New OCSP utility. Allows OCSP requests to be generated or
7821 read. The request can be sent to a responder and the output
7822 parsed, outputed or printed in text form. Not complete yet:
7823 still needs to check the OCSP response validity.
7826 *) New subcommands for 'openssl ca':
7827 'openssl ca -status <serial>' prints the status of the cert with
7828 the given serial number (according to the index file).
7829 'openssl ca -updatedb' updates the expiry status of certificates
7831 [Massimiliano Pala <madwolf@comune.modena.it>]
7833 *) New '-newreq-nodes' command option to CA.pl. This is like
7834 '-newreq', but calls 'openssl req' with the '-nodes' option
7835 so that the resulting key is not encrypted.
7836 [Damien Miller <djm@mindrot.org>]
7838 *) New configuration for the GNU Hurd.
7839 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7841 *) Initial code to implement OCSP basic response verify. This
7842 is currently incomplete. Currently just finds the signer's
7843 certificate and verifies the signature on the response.
7846 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7847 value of OPENSSLDIR. This is available via the new '-d' option
7848 to 'openssl version', and is also included in 'openssl version -a'.
7851 *) Allowing defining memory allocation callbacks that will be given
7852 file name and line number information in additional arguments
7853 (a const char* and an int). The basic functionality remains, as
7854 well as the original possibility to just replace malloc(),
7855 realloc() and free() by functions that do not know about these
7856 additional arguments. To register and find out the current
7857 settings for extended allocation functions, the following
7858 functions are provided:
7860 CRYPTO_set_mem_ex_functions
7861 CRYPTO_set_locked_mem_ex_functions
7862 CRYPTO_get_mem_ex_functions
7863 CRYPTO_get_locked_mem_ex_functions
7865 These work the same way as CRYPTO_set_mem_functions and friends.
7866 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7867 extended allocation function is enabled.
7868 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7869 a conventional allocation function is enabled.
7870 [Richard Levitte, Bodo Moeller]
7872 *) Finish off removing the remaining LHASH function pointer casts.
7873 There should no longer be any prototype-casting required when using
7874 the LHASH abstraction, and any casts that remain are "bugs". See
7875 the callback types and macros at the head of lhash.h for details
7876 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7879 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7880 If /dev/[u]random devices are not available or do not return enough
7881 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7883 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7884 /etc/entropy will be queried once each in this sequence, querying stops
7885 when enough entropy was collected without querying more sockets.
7888 *) Change the Unix RAND_poll() variant to be able to poll several
7889 random devices, as specified by DEVRANDOM, until a sufficient amount
7890 of data has been collected. We spend at most 10 ms on each file
7891 (select timeout) and read in non-blocking mode. DEVRANDOM now
7892 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7893 (previously it was just the string "/dev/urandom"), so on typical
7894 platforms the 10 ms delay will never occur.
7895 Also separate out the Unix variant to its own file, rand_unix.c.
7896 For VMS, there's a currently-empty rand_vms.c.
7899 *) Move OCSP client related routines to ocsp_cl.c. These
7900 provide utility functions which an application needing
7901 to issue a request to an OCSP responder and analyse the
7902 response will typically need: as opposed to those which an
7903 OCSP responder itself would need which will be added later.
7905 OCSP_request_sign() signs an OCSP request with an API similar
7906 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7907 response. OCSP_response_get1_basic() extracts basic response
7908 from response. OCSP_resp_find_status(): finds and extracts status
7909 information from an OCSP_CERTID structure (which will be created
7910 when the request structure is built). These are built from lower
7911 level functions which work on OCSP_SINGLERESP structures but
7912 won't normally be used unless the application wishes to examine
7913 extensions in the OCSP response for example.
7915 Replace nonce routines with a pair of functions.
7916 OCSP_request_add1_nonce() adds a nonce value and optionally
7917 generates a random value. OCSP_check_nonce() checks the
7918 validity of the nonce in an OCSP response.
7921 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7922 This doesn't copy the supplied OCSP_CERTID and avoids the
7923 need to free up the newly created id. Change return type
7924 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7925 This can then be used to add extensions to the request.
7926 Deleted OCSP_request_new(), since most of its functionality
7927 is now in OCSP_REQUEST_new() (and the case insensitive name
7928 clash) apart from the ability to set the request name which
7929 will be added elsewhere.
7932 *) Update OCSP API. Remove obsolete extensions argument from
7933 various functions. Extensions are now handled using the new
7934 OCSP extension code. New simple OCSP HTTP function which
7935 can be used to send requests and parse the response.
7938 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7939 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7940 uses the special reorder version of SET OF to sort the attributes
7941 and reorder them to match the encoded order. This resolves a long
7942 standing problem: a verify on a PKCS7 structure just after signing
7943 it used to fail because the attribute order did not match the
7944 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7945 it uses the received order. This is necessary to tolerate some broken
7946 software that does not order SET OF. This is handled by encoding
7947 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7948 to produce the required SET OF.
7951 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7952 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7953 files to get correct declarations of the ASN.1 item variables.
7956 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7957 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7958 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7959 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7960 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7961 ASN1_ITEM and no wrapper functions.
7964 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7965 replace the old function pointer based I/O routines. Change most of
7966 the *_d2i_bio() and *_d2i_fp() functions to use these.
7969 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7970 lines, recognize more "algorithms" that can be deselected, and make
7971 it complain about algorithm deselection that isn't recognised.
7974 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7975 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7976 to use new functions. Add NO_ASN1_OLD which can be set to remove
7977 some old style ASN1 functions: this can be used to determine if old
7978 code will still work when these eventually go away.
7981 *) New extension functions for OCSP structures, these follow the
7982 same conventions as certificates and CRLs.
7985 *) New function X509V3_add1_i2d(). This automatically encodes and
7986 adds an extension. Its behaviour can be customised with various
7987 flags to append, replace or delete. Various wrappers added for
7988 certificates and CRLs.
7991 *) Fix to avoid calling the underlying ASN1 print routine when
7992 an extension cannot be parsed. Correct a typo in the
7993 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7996 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7997 entries for variables.
8000 *) Add functionality to apps/openssl.c for detecting locking
8001 problems: As the program is single-threaded, all we have
8002 to do is register a locking callback using an array for
8003 storing which locks are currently held by the program.
8006 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8007 SSL_get_ex_data_X509_STORE_idx(), which is used in
8008 ssl_verify_cert_chain() and thus can be called at any time
8009 during TLS/SSL handshakes so that thread-safety is essential.
8010 Unfortunately, the ex_data design is not at all suited
8011 for multi-threaded use, so it probably should be abolished.
8014 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8015 [Broadcom, tweaked and integrated by Geoff Thorpe]
8017 *) Move common extension printing code to new function
8018 X509V3_print_extensions(). Reorganise OCSP print routines and
8019 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8022 *) New function X509_signature_print() to remove duplication in some
8026 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8027 set (this was treated exactly the same as SET OF previously). This
8028 is used to reorder the STACK representing the structure to match the
8029 encoding. This will be used to get round a problem where a PKCS7
8030 structure which was signed could not be verified because the STACK
8031 order did not reflect the encoded order.
8034 *) Reimplement the OCSP ASN1 module using the new code.
8037 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8038 for its ASN1 operations. The old style function pointers still exist
8039 for now but they will eventually go away.
8042 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8043 completely replaces the old ASN1 functionality with a table driven
8044 encoder and decoder which interprets an ASN1_ITEM structure describing
8045 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8046 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8047 has also been converted to the new form.
8050 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8051 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8052 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8053 for negative moduli.
8056 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8057 of not touching the result's sign bit.
8060 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8064 *) Changed the LHASH code to use prototypes for callbacks, and created
8065 macros to declare and implement thin (optionally static) functions
8066 that provide type-safety and avoid function pointer casting for the
8067 type-specific callbacks.
8070 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8072 [Veers Staats <staatsvr@asc.hpc.mil>,
8073 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8075 *) Reformat the FAQ so the different questions and answers can be divided
8076 in sections depending on the subject.
8079 *) Have the zlib compression code load ZLIB.DLL dynamically under
8083 *) New function BN_mod_sqrt for computing square roots modulo a prime
8084 (using the probabilistic Tonelli-Shanks algorithm unless
8085 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8086 be handled deterministically).
8087 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8089 *) Make BN_mod_inverse faster by explicitly handling small quotients
8090 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8091 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8094 *) New function BN_kronecker.
8097 *) Fix BN_gcd so that it works on negative inputs; the result is
8098 positive unless both parameters are zero.
8099 Previously something reasonably close to an infinite loop was
8100 possible because numbers could be growing instead of shrinking
8101 in the implementation of Euclid's algorithm.
8104 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8105 sign of the number in question.
8107 Fix BN_is_word(a,w) to work correctly for w == 0.
8109 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8110 because its test if the absolute value of 'a' equals 'w'.
8111 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8112 it exists mostly for use in the implementations of BN_is_zero(),
8113 BN_is_one(), and BN_is_word().
8116 *) New function BN_swap.
8119 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8120 the exponentiation functions are more likely to produce reasonable
8121 results on negative inputs.
8124 *) Change BN_mod_mul so that the result is always non-negative.
8125 Previously, it could be negative if one of the factors was negative;
8126 I don't think anyone really wanted that behaviour.
8129 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8130 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8131 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8132 and add new functions:
8141 BN_mod_lshift1_quick
8145 These functions always generate non-negative results.
8147 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8148 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8150 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8151 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8152 be reduced modulo m.
8153 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8156 The following entry accidentally appeared in the CHANGES file
8157 distributed with OpenSSL 0.9.7. The modifications described in
8158 it do *not* apply to OpenSSL 0.9.7.
8160 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8161 was actually never needed) and in BN_mul(). The removal in BN_mul()
8162 required a small change in bn_mul_part_recursive() and the addition
8163 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8164 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8165 bn_sub_words() and bn_add_words() except they take arrays with
8170 *) In 'openssl passwd', verify passwords read from the terminal
8171 unless the '-salt' option is used (which usually means that
8172 verification would just waste user's time since the resulting
8173 hash is going to be compared with some given password hash)
8174 or the new '-noverify' option is used.
8176 This is an incompatible change, but it does not affect
8177 non-interactive use of 'openssl passwd' (passwords on the command
8178 line, '-stdin' option, '-in ...' option) and thus should not
8182 *) Remove all references to RSAref, since there's no more need for it.
8185 *) Make DSO load along a path given through an environment variable
8186 (SHLIB_PATH) with shl_load().
8189 *) Constify the ENGINE code as a result of BIGNUM constification.
8190 Also constify the RSA code and most things related to it. In a
8191 few places, most notable in the depth of the ASN.1 code, ugly
8192 casts back to non-const were required (to be solved at a later
8196 *) Make it so the openssl application has all engines loaded by default.
8199 *) Constify the BIGNUM routines a little more.
8202 *) Add the following functions:
8204 ENGINE_load_cswift()
8206 ENGINE_load_atalla()
8208 ENGINE_load_builtin_engines()
8210 That way, an application can itself choose if external engines that
8211 are built-in in OpenSSL shall ever be used or not. The benefit is
8212 that applications won't have to be linked with libdl or other dso
8213 libraries unless it's really needed.
8215 Changed 'openssl engine' to load all engines on demand.
8216 Changed the engine header files to avoid the duplication of some
8217 declarations (they differed!).
8220 *) 'openssl engine' can now list capabilities.
8223 *) Better error reporting in 'openssl engine'.
8226 *) Never call load_dh_param(NULL) in s_server.
8229 *) Add engine application. It can currently list engines by name and
8230 identity, and test if they are actually available.
8233 *) Improve RPM specification file by forcing symbolic linking and making
8234 sure the installed documentation is also owned by root.root.
8235 [Damien Miller <djm@mindrot.org>]
8237 *) Give the OpenSSL applications more possibilities to make use of
8238 keys (public as well as private) handled by engines.
8241 *) Add OCSP code that comes from CertCo.
8244 *) Add VMS support for the Rijndael code.
8247 *) Added untested support for Nuron crypto accelerator.
8250 *) Add support for external cryptographic devices. This code was
8251 previously distributed separately as the "engine" branch.
8252 [Geoff Thorpe, Richard Levitte]
8254 *) Rework the filename-translation in the DSO code. It is now possible to
8255 have far greater control over how a "name" is turned into a filename
8256 depending on the operating environment and any oddities about the
8257 different shared library filenames on each system.
8260 *) Support threads on FreeBSD-elf in Configure.
8263 *) Fix for SHA1 assembly problem with MASM: it produces
8264 warnings about corrupt line number information when assembling
8265 with debugging information. This is caused by the overlapping
8267 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8270 NCONF_get_number() has no error checking at all. As a replacement,
8271 NCONF_get_number_e() is defined (_e for "error checking") and is
8272 promoted strongly. The old NCONF_get_number is kept around for
8273 binary backward compatibility.
8274 Make it possible for methods to load from something other than a BIO,
8275 by providing a function pointer that is given a name instead of a BIO.
8276 For example, this could be used to load configuration data from an
8280 *) Fix for non blocking accept BIOs. Added new I/O special reason
8281 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8282 with non blocking I/O was not possible because no retry code was
8283 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8287 *) Added the beginnings of Rijndael support.
8290 *) Fix for bug in DirectoryString mask setting. Add support for
8291 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8292 to allow certificate printing to more controllable, additional
8293 'certopt' option to 'x509' to allow new printing options to be
8297 *) Clean old EAY MD5 hack from e_os.h.
8300 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8302 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8303 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8304 [Joe Orton, Steve Henson]
8306 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8308 *) Fix additional bug revealed by the NISCC test suite:
8310 Stop bug triggering large recursion when presented with
8311 certain ASN.1 tags (CVE-2003-0851)
8314 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8316 *) Fix various bugs revealed by running the NISCC test suite:
8318 Stop out of bounds reads in the ASN1 code when presented with
8319 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8321 If verify callback ignores invalid public key errors don't try to check
8322 certificate signature with the NULL public key.
8326 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8327 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8331 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8332 extra data after the compression methods not only for TLS 1.0
8333 but also for SSL 3.0 (as required by the specification).
8334 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8336 *) Change X509_certificate_type() to mark the key as exported/exportable
8337 when it's 512 *bits* long, not 512 bytes.
8340 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8342 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8343 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8344 a protocol version number mismatch like a decryption error
8345 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8348 *) Turn on RSA blinding by default in the default implementation
8349 to avoid a timing attack. Applications that don't want it can call
8350 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8351 They would be ill-advised to do so in most cases.
8352 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8354 *) Change RSA blinding code so that it works when the PRNG is not
8355 seeded (in this case, the secret RSA exponent is abused as
8356 an unpredictable seed -- if it is not unpredictable, there
8357 is no point in blinding anyway). Make RSA blinding thread-safe
8358 by remembering the creator's thread ID in rsa->blinding and
8359 having all other threads use local one-time blinding factors
8360 (this requires more computation than sharing rsa->blinding, but
8361 avoids excessive locking; and if an RSA object is not shared
8362 between threads, blinding will still be very fast).
8365 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8367 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8368 via timing by performing a MAC computation even if incorrect
8369 block cipher padding has been found. This is a countermeasure
8370 against active attacks where the attacker has to distinguish
8371 between bad padding and a MAC verification error. (CVE-2003-0078)
8373 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8374 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8375 Martin Vuagnoux (EPFL, Ilion)]
8377 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8379 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8380 memory from it's contents. This is done with a counter that will
8381 place alternating values in each byte. This can be used to solve
8382 two issues: 1) the removal of calls to memset() by highly optimizing
8383 compilers, and 2) cleansing with other values than 0, since those can
8384 be read through on certain media, for example a swap space on disk.
8387 *) Bugfix: client side session caching did not work with external caching,
8388 because the session->cipher setting was not restored when reloading
8389 from the external cache. This problem was masked, when
8390 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8391 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8394 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8395 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8396 [Zeev Lieber <zeev-l@yahoo.com>]
8398 *) Undo an undocumented change introduced in 0.9.6e which caused
8399 repeated calls to OpenSSL_add_all_ciphers() and
8400 OpenSSL_add_all_digests() to be ignored, even after calling
8404 *) Change the default configuration reader to deal with last line not
8405 being properly terminated.
8408 *) Change X509_NAME_cmp() so it applies the special rules on handling
8409 DN values that are of type PrintableString, as well as RDNs of type
8410 emailAddress where the value has the type ia5String.
8411 [stefank@valicert.com via Richard Levitte]
8413 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8414 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8415 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8416 the bitwise-OR of the two for use by the majority of applications
8417 wanting this behaviour, and update the docs. The documented
8418 behaviour and actual behaviour were inconsistent and had been
8419 changing anyway, so this is more a bug-fix than a behavioural
8421 [Geoff Thorpe, diagnosed by Nadav Har'El]
8423 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8424 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8427 *) Fix initialization code race conditions in
8428 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8429 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8430 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8431 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8432 ssl2_get_cipher_by_char(),
8433 ssl3_get_cipher_by_char().
8434 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8436 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8437 the cached sessions are flushed, as the remove_cb() might use ex_data
8438 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8439 (see [openssl.org #212]).
8440 [Geoff Thorpe, Lutz Jaenicke]
8442 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8443 length, instead of the encoding length to d2i_ASN1_OBJECT.
8446 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8448 *) [In 0.9.6g-engine release:]
8449 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8450 [Lynn Gazis <lgazis@rainbow.com>]
8452 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8454 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8455 and get fix the header length calculation.
8456 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8457 Alon Kantor <alonk@checkpoint.com> (and others),
8460 *) Use proper error handling instead of 'assertions' in buffer
8461 overflow checks added in 0.9.6e. This prevents DoS (the
8462 assertions could call abort()).
8463 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8465 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8467 *) Add various sanity checks to asn1_get_length() to reject
8468 the ASN1 length bytes if they exceed sizeof(long), will appear
8469 negative or the content length exceeds the length of the
8471 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8473 *) Fix cipher selection routines: ciphers without encryption had no flags
8474 for the cipher strength set and where therefore not handled correctly
8475 by the selection routines (PR #130).
8478 *) Fix EVP_dsa_sha macro.
8482 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8483 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8484 that was added in OpenSSL 0.9.6d.
8486 As the countermeasure turned out to be incompatible with some
8487 broken SSL implementations, the new option is part of SSL_OP_ALL.
8488 SSL_OP_ALL is usually employed when compatibility with weird SSL
8489 implementations is desired (e.g. '-bugs' option to 's_client' and
8490 's_server'), so the new option is automatically set in many
8494 *) Changes in security patch:
8496 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8497 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8498 Air Force Materiel Command, USAF, under agreement number
8501 *) Add various sanity checks to asn1_get_length() to reject
8502 the ASN1 length bytes if they exceed sizeof(long), will appear
8503 negative or the content length exceeds the length of the
8504 supplied buffer. (CVE-2002-0659)
8505 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8507 *) Assertions for various potential buffer overflows, not known to
8509 [Ben Laurie (CHATS)]
8511 *) Various temporary buffers to hold ASCII versions of integers were
8512 too small for 64 bit platforms. (CVE-2002-0655)
8513 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8515 *) Remote buffer overflow in SSL3 protocol - an attacker could
8516 supply an oversized session ID to a client. (CVE-2002-0656)
8517 [Ben Laurie (CHATS)]
8519 *) Remote buffer overflow in SSL2 protocol - an attacker could
8520 supply an oversized client master key. (CVE-2002-0656)
8521 [Ben Laurie (CHATS)]
8523 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8525 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8526 encoded as NULL) with id-dsa-with-sha1.
8527 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8529 *) Check various X509_...() return values in apps/req.c.
8530 [Nils Larsch <nla@trustcenter.de>]
8532 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8533 an end-of-file condition would erroneously be flagged, when the CRLF
8534 was just at the end of a processed block. The bug was discovered when
8535 processing data through a buffering memory BIO handing the data to a
8536 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8537 <ptsekov@syntrex.com> and Nedelcho Stanev.
8540 *) Implement a countermeasure against a vulnerability recently found
8541 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8542 before application data chunks to avoid the use of known IVs
8543 with data potentially chosen by the attacker.
8546 *) Fix length checks in ssl3_get_client_hello().
8549 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8550 to prevent ssl3_read_internal() from incorrectly assuming that
8551 ssl3_read_bytes() found application data while handshake
8552 processing was enabled when in fact s->s3->in_read_app_data was
8553 merely automatically cleared during the initial handshake.
8554 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8556 *) Fix object definitions for Private and Enterprise: they were not
8557 recognized in their shortname (=lowercase) representation. Extend
8558 obj_dat.pl to issue an error when using undefined keywords instead
8559 of silently ignoring the problem (Svenning Sorensen
8560 <sss@sss.dnsalias.net>).
8563 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8564 generators, i.e. generators other than 2 and 5. (Previously, the
8565 code did not properly initialise the 'add' and 'rem' values to
8566 BN_generate_prime().)
8568 In the new general case, we do not insist that 'generator' is
8569 actually a primitive root: This requirement is rather pointless;
8570 a generator of the order-q subgroup is just as good, if not
8574 *) Map new X509 verification errors to alerts. Discovered and submitted by
8575 Tom Wu <tom@arcot.com>.
8578 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8579 returning non-zero before the data has been completely received
8580 when using non-blocking I/O.
8581 [Bodo Moeller; problem pointed out by John Hughes]
8583 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8584 [Ben Laurie, Lutz Jaenicke]
8586 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8587 Yoram Zahavi <YoramZ@gilian.com>).
8590 *) Add information about CygWin 1.3 and on, and preserve proper
8591 configuration for the versions before that.
8592 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8594 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8595 check whether we deal with a copy of a session and do not delete from
8596 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8597 <izhar@checkpoint.com>.
8600 *) Do not store session data into the internal session cache, if it
8601 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8602 flag is set). Proposed by Aslam <aslam@funk.com>.
8605 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8609 *) [In 0.9.6d-engine release:]
8610 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8611 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8613 *) Add the configuration target linux-s390x.
8614 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8616 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8617 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8618 variable as an indication that a ClientHello message has been
8619 received. As the flag value will be lost between multiple
8620 invocations of ssl3_accept when using non-blocking I/O, the
8621 function may not be aware that a handshake has actually taken
8622 place, thus preventing a new session from being added to the
8625 To avoid this problem, we now set s->new_session to 2 instead of
8626 using a local variable.
8627 [Lutz Jaenicke, Bodo Moeller]
8629 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8630 if the SSL_R_LENGTH_MISMATCH error is detected.
8631 [Geoff Thorpe, Bodo Moeller]
8633 *) New 'shared_ldflag' column in Configure platform table.
8636 *) Fix EVP_CIPHER_mode macro.
8637 ["Dan S. Camper" <dan@bti.net>]
8639 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8640 type, we must throw them away by setting rr->length to 0.
8641 [D P Chang <dpc@qualys.com>]
8643 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8645 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8646 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8647 worked incorrectly for those cases where range = 10..._2 and
8648 3*range is two bits longer than range.)
8651 *) Only add signing time to PKCS7 structures if it is not already
8655 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8656 OBJ_ld_ce should be OBJ_id_ce.
8657 Also some ip-pda OIDs in crypto/objects/objects.txt were
8658 incorrect (cf. RFC 3039).
8659 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8661 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8662 returns early because it has nothing to do.
8663 [Andy Schneider <andy.schneider@bjss.co.uk>]
8665 *) [In 0.9.6c-engine release:]
8666 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8667 [Andy Schneider <andy.schneider@bjss.co.uk>]
8669 *) [In 0.9.6c-engine release:]
8670 Add support for Cryptographic Appliance's keyserver technology.
8671 (Use engine 'keyclient')
8672 [Cryptographic Appliances and Geoff Thorpe]
8674 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8675 is called via tools/c89.sh because arguments have to be
8676 rearranged (all '-L' options must appear before the first object
8678 [Richard Shapiro <rshapiro@abinitio.com>]
8680 *) [In 0.9.6c-engine release:]
8681 Add support for Broadcom crypto accelerator cards, backported
8683 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8685 *) [In 0.9.6c-engine release:]
8686 Add support for SureWare crypto accelerator cards from
8687 Baltimore Technologies. (Use engine 'sureware')
8688 [Baltimore Technologies and Mark Cox]
8690 *) [In 0.9.6c-engine release:]
8691 Add support for crypto accelerator cards from Accelerated
8692 Encryption Processing, www.aep.ie. (Use engine 'aep')
8693 [AEP Inc. and Mark Cox]
8695 *) Add a configuration entry for gcc on UnixWare.
8696 [Gary Benson <gbenson@redhat.com>]
8698 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8699 messages are stored in a single piece (fixed-length part and
8700 variable-length part combined) and fix various bugs found on the way.
8703 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8704 instead. BIO_gethostbyname() does not know what timeouts are
8705 appropriate, so entries would stay in cache even when they have
8707 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8709 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8710 faced with a pathologically small ClientHello fragment that does
8711 not contain client_version: Instead of aborting with an error,
8712 simply choose the highest available protocol version (i.e.,
8713 TLS 1.0 unless it is disabled). In practice, ClientHello
8714 messages are never sent like this, but this change gives us
8715 strictly correct behaviour at least for TLS.
8718 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8719 never resets s->method to s->ctx->method when called from within
8720 one of the SSL handshake functions.
8721 [Bodo Moeller; problem pointed out by Niko Baric]
8723 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8724 (sent using the client's version number) if client_version is
8725 smaller than the protocol version in use. Also change
8726 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8727 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8728 the client will at least see that alert.
8731 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8735 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8736 client receives HelloRequest while in a handshake.
8737 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8739 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8740 should end in 'break', not 'goto end' which circumvents various
8741 cleanups done in state SSL_ST_OK. But session related stuff
8742 must be disabled for SSL_ST_OK in the case that we just sent a
8745 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8746 before just sending a HelloRequest.
8747 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8749 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8750 reveal whether illegal block cipher padding was found or a MAC
8751 verification error occurred. (Neither SSLerr() codes nor alerts
8752 are directly visible to potential attackers, but the information
8753 may leak via logfiles.)
8755 Similar changes are not required for the SSL 2.0 implementation
8756 because the number of padding bytes is sent in clear for SSL 2.0,
8757 and the extra bytes are just ignored. However ssl/s2_pkt.c
8758 failed to verify that the purported number of padding bytes is in
8762 *) Add OpenUNIX-8 support including shared libraries
8763 (Boyd Lynn Gerber <gerberb@zenez.com>).
8766 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8767 'wristwatch attack' using huge encoding parameters (cf.
8768 James H. Manger's CRYPTO 2001 paper). Note that the
8769 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8770 encoding parameters and hence was not vulnerable.
8773 *) BN_sqr() bug fix.
8774 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8776 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8777 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8778 followed by modular reduction.
8779 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8781 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8782 equivalent based on BN_pseudo_rand() instead of BN_rand().
8785 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8786 This function was broken, as the check for a new client hello message
8787 to handle SGC did not allow these large messages.
8788 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8791 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8794 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8795 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8798 *) Rework the configuration and shared library support for Tru64 Unix.
8799 The configuration part makes use of modern compiler features and
8800 still retains old compiler behavior for those that run older versions
8801 of the OS. The shared library support part includes a variant that
8802 uses the RPATH feature, and is available through the special
8803 configuration target "alpha-cc-rpath", which will never be selected
8805 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8807 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8808 with the same message size as in ssl3_get_certificate_request().
8809 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8810 messages might inadvertently be reject as too long.
8811 [Petr Lampa <lampa@fee.vutbr.cz>]
8813 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8816 *) Modified SSL library such that the verify_callback that has been set
8817 specificly for an SSL object with SSL_set_verify() is actually being
8818 used. Before the change, a verify_callback set with this function was
8819 ignored and the verify_callback() set in the SSL_CTX at the time of
8820 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8821 to allow the necessary settings.
8824 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8825 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8826 done automatically (in contradiction to the requirements of the C
8827 standard). This made problems when used from OpenSSH.
8830 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8831 dh->length and always used
8833 BN_rand_range(priv_key, dh->p).
8835 BN_rand_range() is not necessary for Diffie-Hellman, and this
8836 specific range makes Diffie-Hellman unnecessarily inefficient if
8837 dh->length (recommended exponent length) is much smaller than the
8838 length of dh->p. We could use BN_rand_range() if the order of
8839 the subgroup was stored in the DH structure, but we only have
8844 BN_rand(priv_key, l, ...)
8846 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8852 RSA_eay_public_encrypt
8853 RSA_eay_private_decrypt
8854 RSA_eay_private_encrypt (signing)
8855 RSA_eay_public_decrypt (signature verification)
8857 (default implementations for RSA_public_encrypt,
8858 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8859 always reject numbers >= n.
8862 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8863 to synchronize access to 'locking_thread'. This is necessary on
8864 systems where access to 'locking_thread' (an 'unsigned long'
8865 variable) is not atomic.
8868 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8869 *before* setting the 'crypto_lock_rand' flag. The previous code had
8870 a race condition if 0 is a valid thread ID.
8871 [Travis Vitek <vitek@roguewave.com>]
8873 *) Add support for shared libraries under Irix.
8874 [Albert Chin-A-Young <china@thewrittenword.com>]
8876 *) Add configuration option to build on Linux on both big-endian and
8878 [Ralf Baechle <ralf@uni-koblenz.de>]
8880 *) Add the possibility to create shared libraries on HP-UX.
8883 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8885 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8886 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8887 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8888 PRNG state recovery was possible based on the output of
8889 one PRNG request appropriately sized to gain knowledge on
8890 'md' followed by enough consecutive 1-byte PRNG requests
8891 to traverse all of 'state'.
8893 1. When updating 'md_local' (the current thread's copy of 'md')
8894 during PRNG output generation, hash all of the previous
8895 'md_local' value, not just the half used for PRNG output.
8897 2. Make the number of bytes from 'state' included into the hash
8898 independent from the number of PRNG bytes requested.
8900 The first measure alone would be sufficient to avoid
8901 Markku-Juhani's attack. (Actually it had never occurred
8902 to me that the half of 'md_local' used for chaining was the
8903 half from which PRNG output bytes were taken -- I had always
8904 assumed that the secret half would be used.) The second
8905 measure makes sure that additional data from 'state' is never
8906 mixed into 'md_local' in small portions; this heuristically
8907 further strengthens the PRNG.
8910 *) Fix crypto/bn/asm/mips3.s.
8913 *) When only the key is given to "enc", the IV is undefined. Print out
8914 an error message in this case.
8917 *) Handle special case when X509_NAME is empty in X509 printing routines.
8920 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8921 positive and less than q.
8924 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8925 used: it isn't thread safe and the add_lock_callback should handle
8927 [Paul Rose <Paul.Rose@bridge.com>]
8929 *) Verify that incoming data obeys the block size in
8930 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8934 [Ulf Möller, Bodo Möller]
8936 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8937 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8938 when fixing the server behaviour for backwards-compatible 'client
8939 hello' messages. (Note that the attack is impractical against
8940 SSL 3.0 and TLS 1.0 anyway because length and version checking
8941 means that the probability of guessing a valid ciphertext is
8942 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8945 Before 0.9.5, the countermeasure (hide the error by generating a
8946 random 'decryption result') did not work properly because
8947 ERR_clear_error() was missing, meaning that SSL_get_error() would
8948 detect the supposedly ignored error.
8950 Both problems are now fixed.
8953 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8954 (previously it was 1024).
8957 *) Fix for compatibility mode trust settings: ignore trust settings
8958 unless some valid trust or reject settings are present.
8961 *) Fix for blowfish EVP: its a variable length cipher.
8964 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8965 parameters in DSA public key structures and return an error in the
8966 DSA routines if parameters are absent.
8969 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8970 in the current directory if neither $RANDFILE nor $HOME was set.
8971 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8972 caused some confusion to Windows users who haven't defined $HOME.
8973 Thus RAND_file_name() is changed again: e_os.h can define a
8974 DEFAULT_HOME, which will be used if $HOME is not set.
8975 For Windows, we use "C:"; on other platforms, we still require
8976 environment variables.
8978 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8979 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8980 having multiple threads call RAND_poll() concurrently.
8983 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8984 combination of a flag and a thread ID variable.
8985 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8986 flag), *other* threads can enter ssleay_add_bytes without obeying
8987 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8988 that they do not hold after the first thread unsets add_do_not_lock).
8991 *) Change bctest again: '-x' expressions are not available in all
8995 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8997 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8998 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9000 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9001 the default extension for executables, if any. Also, make the perl
9002 scripts that use symlink() to test if it really exists and use "cp"
9003 if it doesn't. All this made OpenSSL compilable and installable in
9007 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9008 If SEQUENCE is length is indefinite just set c->slen to the total
9009 amount of data available.
9010 [Steve Henson, reported by shige@FreeBSD.org]
9011 [This change does not apply to 0.9.7.]
9013 *) Change bctest to avoid here-documents inside command substitution
9014 (workaround for FreeBSD /bin/sh bug).
9015 For compatibility with Ultrix, avoid shell functions (introduced
9016 in the bctest version that searches along $PATH).
9019 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9020 with des_encrypt() defined on some operating systems, like Solaris
9024 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9025 On the Importance of Eliminating Errors in Cryptographic
9026 Computations, J. Cryptology 14 (2001) 2, 101-119,
9027 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9030 *) MIPS assembler BIGNUM division bug fix.
9033 *) Disabled incorrect Alpha assembler code.
9036 *) Fix PKCS#7 decode routines so they correctly update the length
9037 after reading an EOC for the EXPLICIT tag.
9039 [This change does not apply to 0.9.7.]
9041 *) Fix bug in PKCS#12 key generation routines. This was triggered
9042 if a 3DES key was generated with a 0 initial byte. Include
9043 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9044 (but broken) behaviour.
9047 *) Enhance bctest to search for a working bc along $PATH and print
9049 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9051 *) Fix memory leaks in err.c: free err_data string if necessary;
9052 don't write to the wrong index in ERR_set_error_data.
9055 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9059 *) Replace rdtsc with _emit statements for VC++ version 5.
9060 [Jeremy Cooper <jeremy@baymoo.org>]
9062 *) Make it possible to reuse SSLv2 sessions.
9065 *) In copy_email() check for >= 0 as a return value for
9066 X509_NAME_get_index_by_NID() since 0 is a valid index.
9067 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9069 *) Avoid coredump with unsupported or invalid public keys by checking if
9070 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9071 PKCS7_verify() fails with non detached data.
9074 *) Don't use getenv in library functions when run as setuid/setgid.
9075 New function OPENSSL_issetugid().
9078 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9079 due to incorrect handling of multi-threading:
9081 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9083 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9085 3. Count how many times MemCheck_off() has been called so that
9086 nested use can be treated correctly. This also avoids
9087 inband-signalling in the previous code (which relied on the
9088 assumption that thread ID 0 is impossible).
9091 *) Add "-rand" option also to s_client and s_server.
9094 *) Fix CPU detection on Irix 6.x.
9095 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9096 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9098 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9101 [This change does not apply to 0.9.7.]
9103 *) Use the cached encoding of an X509_NAME structure rather than
9104 copying it. This is apparently the reason for the libsafe "errors"
9105 but the code is actually correct.
9108 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9109 Bleichenbacher's DSA attack.
9110 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9111 to be set and top=0 forces the highest bit to be set; top=-1 is new
9112 and leaves the highest bit random.
9113 [Ulf Moeller, Bodo Moeller]
9115 *) In the NCONF_...-based implementations for CONF_... queries
9116 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9117 a temporary CONF structure with the data component set to NULL
9118 (which gives segmentation faults in lh_retrieve).
9119 Instead, use NULL for the CONF pointer in CONF_get_string and
9120 CONF_get_number (which may use environment variables) and directly
9121 return NULL from CONF_get_section.
9124 *) Fix potential buffer overrun for EBCDIC.
9127 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9128 keyUsage if basicConstraints absent for a CA.
9131 *) Make SMIME_write_PKCS7() write mail header values with a format that
9132 is more generally accepted (no spaces before the semicolon), since
9133 some programs can't parse those values properly otherwise. Also make
9134 sure BIO's that break lines after each write do not create invalid
9138 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9139 macros previously used would not encode an empty SEQUENCE OF
9140 and break the signature.
9142 [This change does not apply to 0.9.7.]
9144 *) Zero the premaster secret after deriving the master secret in
9148 *) Add some EVP_add_digest_alias registrations (as found in
9149 OpenSSL_add_all_digests()) to SSL_library_init()
9150 aka OpenSSL_add_ssl_algorithms(). This provides improved
9151 compatibility with peers using X.509 certificates
9152 with unconventional AlgorithmIdentifier OIDs.
9155 *) Fix for Irix with NO_ASM.
9156 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9158 *) ./config script fixes.
9159 [Ulf Moeller, Richard Levitte]
9161 *) Fix 'openssl passwd -1'.
9164 *) Change PKCS12_key_gen_asc() so it can cope with non null
9165 terminated strings whose length is passed in the passlen
9166 parameter, for example from PEM callbacks. This was done
9167 by adding an extra length parameter to asc2uni().
9168 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9170 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9171 call failed, free the DSA structure.
9174 *) Fix to uni2asc() to cope with zero length Unicode strings.
9175 These are present in some PKCS#12 files.
9178 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9179 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9180 when writing a 32767 byte record.
9181 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9183 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9184 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9186 (RSA objects have a reference count access to which is protected
9187 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9188 so they are meant to be shared between threads.)
9189 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9190 "Reddie, Steven" <Steven.Reddie@ca.com>]
9192 *) Fix a deadlock in CRYPTO_mem_leaks().
9195 *) Use better test patterns in bntest.
9198 *) rand_win.c fix for Borland C.
9201 *) BN_rshift bugfix for n == 0.
9204 *) Add a 'bctest' script that checks for some known 'bc' bugs
9205 so that 'make test' does not abort just because 'bc' is broken.
9208 *) Store verify_result within SSL_SESSION also for client side to
9209 avoid potential security hole. (Re-used sessions on the client side
9210 always resulted in verify_result==X509_V_OK, not using the original
9211 result of the server certificate verification.)
9214 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9215 SSL3_RT_APPLICATION_DATA, return 0.
9216 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9220 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9221 releases, have been re-implemented by renaming the previous
9222 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9223 and ssl3_read_internal, respectively, and adding 'peek' parameters
9224 to them. The new ssl[23]_{read,peek} functions are calls to
9225 ssl[23]_read_internal with the 'peek' flag set appropriately.
9226 A 'peek' parameter has also been added to ssl3_read_bytes, which
9227 does the actual work for ssl3_read_internal.
9230 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9231 the method-specific "init()" handler. Also clean up ex_data after
9232 calling the method-specific "finish()" handler. Previously, this was
9233 happening the other way round.
9236 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9237 The previous value, 12, was not always sufficient for BN_mod_exp().
9240 *) Make sure that shared libraries get the internal name engine with
9241 the full version number and not just 0. This should mark the
9242 shared libraries as not backward compatible. Of course, this should
9243 be changed again when we can guarantee backward binary compatibility.
9246 *) Fix typo in get_cert_by_subject() in by_dir.c
9247 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9249 *) Rework the system to generate shared libraries:
9251 - Make note of the expected extension for the shared libraries and
9252 if there is a need for symbolic links from for example libcrypto.so.0
9253 to libcrypto.so.0.9.7. There is extended info in Configure for
9256 - Make as few rebuilds of the shared libraries as possible.
9258 - Still avoid linking the OpenSSL programs with the shared libraries.
9260 - When installing, install the shared libraries separately from the
9264 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9266 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9267 and not in SSL_clear because the latter is also used by the
9268 accept/connect functions; previously, the settings made by
9269 SSL_set_read_ahead would be lost during the handshake.
9270 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9272 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9273 Previously, it would create entries for disabled algorithms no
9277 *) Added several new manual pages for SSL_* function.
9280 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9282 *) In ssl23_get_client_hello, generate an error message when faced
9283 with an initial SSL 3.0/TLS record that is too small to contain the
9284 first two bytes of the ClientHello message, i.e. client_version.
9285 (Note that this is a pathologic case that probably has never happened
9286 in real life.) The previous approach was to use the version number
9287 from the record header as a substitute; but our protocol choice
9288 should not depend on that one because it is not authenticated
9289 by the Finished messages.
9292 *) More robust randomness gathering functions for Windows.
9293 [Jeffrey Altman <jaltman@columbia.edu>]
9295 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9296 not set then we don't setup the error code for issuer check errors
9297 to avoid possibly overwriting other errors which the callback does
9298 handle. If an application does set the flag then we assume it knows
9299 what it is doing and can handle the new informational codes
9303 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9304 a general "ANY" type, as such it should be able to decode anything
9305 including tagged types. However it didn't check the class so it would
9306 wrongly interpret tagged types in the same way as their universal
9307 counterpart and unknown types were just rejected. Changed so that the
9308 tagged and unknown types are handled in the same way as a SEQUENCE:
9309 that is the encoding is stored intact. There is also a new type
9310 "V_ASN1_OTHER" which is used when the class is not universal, in this
9311 case we have no idea what the actual type is so we just lump them all
9315 *) On VMS, stdout may very well lead to a file that is written to
9316 in a record-oriented fashion. That means that every write() will
9317 write a separate record, which will be read separately by the
9318 programs trying to read from it. This can be very confusing.
9320 The solution is to put a BIO filter in the way that will buffer
9321 text until a linefeed is reached, and then write everything a
9322 line at a time, so every record written will be an actual line,
9323 not chunks of lines and not (usually doesn't happen, but I've
9324 seen it once) several lines in one record. BIO_f_linebuffer() is
9327 Currently, it's a VMS-only method, because that's where it has
9328 been tested well enough.
9331 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9332 it can return incorrect results.
9333 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9334 but it was in 0.9.6-beta[12].)
9337 *) Disable the check for content being present when verifying detached
9338 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9339 include zero length content when signing messages.
9342 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9343 BIO_ctrl (for BIO pairs).
9346 *) Add DSO method for VMS.
9349 *) Bug fix: Montgomery multiplication could produce results with the
9353 *) Add RPM specification openssl.spec and modify it to build three
9354 packages. The default package contains applications, application
9355 documentation and run-time libraries. The devel package contains
9356 include files, static libraries and function documentation. The
9357 doc package contains the contents of the doc directory. The original
9358 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9361 *) Add a large number of documentation files for many SSL routines.
9362 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9364 *) Add a configuration entry for Sony News 4.
9365 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9367 *) Don't set the two most significant bits to one when generating a
9368 random number < q in the DSA library.
9371 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9372 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9373 the underlying transport is blocking) if a handshake took place.
9374 (The default behaviour is needed by applications such as s_client
9375 and s_server that use select() to determine when to use SSL_read;
9376 but for applications that know in advance when to expect data, it
9377 just makes things more complicated.)
9380 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9384 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9385 work better on such systems.
9386 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9388 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9389 Update PKCS12_parse() so it copies the friendlyName and the
9390 keyid to the certificates aux info.
9393 *) Fix bug in PKCS7_verify() which caused an infinite loop
9394 if there was more than one signature.
9395 [Sven Uszpelkat <su@celocom.de>]
9397 *) Major change in util/mkdef.pl to include extra information
9398 about each symbol, as well as presenting variables as well
9399 as functions. This change means that there's n more need
9400 to rebuild the .num files when some algorithms are excluded.
9403 *) Allow the verify time to be set by an application,
9404 rather than always using the current time.
9407 *) Phase 2 verify code reorganisation. The certificate
9408 verify code now looks up an issuer certificate by a
9409 number of criteria: subject name, authority key id
9410 and key usage. It also verifies self signed certificates
9411 by the same criteria. The main comparison function is
9412 X509_check_issued() which performs these checks.
9414 Lot of changes were necessary in order to support this
9415 without completely rewriting the lookup code.
9417 Authority and subject key identifier are now cached.
9419 The LHASH 'certs' is X509_STORE has now been replaced
9420 by a STACK_OF(X509_OBJECT). This is mainly because an
9421 LHASH can't store or retrieve multiple objects with
9422 the same hash value.
9424 As a result various functions (which were all internal
9425 use only) have changed to handle the new X509_STORE
9426 structure. This will break anything that messed round
9427 with X509_STORE internally.
9429 The functions X509_STORE_add_cert() now checks for an
9430 exact match, rather than just subject name.
9432 The X509_STORE API doesn't directly support the retrieval
9433 of multiple certificates matching a given criteria, however
9434 this can be worked round by performing a lookup first
9435 (which will fill the cache with candidate certificates)
9436 and then examining the cache for matches. This is probably
9437 the best we can do without throwing out X509_LOOKUP
9438 entirely (maybe later...).
9440 The X509_VERIFY_CTX structure has been enhanced considerably.
9442 All certificate lookup operations now go via a get_issuer()
9443 callback. Although this currently uses an X509_STORE it
9444 can be replaced by custom lookups. This is a simple way
9445 to bypass the X509_STORE hackery necessary to make this
9446 work and makes it possible to use more efficient techniques
9447 in future. A very simple version which uses a simple
9448 STACK for its trusted certificate store is also provided
9449 using X509_STORE_CTX_trusted_stack().
9451 The verify_cb() and verify() callbacks now have equivalents
9452 in the X509_STORE_CTX structure.
9454 X509_STORE_CTX also has a 'flags' field which can be used
9455 to customise the verify behaviour.
9458 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9459 excludes S/MIME capabilities.
9462 *) When a certificate request is read in keep a copy of the
9463 original encoding of the signed data and use it when outputting
9464 again. Signatures then use the original encoding rather than
9465 a decoded, encoded version which may cause problems if the
9466 request is improperly encoded.
9469 *) For consistency with other BIO_puts implementations, call
9470 buffer_write(b, ...) directly in buffer_puts instead of calling
9473 In BIO_puts, increment b->num_write as in BIO_write.
9474 [Peter.Sylvester@EdelWeb.fr]
9476 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9477 BN_zero, we may not return a BIGNUM with an array consisting of
9481 *) Avoid calling abort() from within the library when problems are
9482 detected, except if preprocessor symbols have been defined
9483 (such as REF_CHECK, BN_DEBUG etc.).
9486 *) New openssl application 'rsautl'. This utility can be
9487 used for low level RSA operations. DER public key
9488 BIO/fp routines also added.
9491 *) New Configure entry and patches for compiling on QNX 4.
9492 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9494 *) A demo state-machine implementation was sponsored by
9495 Nuron (http://www.nuron.com/) and is now available in
9496 demos/state_machine.
9499 *) New options added to the 'dgst' utility for signature
9500 generation and verification.
9503 *) Unrecognized PKCS#7 content types are now handled via a
9504 catch all ASN1_TYPE structure. This allows unsupported
9505 types to be stored as a "blob" and an application can
9506 encode and decode it manually.
9509 *) Fix various signed/unsigned issues to make a_strex.c
9511 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9513 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9514 length if passed a buffer. ASN1_INTEGER_to_BN failed
9515 if passed a NULL BN and its argument was negative.
9516 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9518 *) Modification to PKCS#7 encoding routines to output definite
9519 length encoding. Since currently the whole structures are in
9520 memory there's not real point in using indefinite length
9521 constructed encoding. However if OpenSSL is compiled with
9522 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9525 *) Added BIO_vprintf() and BIO_vsnprintf().
9528 *) Added more prefixes to parse for in the strings written
9529 through a logging bio, to cover all the levels that are available
9530 through syslog. The prefixes are now:
9532 PANIC, EMERG, EMR => LOG_EMERG
9533 ALERT, ALR => LOG_ALERT
9534 CRIT, CRI => LOG_CRIT
9535 ERROR, ERR => LOG_ERR
9536 WARNING, WARN, WAR => LOG_WARNING
9537 NOTICE, NOTE, NOT => LOG_NOTICE
9538 INFO, INF => LOG_INFO
9539 DEBUG, DBG => LOG_DEBUG
9541 and as before, if none of those prefixes are present at the
9542 beginning of the string, LOG_ERR is chosen.
9544 On Win32, the LOG_* levels are mapped according to this:
9546 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9547 LOG_WARNING => EVENTLOG_WARNING_TYPE
9548 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9552 *) Made it possible to reconfigure with just the configuration
9553 argument "reconf" or "reconfigure". The command line arguments
9554 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9555 and are retrieved from there when reconfiguring.
9559 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9561 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9564 *) The obj_dat.pl script was messing up the sorting of object
9565 names. The reason was that it compared the quoted version
9566 of strings as a result "OCSP" > "OCSP Signing" because
9567 " > SPACE. Changed script to store unquoted versions of
9568 names and add quotes on output. It was also omitting some
9569 names from the lookup table if they were given a default
9570 value (that is if SN is missing it is given the same
9571 value as LN and vice versa), these are now added on the
9572 grounds that if an object has a name we should be able to
9573 look it up. Finally added warning output when duplicate
9574 short or long names are found.
9577 *) Changes needed for Tandem NSK.
9578 [Scott Uroff <scott@xypro.com>]
9580 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9581 RSA_padding_check_SSLv23(), special padding was never detected
9582 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9583 version rollback attacks was not effective.
9585 In s23_clnt.c, don't use special rollback-attack detection padding
9586 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9587 client; similarly, in s23_srvr.c, don't do the rollback check if
9588 SSL 2.0 is the only protocol enabled in the server.
9591 *) Make it possible to get hexdumps of unprintable data with 'openssl
9592 asn1parse'. By implication, the functions ASN1_parse_dump() and
9593 BIO_dump_indent() are added.
9596 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9597 these print out strings and name structures based on various
9598 flags including RFC2253 support and proper handling of
9599 multibyte characters. Added options to the 'x509' utility
9600 to allow the various flags to be set.
9603 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9604 Also change the functions X509_cmp_current_time() and
9605 X509_gmtime_adj() work with an ASN1_TIME structure,
9606 this will enable certificates using GeneralizedTime in validity
9607 dates to be checked.
9610 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9611 negative public key encodings) on by default,
9612 NO_NEG_PUBKEY_BUG can be set to disable it.
9615 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9616 content octets. An i2c_ASN1_OBJECT is unnecessary because
9617 the encoding can be trivially obtained from the structure.
9620 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9621 not read locks (CRYPTO_r_[un]lock).
9624 *) A first attempt at creating official support for shared
9625 libraries through configuration. I've kept it so the
9626 default is static libraries only, and the OpenSSL programs
9627 are always statically linked for now, but there are
9628 preparations for dynamic linking in place.
9629 This has been tested on Linux and Tru64.
9632 *) Randomness polling function for Win9x, as described in:
9633 Peter Gutmann, Software Generation of Practically Strong
9637 *) Fix so PRNG is seeded in req if using an already existing
9641 *) New options to smime application. -inform and -outform
9642 allow alternative formats for the S/MIME message including
9643 PEM and DER. The -content option allows the content to be
9644 specified separately. This should allow things like Netscape
9645 form signing output easier to verify.
9648 *) Fix the ASN1 encoding of tags using the 'long form'.
9651 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9652 STRING types. These convert content octets to and from the
9653 underlying type. The actual tag and length octets are
9654 already assumed to have been read in and checked. These
9655 are needed because all other string types have virtually
9656 identical handling apart from the tag. By having versions
9657 of the ASN1 functions that just operate on content octets
9658 IMPLICIT tagging can be handled properly. It also allows
9659 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9660 and ASN1_INTEGER are identical apart from the tag.
9663 *) Change the handling of OID objects as follows:
9665 - New object identifiers are inserted in objects.txt, following
9666 the syntax given in objects.README.
9667 - objects.pl is used to process obj_mac.num and create a new
9669 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9672 This is currently kind of a hack, and the perl code in objects.pl
9673 isn't very elegant, but it works as I intended. The simplest way
9674 to check that it worked correctly is to look in obj_dat.h and
9675 check the array nid_objs and make sure the objects haven't moved
9676 around (this is important!). Additions are OK, as well as
9677 consistent name changes.
9680 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9683 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9684 The given file adds to whatever has already been seeded into the
9685 random pool through the RANDFILE configuration file option or
9686 environment variable, or the default random state file.
9689 *) mkstack.pl now sorts each macro group into lexical order.
9690 Previously the output order depended on the order the files
9691 appeared in the directory, resulting in needless rewriting
9695 *) Patches to make OpenSSL compile under Win32 again. Mostly
9696 work arounds for the VC++ problem that it treats func() as
9697 func(void). Also stripped out the parts of mkdef.pl that
9698 added extra typesafe functions: these no longer exist.
9701 *) Reorganisation of the stack code. The macros are now all
9702 collected in safestack.h . Each macro is defined in terms of
9703 a "stack macro" of the form SKM_<name>(type, a, b). The
9704 DEBUG_SAFESTACK is now handled in terms of function casts,
9705 this has the advantage of retaining type safety without the
9706 use of additional functions. If DEBUG_SAFESTACK is not defined
9707 then the non typesafe macros are used instead. Also modified the
9708 mkstack.pl script to handle the new form. Needs testing to see
9709 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9710 the default if no major problems. Similar behaviour for ASN1_SET_OF
9711 and PKCS12_STACK_OF.
9714 *) When some versions of IIS use the 'NET' form of private key the
9715 key derivation algorithm is different. Normally MD5(password) is
9716 used as a 128 bit RC4 key. In the modified case
9717 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9718 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9719 as the old Netscape_RSA functions except they have an additional
9720 'sgckey' parameter which uses the modified algorithm. Also added
9721 an -sgckey command line option to the rsa utility. Thanks to
9722 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9723 algorithm to openssl-dev.
9726 *) The evp_local.h macros were using 'c.##kname' which resulted in
9727 invalid expansion on some systems (SCO 5.0.5 for example).
9728 Corrected to 'c.kname'.
9729 [Phillip Porch <root@theporch.com>]
9731 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9732 a STACK of email addresses from a certificate or request, these look
9733 in the subject name and the subject alternative name extensions and
9734 omit any duplicate addresses.
9737 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9738 This makes DSA verification about 2 % faster.
9741 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9742 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9743 plus overhead for 1024 bit moduli).
9744 This makes exponentiations about 0.5 % faster for 1024 bit
9745 exponents (as measured by "openssl speed rsa2048").
9748 *) Rename memory handling macros to avoid conflicts with other
9750 Malloc => OPENSSL_malloc
9751 Malloc_locked => OPENSSL_malloc_locked
9752 Realloc => OPENSSL_realloc
9753 Free => OPENSSL_free
9756 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9757 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9760 *) CygWin32 support.
9761 [John Jarvie <jjarvie@newsguy.com>]
9763 *) The type-safe stack code has been rejigged. It is now only compiled
9764 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9765 by default all type-specific stack functions are "#define"d back to
9766 standard stack functions. This results in more streamlined output
9767 but retains the type-safety checking possibilities of the original
9771 *) The STACK code has been cleaned up, and certain type declarations
9772 that didn't make a lot of sense have been brought in line. This has
9773 also involved a cleanup of sorts in safestack.h to more correctly
9774 map type-safe stack functions onto their plain stack counterparts.
9775 This work has also resulted in a variety of "const"ifications of
9776 lots of the code, especially "_cmp" operations which should normally
9777 be prototyped with "const" parameters anyway.
9780 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9781 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9782 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9783 where all of 'md' is used each time the PRNG is used, but 'state'
9784 is used only indexed by a cyclic counter. As entropy may not be
9785 well distributed from the beginning, 'md' is important as a
9786 chaining variable. However, the output function chains only half
9787 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9788 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9789 in all of 'state' being rewritten, with the new values depending
9790 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9793 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9794 the handshake is continued after ssl_verify_cert_chain();
9795 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9796 can lead to 'unexplainable' connection aborts later.
9797 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9799 *) Major EVP API cipher revision.
9800 Add hooks for extra EVP features. This allows various cipher
9801 parameters to be set in the EVP interface. Support added for variable
9802 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9803 setting of RC2 and RC5 parameters.
9805 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9808 Remove lots of duplicated code from the EVP library. For example *every*
9809 cipher init() function handles the 'iv' in the same way according to the
9810 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9811 for CFB and OFB modes they zero ctx->num.
9813 New functionality allows removal of S/MIME code RC2 hack.
9815 Most of the routines have the same form and so can be declared in terms
9818 By shifting this to the top level EVP_CipherInit() it can be removed from
9819 all individual ciphers. If the cipher wants to handle IVs or keys
9820 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9823 Change lots of functions like EVP_EncryptUpdate() to now return a
9824 value: although software versions of the algorithms cannot fail
9825 any installed hardware versions can.
9828 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9829 this option is set, tolerate broken clients that send the negotiated
9830 protocol version number instead of the requested protocol version
9834 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9835 i.e. non-zero for export ciphersuites, zero otherwise.
9836 Previous versions had this flag inverted, inconsistent with
9837 rsa_tmp_cb (..._TMP_RSA_CB).
9838 [Bodo Moeller; problem reported by Amit Chopra]
9840 *) Add missing DSA library text string. Work around for some IIS
9841 key files with invalid SEQUENCE encoding.
9844 *) Add a document (doc/standards.txt) that list all kinds of standards
9845 and so on that are implemented in OpenSSL.
9848 *) Enhance c_rehash script. Old version would mishandle certificates
9849 with the same subject name hash and wouldn't handle CRLs at all.
9850 Added -fingerprint option to crl utility, to support new c_rehash
9854 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9857 *) Fix for SSL server purpose checking. Server checking was
9858 rejecting certificates which had extended key usage present
9859 but no ssl client purpose.
9860 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9862 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9863 is a little unclear about how a blank password is handled.
9864 Since the password in encoded as a BMPString with terminating
9865 double NULL a zero length password would end up as just the
9866 double NULL. However no password at all is different and is
9867 handled differently in the PKCS#12 key generation code. NS
9868 treats a blank password as zero length. MSIE treats it as no
9869 password on export: but it will try both on import. We now do
9870 the same: PKCS12_parse() tries zero length and no password if
9871 the password is set to "" or NULL (NULL is now a valid password:
9872 it wasn't before) as does the pkcs12 application.
9875 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9876 perror when PEM_read_bio_X509_REQ fails, the error message must
9877 be obtained from the error queue.
9880 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9881 it in ERR_remove_state if appropriate, and change ERR_get_state
9882 accordingly to avoid race conditions (this is necessary because
9883 thread_hash is no longer constant once set).
9886 *) Bugfix for linux-elf makefile.one.
9889 *) RSA_get_default_method() will now cause a default
9890 RSA_METHOD to be chosen if one doesn't exist already.
9891 Previously this was only set during a call to RSA_new()
9892 or RSA_new_method(NULL) meaning it was possible for
9893 RSA_get_default_method() to return NULL.
9896 *) Added native name translation to the existing DSO code
9897 that will convert (if the flag to do so is set) filenames
9898 that are sufficiently small and have no path information
9899 into a canonical native form. Eg. "blah" converted to
9900 "libblah.so" or "blah.dll" etc.
9903 *) New function ERR_error_string_n(e, buf, len) which is like
9904 ERR_error_string(e, buf), but writes at most 'len' bytes
9905 including the 0 terminator. For ERR_error_string_n, 'buf'
9907 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9909 *) CONF library reworked to become more general. A new CONF
9910 configuration file reader "class" is implemented as well as a
9911 new functions (NCONF_*, for "New CONF") to handle it. The now
9912 old CONF_* functions are still there, but are reimplemented to
9913 work in terms of the new functions. Also, a set of functions
9914 to handle the internal storage of the configuration data is
9915 provided to make it easier to write new configuration file
9916 reader "classes" (I can definitely see something reading a
9917 configuration file in XML format, for example), called _CONF_*,
9918 or "the configuration storage API"...
9920 The new configuration file reading functions are:
9922 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9923 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9925 NCONF_default, NCONF_WIN32
9927 NCONF_dump_fp, NCONF_dump_bio
9929 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9930 NCONF_new creates a new CONF object. This works in the same way
9931 as other interfaces in OpenSSL, like the BIO interface.
9932 NCONF_dump_* dump the internal storage of the configuration file,
9933 which is useful for debugging. All other functions take the same
9934 arguments as the old CONF_* functions wth the exception of the
9935 first that must be a `CONF *' instead of a `LHASH *'.
9937 To make it easer to use the new classes with the old CONF_* functions,
9938 the function CONF_set_default_method is provided.
9941 *) Add '-tls1' option to 'openssl ciphers', which was already
9942 mentioned in the documentation but had not been implemented.
9943 (This option is not yet really useful because even the additional
9944 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9947 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9948 OpenSSL-based applications) load shared libraries and bind to
9949 them in a portable way.
9950 [Geoff Thorpe, with contributions from Richard Levitte]
9952 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9954 *) Make sure _lrotl and _lrotr are only used with MSVC.
9956 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9957 (the default implementation of RAND_status).
9959 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9960 to '-clrext' (= clear extensions), as intended and documented.
9961 [Bodo Moeller; inconsistency pointed out by Michael Attili
9964 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9965 was larger than the MD block size.
9966 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9968 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9969 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9970 using the passed key: if the passed key was a private key the result
9971 of X509_print(), for example, would be to print out all the private key
9975 *) des_quad_cksum() byte order bug fix.
9976 [Ulf Möller, using the problem description in krb4-0.9.7, where
9977 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9979 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9981 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9983 *) For easily testing in shell scripts whether some command
9984 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9985 returns with exit code 0 iff no command of the given name is available.
9986 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9987 the output goes to stdout and nothing is printed to stderr.
9988 Additional arguments are always ignored.
9990 Since for each cipher there is a command of the same name,
9991 the 'no-cipher' compilation switches can be tested this way.
9993 ('openssl no-XXX' is not able to detect pseudo-commands such
9994 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9997 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10000 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10001 is set; it will be thrown away anyway because each handshake creates
10003 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10004 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10005 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10006 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10009 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10010 'Q' and 'R' lose their special meanings (quit/renegotiate).
10011 This is part of what -quiet does; unlike -quiet, -ign_eof
10012 does not suppress any output.
10015 *) Add compatibility options to the purpose and trust code. The
10016 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10017 accepts a certificate or CA, this was the previous behaviour,
10018 with all the associated security issues.
10020 X509_TRUST_COMPAT is the old trust behaviour: only and
10021 automatically trust self signed roots in certificate store. A
10022 new trust setting X509_TRUST_DEFAULT is used to specify that
10023 a purpose has no associated trust setting and it should instead
10024 use the value in the default purpose.
10027 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10028 and fix a memory leak.
10031 *) In util/mkerr.pl (which implements 'make errors'), preserve
10032 reason strings from the previous version of the .c file, as
10033 the default to have only downcase letters (and digits) in
10034 automatically generated reasons codes is not always appropriate.
10037 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10038 using strerror. Previously, ERR_reason_error_string() returned
10039 library names as reason strings for SYSerr; but SYSerr is a special
10040 case where small numbers are errno values, not library numbers.
10043 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10044 converts DSA parameters into DH parameters. (When creating parameters,
10045 DSA_generate_parameters is used.)
10048 *) Include 'length' (recommended exponent length) in C code generated
10049 by 'openssl dhparam -C'.
10052 *) The second argument to set_label in perlasm was already being used
10053 so couldn't be used as a "file scope" flag. Moved to third argument
10057 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10058 instead of RAND_bytes for encryption IVs and salts.
10061 *) Include RAND_status() into RAND_METHOD instead of implementing
10062 it only for md_rand.c Otherwise replacing the PRNG by calling
10063 RAND_set_rand_method would be impossible.
10066 *) Don't let DSA_generate_key() enter an infinite loop if the random
10067 number generation fails.
10070 *) New 'rand' application for creating pseudo-random output.
10073 *) Added configuration support for Linux/IA64
10074 [Rolf Haberrecker <rolf@suse.de>]
10076 *) Assembler module support for Mingw32.
10079 *) Shared library support for HPUX (in shlib/).
10080 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10082 *) Shared library support for Solaris gcc.
10083 [Lutz Behnke <behnke@trustcenter.de>]
10085 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10087 *) PKCS7_encrypt() was adding text MIME headers twice because they
10088 were added manually and by SMIME_crlf_copy().
10091 *) In bntest.c don't call BN_rand with zero bits argument.
10092 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10094 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10095 case was implemented. This caused BN_div_recp() to fail occasionally.
10098 *) Add an optional second argument to the set_label() in the perl
10099 assembly language builder. If this argument exists and is set
10100 to 1 it signals that the assembler should use a symbol whose
10101 scope is the entire file, not just the current function. This
10102 is needed with MASM which uses the format label:: for this scope.
10103 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10105 *) Change the ASN1 types so they are typedefs by default. Before
10106 almost all types were #define'd to ASN1_STRING which was causing
10107 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10111 *) Change names of new functions to the new get1/get0 naming
10112 convention: After 'get1', the caller owns a reference count
10113 and has to call ..._free; 'get0' returns a pointer to some
10114 data structure without incrementing reference counters.
10115 (Some of the existing 'get' functions increment a reference
10116 counter, some don't.)
10117 Similarly, 'set1' and 'add1' functions increase reference
10118 counters or duplicate objects.
10121 *) Allow for the possibility of temp RSA key generation failure:
10122 the code used to assume it always worked and crashed on failure.
10125 *) Fix potential buffer overrun problem in BIO_printf().
10126 [Ulf Möller, using public domain code by Patrick Powell; problem
10127 pointed out by David Sacerdote <das33@cornell.edu>]
10129 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10130 RAND_egd() and RAND_status(). In the command line application,
10131 the EGD socket can be specified like a seed file using RANDFILE
10135 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10136 Some CAs (e.g. Verisign) distribute certificates in this form.
10139 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10140 list to exclude them. This means that no special compilation option
10141 is needed to use anonymous DH: it just needs to be included in the
10145 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10146 EVP_MD_type. The old functionality is available in a new macro called
10147 EVP_MD_md(). Change code that uses it and update docs.
10150 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10151 where the 'void *' argument is replaced by a function pointer argument.
10152 Previously 'void *' was abused to point to functions, which works on
10153 many platforms, but is not correct. As these functions are usually
10154 called by macros defined in OpenSSL header files, most source code
10155 should work without changes.
10158 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10159 sections with information on -D... compiler switches used for
10160 compiling the library so that applications can see them. To enable
10161 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10162 must be defined. E.g.,
10163 #define OPENSSL_ALGORITHM_DEFINES
10164 #include <openssl/opensslconf.h>
10165 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10166 [Richard Levitte, Ulf and Bodo Möller]
10168 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10172 *) Change the 'other' type in certificate aux info to a STACK_OF
10173 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10174 the required ASN1 format: arbitrary types determined by an OID.
10177 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10178 argument to 'req'. This is not because the function is newer or
10179 better than others it just uses the work 'NEW' in the certificate
10180 request header lines. Some software needs this.
10183 *) Reorganise password command line arguments: now passwords can be
10184 obtained from various sources. Delete the PEM_cb function and make
10185 it the default behaviour: i.e. if the callback is NULL and the
10186 usrdata argument is not NULL interpret it as a null terminated pass
10187 phrase. If usrdata and the callback are NULL then the pass phrase
10188 is prompted for as usual.
10191 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10192 the support is automatically enabled. The resulting binaries will
10193 autodetect the card and use it if present.
10194 [Ben Laurie and Compaq Inc.]
10196 *) Work around for Netscape hang bug. This sends certificate request
10197 and server done in one record. Since this is perfectly legal in the
10198 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10199 the bugs/SSLv3 entry for more info.
10202 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10205 *) Add -rand argument to smime and pkcs12 applications and read/write
10209 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10212 *) Add command line password options to the remaining applications.
10215 *) Bug fix for BN_div_recp() for numerators with an even number of
10219 *) More tests in bntest.c, and changed test_bn output.
10222 *) ./config recognizes MacOS X now.
10225 *) Bug fix for BN_div() when the first words of num and divisor are
10226 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10229 *) Add support for various broken PKCS#8 formats, and command line
10230 options to produce them.
10233 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10234 get temporary BIGNUMs from a BN_CTX.
10237 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10241 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10242 include a #define from the old name to the new. The original intent
10243 was that statically linked binaries could for example just call
10244 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10245 link with digests. This never worked because SSLeay_add_all_digests()
10246 and SSLeay_add_all_ciphers() were in the same source file so calling
10247 one would link with the other. They are now in separate source files.
10250 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10253 *) Use a less unusual form of the Miller-Rabin primality test (it used
10254 a binary algorithm for exponentiation integrated into the Miller-Rabin
10255 loop, our standard modexp algorithms are faster).
10258 *) Support for the EBCDIC character set completed.
10259 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10261 *) Source code cleanups: use const where appropriate, eliminate casts,
10262 use void * instead of char * in lhash.
10265 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10266 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10267 this the server could overwrite ephemeral keys that the client
10271 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10272 using 50 iterations of the Rabin-Miller test.
10274 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10275 iterations of the Rabin-Miller test as required by the appendix
10276 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10277 As BN_is_prime_fasttest includes trial division, DSA parameter
10278 generation becomes much faster.
10280 This implies a change for the callback functions in DSA_is_prime
10281 and DSA_generate_parameters: The callback function is called once
10282 for each positive witness in the Rabin-Miller test, not just
10283 occasionally in the inner loop; and the parameters to the
10284 callback function now provide an iteration count for the outer
10285 loop rather than for the current invocation of the inner loop.
10286 DSA_generate_parameters additionally can call the callback
10287 function with an 'iteration count' of -1, meaning that a
10288 candidate has passed the trial division test (when q is generated
10289 from an application-provided seed, trial division is skipped).
10292 *) New function BN_is_prime_fasttest that optionally does trial
10293 division before starting the Rabin-Miller test and has
10294 an additional BN_CTX * argument (whereas BN_is_prime always
10295 has to allocate at least one BN_CTX).
10296 'callback(1, -1, cb_arg)' is called when a number has passed the
10297 trial division stage.
10300 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10304 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10307 *) New function BN_pseudo_rand().
10310 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10311 bignum version of BN_from_montgomery() with the working code from
10312 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10316 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10317 made it impossible to use the same SSL_SESSION data structure in
10318 SSL2 clients in multiple threads.
10321 *) The return value of RAND_load_file() no longer counts bytes obtained
10322 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10323 to seed the PRNG (previously an explicit byte count was required).
10324 [Ulf Möller, Bodo Möller]
10326 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10327 used (char *) instead of (void *) and had casts all over the place.
10330 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10333 *) Retain source code compatibility for BN_prime_checks macro:
10334 BN_is_prime(..., BN_prime_checks, ...) now uses
10335 BN_prime_checks_for_size to determine the appropriate number of
10336 Rabin-Miller iterations.
10339 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10340 DH_CHECK_P_NOT_SAFE_PRIME.
10341 (Check if this is true? OpenPGP calls them "strong".)
10344 *) Merge the functionality of "dh" and "gendh" programs into a new program
10345 "dhparam". The old programs are retained for now but will handle DH keys
10346 (instead of parameters) in future.
10349 *) Make the ciphers, s_server and s_client programs check the return values
10350 when a new cipher list is set.
10353 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10354 ciphers. Before when the 56bit ciphers were enabled the sorting was
10357 The syntax for the cipher sorting has been extended to support sorting by
10358 cipher-strength (using the strength_bits hard coded in the tables).
10359 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10361 Fix a bug in the cipher-command parser: when supplying a cipher command
10362 string with an "undefined" symbol (neither command nor alphanumeric
10363 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10364 an error is flagged.
10366 Due to the strength-sorting extension, the code of the
10367 ssl_create_cipher_list() function was completely rearranged. I hope that
10368 the readability was also increased :-)
10369 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10371 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10372 for the first serial number and places 2 in the serial number file. This
10373 avoids problems when the root CA is created with serial number zero and
10374 the first user certificate has the same issuer name and serial number
10378 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10379 the new code. Add documentation for this stuff.
10382 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10383 X509_*() to X509at_*() on the grounds that they don't handle X509
10384 structures and behave in an analogous way to the X509v3 functions:
10385 they shouldn't be called directly but wrapper functions should be used
10388 So we also now have some wrapper functions that call the X509at functions
10389 when passed certificate requests. (TO DO: similar things can be done with
10390 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10391 things. Some of these need some d2i or i2d and print functionality
10392 because they handle more complex structures.)
10395 *) Add missing #ifndefs that caused missing symbols when building libssl
10396 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10397 NO_RSA in ssl/s2*.c.
10398 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10400 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10401 has a return value which indicates the quality of the random data
10402 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10403 error queue. New function RAND_pseudo_bytes() generates output that is
10404 guaranteed to be unique but not unpredictable. RAND_add is like
10405 RAND_seed, but takes an extra argument for an entropy estimate
10406 (RAND_seed always assumes full entropy).
10409 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10410 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10411 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10412 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10413 false-positive rate of at most 2^-80 for random input.
10416 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10419 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10420 in the 0.9.5 release), this returns the chain
10421 from an X509_CTX structure with a dup of the stack and all
10422 the X509 reference counts upped: so the stack will exist
10423 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10426 Also make SSL_SESSION_print() print out the verify return
10430 *) Add manpage for the pkcs12 command. Also change the default
10431 behaviour so MAC iteration counts are used unless the new
10432 -nomaciter option is used. This improves file security and
10433 only older versions of MSIE (4.0 for example) need it.
10436 *) Honor the no-xxx Configure options when creating .DEF files.
10439 *) Add PKCS#10 attributes to field table: challengePassword,
10440 unstructuredName and unstructuredAddress. These are taken from
10441 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10442 international characters are used.
10444 More changes to X509_ATTRIBUTE code: allow the setting of types
10445 based on strings. Remove the 'loc' parameter when adding
10446 attributes because these will be a SET OF encoding which is sorted
10450 *) Initial changes to the 'req' utility to allow request generation
10451 automation. This will allow an application to just generate a template
10452 file containing all the field values and have req construct the
10455 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10456 used all over the place including certificate requests and PKCS#7
10457 structures. They are currently handled manually where necessary with
10458 some primitive wrappers for PKCS#7. The new functions behave in a
10459 manner analogous to the X509 extension functions: they allow
10460 attributes to be looked up by NID and added.
10462 Later something similar to the X509V3 code would be desirable to
10463 automatically handle the encoding, decoding and printing of the
10464 more complex types. The string types like challengePassword can
10465 be handled by the string table functions.
10467 Also modified the multi byte string table handling. Now there is
10468 a 'global mask' which masks out certain types. The table itself
10469 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10470 is useful when for example there is only one permissible type
10471 (as in countryName) and using the mask might result in no valid
10475 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10476 SSL_get_peer_finished to allow applications to obtain the latest
10477 Finished messages sent to the peer or expected from the peer,
10478 respectively. (SSL_get_peer_finished is usually the Finished message
10479 actually received from the peer, otherwise the protocol will be aborted.)
10481 As the Finished message are message digests of the complete handshake
10482 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10483 be used for external authentication procedures when the authentication
10484 provided by SSL/TLS is not desired or is not enough.
10487 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10488 the host supports BWX extension and if Compaq C is present on the
10489 $PATH. Just exploiting of the BWX extension results in 20-30%
10490 performance kick for some algorithms, e.g. DES and RC4 to mention
10491 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10495 *) Add support for MS "fast SGC". This is arguably a violation of the
10496 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10497 weak crypto and after checking the certificate is SGC a second one
10498 with strong crypto. MS SGC stops the first handshake after receiving
10499 the server certificate message and sends a second client hello. Since
10500 a server will typically do all the time consuming operations before
10501 expecting any further messages from the client (server key exchange
10502 is the most expensive) there is little difference between the two.
10504 To get OpenSSL to support MS SGC we have to permit a second client
10505 hello message after we have sent server done. In addition we have to
10506 reset the MAC if we do get this second client hello.
10509 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10510 if a DER encoded private key is RSA or DSA traditional format. Changed
10511 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10512 format DER encoded private key. Newer code should use PKCS#8 format which
10513 has the key type encoded in the ASN1 structure. Added DER private key
10514 support to pkcs8 application.
10517 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10518 ciphersuites has been selected (as required by the SSL 3/TLS 1
10519 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10520 is set, we interpret this as a request to violate the specification
10521 (the worst that can happen is a handshake failure, and 'correct'
10522 behaviour would result in a handshake failure anyway).
10525 *) In SSL_CTX_add_session, take into account that there might be multiple
10526 SSL_SESSION structures with the same session ID (e.g. when two threads
10527 concurrently obtain them from an external cache).
10528 The internal cache can handle only one SSL_SESSION with a given ID,
10529 so if there's a conflict, we now throw out the old one to achieve
10533 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10534 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10535 some routines that use cipher OIDs: some ciphers do not have OIDs
10536 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10540 *) Simplify the trust setting structure and code. Now we just have
10541 two sequences of OIDs for trusted and rejected settings. These will
10542 typically have values the same as the extended key usage extension
10543 and any application specific purposes.
10545 The trust checking code now has a default behaviour: it will just
10546 check for an object with the same NID as the passed id. Functions can
10547 be provided to override either the default behaviour or the behaviour
10548 for a given id. SSL client, server and email already have functions
10549 in place for compatibility: they check the NID and also return "trusted"
10550 if the certificate is self signed.
10553 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10554 traditional format into an EVP_PKEY structure.
10557 *) Add a password callback function PEM_cb() which either prompts for
10558 a password if usr_data is NULL or otherwise assumes it is a null
10559 terminated password. Allow passwords to be passed on command line
10560 environment or config files in a few more utilities.
10563 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10564 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10565 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10566 Update documentation.
10569 *) Support for ASN1 "NULL" type. This could be handled before by using
10570 ASN1_TYPE but there wasn't any function that would try to read a NULL
10571 and produce an error if it couldn't. For compatibility we also have
10572 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10573 don't allocate anything because they don't need to.
10576 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10578 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10580 *) Rebuild of the memory allocation routines used by OpenSSL code and
10581 possibly others as well. The purpose is to make an interface that
10582 provide hooks so anyone can build a separate set of allocation and
10583 deallocation routines to be used by OpenSSL, for example memory
10584 pool implementations, or something else, which was previously hard
10585 since Malloc(), Realloc() and Free() were defined as macros having
10586 the values malloc, realloc and free, respectively (except for Win32
10587 compilations). The same is provided for memory debugging code.
10588 OpenSSL already comes with functionality to find memory leaks, but
10589 this gives people a chance to debug other memory problems.
10591 With these changes, a new set of functions and macros have appeared:
10593 CRYPTO_set_mem_debug_functions() [F]
10594 CRYPTO_get_mem_debug_functions() [F]
10595 CRYPTO_dbg_set_options() [F]
10596 CRYPTO_dbg_get_options() [F]
10597 CRYPTO_malloc_debug_init() [M]
10599 The memory debug functions are NULL by default, unless the library
10600 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10601 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10602 gives the standard debugging functions that come with OpenSSL) or
10603 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10604 provided by the library user) must be used. When the standard
10605 debugging functions are used, CRYPTO_dbg_set_options can be used to
10606 request additional information:
10607 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10608 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10610 Also, things like CRYPTO_set_mem_functions will always give the
10611 expected result (the new set of functions is used for allocation
10612 and deallocation) at all times, regardless of platform and compiler
10615 To finish it up, some functions that were never use in any other
10616 way than through macros have a new API and new semantic:
10618 CRYPTO_dbg_malloc()
10619 CRYPTO_dbg_realloc()
10622 All macros of value have retained their old syntax.
10623 [Richard Levitte and Bodo Moeller]
10625 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10626 ordering of SMIMECapabilities wasn't in "strength order" and there
10627 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10631 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10632 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10633 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10635 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10636 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10637 functionality to handle multipart/signed properly) and a utility
10638 called 'smime' to call all this stuff. This is based on code I
10639 originally wrote for Celo who have kindly allowed it to be
10640 included in OpenSSL.
10643 *) Add variants des_set_key_checked and des_set_key_unchecked of
10644 des_set_key (aka des_key_sched). Global variable des_check_key
10645 decides which of these is called by des_set_key; this way
10646 des_check_key behaves as it always did, but applications and
10647 the library itself, which was buggy for des_check_key == 1,
10648 have a cleaner way to pick the version they need.
10651 *) New function PKCS12_newpass() which changes the password of a
10655 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10656 dynamic mix. In both cases the ids can be used as an index into the
10657 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10658 functions so they accept a list of the field values and the
10659 application doesn't need to directly manipulate the X509_TRUST
10663 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10667 *) Modify the way the V3 extension code looks up extensions. This now
10668 works in a similar way to the object code: we have some "standard"
10669 extensions in a static table which is searched with OBJ_bsearch()
10670 and the application can add dynamic ones if needed. The file
10671 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10672 updated whenever a new extension is added to the core code and kept
10673 in ext_nid order. There is a simple program 'tabtest.c' which checks
10674 this. New extensions are not added too often so this file can readily
10675 be maintained manually.
10677 There are two big advantages in doing things this way. The extensions
10678 can be looked up immediately and no longer need to be "added" using
10679 X509V3_add_standard_extensions(): this function now does nothing.
10680 [Side note: I get *lots* of email saying the extension code doesn't
10681 work because people forget to call this function]
10682 Also no dynamic allocation is done unless new extensions are added:
10683 so if we don't add custom extensions there is no need to call
10684 X509V3_EXT_cleanup().
10687 *) Modify enc utility's salting as follows: make salting the default. Add a
10688 magic header, so unsalted files fail gracefully instead of just decrypting
10689 to garbage. This is because not salting is a big security hole, so people
10690 should be discouraged from doing it.
10693 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10694 digest to be passed on the command line but it only used this
10695 parameter when signing a certificate. Modified so all relevant
10696 operations are affected by the digest parameter including the
10697 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10698 DSA key was used because it didn't fix the digest.
10701 *) Initial certificate chain verify code. Currently tests the untrusted
10702 certificates for consistency with the verify purpose (which is set
10703 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10705 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10706 this is because it will reject chains with invalid extensions whereas
10707 every previous version of OpenSSL and SSLeay made no checks at all.
10709 Trust code: checks the root CA for the relevant trust settings. Trust
10710 settings have an initial value consistent with the verify purpose: e.g.
10711 if the verify purpose is for SSL client use it expects the CA to be
10712 trusted for SSL client use. However the default value can be changed to
10713 permit custom trust settings: one example of this would be to only trust
10714 certificates from a specific "secure" set of CAs.
10716 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10717 which should be used for version portability: especially since the
10718 verify structure is likely to change more often now.
10720 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10721 to set them. If not set then assume SSL clients will verify SSL servers
10724 Two new options to the verify program: -untrusted allows a set of
10725 untrusted certificates to be passed in and -purpose which sets the
10726 intended purpose of the certificate. If a purpose is set then the
10727 new chain verify code is used to check extension consistency.
10730 *) Support for the authority information access extension.
10733 *) Modify RSA and DSA PEM read routines to transparently handle
10734 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10735 public keys in a format compatible with certificate
10736 SubjectPublicKeyInfo structures. Unfortunately there were already
10737 functions called *_PublicKey_* which used various odd formats so
10738 these are retained for compatibility: however the DSA variants were
10739 never in a public release so they have been deleted. Changed dsa/rsa
10740 utilities to handle the new format: note no releases ever handled public
10741 keys so we should be OK.
10743 The primary motivation for this change is to avoid the same fiasco
10744 that dogs private keys: there are several incompatible private key
10745 formats some of which are standard and some OpenSSL specific and
10746 require various evil hacks to allow partial transparent handling and
10747 even then it doesn't work with DER formats. Given the option anything
10748 other than PKCS#8 should be dumped: but the other formats have to
10749 stay in the name of compatibility.
10751 With public keys and the benefit of hindsight one standard format
10752 is used which works with EVP_PKEY, RSA or DSA structures: though
10753 it clearly returns an error if you try to read the wrong kind of key.
10755 Added a -pubkey option to the 'x509' utility to output the public key.
10756 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10757 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10758 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10759 that do the same as the EVP_PKEY_assign_*() except they up the
10760 reference count of the added key (they don't "swallow" the
10764 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10765 CRLs would fail if the file contained no certificates or no CRLs:
10766 added a new function to read in both types and return the number
10767 read: this means that if none are read it will be an error. The
10768 DER versions of the certificate and CRL reader would always fail
10769 because it isn't possible to mix certificates and CRLs in DER format
10770 without choking one or the other routine. Changed this to just read
10771 a certificate: this is the best we can do. Also modified the code
10772 in apps/verify.c to take notice of return codes: it was previously
10773 attempting to read in certificates from NULL pointers and ignoring
10774 any errors: this is one reason why the cert and CRL reader seemed
10775 to work. It doesn't check return codes from the default certificate
10776 routines: these may well fail if the certificates aren't installed.
10779 *) Code to support otherName option in GeneralName.
10782 *) First update to verify code. Change the verify utility
10783 so it warns if it is passed a self signed certificate:
10784 for consistency with the normal behaviour. X509_verify
10785 has been modified to it will now verify a self signed
10786 certificate if *exactly* the same certificate appears
10787 in the store: it was previously impossible to trust a
10788 single self signed certificate. This means that:
10789 openssl verify ss.pem
10790 now gives a warning about a self signed certificate but
10791 openssl verify -CAfile ss.pem ss.pem
10795 *) For servers, store verify_result in SSL_SESSION data structure
10796 (and add it to external session representation).
10797 This is needed when client certificate verifications fails,
10798 but an application-provided verification callback (set by
10799 SSL_CTX_set_cert_verify_callback) allows accepting the session
10800 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10801 but returns 1): When the session is reused, we have to set
10802 ssl->verify_result to the appropriate error code to avoid
10804 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10806 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10807 case in PKCS7_dataInit() where the signed PKCS7 structure
10808 didn't contain any existing data because it was being created.
10809 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10811 *) Add a salt to the key derivation routines in enc.c. This
10812 forms the first 8 bytes of the encrypted file. Also add a
10813 -S option to allow a salt to be input on the command line.
10816 *) New function X509_cmp(). Oddly enough there wasn't a function
10817 to compare two certificates. We do this by working out the SHA1
10818 hash and comparing that. X509_cmp() will be needed by the trust
10822 *) SSL_get1_session() is like SSL_get_session(), but increments
10823 the reference count in the SSL_SESSION returned.
10824 [Geoff Thorpe <geoff@eu.c2.net>]
10826 *) Fix for 'req': it was adding a null to request attributes.
10827 Also change the X509_LOOKUP and X509_INFO code to handle
10828 certificate auxiliary information.
10831 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10835 *) Add the possibility to add extra information to the memory leak
10836 detecting output, to form tracebacks, showing from where each
10837 allocation was originated: CRYPTO_push_info("constant string") adds
10838 the string plus current file name and line number to a per-thread
10839 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10840 is like calling CYRPTO_pop_info() until the stack is empty.
10841 Also updated memory leak detection code to be multi-thread-safe.
10844 *) Add options -text and -noout to pkcs7 utility and delete the
10845 encryption options which never did anything. Update docs.
10848 *) Add options to some of the utilities to allow the pass phrase
10849 to be included on either the command line (not recommended on
10850 OSes like Unix) or read from the environment. Update the
10851 manpages and fix a few bugs.
10854 *) Add a few manpages for some of the openssl commands.
10857 *) Fix the -revoke option in ca. It was freeing up memory twice,
10858 leaking and not finding already revoked certificates.
10861 *) Extensive changes to support certificate auxiliary information.
10862 This involves the use of X509_CERT_AUX structure and X509_AUX
10863 functions. An X509_AUX function such as PEM_read_X509_AUX()
10864 can still read in a certificate file in the usual way but it
10865 will also read in any additional "auxiliary information". By
10866 doing things this way a fair degree of compatibility can be
10867 retained: existing certificates can have this information added
10868 using the new 'x509' options.
10870 Current auxiliary information includes an "alias" and some trust
10871 settings. The trust settings will ultimately be used in enhanced
10872 certificate chain verification routines: currently a certificate
10873 can only be trusted if it is self signed and then it is trusted
10877 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10878 The problem was that one of the replacement routines had not been working
10879 since SSLeay releases. For now the offending routine has been replaced
10880 with non-optimised assembler. Even so, this now gives around 95%
10881 performance improvement for 1024 bit RSA signs.
10884 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10885 handling. Most clients have the effective key size in bits equal to
10886 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10887 A few however don't do this and instead use the size of the decrypted key
10888 to determine the RC2 key length and the AlgorithmIdentifier to determine
10889 the effective key length. In this case the effective key length can still
10890 be 40 bits but the key length can be 168 bits for example. This is fixed
10891 by manually forcing an RC2 key into the EVP_PKEY structure because the
10892 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10893 the key length and effective key length are equal.
10896 *) Add a bunch of functions that should simplify the creation of
10897 X509_NAME structures. Now you should be able to do:
10898 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10899 and have it automatically work out the correct field type and fill in
10900 the structures. The more adventurous can try:
10901 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10902 and it will (hopefully) work out the correct multibyte encoding.
10905 *) Change the 'req' utility to use the new field handling and multibyte
10906 copy routines. Before the DN field creation was handled in an ad hoc
10907 way in req, ca, and x509 which was rather broken and didn't support
10908 BMPStrings or UTF8Strings. Since some software doesn't implement
10909 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10910 using the dirstring_type option. See the new comment in the default
10911 openssl.cnf for more info.
10914 *) Make crypto/rand/md_rand.c more robust:
10915 - Assure unique random numbers after fork().
10916 - Make sure that concurrent threads access the global counter and
10917 md serializably so that we never lose entropy in them
10918 or use exactly the same state in multiple threads.
10919 Access to the large state is not always serializable because
10920 the additional locking could be a performance killer, and
10921 md should be large enough anyway.
10924 *) New file apps/app_rand.c with commonly needed functionality
10925 for handling the random seed file.
10927 Use the random seed file in some applications that previously did not:
10929 dsaparam -genkey (which also ignored its '-rand' option),
10932 x509 (when signing).
10933 Except on systems with /dev/urandom, it is crucial to have a random
10934 seed file at least for key creation, DSA signing, and for DH exchanges;
10935 for RSA signatures we could do without one.
10937 gendh and gendsa (unlike genrsa) used to read only the first byte
10938 of each file listed in the '-rand' option. The function as previously
10939 found in genrsa is now in app_rand.c and is used by all programs
10940 that support '-rand'.
10943 *) In RAND_write_file, use mode 0600 for creating files;
10944 don't just chmod when it may be too late.
10947 *) Report an error from X509_STORE_load_locations
10948 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10951 *) New function ASN1_mbstring_copy() this copies a string in either
10952 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10953 into an ASN1_STRING type. A mask of permissible types is passed
10954 and it chooses the "minimal" type to use or an error if not type
10958 *) Add function equivalents to the various macros in asn1.h. The old
10959 macros are retained with an M_ prefix. Code inside the library can
10960 use the M_ macros. External code (including the openssl utility)
10961 should *NOT* in order to be "shared library friendly".
10964 *) Add various functions that can check a certificate's extensions
10965 to see if it usable for various purposes such as SSL client,
10966 server or S/MIME and CAs of these types. This is currently
10967 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10968 verification. Also added a -purpose flag to x509 utility to
10969 print out all the purposes.
10972 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10976 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10977 for, obtain and decode and extension and obtain its critical flag.
10978 This allows all the necessary extension code to be handled in a
10979 single function call.
10982 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10983 platforms. See crypto/rc4/rc4_enc.c for further details.
10986 *) New -noout option to asn1parse. This causes no output to be produced
10987 its main use is when combined with -strparse and -out to extract data
10988 from a file (which may not be in ASN.1 format).
10991 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10992 when producing the local key id.
10993 [Richard Levitte <levitte@stacken.kth.se>]
10995 *) New option -dhparam in s_server. This allows a DH parameter file to be
10996 stated explicitly. If it is not stated then it tries the first server
10997 certificate file. The previous behaviour hard coded the filename
11001 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11002 a public key to be input or output. For example:
11003 openssl rsa -in key.pem -pubout -out pubkey.pem
11004 Also added necessary DSA public key functions to handle this.
11007 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11008 in the message. This was handled by allowing
11009 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11010 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11012 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11013 to the end of the strings whereas this didn't. This would cause problems
11014 if strings read with d2i_ASN1_bytes() were later modified.
11015 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11017 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11018 data and it contains EOF it will end up returning an error. This is
11019 caused by input 46 bytes long. The cause is due to the way base64
11020 BIOs find the start of base64 encoded data. They do this by trying a
11021 trial decode on each line until they find one that works. When they
11022 do a flag is set and it starts again knowing it can pass all the
11023 data directly through the decoder. Unfortunately it doesn't reset
11024 the context it uses. This means that if EOF is reached an attempt
11025 is made to pass two EOFs through the context and this causes the
11026 resulting error. This can also cause other problems as well. As is
11027 usual with these problems it takes *ages* to find and the fix is
11028 trivial: move one line.
11029 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11031 *) Ugly workaround to get s_client and s_server working under Windows. The
11032 old code wouldn't work because it needed to select() on sockets and the
11033 tty (for keypresses and to see if data could be written). Win32 only
11034 supports select() on sockets so we select() with a 1s timeout on the
11035 sockets and then see if any characters are waiting to be read, if none
11036 are present then we retry, we also assume we can always write data to
11037 the tty. This isn't nice because the code then blocks until we've
11038 received a complete line of data and it is effectively polling the
11039 keyboard at 1s intervals: however it's quite a bit better than not
11040 working at all :-) A dedicated Windows application might handle this
11041 with an event loop for example.
11044 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11045 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11046 will be called when RSA_sign() and RSA_verify() are used. This is useful
11047 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11048 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11049 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11050 This necessitated the support of an extra signature type NID_md5_sha1
11051 for SSL signatures and modifications to the SSL library to use it instead
11052 of calling RSA_public_decrypt() and RSA_private_encrypt().
11055 *) Add new -verify -CAfile and -CApath options to the crl program, these
11056 will lookup a CRL issuers certificate and verify the signature in a
11057 similar way to the verify program. Tidy up the crl program so it
11058 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11059 less strict. It will now permit CRL extensions even if it is not
11060 a V2 CRL: this will allow it to tolerate some broken CRLs.
11063 *) Initialize all non-automatic variables each time one of the openssl
11064 sub-programs is started (this is necessary as they may be started
11065 multiple times from the "OpenSSL>" prompt).
11066 [Lennart Bang, Bodo Moeller]
11068 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11069 removing all other RSA functionality (this is what NO_RSA does). This
11070 is so (for example) those in the US can disable those operations covered
11071 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11075 *) Non-copying interface to BIO pairs.
11076 (still largely untested)
11079 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11080 ASCII string. This was handled independently in various places before.
11083 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11084 UTF8 strings a character at a time.
11087 *) Use client_version from client hello to select the protocol
11088 (s23_srvr.c) and for RSA client key exchange verification
11089 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11092 *) Add various utility functions to handle SPKACs, these were previously
11093 handled by poking round in the structure internals. Added new function
11094 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11095 print, verify and generate SPKACs. Based on an original idea from
11096 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11099 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11102 *) Allow the config file extension section to be overwritten on the
11103 command line. Based on an original idea from Massimiliano Pala
11104 <madwolf@comune.modena.it>. The new option is called -extensions
11105 and can be applied to ca, req and x509. Also -reqexts to override
11106 the request extensions in req and -crlexts to override the crl extensions
11110 *) Add new feature to the SPKAC handling in ca. Now you can include
11111 the same field multiple times by preceding it by "XXXX." for example:
11114 this is the same syntax as used in the req config file.
11117 *) Allow certificate extensions to be added to certificate requests. These
11118 are specified in a 'req_extensions' option of the req section of the
11119 config file. They can be printed out with the -text option to req but
11120 are otherwise ignored at present.
11123 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11124 data read consists of only the final block it would not decrypted because
11125 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11126 A misplaced 'break' also meant the decrypted final block might not be
11127 copied until the next read.
11130 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11131 a few extra parameters to the DH structure: these will be useful if
11132 for example we want the value of 'q' or implement X9.42 DH.
11135 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11136 provides hooks that allow the default DSA functions or functions on a
11137 "per key" basis to be replaced. This allows hardware acceleration and
11138 hardware key storage to be handled without major modification to the
11139 library. Also added low level modexp hooks and CRYPTO_EX structure and
11140 associated functions.
11143 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11144 as "read only": it can't be written to and the buffer it points to will
11145 not be freed. Reading from a read only BIO is much more efficient than
11146 a normal memory BIO. This was added because there are several times when
11147 an area of memory needs to be read from a BIO. The previous method was
11148 to create a memory BIO and write the data to it, this results in two
11149 copies of the data and an O(n^2) reading algorithm. There is a new
11150 function BIO_new_mem_buf() which creates a read only memory BIO from
11151 an area of memory. Also modified the PKCS#7 routines to use read only
11155 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11156 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11157 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11158 but a retry condition occurred while trying to read the rest.
11161 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11162 NID_pkcs7_encrypted by default: this was wrong since this should almost
11163 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11164 the encrypted data type: this is a more sensible place to put it and it
11165 allows the PKCS#12 code to be tidied up that duplicated this
11169 *) Changed obj_dat.pl script so it takes its input and output files on
11170 the command line. This should avoid shell escape redirection problems
11174 *) Initial support for certificate extension requests, these are included
11175 in things like Xenroll certificate requests. Included functions to allow
11176 extensions to be obtained and added.
11179 *) -crlf option to s_client and s_server for sending newlines as
11180 CRLF (as required by many protocols).
11183 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11185 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11186 [Ralf S. Engelschall]
11188 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11189 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11191 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11195 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11196 DH parameters/keys (q is lost during that conversion, but the resulting
11197 DH parameters contain its length).
11199 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11200 much faster than DH_generate_parameters (which creates parameters
11201 where p = 2*q + 1), and also the smaller q makes DH computations
11202 much more efficient (160-bit exponentiation instead of 1024-bit
11203 exponentiation); so this provides a convenient way to support DHE
11204 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11205 utter importance to use
11206 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11208 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11209 when such DH parameters are used, because otherwise small subgroup
11210 attacks may become possible!
11213 *) Avoid memory leak in i2d_DHparams.
11216 *) Allow the -k option to be used more than once in the enc program:
11217 this allows the same encrypted message to be read by multiple recipients.
11220 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11221 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11222 it will always use the numerical form of the OID, even if it has a short
11226 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11227 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11228 otherwise bn_mod_exp was called. In the case of hardware keys for example
11229 no private key components need be present and it might store extra data
11230 in the RSA structure, which cannot be accessed from bn_mod_exp.
11231 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11232 private key operations.
11235 *) Added support for SPARC Linux.
11238 *) pem_password_cb function type incompatibly changed from
11239 typedef int pem_password_cb(char *buf, int size, int rwflag);
11241 ....(char *buf, int size, int rwflag, void *userdata);
11242 so that applications can pass data to their callbacks:
11243 The PEM[_ASN1]_{read,write}... functions and macros now take an
11244 additional void * argument, which is just handed through whenever
11245 the password callback is called.
11246 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11248 New function SSL_CTX_set_default_passwd_cb_userdata.
11250 Compatibility note: As many C implementations push function arguments
11251 onto the stack in reverse order, the new library version is likely to
11252 interoperate with programs that have been compiled with the old
11253 pem_password_cb definition (PEM_whatever takes some data that
11254 happens to be on the stack as its last argument, and the callback
11255 just ignores this garbage); but there is no guarantee whatsoever that
11258 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11259 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11260 problems not only on Windows, but also on some Unix platforms.
11261 To avoid problematic command lines, these definitions are now in an
11262 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11263 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11266 *) MIPS III/IV assembler module is reimplemented.
11269 *) More DES library cleanups: remove references to srand/rand and
11270 delete an unused file.
11273 *) Add support for the free Netwide assembler (NASM) under Win32,
11274 since not many people have MASM (ml) and it can be hard to obtain.
11275 This is currently experimental but it seems to work OK and pass all
11276 the tests. Check out INSTALL.W32 for info.
11279 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11280 without temporary keys kept an extra copy of the server key,
11281 and connections with temporary keys did not free everything in case
11285 *) New function RSA_check_key and new openssl rsa option -check
11286 for verifying the consistency of RSA keys.
11287 [Ulf Moeller, Bodo Moeller]
11289 *) Various changes to make Win32 compile work:
11290 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11291 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11292 comparison" warnings.
11293 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11296 *) Add a debugging option to PKCS#5 v2 key generation function: when
11297 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11298 derived keys are printed to stderr.
11301 *) Copy the flags in ASN1_STRING_dup().
11302 [Roman E. Pavlov <pre@mo.msk.ru>]
11304 *) The x509 application mishandled signing requests containing DSA
11305 keys when the signing key was also DSA and the parameters didn't match.
11307 It was supposed to omit the parameters when they matched the signing key:
11308 the verifying software was then supposed to automatically use the CA's
11309 parameters if they were absent from the end user certificate.
11311 Omitting parameters is no longer recommended. The test was also
11312 the wrong way round! This was probably due to unusual behaviour in
11313 EVP_cmp_parameters() which returns 1 if the parameters match.
11314 This meant that parameters were omitted when they *didn't* match and
11315 the certificate was useless. Certificates signed with 'ca' didn't have
11317 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11319 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11320 The interface is as follows:
11321 Applications can use
11322 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11323 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11324 "off" is now the default.
11325 The library internally uses
11326 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11327 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11328 to disable memory-checking temporarily.
11330 Some inconsistent states that previously were possible (and were
11331 even the default) are now avoided.
11333 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11334 with each memory chunk allocated; this is occasionally more helpful
11335 than just having a counter.
11337 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11339 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11343 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11344 which largely parallels "options", but is for changing API behaviour,
11345 whereas "options" are about protocol behaviour.
11346 Initial "mode" flags are:
11348 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11349 a single record has been written.
11350 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11351 retries use the same buffer location.
11352 (But all of the contents must be
11356 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11359 *) Fix problems with no-hmac etc.
11360 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11362 *) New functions RSA_get_default_method(), RSA_set_method() and
11363 RSA_get_method(). These allows replacement of RSA_METHODs without having
11364 to mess around with the internals of an RSA structure.
11367 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11368 Also really enable memory leak checks in openssl.c and in some
11370 [Chad C. Mulligan, Bodo Moeller]
11372 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11373 up the length of negative integers. This has now been simplified to just
11374 store the length when it is first determined and use it later, rather
11375 than trying to keep track of where data is copied and updating it to
11377 [Steve Henson, reported by Brien Wheeler
11378 <bwheeler@authentica-security.com>]
11380 *) Add a new function PKCS7_signatureVerify. This allows the verification
11381 of a PKCS#7 signature but with the signing certificate passed to the
11382 function itself. This contrasts with PKCS7_dataVerify which assumes the
11383 certificate is present in the PKCS#7 structure. This isn't always the
11384 case: certificates can be omitted from a PKCS#7 structure and be
11385 distributed by "out of band" means (such as a certificate database).
11388 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11389 function prototypes in pem.h, also change util/mkdef.pl to add the
11390 necessary function names.
11393 *) mk1mf.pl (used by Windows builds) did not properly read the
11394 options set by Configure in the top level Makefile, and Configure
11395 was not even able to write more than one option correctly.
11396 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11399 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11400 file to be loaded from a BIO or FILE pointer. The BIO version will
11401 for example allow memory BIOs to contain config info.
11404 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11405 Whoever hopes to achieve shared-library compatibility across versions
11406 must use this, not the compile-time macro.
11407 (Exercise 0.9.4: Which is the minimum library version required by
11409 Note: All this applies only to multi-threaded programs, others don't
11413 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11414 through a BIO pair triggered the default case, i.e.
11415 SSLerr(...,SSL_R_UNKNOWN_STATE).
11418 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11419 can use the SSL library even if none of the specific BIOs is
11423 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11424 for the encoded length.
11425 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11427 *) Add initial documentation of the X509V3 functions.
11430 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11431 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11432 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11433 secure PKCS#8 private key format with a high iteration count.
11436 *) Fix determination of Perl interpreter: A perl or perl5
11437 _directory_ in $PATH was also accepted as the interpreter.
11438 [Ralf S. Engelschall]
11440 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11441 wrong with it but it was very old and did things like calling
11442 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11443 unusual formatting.
11446 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11447 to use the new extension code.
11450 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11451 with macros. This should make it easier to change their form, add extra
11452 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11456 *) Add to configuration table a new entry that can specify an alternative
11457 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11458 according to Mark Crispin <MRC@Panda.COM>.
11462 *) DES CBC did not update the IV. Weird.
11465 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11466 Changing the behaviour of the former might break existing programs --
11467 where IV updating is needed, des_ncbc_encrypt can be used.
11470 *) When bntest is run from "make test" it drives bc to check its
11471 calculations, as well as internally checking them. If an internal check
11472 fails, it needs to cause bc to give a non-zero result or make test carries
11473 on without noticing the failure. Fixed.
11476 *) DES library cleanups.
11479 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11480 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11481 ciphers. NOTE: although the key derivation function has been verified
11482 against some published test vectors it has not been extensively tested
11483 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11487 *) Instead of "mkdir -p", which is not fully portable, use new
11488 Perl script "util/mkdir-p.pl".
11491 *) Rewrite the way password based encryption (PBE) is handled. It used to
11492 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11493 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11494 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11495 the 'parameter' field of the AlgorithmIdentifier is passed to the
11496 underlying key generation function so it must do its own ASN1 parsing.
11497 This has also changed the EVP_PBE_CipherInit() function which now has a
11498 'parameter' argument instead of literal salt and iteration count values
11499 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11502 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11503 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11504 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11505 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11506 value was just used as a "magic string" and not used directly its
11507 value doesn't matter.
11510 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11514 *) "linux-sparc64" configuration (ultrapenguin).
11515 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11516 "linux-sparc" configuration.
11517 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11519 *) config now generates no-xxx options for missing ciphers.
11522 *) Support the EBCDIC character set (work in progress).
11523 File ebcdic.c not yet included because it has a different license.
11524 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11526 *) Support BS2000/OSD-POSIX.
11527 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11529 *) Make callbacks for key generation use void * instead of char *.
11532 *) Make S/MIME samples compile (not yet tested).
11535 *) Additional typesafe stacks.
11538 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11542 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11544 *) New configuration variant "sco5-gcc".
11546 *) Updated some demos.
11547 [Sean O Riordain, Wade Scholine]
11549 *) Add missing BIO_free at exit of pkcs12 application.
11552 *) Fix memory leak in conf.c.
11555 *) Updates for Win32 to assembler version of MD5.
11558 *) Set #! path to perl in apps/der_chop to where we found it
11559 instead of using a fixed path.
11562 *) SHA library changes for irix64-mips4-cc.
11565 *) Improvements for VMS support.
11569 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11571 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11572 This also avoids the problems with SC4.2 and unpatched SC5.
11573 [Andy Polyakov <appro@fy.chalmers.se>]
11575 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11576 These are required because of the typesafe stack would otherwise break
11577 existing code. If old code used a structure member which used to be STACK
11578 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11579 sk_num or sk_value it would produce an error because the num, data members
11580 are not present in STACK_OF. Now it just produces a warning. sk_set
11581 replaces the old method of assigning a value to sk_value
11582 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11583 that does this will no longer work (and should use sk_set instead) but
11584 this could be regarded as a "questionable" behaviour anyway.
11587 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11588 correctly handle encrypted S/MIME data.
11591 *) Change type of various DES function arguments from des_cblock
11592 (which means, in function argument declarations, pointer to char)
11593 to des_cblock * (meaning pointer to array with 8 char elements),
11594 which allows the compiler to do more typechecking; it was like
11595 that back in SSLeay, but with lots of ugly casts.
11597 Introduce new type const_des_cblock.
11600 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11601 problems: find RecipientInfo structure that matches recipient certificate
11602 and initialise the ASN1 structures properly based on passed cipher.
11605 *) Belatedly make the BN tests actually check the results.
11608 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11609 to and from BNs: it was completely broken. New compilation option
11610 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11611 key elements as negative integers.
11614 *) Reorganize and speed up MD5.
11615 [Andy Polyakov <appro@fy.chalmers.se>]
11618 [Richard Levitte <richard@levitte.org>]
11620 *) New option -out to asn1parse to allow the parsed structure to be
11621 output to a file. This is most useful when combined with the -strparse
11622 option to examine the output of things like OCTET STRINGS.
11625 *) Make SSL library a little more fool-proof by not requiring any longer
11626 that SSL_set_{accept,connect}_state be called before
11627 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11628 in many applications because usually everything *appeared* to work as
11629 intended anyway -- now it really works as intended).
11632 *) Move openssl.cnf out of lib/.
11635 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11636 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11637 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11638 [Ralf S. Engelschall]
11640 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11641 handle PKCS#7 enveloped data properly.
11642 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11644 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11645 copying pointers. The cert_st handling is changed by this in
11646 various ways (and thus what used to be known as ctx->default_cert
11647 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11648 any longer when s->cert does not give us what we need).
11649 ssl_cert_instantiate becomes obsolete by this change.
11650 As soon as we've got the new code right (possibly it already is?),
11651 we have solved a couple of bugs of the earlier code where s->cert
11652 was used as if it could not have been shared with other SSL structures.
11654 Note that using the SSL API in certain dirty ways now will result
11655 in different behaviour than observed with earlier library versions:
11656 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11657 does not influence s as it used to.
11659 In order to clean up things more thoroughly, inside SSL_SESSION
11660 we don't use CERT any longer, but a new structure SESS_CERT
11661 that holds per-session data (if available); currently, this is
11662 the peer's certificate chain and, for clients, the server's certificate
11663 and temporary key. CERT holds only those values that can have
11664 meaningful defaults in an SSL_CTX.
11667 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11668 from the internal representation. Various PKCS#7 fixes: remove some
11669 evil casts and set the enc_dig_alg field properly based on the signing
11673 *) Allow PKCS#12 password to be set from the command line or the
11674 environment. Let 'ca' get its config file name from the environment
11675 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11679 *) Allow certificate policies extension to use an IA5STRING for the
11680 organization field. This is contrary to the PKIX definition but
11681 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11685 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11686 without disallowing inline assembler and the like for non-pedantic builds.
11689 *) Support Borland C++ builder.
11690 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11692 *) Support Mingw32.
11695 *) SHA-1 cleanups and performance enhancements.
11696 [Andy Polyakov <appro@fy.chalmers.se>]
11698 *) Sparc v8plus assembler for the bignum library.
11699 [Andy Polyakov <appro@fy.chalmers.se>]
11701 *) Accept any -xxx and +xxx compiler options in Configure.
11704 *) Update HPUX configuration.
11707 *) Add missing sk_<type>_unshift() function to safestack.h
11708 [Ralf S. Engelschall]
11710 *) New function SSL_CTX_use_certificate_chain_file that sets the
11711 "extra_cert"s in addition to the certificate. (This makes sense
11712 only for "PEM" format files, as chains as a whole are not
11716 *) Support verify_depth from the SSL API.
11717 x509_vfy.c had what can be considered an off-by-one-error:
11718 Its depth (which was not part of the external interface)
11719 was actually counting the number of certificates in a chain;
11720 now it really counts the depth.
11723 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11724 instead of X509err, which often resulted in confusing error
11725 messages since the error codes are not globally unique
11726 (e.g. an alleged error in ssl3_accept when a certificate
11727 didn't match the private key).
11729 *) New function SSL_CTX_set_session_id_context that allows to set a default
11730 value (so that you don't need SSL_set_session_id_context for each
11731 connection using the SSL_CTX).
11734 *) OAEP decoding bug fix.
11737 *) Support INSTALL_PREFIX for package builders, as proposed by
11741 *) New Configure options "threads" and "no-threads". For systems
11742 where the proper compiler options are known (currently Solaris
11743 and Linux), "threads" is the default.
11746 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11749 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11750 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11751 such as /usr/local/bin.
11754 *) "make linux-shared" to build shared libraries.
11755 [Niels Poppe <niels@netbox.org>]
11757 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11760 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11761 extension adding in x509 utility.
11764 *) Remove NOPROTO sections and error code comments.
11767 *) Partial rewrite of the DEF file generator to now parse the ANSI
11771 *) New Configure options --prefix=DIR and --openssldir=DIR.
11774 *) Complete rewrite of the error code script(s). It is all now handled
11775 by one script at the top level which handles error code gathering,
11776 header rewriting and C source file generation. It should be much better
11777 than the old method: it now uses a modified version of Ulf's parser to
11778 read the ANSI prototypes in all header files (thus the old K&R definitions
11779 aren't needed for error creation any more) and do a better job of
11780 translating function codes into names. The old 'ASN1 error code imbedded
11781 in a comment' is no longer necessary and it doesn't use .err files which
11782 have now been deleted. Also the error code call doesn't have to appear all
11783 on one line (which resulted in some large lines...).
11786 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11789 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11790 0 (which usually indicates a closed connection), but continue reading.
11793 *) Fix some race conditions.
11796 *) Add support for CRL distribution points extension. Add Certificate
11797 Policies and CRL distribution points documentation.
11800 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11803 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11804 8 of keying material. Merlin has also confirmed interop with this fix
11805 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11806 [Merlin Hughes <merlin@baltimore.ie>]
11808 *) Fix lots of warnings.
11809 [Richard Levitte <levitte@stacken.kth.se>]
11811 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11812 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11813 [Richard Levitte <levitte@stacken.kth.se>]
11815 *) Fix problems with sizeof(long) == 8.
11816 [Andy Polyakov <appro@fy.chalmers.se>]
11818 *) Change functions to ANSI C.
11821 *) Fix typos in error codes.
11822 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11824 *) Remove defunct assembler files from Configure.
11827 *) SPARC v8 assembler BIGNUM implementation.
11828 [Andy Polyakov <appro@fy.chalmers.se>]
11830 *) Support for Certificate Policies extension: both print and set.
11831 Various additions to support the r2i method this uses.
11834 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11835 return a const string when you are expecting an allocated buffer.
11838 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11839 types DirectoryString and DisplayText.
11842 *) Add code to allow r2i extensions to access the configuration database,
11843 add an LHASH database driver and add several ctx helper functions.
11846 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11847 fail when they extended the size of a BIGNUM.
11850 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11851 support typesafe stack.
11854 *) Fix typo in SSL_[gs]et_options().
11855 [Nils Frostberg <nils@medcom.se>]
11857 *) Delete various functions and files that belonged to the (now obsolete)
11858 old X509V3 handling code.
11861 *) New Configure option "rsaref".
11864 *) Don't auto-generate pem.h.
11867 *) Introduce type-safe ASN.1 SETs.
11870 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11871 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11873 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11874 that links with OpenSSL (well at least cause lots of warnings), but fear
11875 not: the conversion is trivial, and it eliminates loads of evil casts. A
11876 few STACKed things have been converted already. Feel free to convert more.
11877 In the fullness of time, I'll do away with the STACK type altogether.
11880 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11881 specified in <certfile> by updating the entry in the index.txt file.
11882 This way one no longer has to edit the index.txt file manually for
11883 revoking a certificate. The -revoke option does the gory details now.
11884 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11886 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11887 `-text' option at all and this way the `-noout -text' combination was
11888 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11889 [Ralf S. Engelschall]
11891 *) Make sure a corresponding plain text error message exists for the
11892 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11893 verify callback function determined that a certificate was revoked.
11894 [Ralf S. Engelschall]
11896 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11897 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11898 all available ciphers including rc5, which was forgotten until now.
11899 In order to let the testing shell script know which algorithms
11900 are available, a new (up to now undocumented) command
11901 "openssl list-cipher-commands" is used.
11904 *) Bugfix: s_client occasionally would sleep in select() when
11905 it should have checked SSL_pending() first.
11908 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11909 the raw DSA values prior to ASN.1 encoding.
11912 *) Tweaks to Configure
11913 [Niels Poppe <niels@netbox.org>]
11915 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11919 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11922 *) New config option to avoid instructions that are illegal on the 80386.
11923 The default code is faster, but requires at least a 486.
11926 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11927 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11928 same as SSL2_VERSION anyway.
11931 *) New "-showcerts" option for s_client.
11934 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11935 application. Various cleanups and fixes.
11938 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11939 modify error routines to work internally. Add error codes and PBE init
11940 to library startup routines.
11943 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11944 packing functions to asn1 and evp. Changed function names and error
11945 codes along the way.
11948 *) PKCS12 integration: and so it begins... First of several patches to
11949 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11950 objects to objects.h
11953 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11954 and display support for Thawte strong extranet extension.
11957 *) Add LinuxPPC support.
11958 [Jeff Dubrule <igor@pobox.org>]
11960 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11961 bn_div_words in alpha.s.
11962 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11964 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11965 OAEP isn't supported when OpenSSL is built with RSAref.
11966 [Ulf Moeller <ulf@fitug.de>]
11968 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11969 so they no longer are missing under -DNOPROTO.
11970 [Soren S. Jorvang <soren@t.dk>]
11973 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11975 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11976 doesn't work when the session is reused. Coming soon!
11979 *) Fix a security hole, that allows sessions to be reused in the wrong
11980 context thus bypassing client cert protection! All software that uses
11981 client certs and session caches in multiple contexts NEEDS PATCHING to
11982 allow session reuse! A fuller solution is in the works.
11983 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11985 *) Some more source tree cleanups (removed obsolete files
11986 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11987 permission on "config" script to be executable) and a fix for the INSTALL
11989 [Ulf Moeller <ulf@fitug.de>]
11991 *) Remove some legacy and erroneous uses of malloc, free instead of
11993 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11995 *) Make rsa_oaep_test return non-zero on error.
11996 [Ulf Moeller <ulf@fitug.de>]
11998 *) Add support for native Solaris shared libraries. Configure
11999 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12000 if someone would make that last step automatic.
12001 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12003 *) ctx_size was not built with the right compiler during "make links". Fixed.
12006 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12007 except NULL ciphers". This means the default cipher list will no longer
12008 enable NULL ciphers. They need to be specifically enabled e.g. with
12009 the string "DEFAULT:eNULL".
12012 *) Fix to RSA private encryption routines: if p < q then it would
12013 occasionally produce an invalid result. This will only happen with
12014 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12017 *) Be less restrictive and allow also `perl util/perlpath.pl
12018 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12019 because this way one can also use an interpreter named `perl5' (which is
12020 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12021 installed as `perl').
12022 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12024 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12025 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12027 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12028 advapi32.lib to Win32 build and change the pem test comparison
12029 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12030 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12031 and crypto/des/ede_cbcm_enc.c.
12034 *) DES quad checksum was broken on big-endian architectures. Fixed.
12037 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12038 Win32 test batch file so it (might) work again. The Win32 test batch file
12039 is horrible: I feel ill....
12042 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12043 in e_os.h. Audit of header files to check ANSI and non ANSI
12044 sections: 10 functions were absent from non ANSI section and not exported
12045 from Windows DLLs. Fixed up libeay.num for new functions.
12048 *) Make `openssl version' output lines consistent.
12049 [Ralf S. Engelschall]
12051 *) Fix Win32 symbol export lists for BIO functions: Added
12052 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12053 to ms/libeay{16,32}.def.
12054 [Ralf S. Engelschall]
12056 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12057 fine under Unix and passes some trivial tests I've now added. But the
12058 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12059 added to make sure no one expects that this stuff really works in the
12060 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12061 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12063 [Ralf S. Engelschall]
12065 *) Fix the generation of two part addresses in perl.
12066 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12068 *) Add config entry for Linux on MIPS.
12069 [John Tobey <jtobey@channel1.com>]
12071 *) Make links whenever Configure is run, unless we are on Windoze.
12074 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12075 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12079 *) Add a useful kludge to allow package maintainers to specify compiler and
12080 other platforms details on the command line without having to patch the
12081 Configure script everytime: One now can use ``perl Configure
12082 <id>:<details>'', i.e. platform ids are allowed to have details appended
12083 to them (separated by colons). This is treated as there would be a static
12084 pre-configured entry in Configure's %table under key <id> with value
12085 <details> and ``perl Configure <id>'' is called. So, when you want to
12086 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12087 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12088 now, which overrides the FreeBSD-elf entry on-the-fly.
12089 [Ralf S. Engelschall]
12091 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12094 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12095 on the `perl Configure ...' command line. This way one can compile
12096 OpenSSL libraries with Position Independent Code (PIC) which is needed
12097 for linking it into DSOs.
12098 [Ralf S. Engelschall]
12100 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12104 *) Cleaned up the LICENSE document: The official contact for any license
12105 questions now is the OpenSSL core team under openssl-core@openssl.org.
12106 And add a paragraph about the dual-license situation to make sure people
12107 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12108 to the OpenSSL toolkit.
12109 [Ralf S. Engelschall]
12111 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12112 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12113 Additionally cleaned up the `make links' target: Remove unnecessary
12114 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12115 to speed processing and no longer clutter the display with confusing
12116 stuff. Instead only the actually done links are displayed.
12117 [Ralf S. Engelschall]
12119 *) Permit null encryption ciphersuites, used for authentication only. It used
12120 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12121 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12125 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12126 signed attributes when verifying signatures (this would break them),
12127 the detached data encoding was wrong and public keys obtained using
12128 X509_get_pubkey() weren't freed.
12131 *) Add text documentation for the BUFFER functions. Also added a work around
12132 to a Win95 console bug. This was triggered by the password read stuff: the
12133 last character typed gets carried over to the next fread(). If you were
12134 generating a new cert request using 'req' for example then the last
12135 character of the passphrase would be CR which would then enter the first
12139 *) Added the new `Includes OpenSSL Cryptography Software' button as
12140 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12141 button and can be used by applications based on OpenSSL to show the
12142 relationship to the OpenSSL project.
12143 [Ralf S. Engelschall]
12145 *) Remove confusing variables in function signatures in files
12146 ssl/ssl_lib.c and ssl/ssl.h.
12147 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12149 *) Don't install bss_file.c under PREFIX/include/
12150 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12152 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12153 functions that return function pointers and has support for NT specific
12154 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12155 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12156 unsigned to signed types: this was killing the Win32 compile.
12159 *) Add new certificate file to stack functions,
12160 SSL_add_dir_cert_subjects_to_stack() and
12161 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12162 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12163 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12164 This means that Apache-SSL and similar packages don't have to mess around
12165 to add as many CAs as they want to the preferred list.
12168 *) Experiment with doxygen documentation. Currently only partially applied to
12170 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12171 openssl.doxy as the configuration file.
12174 *) Get rid of remaining C++-style comments which strict C compilers hate.
12175 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12177 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12178 compiled in by default: it has problems with large keys.
12181 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12182 DH private keys and/or callback functions which directly correspond to
12183 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12184 is needed for applications which have to configure certificates on a
12185 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12187 For the RSA certificate situation is makes no difference, but
12188 for the DSA certificate situation this fixes the "no shared cipher"
12189 problem where the OpenSSL cipher selection procedure failed because the
12190 temporary keys were not overtaken from the context and the API provided
12191 no way to reconfigure them.
12192 The new functions now let applications reconfigure the stuff and they
12193 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12194 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12195 non-public-API function ssl_cert_instantiate() is used as a helper
12196 function and also to reduce code redundancy inside ssl_rsa.c.
12197 [Ralf S. Engelschall]
12199 *) Move s_server -dcert and -dkey options out of the undocumented feature
12200 area because they are useful for the DSA situation and should be
12201 recognized by the users.
12202 [Ralf S. Engelschall]
12204 *) Fix the cipher decision scheme for export ciphers: the export bits are
12205 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12206 SSL_EXP_MASK. So, the original variable has to be used instead of the
12207 already masked variable.
12208 [Richard Levitte <levitte@stacken.kth.se>]
12210 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12211 [Richard Levitte <levitte@stacken.kth.se>]
12213 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12214 from `int' to `unsigned int' because it's a length and initialized by
12215 EVP_DigestFinal() which expects an `unsigned int *'.
12216 [Richard Levitte <levitte@stacken.kth.se>]
12218 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12219 script. Instead use the usual Shell->Perl transition trick.
12220 [Ralf S. Engelschall]
12222 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12223 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12224 -noout -modulus' as it's already the case for `openssl rsa -noout
12225 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12226 currently the public key is printed (a decision which was already done by
12227 `openssl dsa -modulus' in the past) which serves a similar purpose.
12228 Additionally the NO_RSA no longer completely removes the whole -modulus
12229 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12231 [Ralf S. Engelschall]
12233 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12234 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12235 [Arne Ansper <arne@ats.cyber.ee>]
12237 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12238 to be added. Now both 'req' and 'ca' can use new objects defined in the
12242 *) Add cool BIO that does syslog (or event log on NT).
12243 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12245 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12246 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12247 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12248 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12251 *) Add preliminary config info for new extension code.
12254 *) Make RSA_NO_PADDING really use no padding.
12255 [Ulf Moeller <ulf@fitug.de>]
12257 *) Generate errors when private/public key check is done.
12260 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12261 for some CRL extensions and new objects added.
12264 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12265 key usage extension and fuller support for authority key id.
12268 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12269 padding method for RSA, which is recommended for new applications in PKCS
12270 #1 v2.0 (RFC 2437, October 1998).
12271 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12272 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12273 against Bleichbacher's attack on RSA.
12274 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12277 *) Updates to the new SSL compression code
12278 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12280 *) Fix so that the version number in the master secret, when passed
12281 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12282 (because the server will not accept higher), that the version number
12283 is 0x03,0x01, not 0x03,0x00
12284 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12286 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12287 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12288 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12291 *) Support for RAW extensions where an arbitrary extension can be
12292 created by including its DER encoding. See apps/openssl.cnf for
12296 *) Make sure latest Perl versions don't interpret some generated C array
12297 code as Perl array code in the crypto/err/err_genc.pl script.
12298 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12300 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12301 not many people have the assembler. Various Win32 compilation fixes and
12302 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12303 build instructions.
12306 *) Modify configure script 'Configure' to automatically create crypto/date.h
12307 file under Win32 and also build pem.h from pem.org. New script
12308 util/mkfiles.pl to create the MINFO file on environments that can't do a
12309 'make files': perl util/mkfiles.pl >MINFO should work.
12312 *) Major rework of DES function declarations, in the pursuit of correctness
12313 and purity. As a result, many evil casts evaporated, and some weirdness,
12314 too. You may find this causes warnings in your code. Zapping your evil
12315 casts will probably fix them. Mostly.
12318 *) Fix for a typo in asn1.h. Bug fix to object creation script
12319 obj_dat.pl. It considered a zero in an object definition to mean
12320 "end of object": none of the objects in objects.h have any zeros
12321 so it wasn't spotted.
12322 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12324 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12325 Masking (CBCM). In the absence of test vectors, the best I have been able
12326 to do is check that the decrypt undoes the encrypt, so far. Send me test
12327 vectors if you have them.
12330 *) Correct calculation of key length for export ciphers (too much space was
12331 allocated for null ciphers). This has not been tested!
12334 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12335 message is now correct (it understands "crypto" and "ssl" on its
12336 command line). There is also now an "update" option. This will update
12337 the util/ssleay.num and util/libeay.num files with any new functions.
12339 perl util/mkdef.pl crypto ssl update
12340 it will update them.
12343 *) Overhauled the Perl interface (perl/*):
12344 - ported BN stuff to OpenSSL's different BN library
12345 - made the perl/ source tree CVS-aware
12346 - renamed the package from SSLeay to OpenSSL (the files still contain
12347 their history because I've copied them in the repository)
12348 - removed obsolete files (the test scripts will be replaced
12349 by better Test::Harness variants in the future)
12350 [Ralf S. Engelschall]
12352 *) First cut for a very conservative source tree cleanup:
12353 1. merge various obsolete readme texts into doc/ssleay.txt
12354 where we collect the old documents and readme texts.
12355 2. remove the first part of files where I'm already sure that we no
12356 longer need them because of three reasons: either they are just temporary
12357 files which were left by Eric or they are preserved original files where
12358 I've verified that the diff is also available in the CVS via "cvs diff
12359 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12360 the crypto/md/ stuff).
12361 [Ralf S. Engelschall]
12363 *) More extension code. Incomplete support for subject and issuer alt
12364 name, issuer and authority key id. Change the i2v function parameters
12365 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12366 what that's for :-) Fix to ASN1 macro which messed up
12367 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12370 *) Preliminary support for ENUMERATED type. This is largely copied from the
12374 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12375 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12377 *) Make sure `make rehash' target really finds the `openssl' program.
12378 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12380 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12381 like to hear about it if this slows down other processors.
12384 *) Add CygWin32 platform information to Configure script.
12385 [Alan Batie <batie@aahz.jf.intel.com>]
12387 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12388 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12390 *) New program nseq to manipulate netscape certificate sequences
12393 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12397 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12398 but the BN code had some problems that would cause failures when
12399 doing certificate verification and some other functions.
12400 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12402 *) Add ASN1 and PEM code to support netscape certificate sequences.
12405 *) Add ASN1 and PEM code to support netscape certificate sequences.
12408 *) Add several PKIX and private extended key usage OIDs.
12411 *) Modify the 'ca' program to handle the new extension code. Modify
12412 openssl.cnf for new extension format, add comments.
12415 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12416 and add a sample to openssl.cnf so req -x509 now adds appropriate
12420 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12421 error code, add initial support to X509_print() and x509 application.
12424 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12425 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12426 stuff is currently isolated and isn't even compiled yet.
12429 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12430 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12431 Removed the versions check from X509 routines when loading extensions:
12432 this allows certain broken certificates that don't set the version
12433 properly to be processed.
12436 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12437 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12438 can still be regenerated with "make depend".
12441 *) Spelling mistake in C version of CAST-128.
12442 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12444 *) Changes to the error generation code. The perl script err-code.pl
12445 now reads in the old error codes and retains the old numbers, only
12446 adding new ones if necessary. It also only changes the .err files if new
12447 codes are added. The makefiles have been modified to only insert errors
12448 when needed (to avoid needlessly modifying header files). This is done
12449 by only inserting errors if the .err file is newer than the auto generated
12450 C file. To rebuild all the error codes from scratch (the old behaviour)
12451 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12452 or delete all the .err files.
12455 *) CAST-128 was incorrectly implemented for short keys. The C version has
12456 been fixed, but is untested. The assembler versions are also fixed, but
12457 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12458 to regenerate it if needed.
12459 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12460 Hagino <itojun@kame.net>]
12462 *) File was opened incorrectly in randfile.c.
12463 [Ulf Möller <ulf@fitug.de>]
12465 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12466 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12467 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12468 al: it's just almost always a UTCTime. Note this patch adds new error
12469 codes so do a "make errors" if there are problems.
12472 *) Correct Linux 1 recognition in config.
12473 [Ulf Möller <ulf@fitug.de>]
12475 *) Remove pointless MD5 hash when using DSA keys in ca.
12476 [Anonymous <nobody@replay.com>]
12478 *) Generate an error if given an empty string as a cert directory. Also
12479 generate an error if handed NULL (previously returned 0 to indicate an
12480 error, but didn't set one).
12481 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12483 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12486 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12487 parameters. This was causing a warning which killed off the Win32 compile.
12490 *) Remove C++ style comments from crypto/bn/bn_local.h.
12491 [Neil Costigan <neil.costigan@celocom.com>]
12493 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12494 based on a text string, looking up short and long names and finally
12495 "dot" format. The "dot" format stuff didn't work. Added new function
12496 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12497 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12498 OID is not part of the table.
12501 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12502 X509_LOOKUP_by_alias().
12505 *) Sort openssl functions by name.
12508 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12509 encryption from sample DSA keys (in case anyone is interested the password
12513 *) Make _all_ *_free functions accept a NULL pointer.
12514 [Frans Heymans <fheymans@isaserver.be>]
12516 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12518 [Anonymous <nobody@replay.com>]
12520 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12521 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12523 *) Don't blow it for numeric -newkey arguments to apps/req.
12524 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12526 *) Temp key "for export" tests were wrong in s3_srvr.c.
12527 [Anonymous <nobody@replay.com>]
12529 *) Add prototype for temp key callback functions
12530 SSL_CTX_set_tmp_{rsa,dh}_callback().
12533 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12534 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12537 *) X509_name_add_entry() freed the wrong thing after an error.
12538 [Arne Ansper <arne@ats.cyber.ee>]
12540 *) rsa_eay.c would attempt to free a NULL context.
12541 [Arne Ansper <arne@ats.cyber.ee>]
12543 *) BIO_s_socket() had a broken should_retry() on Windoze.
12544 [Arne Ansper <arne@ats.cyber.ee>]
12546 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12547 [Arne Ansper <arne@ats.cyber.ee>]
12549 *) Make sure the already existing X509_STORE->depth variable is initialized
12550 in X509_STORE_new(), but document the fact that this variable is still
12551 unused in the certificate verification process.
12552 [Ralf S. Engelschall]
12554 *) Fix the various library and apps files to free up pkeys obtained from
12555 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12558 *) Fix reference counting in X509_PUBKEY_get(). This makes
12559 demos/maurice/example2.c work, amongst others, probably.
12560 [Steve Henson and Ben Laurie]
12562 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12563 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12564 are no longer created. This way we have a single and consistent command
12565 line interface `openssl <command>', similar to `cvs <command>'.
12566 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12568 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12569 BIT STRING wrapper always have zero unused bits.
12572 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12575 *) Make the top-level INSTALL documentation easier to understand.
12578 *) Makefiles updated to exit if an error occurs in a sub-directory
12579 make (including if user presses ^C) [Paul Sutton]
12581 *) Make Montgomery context stuff explicit in RSA data structure.
12584 *) Fix build order of pem and err to allow for generated pem.h.
12587 *) Fix renumbering bug in X509_NAME_delete_entry().
12590 *) Enhanced the err-ins.pl script so it makes the error library number
12591 global and can add a library name. This is needed for external ASN1 and
12592 other error libraries.
12595 *) Fixed sk_insert which never worked properly.
12598 *) Fix ASN1 macros so they can handle indefinite length constructed
12599 EXPLICIT tags. Some non standard certificates use these: they can now
12603 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12604 into a single doc/ssleay.txt bundle. This way the information is still
12605 preserved but no longer messes up this directory. Now it's new room for
12606 the new set of documentation files.
12607 [Ralf S. Engelschall]
12609 *) SETs were incorrectly DER encoded. This was a major pain, because they
12610 shared code with SEQUENCEs, which aren't coded the same. This means that
12611 almost everything to do with SETs or SEQUENCEs has either changed name or
12612 number of arguments.
12613 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12615 *) Fix test data to work with the above.
12618 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12619 was already fixed by Eric for 0.9.1 it seems.
12620 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12622 *) Autodetect FreeBSD3.
12625 *) Fix various bugs in Configure. This affects the following platforms:
12629 unixware-2.0-pentium
12633 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12634 before they are needed.
12637 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12641 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12643 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12644 changed SSLeay to OpenSSL in version strings.
12645 [Ralf S. Engelschall]
12647 *) Some fixups to the top-level documents.
12650 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12651 because the symlink to include/ was missing.
12652 [Ralf S. Engelschall]
12654 *) Incorporated the popular no-RSA/DSA-only patches
12655 which allow to compile a RSA-free SSLeay.
12656 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12658 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12659 when "ssleay" is still not found.
12660 [Ralf S. Engelschall]
12662 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12663 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12665 *) Updated the README file.
12666 [Ralf S. Engelschall]
12668 *) Added various .cvsignore files in the CVS repository subdirs
12669 to make a "cvs update" really silent.
12670 [Ralf S. Engelschall]
12672 *) Recompiled the error-definition header files and added
12673 missing symbols to the Win32 linker tables.
12674 [Ralf S. Engelschall]
12676 *) Cleaned up the top-level documents;
12677 o new files: CHANGES and LICENSE
12678 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12679 o merged COPYRIGHT into LICENSE
12680 o removed obsolete TODO file
12681 o renamed MICROSOFT to INSTALL.W32
12682 [Ralf S. Engelschall]
12684 *) Removed dummy files from the 0.9.1b source tree:
12685 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12686 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12687 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12688 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12689 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12690 [Ralf S. Engelschall]
12692 *) Added various platform portability fixes.
12695 *) The Genesis of the OpenSSL rpject:
12696 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12697 Young and Tim J. Hudson created while they were working for C2Net until
12699 [The OpenSSL Project]
12702 Changes between 0.9.0b and 0.9.1b [not released]
12704 *) Updated a few CA certificates under certs/
12707 *) Changed some BIGNUM api stuff.
12710 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12711 DGUX x86, Linux Alpha, etc.
12714 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12715 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12719 *) Add -strparse option to asn1pars program which parses nested
12721 [Dr Stephen Henson <shenson@bigfoot.com>]
12723 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12726 *) DSA fix for "ca" program.
12729 *) Added "-genkey" option to "dsaparam" program.
12732 *) Added RIPE MD160 (rmd160) message digest.
12735 *) Added -a (all) option to "ssleay version" command.
12738 *) Added PLATFORM define which is the id given to Configure.
12741 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12744 *) Extended the ASN.1 parser routines.
12747 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12750 *) Added a BN_CTX to the BN library.
12753 *) Fixed the weak key values in DES library
12756 *) Changed API in EVP library for cipher aliases.
12759 *) Added support for RC2/64bit cipher.
12762 *) Converted the lhash library to the crypto/mem.c functions.
12765 *) Added more recognized ASN.1 object ids.
12768 *) Added more RSA padding checks for SSL/TLS.
12771 *) Added BIO proxy/filter functionality.
12774 *) Added extra_certs to SSL_CTX which can be used
12775 send extra CA certificates to the client in the CA cert chain sending
12776 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12779 *) Now Fortezza is denied in the authentication phase because
12780 this is key exchange mechanism is not supported by SSLeay at all.
12783 *) Additional PKCS1 checks.
12786 *) Support the string "TLSv1" for all TLS v1 ciphers.
12789 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12790 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12793 *) Fixed a few memory leaks.
12796 *) Fixed various code and comment typos.
12799 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12800 bytes sent in the client random.
12801 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob