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 1.1.0i and 1.1.0j [xx XXX xxxx]
14 Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
16 *) Client DoS due to large DH parameter
18 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
19 malicious server can send a very large prime value to the client. This will
20 cause the client to spend an unreasonably long period of time generating a
21 key for this prime resulting in a hang until the client has finished. This
22 could be exploited in a Denial Of Service attack.
24 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
28 *) Cache timing vulnerability in RSA Key Generation
30 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
31 a cache timing side channel attack. An attacker with sufficient access to
32 mount cache timing attacks during the RSA key generation process could
33 recover the private key.
35 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
36 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
40 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
41 parameter is no longer accepted, as it leads to a corrupt table. NULL
42 pem_str is reserved for alias entries only.
45 *) Revert blinding in ECDSA sign and instead make problematic addition
46 length-invariant. Switch even to fixed-length Montgomery multiplication.
49 *) Change generating and checking of primes so that the error rate of not
50 being prime depends on the intended use based on the size of the input.
51 For larger primes this will result in more rounds of Miller-Rabin.
52 The maximal error rate for primes with more than 1080 bits is lowered
54 [Kurt Roeckx, Annie Yousar]
56 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
59 *) Add blinding to ECDSA and DSA signatures to protect against side channel
60 attacks discovered by Keegan Ryan (NCC Group).
63 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
64 now allow empty (zero character) pass phrases.
67 *) Certificate time validation (X509_cmp_time) enforces stricter
68 compliance with RFC 5280. Fractional seconds and timezone offsets
69 are no longer allowed.
72 *) Fixed a text canonicalisation bug in CMS
74 Where a CMS detached signature is used with text content the text goes
75 through a canonicalisation process first prior to signing or verifying a
76 signature. This process strips trailing space at the end of lines, converts
77 line terminators to CRLF and removes additional trailing line terminators
78 at the end of a file. A bug in the canonicalisation process meant that
79 some characters, such as form-feed, were incorrectly treated as whitespace
80 and removed. This is contrary to the specification (RFC5485). This fix
81 could mean that detached text data signed with an earlier version of
82 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
83 signed with a fixed OpenSSL may fail to verify with an earlier version of
84 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
85 and use the "-binary" flag (for the "cms" command line application) or set
86 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
89 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
91 *) Constructed ASN.1 types with a recursive definition could exceed the stack
93 Constructed ASN.1 types with a recursive definition (such as can be found
94 in PKCS7) could eventually exceed the stack given malicious input with
95 excessive recursion. This could result in a Denial Of Service attack. There
96 are no such structures used within SSL/TLS that come from untrusted sources
97 so this is considered safe.
99 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
104 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
106 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
107 effectively reduced to only comparing the least significant bit of each
108 byte. This allows an attacker to forge messages that would be considered as
109 authenticated in an amount of tries lower than that guaranteed by the
110 security claims of the scheme. The module can only be compiled by the
111 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
113 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
118 *) Add a build target 'build_all_generated', to build all generated files
119 and only that. This can be used to prepare everything that requires
120 things like perl for a system that lacks perl and then move everything
121 to that system and do the rest of the build there.
124 *) Backport SSL_OP_NO_RENGOTIATION
126 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
127 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
128 changes this is no longer possible in 1.1.0. Therefore the new
129 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
130 1.1.0 to provide equivalent functionality.
132 Note that if an application built against 1.1.0h headers (or above) is run
133 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
134 accepted but nothing will happen, i.e. renegotiation will not be prevented.
137 *) Removed the OS390-Unix config target. It relied on a script that doesn't
141 *) rsaz_1024_mul_avx2 overflow bug on x86_64
143 There is an overflow bug in the AVX2 Montgomery multiplication procedure
144 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
145 Analysis suggests that attacks against RSA and DSA as a result of this
146 defect would be very difficult to perform and are not believed likely.
147 Attacks against DH1024 are considered just feasible, because most of the
148 work necessary to deduce information about a private key may be performed
149 offline. The amount of resources required for such an attack would be
150 significant. However, for an attack on TLS to be meaningful, the server
151 would have to share the DH1024 private key among multiple clients, which is
152 no longer an option since CVE-2016-0701.
154 This only affects processors that support the AVX2 but not ADX extensions
155 like Intel Haswell (4th generation).
157 This issue was reported to OpenSSL by David Benjamin (Google). The issue
158 was originally found via the OSS-Fuzz project.
162 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
164 *) bn_sqrx8x_internal carry bug on x86_64
166 There is a carry propagating bug in the x86_64 Montgomery squaring
167 procedure. No EC algorithms are affected. Analysis suggests that attacks
168 against RSA and DSA as a result of this defect would be very difficult to
169 perform and are not believed likely. Attacks against DH are considered just
170 feasible (although very difficult) because most of the work necessary to
171 deduce information about a private key may be performed offline. The amount
172 of resources required for such an attack would be very significant and
173 likely only accessible to a limited number of attackers. An attacker would
174 additionally need online access to an unpatched system using the target
175 private key in a scenario with persistent DH parameters and a private
176 key that is shared between multiple clients.
178 This only affects processors that support the BMI1, BMI2 and ADX extensions
179 like Intel Broadwell (5th generation) and later or AMD Ryzen.
181 This issue was reported to OpenSSL by the OSS-Fuzz project.
185 *) Malformed X.509 IPAddressFamily could cause OOB read
187 If an X.509 certificate has a malformed IPAddressFamily extension,
188 OpenSSL could do a one-byte buffer overread. The most likely result
189 would be an erroneous display of the certificate in text format.
191 This issue was reported to OpenSSL by the OSS-Fuzz project.
195 *) Ignore the '-named_curve auto' value for compatibility of applications
197 [Tomas Mraz <tmraz@fedoraproject.org>]
199 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
202 Changes between 1.1.0e and 1.1.0f [25 May 2017]
204 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
205 platform rather than 'mingw'.
208 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
209 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
210 which is the minimum version we support.
213 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
215 *) Encrypt-Then-Mac renegotiation crash
217 During a renegotiation handshake if the Encrypt-Then-Mac extension is
218 negotiated where it was not in the original handshake (or vice-versa) then
219 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
220 and servers are affected.
222 This issue was reported to OpenSSL by Joe Orton (Red Hat).
226 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
228 *) Truncated packet could crash via OOB read
230 If one side of an SSL/TLS path is running on a 32-bit host and a specific
231 cipher is being used, then a truncated packet can cause that host to
232 perform an out-of-bounds read, usually resulting in a crash.
234 This issue was reported to OpenSSL by Robert Święcki of Google.
238 *) Bad (EC)DHE parameters cause a client crash
240 If a malicious server supplies bad parameters for a DHE or ECDHE key
241 exchange then this can result in the client attempting to dereference a
242 NULL pointer leading to a client crash. This could be exploited in a Denial
245 This issue was reported to OpenSSL by Guido Vranken.
249 *) BN_mod_exp may produce incorrect results on x86_64
251 There is a carry propagating bug in the x86_64 Montgomery squaring
252 procedure. No EC algorithms are affected. Analysis suggests that attacks
253 against RSA and DSA as a result of this defect would be very difficult to
254 perform and are not believed likely. Attacks against DH are considered just
255 feasible (although very difficult) because most of the work necessary to
256 deduce information about a private key may be performed offline. The amount
257 of resources required for such an attack would be very significant and
258 likely only accessible to a limited number of attackers. An attacker would
259 additionally need online access to an unpatched system using the target
260 private key in a scenario with persistent DH parameters and a private
261 key that is shared between multiple clients. For example this can occur by
262 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
263 similar to CVE-2015-3193 but must be treated as a separate problem.
265 This issue was reported to OpenSSL by the OSS-Fuzz project.
269 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
271 *) ChaCha20/Poly1305 heap-buffer-overflow
273 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
274 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
275 crash. This issue is not considered to be exploitable beyond a DoS.
277 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
281 *) CMS Null dereference
283 Applications parsing invalid CMS structures can crash with a NULL pointer
284 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
285 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
286 structure callback if an attempt is made to free certain invalid encodings.
287 Only CHOICE structures using a callback which do not handle NULL value are
290 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
294 *) Montgomery multiplication may produce incorrect results
296 There is a carry propagating bug in the Broadwell-specific Montgomery
297 multiplication procedure that handles input lengths divisible by, but
298 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
299 and DH private keys are impossible. This is because the subroutine in
300 question is not used in operations with the private key itself and an input
301 of the attacker's direct choice. Otherwise the bug can manifest itself as
302 transient authentication and key negotiation failures or reproducible
303 erroneous outcome of public-key operations with specially crafted input.
304 Among EC algorithms only Brainpool P-512 curves are affected and one
305 presumably can attack ECDH key negotiation. Impact was not analyzed in
306 detail, because pre-requisites for attack are considered unlikely. Namely
307 multiple clients have to choose the curve in question and the server has to
308 share the private key among them, neither of which is default behaviour.
309 Even then only clients that chose the curve will be affected.
311 This issue was publicly reported as transient failures and was not
312 initially recognized as a security issue. Thanks to Richard Morgan for
313 providing reproducible case.
317 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
318 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
319 prevent issues where no progress is being made and the peer continually
320 sends unrecognised record types, using up resources processing them.
323 *) Removed automatic addition of RPATH in shared libraries and executables,
324 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
327 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
329 *) Fix Use After Free for large message sizes
331 The patch applied to address CVE-2016-6307 resulted in an issue where if a
332 message larger than approx 16k is received then the underlying buffer to
333 store the incoming message is reallocated and moved. Unfortunately a
334 dangling pointer to the old location is left which results in an attempt to
335 write to the previously freed location. This is likely to result in a
336 crash, however it could potentially lead to execution of arbitrary code.
338 This issue only affects OpenSSL 1.1.0a.
340 This issue was reported to OpenSSL by Robert Święcki.
344 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
346 *) OCSP Status Request extension unbounded memory growth
348 A malicious client can send an excessively large OCSP Status Request
349 extension. If that client continually requests renegotiation, sending a
350 large OCSP Status Request extension each time, then there will be unbounded
351 memory growth on the server. This will eventually lead to a Denial Of
352 Service attack through memory exhaustion. Servers with a default
353 configuration are vulnerable even if they do not support OCSP. Builds using
354 the "no-ocsp" build time option are not affected.
356 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
360 *) SSL_peek() hang on empty record
362 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
363 sends an empty record. This could be exploited by a malicious peer in a
364 Denial Of Service attack.
366 This issue was reported to OpenSSL by Alex Gaynor.
370 *) Excessive allocation of memory in tls_get_message_header() and
371 dtls1_preprocess_fragment()
373 A (D)TLS message includes 3 bytes for its length in the header for the
374 message. This would allow for messages up to 16Mb in length. Messages of
375 this length are excessive and OpenSSL includes a check to ensure that a
376 peer is sending reasonably sized messages in order to avoid too much memory
377 being consumed to service a connection. A flaw in the logic of version
378 1.1.0 means that memory for the message is allocated too early, prior to
379 the excessive message length check. Due to way memory is allocated in
380 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
381 to service a connection. This could lead to a Denial of Service through
382 memory exhaustion. However, the excessive message length check still takes
383 place, and this would cause the connection to immediately fail. Assuming
384 that the application calls SSL_free() on the failed connection in a timely
385 manner then the 21Mb of allocated memory will then be immediately freed
386 again. Therefore the excessive memory allocation will be transitory in
387 nature. This then means that there is only a security impact if:
389 1) The application does not call SSL_free() in a timely manner in the event
390 that the connection fails
392 2) The application is working in a constrained environment where there is
393 very little free memory
395 3) The attacker initiates multiple connection attempts such that there are
396 multiple connections in a state where memory has been allocated for the
397 connection; SSL_free() has not yet been called; and there is insufficient
398 memory to service the multiple requests.
400 Except in the instance of (1) above any Denial Of Service is likely to be
401 transitory because as soon as the connection fails the memory is
402 subsequently freed again in the SSL_free() call. However there is an
403 increased risk during this period of application crashes due to the lack of
404 memory - which would then mean a more serious Denial of Service.
406 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
407 (CVE-2016-6307 and CVE-2016-6308)
410 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
411 had to be removed. Primary reason is that vendor assembler can't
412 assemble our modules with -KPIC flag. As result it, assembly
413 support, was not even available as option. But its lack means
414 lack of side-channel resistant code, which is incompatible with
415 security by todays standards. Fortunately gcc is readily available
416 prepackaged option, which we firmly point at...
419 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
421 *) Windows command-line tool supports UTF-8 opt-in option for arguments
422 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
423 (to any value) allows Windows user to access PKCS#12 file generated
424 with Windows CryptoAPI and protected with non-ASCII password, as well
425 as files generated under UTF-8 locale on Linux also protected with
429 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
430 have been disabled by default and removed from DEFAULT, just like RC4.
431 See the RC4 item below to re-enable both.
434 *) The method for finding the storage location for the Windows RAND seed file
435 has changed. First we check %RANDFILE%. If that is not set then we check
436 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
437 all else fails we fall back to C:\.
440 *) The EVP_EncryptUpdate() function has had its return type changed from void
441 to int. A return of 0 indicates and error while a return of 1 indicates
445 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
446 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
447 off the constant time implementation for RSA, DSA and DH have been made
448 no-ops and deprecated.
451 *) Windows RAND implementation was simplified to only get entropy by
452 calling CryptGenRandom(). Various other RAND-related tickets
454 [Joseph Wylie Yandle, Rich Salz]
456 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
457 and OPENSSL_LH_, respectively. The old names are available
458 with API compatibility. They new names are now completely documented.
461 *) Unify TYPE_up_ref(obj) methods signature.
462 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
463 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
464 int (instead of void) like all others TYPE_up_ref() methods.
465 So now these methods also check the return value of CRYPTO_atomic_add(),
466 and the validity of object reference counter.
467 [fdasilvayy@gmail.com]
469 *) With Windows Visual Studio builds, the .pdb files are installed
470 alongside the installed libraries and executables. For a static
471 library installation, ossl_static.pdb is the associate compiler
472 generated .pdb file to be used when linking programs.
475 *) Remove openssl.spec. Packaging files belong with the packagers.
478 *) Automatic Darwin/OSX configuration has had a refresh, it will now
479 recognise x86_64 architectures automatically. You can still decide
480 to build for a different bitness with the environment variable
481 KERNEL_BITS (can be 32 or 64), for example:
483 KERNEL_BITS=32 ./config
487 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
488 256 bit AES and HMAC with SHA256.
491 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
494 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
497 *) To enable users to have their own config files and build file templates,
498 Configure looks in the directory indicated by the environment variable
499 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
500 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
501 name and is used as is.
504 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
505 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
506 X509_CERT_FILE_CTX was removed.
509 *) "shared" builds are now the default. To create only static libraries use
510 the "no-shared" Configure option.
513 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
514 All of these option have not worked for some while and are fundamental
518 *) Make various cleanup routines no-ops and mark them as deprecated. Most
519 global cleanup functions are no longer required because they are handled
520 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
521 Explicitly de-initing can cause problems (e.g. where a library that uses
522 OpenSSL de-inits, but an application is still using it). The affected
523 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
524 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
525 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
529 *) --strict-warnings no longer enables runtime debugging options
530 such as REF_DEBUG. Instead, debug options are automatically
531 enabled with '--debug' builds.
532 [Andy Polyakov, Emilia Käsper]
534 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
535 have been moved out of the public header files. New functions for managing
536 these have been added.
539 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
540 objects have been moved out of the public header files. New
541 functions for managing these have been added.
544 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
545 have been moved out of the public header files. New functions for managing
546 these have been added.
549 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
550 moved out of the public header files. New functions for managing these
554 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
557 *) Removed the mk1mf build scripts.
560 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
561 it is always safe to #include a header now.
564 *) Removed the aged BC-32 config and all its supporting scripts
567 *) Removed support for Ultrix, Netware, and OS/2.
570 *) Add support for HKDF.
573 *) Add support for blake2b and blake2s
576 *) Added support for "pipelining". Ciphers that have the
577 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
578 encryptions/decryptions simultaneously. There are currently no built-in
579 ciphers with this property but the expectation is that engines will be able
580 to offer it to significantly improve throughput. Support has been extended
581 into libssl so that multiple records for a single connection can be
582 processed in one go (for >=TLS 1.1).
585 *) Added the AFALG engine. This is an async capable engine which is able to
586 offload work to the Linux kernel. In this initial version it only supports
587 AES128-CBC. The kernel must be version 4.1.0 or greater.
590 *) OpenSSL now uses a new threading API. It is no longer necessary to
591 set locking callbacks to use OpenSSL in a multi-threaded environment. There
592 are two supported threading models: pthreads and windows threads. It is
593 also possible to configure OpenSSL at compile time for "no-threads". The
594 old threading API should no longer be used. The functions have been
595 replaced with "no-op" compatibility macros.
596 [Alessandro Ghedini, Matt Caswell]
598 *) Modify behavior of ALPN to invoke callback after SNI/servername
599 callback, such that updates to the SSL_CTX affect ALPN.
602 *) Add SSL_CIPHER queries for authentication and key-exchange.
605 *) Changes to the DEFAULT cipherlist:
606 - Prefer (EC)DHE handshakes over plain RSA.
607 - Prefer AEAD ciphers over legacy ciphers.
608 - Prefer ECDSA over RSA when both certificates are available.
609 - Prefer TLSv1.2 ciphers/PRF.
610 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
614 *) Change the ECC default curve list to be this, in order: x25519,
615 secp256r1, secp521r1, secp384r1.
618 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
619 disabled by default. They can be re-enabled using the
620 enable-weak-ssl-ciphers option to Configure.
623 *) If the server has ALPN configured, but supports no protocols that the
624 client advertises, send a fatal "no_application_protocol" alert.
625 This behaviour is SHALL in RFC 7301, though it isn't universally
626 implemented by other servers.
629 *) Add X25519 support.
630 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
631 for public and private key encoding using the format documented in
632 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
633 key generation and key derivation.
635 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
639 *) Deprecate SRP_VBASE_get_by_user.
640 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
641 In order to fix an unavoidable memory leak (CVE-2016-0798),
642 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
643 seed, even if the seed is configured.
645 Users should use SRP_VBASE_get1_by_user instead. Note that in
646 SRP_VBASE_get1_by_user, caller must free the returned value. Note
647 also that even though configuring the SRP seed attempts to hide
648 invalid usernames by continuing the handshake with fake
649 credentials, this behaviour is not constant time and no strong
650 guarantees are made that the handshake is indistinguishable from
651 that of a valid user.
654 *) Configuration change; it's now possible to build dynamic engines
655 without having to build shared libraries and vice versa. This
656 only applies to the engines in engines/, those in crypto/engine/
657 will always be built into libcrypto (i.e. "static").
659 Building dynamic engines is enabled by default; to disable, use
660 the configuration option "disable-dynamic-engine".
662 The only requirements for building dynamic engines are the
663 presence of the DSO module and building with position independent
664 code, so they will also automatically be disabled if configuring
665 with "disable-dso" or "disable-pic".
667 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
668 are also taken away from openssl/opensslconf.h, as they are
672 *) Configuration change; if there is a known flag to compile
673 position independent code, it will always be applied on the
674 libcrypto and libssl object files, and never on the application
675 object files. This means other libraries that use routines from
676 libcrypto / libssl can be made into shared libraries regardless
677 of how OpenSSL was configured.
679 If this isn't desirable, the configuration options "disable-pic"
680 or "no-pic" can be used to disable the use of PIC. This will
681 also disable building shared libraries and dynamic engines.
684 *) Removed JPAKE code. It was experimental and has no wide use.
687 *) The INSTALL_PREFIX Makefile variable has been renamed to
688 DESTDIR. That makes for less confusion on what this variable
689 is for. Also, the configuration option --install_prefix is
693 *) Heartbeat for TLS has been removed and is disabled by default
694 for DTLS; configure with enable-heartbeats. Code that uses the
695 old #define's might need to be updated.
696 [Emilia Käsper, Rich Salz]
698 *) Rename REF_CHECK to REF_DEBUG.
701 *) New "unified" build system
703 The "unified" build system is aimed to be a common system for all
704 platforms we support. With it comes new support for VMS.
706 This system builds supports building in a different directory tree
707 than the source tree. It produces one Makefile (for unix family
708 or lookalikes), or one descrip.mms (for VMS).
710 The source of information to make the Makefile / descrip.mms is
711 small files called 'build.info', holding the necessary
712 information for each directory with source to compile, and a
713 template in Configurations, like unix-Makefile.tmpl or
716 With this change, the library names were also renamed on Windows
717 and on VMS. They now have names that are closer to the standard
718 on Unix, and include the major version number, and in certain
719 cases, the architecture they are built for. See "Notes on shared
720 libraries" in INSTALL.
722 We rely heavily on the perl module Text::Template.
725 *) Added support for auto-initialisation and de-initialisation of the library.
726 OpenSSL no longer requires explicit init or deinit routines to be called,
727 except in certain circumstances. See the OPENSSL_init_crypto() and
728 OPENSSL_init_ssl() man pages for further information.
731 *) The arguments to the DTLSv1_listen function have changed. Specifically the
732 "peer" argument is now expected to be a BIO_ADDR object.
734 *) Rewrite of BIO networking library. The BIO library lacked consistent
735 support of IPv6, and adding it required some more extensive
736 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
737 which hold all types of addresses and chains of address information.
738 It also introduces a new API, with functions like BIO_socket,
739 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
740 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
741 have been adapted accordingly.
744 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
748 *) CRIME protection: disable compression by default, even if OpenSSL is
749 compiled with zlib enabled. Applications can still enable compression
750 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
751 using the SSL_CONF library to configure compression.
754 *) The signature of the session callback configured with
755 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
756 was explicitly marked as 'const unsigned char*' instead of
760 *) Always DPURIFY. Remove the use of uninitialized memory in the
761 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
764 *) Removed many obsolete configuration items, including
765 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
766 MD2_CHAR, MD2_INT, MD2_LONG
768 IDEA_SHORT, IDEA_LONG
769 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
770 [Rich Salz, with advice from Andy Polyakov]
772 *) Many BN internals have been moved to an internal header file.
773 [Rich Salz with help from Andy Polyakov]
775 *) Configuration and writing out the results from it has changed.
776 Files such as Makefile include/openssl/opensslconf.h and are now
777 produced through general templates, such as Makefile.in and
778 crypto/opensslconf.h.in and some help from the perl module
781 Also, the center of configuration information is no longer
782 Makefile. Instead, Configure produces a perl module in
783 configdata.pm which holds most of the config data (in the hash
784 table %config), the target data that comes from the target
785 configuration in one of the Configurations/*.conf files (in
789 *) To clarify their intended purposes, the Configure options
790 --prefix and --openssldir change their semantics, and become more
791 straightforward and less interdependent.
793 --prefix shall be used exclusively to give the location INSTALLTOP
794 where programs, scripts, libraries, include files and manuals are
795 going to be installed. The default is now /usr/local.
797 --openssldir shall be used exclusively to give the default
798 location OPENSSLDIR where certificates, private keys, CRLs are
799 managed. This is also where the default openssl.cnf gets
801 If the directory given with this option is a relative path, the
802 values of both the --prefix value and the --openssldir value will
803 be combined to become OPENSSLDIR.
804 The default for --openssldir is INSTALLTOP/ssl.
806 Anyone who uses --openssldir to specify where OpenSSL is to be
807 installed MUST change to use --prefix instead.
810 *) The GOST engine was out of date and therefore it has been removed. An up
811 to date GOST engine is now being maintained in an external repository.
812 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
813 support for GOST ciphersuites (these are only activated if a GOST engine
817 *) EGD is no longer supported by default; use enable-egd when
819 [Ben Kaduk and Rich Salz]
821 *) The distribution now has Makefile.in files, which are used to
822 create Makefile's when Configure is run. *Configure must be run
823 before trying to build now.*
826 *) The return value for SSL_CIPHER_description() for error conditions
830 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
832 Obtaining and performing DNSSEC validation of TLSA records is
833 the application's responsibility. The application provides
834 the TLSA records of its choice to OpenSSL, and these are then
835 used to authenticate the peer.
837 The TLSA records need not even come from DNS. They can, for
838 example, be used to implement local end-entity certificate or
839 trust-anchor "pinning", where the "pin" data takes the form
840 of TLSA records, which can augment or replace verification
841 based on the usual WebPKI public certification authorities.
844 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
845 continues to support deprecated interfaces in default builds.
846 However, applications are strongly advised to compile their
847 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
848 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
849 or the 1.1.0 releases.
851 In environments in which all applications have been ported to
852 not use any deprecated interfaces OpenSSL's Configure script
853 should be used with the --api=1.1.0 option to entirely remove
854 support for the deprecated features from the library and
855 unconditionally disable them in the installed headers.
856 Essentially the same effect can be achieved with the "no-deprecated"
857 argument to Configure, except that this will always restrict
858 the build to just the latest API, rather than a fixed API
861 As applications are ported to future revisions of the API,
862 they should update their compile-time OPENSSL_API_COMPAT define
863 accordingly, but in most cases should be able to continue to
864 compile with later releases.
866 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
867 0x10000000L and 0x00908000L, respectively. However those
868 versions did not support the OPENSSL_API_COMPAT feature, and
869 so applications are not typically tested for explicit support
870 of just the undeprecated features of either release.
873 *) Add support for setting the minimum and maximum supported protocol.
874 It can bet set via the SSL_set_min_proto_version() and
875 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
876 MaxProtcol. It's recommended to use the new APIs to disable
877 protocols instead of disabling individual protocols using
878 SSL_set_options() or SSL_CONF's Protocol. This change also
879 removes support for disabling TLS 1.2 in the OpenSSL TLS
880 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
883 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
886 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
887 and integrates ECDSA and ECDH functionality into EC. Implementations can
888 now redirect key generation and no longer need to convert to or from
891 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
892 include the ec.h header file instead.
895 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
896 ciphers who are no longer supported and drops support the ephemeral RSA key
897 exchange. The LOW ciphers currently doesn't have any ciphers in it.
900 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
901 opaque. For HMAC_CTX, the following constructors and destructors
904 HMAC_CTX *HMAC_CTX_new(void);
905 void HMAC_CTX_free(HMAC_CTX *ctx);
907 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
908 destroy such methods has been added. See EVP_MD_meth_new(3) and
909 EVP_CIPHER_meth_new(3) for documentation.
912 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
913 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
914 EVP_MD_CTX_reset() should be called instead to reinitialise
915 an already created structure.
916 2) For consistency with the majority of our object creators and
917 destructors, EVP_MD_CTX_(create|destroy) were renamed to
918 EVP_MD_CTX_(new|free). The old names are retained as macros
919 for deprecated builds.
922 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
923 cryptographic operations to be performed asynchronously as long as an
924 asynchronous capable engine is used. See the ASYNC_start_job() man page for
925 further details. Libssl has also had this capability integrated with the
926 introduction of the new mode SSL_MODE_ASYNC and associated error
927 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
928 pages. This work was developed in partnership with Intel Corp.
931 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
932 always enabled now. If you want to disable the support you should
933 exclude it using the list of supported ciphers. This also means that the
934 "-no_ecdhe" option has been removed from s_server.
937 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
938 SSL_{CTX_}set1_curves() which can set a list.
941 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
942 curve you want to support using SSL_{CTX_}set1_curves().
945 *) State machine rewrite. The state machine code has been significantly
946 refactored in order to remove much duplication of code and solve issues
947 with the old code (see ssl/statem/README for further details). This change
948 does have some associated API changes. Notably the SSL_state() function
949 has been removed and replaced by SSL_get_state which now returns an
950 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
951 altogether. The previous handshake states defined in ssl.h and ssl3.h have
955 *) All instances of the string "ssleay" in the public API were replaced
956 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
957 Some error codes related to internal RSA_eay API's were renamed.
960 *) The demo files in crypto/threads were moved to demo/threads.
963 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
965 [Matt Caswell, Rich Salz]
967 *) New ASN.1 embed macro.
969 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
970 structure is not allocated: it is part of the parent. That is instead of
978 This reduces memory fragmentation and make it impossible to accidentally
979 set a mandatory field to NULL.
981 This currently only works for some fields specifically a SEQUENCE, CHOICE,
982 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
983 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
987 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
990 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
991 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
992 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
993 DES and RC4 ciphersuites.
996 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
997 This changes the decoding behaviour for some invalid messages,
998 though the change is mostly in the more lenient direction, and
999 legacy behaviour is preserved as much as possible.
1002 *) Fix no-stdio build.
1003 [ David Woodhouse <David.Woodhouse@intel.com> and also
1004 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1006 *) New testing framework
1007 The testing framework has been largely rewritten and is now using
1008 perl and the perl modules Test::Harness and an extended variant of
1009 Test::More called OpenSSL::Test to do its work. All test scripts in
1010 test/ have been rewritten into test recipes, and all direct calls to
1011 executables in test/Makefile have become individual recipes using the
1012 simplified testing OpenSSL::Test::Simple.
1014 For documentation on our testing modules, do:
1016 perldoc test/testlib/OpenSSL/Test/Simple.pm
1017 perldoc test/testlib/OpenSSL/Test.pm
1021 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1022 are used; the latter aborts on memory leaks (usually checked on exit).
1023 Some undocumented "set malloc, etc., hooks" functions were removed
1024 and others were changed. All are now documented.
1027 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1029 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1031 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1032 from RFC4279, RFC4785, RFC5487, RFC5489.
1034 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1035 original RSA_PSK patch.
1038 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1039 era flag was never set throughout the codebase (only read). Also removed
1040 SSL3_FLAGS_POP_BUFFER which was only used if
1041 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1044 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1045 to be "oneline" instead of "compat".
1048 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1049 not aware of clients that still exhibit this bug, and the workaround
1050 hasn't been working properly for a while.
1053 *) The return type of BIO_number_read() and BIO_number_written() as well as
1054 the corresponding num_read and num_write members in the BIO structure has
1055 changed from unsigned long to uint64_t. On platforms where an unsigned
1056 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1060 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1061 OpenSSL without support for them. It also means that maintaining
1062 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1063 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1066 *) Removed support for the two export grade static DH ciphersuites
1067 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1068 were newly added (along with a number of other static DH ciphersuites) to
1069 1.0.2. However the two export ones have *never* worked since they were
1070 introduced. It seems strange in any case to be adding new export
1071 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1074 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1075 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1076 and turned into macros which simply call the new preferred function names
1077 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1078 should use the new names instead. Also as part of this change the ssl23.h
1079 header file has been removed.
1082 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1083 code and the associated standard is no longer considered fit-for-purpose.
1086 *) RT2547 was closed. When generating a private key, try to make the
1087 output file readable only by the owner. This behavior change might
1088 be noticeable when interacting with other software.
1090 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1094 *) Added HTTP GET support to the ocsp command.
1097 *) Changed default digest for the dgst and enc commands from MD5 to
1101 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1104 *) Added support for TLS extended master secret from
1105 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1106 initial patch which was a great help during development.
1109 *) All libssl internal structures have been removed from the public header
1110 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1111 now redundant). Users should not attempt to access internal structures
1112 directly. Instead they should use the provided API functions.
1115 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1116 Access to deprecated functions can be re-enabled by running config with
1117 "enable-deprecated". In addition applications wishing to use deprecated
1118 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1119 will, by default, disable some transitive includes that previously existed
1120 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1123 *) Added support for OCB mode. OpenSSL has been granted a patent license
1124 compatible with the OpenSSL license for use of OCB. Details are available
1125 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1126 for OCB can be removed by calling config with no-ocb.
1129 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1130 compatible client hello.
1133 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1134 done while fixing the error code for the key-too-small case.
1135 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1137 *) CA.sh has been removed; use CA.pl instead.
1140 *) Removed old DES API.
1143 *) Remove various unsupported platforms:
1149 Sinix/ReliantUNIX RM400
1154 16-bit platforms such as WIN16
1157 *) Clean up OPENSSL_NO_xxx #define's
1158 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1159 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1160 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1161 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1162 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1163 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1164 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1165 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1166 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1167 Remove MS_STATIC; it's a relic from platforms <32 bits.
1170 *) Cleaned up dead code
1171 Remove all but one '#ifdef undef' which is to be looked at.
1174 *) Clean up calling of xxx_free routines.
1175 Just like free(), fix most of the xxx_free routines to accept
1176 NULL. Remove the non-null checks from callers. Save much code.
1179 *) Add secure heap for storage of private keys (when possible).
1180 Add BIO_s_secmem(), CBIGNUM, etc.
1181 Contributed by Akamai Technologies under our Corporate CLA.
1184 *) Experimental support for a new, fast, unbiased prime candidate generator,
1185 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1186 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1188 *) New output format NSS in the sess_id command line tool. This allows
1189 exporting the session id and the master key in NSS keylog format.
1190 [Martin Kaiser <martin@kaiser.cx>]
1192 *) Harmonize version and its documentation. -f flag is used to display
1194 [mancha <mancha1@zoho.com>]
1196 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1197 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1198 [mancha <mancha1@zoho.com>]
1200 *) Fix some double frees. These are not thought to be exploitable.
1201 [mancha <mancha1@zoho.com>]
1203 *) A missing bounds check in the handling of the TLS heartbeat extension
1204 can be used to reveal up to 64k of memory to a connected client or
1207 Thanks for Neel Mehta of Google Security for discovering this bug and to
1208 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1209 preparing the fix (CVE-2014-0160)
1210 [Adam Langley, Bodo Moeller]
1212 *) Fix for the attack described in the paper "Recovering OpenSSL
1213 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1214 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1215 http://eprint.iacr.org/2014/140
1217 Thanks to Yuval Yarom and Naomi Benger for discovering this
1218 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1219 [Yuval Yarom and Naomi Benger]
1221 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1222 this fixes a limitation in previous versions of OpenSSL.
1225 *) Experimental encrypt-then-mac support.
1227 Experimental support for encrypt then mac from
1228 draft-gutmann-tls-encrypt-then-mac-02.txt
1230 To enable it set the appropriate extension number (0x42 for the test
1231 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1233 For non-compliant peers (i.e. just about everything) this should have no
1236 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1240 *) Add EVP support for key wrapping algorithms, to avoid problems with
1241 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1242 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1243 algorithms and include tests cases.
1246 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1250 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1251 MGF1 digest and OAEP label.
1254 *) Make openssl verify return errors.
1255 [Chris Palmer <palmer@google.com> and Ben Laurie]
1257 *) New function ASN1_TIME_diff to calculate the difference between two
1258 ASN1_TIME structures or one structure and the current time.
1261 *) Update fips_test_suite to support multiple command line options. New
1262 test to induce all self test errors in sequence and check expected
1266 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1267 sign or verify all in one operation.
1270 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1271 test programs and fips_test_suite. Includes functionality to parse
1272 the minimal script output of fipsalgest.pl directly.
1275 *) Add authorisation parameter to FIPS_module_mode_set().
1278 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1281 *) Use separate DRBG fields for internal and external flags. New function
1282 FIPS_drbg_health_check() to perform on demand health checking. Add
1283 generation tests to fips_test_suite with reduced health check interval to
1284 demonstrate periodic health checking. Add "nodh" option to
1285 fips_test_suite to skip very slow DH test.
1288 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1292 *) More extensive health check for DRBG checking many more failure modes.
1293 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1294 combination: call this in fips_test_suite.
1297 *) Add support for canonical generation of DSA parameter 'g'. See
1300 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1301 POST to handle HMAC cases.
1304 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1305 to return numerical and string versions of the FIPS module number.
1308 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1309 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1310 outside the validated module in the FIPS capable OpenSSL.
1313 *) Minor change to DRBG entropy callback semantics. In some cases
1314 there is no multiple of the block length between min_len and
1315 max_len. Allow the callback to return more than max_len bytes
1316 of entropy but discard any extra: it is the callback's responsibility
1317 to ensure that the extra data discarded does not impact the
1318 requested amount of entropy.
1321 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1322 information in FIPS186-3, SP800-57 and SP800-131A.
1325 *) CCM support via EVP. Interface is very similar to GCM case except we
1326 must supply all data in one chunk (i.e. no update, final) and the
1327 message length must be supplied if AAD is used. Add algorithm test
1331 *) Initial version of POST overhaul. Add POST callback to allow the status
1332 of POST to be monitored and/or failures induced. Modify fips_test_suite
1333 to use callback. Always run all selftests even if one fails.
1336 *) XTS support including algorithm test driver in the fips_gcmtest program.
1337 Note: this does increase the maximum key length from 32 to 64 bytes but
1338 there should be no binary compatibility issues as existing applications
1339 will never use XTS mode.
1342 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1343 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1344 performs algorithm blocking for unapproved PRNG types. Also do not
1345 set PRNG type in FIPS_mode_set(): leave this to the application.
1346 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1347 the standard OpenSSL PRNG: set additional data to a date time vector.
1350 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1351 This shouldn't present any incompatibility problems because applications
1352 shouldn't be using these directly and any that are will need to rethink
1353 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1356 *) Extensive self tests and health checking required by SP800-90 DRBG.
1357 Remove strength parameter from FIPS_drbg_instantiate and always
1358 instantiate at maximum supported strength.
1361 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1364 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1367 *) New function DH_compute_key_padded() to compute a DH key and pad with
1368 leading zeroes if needed: this complies with SP800-56A et al.
1371 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1372 anything, incomplete, subject to change and largely untested at present.
1375 *) Modify fipscanisteronly build option to only build the necessary object
1376 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1379 *) Add experimental option FIPSSYMS to give all symbols in
1380 fipscanister.o and FIPS or fips prefix. This will avoid
1381 conflicts with future versions of OpenSSL. Add perl script
1382 util/fipsas.pl to preprocess assembly language source files
1383 and rename any affected symbols.
1386 *) Add selftest checks and algorithm block of non-fips algorithms in
1387 FIPS mode. Remove DES2 from selftests.
1390 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1391 return internal method without any ENGINE dependencies. Add new
1392 tiny fips sign and verify functions.
1395 *) New build option no-ec2m to disable characteristic 2 code.
1398 *) New build option "fipscanisteronly". This only builds fipscanister.o
1399 and (currently) associated fips utilities. Uses the file Makefile.fips
1400 instead of Makefile.org as the prototype.
1403 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1404 Update fips_gcmtest to use IV generator.
1407 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1408 setting output buffer to NULL. The *Final function must be
1409 called although it will not retrieve any additional data. The tag
1410 can be set or retrieved with a ctrl. The IV length is by default 12
1411 bytes (96 bits) but can be set to an alternative value. If the IV
1412 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1416 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1417 underlying do_cipher function handles all cipher semantics itself
1418 including padding and finalisation. This is useful if (for example)
1419 an ENGINE cipher handles block padding itself. The behaviour of
1420 do_cipher is subtly changed if this flag is set: the return value
1421 is the number of characters written to the output buffer (zero is
1422 no longer an error code) or a negative error code. Also if the
1423 input buffer is NULL and length 0 finalisation should be performed.
1426 *) If a candidate issuer certificate is already part of the constructed
1427 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1430 *) Improve forward-security support: add functions
1432 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1433 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1435 for use by SSL/TLS servers; the callback function will be called whenever a
1436 new session is created, and gets to decide whether the session may be
1437 cached to make it resumable (return 0) or not (return 1). (As by the
1438 SSL/TLS protocol specifications, the session_id sent by the server will be
1439 empty to indicate that the session is not resumable; also, the server will
1440 not generate RFC 4507 (RFC 5077) session tickets.)
1442 A simple reasonable callback implementation is to return is_forward_secure.
1443 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1444 by the SSL/TLS server library, indicating whether it can provide forward
1446 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1448 *) New -verify_name option in command line utilities to set verification
1452 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1453 Add CMAC pkey methods.
1456 *) Experimental renegotiation in s_server -www mode. If the client
1457 browses /reneg connection is renegotiated. If /renegcert it is
1458 renegotiated requesting a certificate.
1461 *) Add an "external" session cache for debugging purposes to s_server. This
1462 should help trace issues which normally are only apparent in deployed
1463 multi-process servers.
1466 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1467 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1468 BIO_set_cipher() and some obscure PEM functions were changed so they
1469 can now return an error. The RAND changes required a change to the
1470 RAND_METHOD structure.
1473 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1474 a gcc attribute to warn if the result of a function is ignored. This
1475 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1476 whose return value is often ignored.
1479 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1480 These allow SCTs (signed certificate timestamps) to be requested and
1481 validated when establishing a connection.
1482 [Rob Percival <robpercival@google.com>]
1484 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1486 *) Prevent padding oracle in AES-NI CBC MAC check
1488 A MITM attacker can use a padding oracle attack to decrypt traffic
1489 when the connection uses an AES CBC cipher and the server support
1492 This issue was introduced as part of the fix for Lucky 13 padding
1493 attack (CVE-2013-0169). The padding check was rewritten to be in
1494 constant time by making sure that always the same bytes are read and
1495 compared against either the MAC or padding bytes. But it no longer
1496 checked that there was enough data to have both the MAC and padding
1499 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1503 *) Fix EVP_EncodeUpdate overflow
1505 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1506 Base64 encoding of binary data. If an attacker is able to supply very large
1507 amounts of input data then a length check can overflow resulting in a heap
1510 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1511 the PEM_write_bio* family of functions. These are mainly used within the
1512 OpenSSL command line applications, so any application which processes data
1513 from an untrusted source and outputs it as a PEM file should be considered
1514 vulnerable to this issue. User applications that call these APIs directly
1515 with large amounts of untrusted data may also be vulnerable.
1517 This issue was reported by Guido Vranken.
1521 *) Fix EVP_EncryptUpdate overflow
1523 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1524 is able to supply very large amounts of input data after a previous call to
1525 EVP_EncryptUpdate() with a partial block then a length check can overflow
1526 resulting in a heap corruption. Following an analysis of all OpenSSL
1527 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1528 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1529 the first called function after an EVP_EncryptInit(), and therefore that
1530 specific call must be safe. The second form is where the length passed to
1531 EVP_EncryptUpdate() can be seen from the code to be some small value and
1532 therefore there is no possibility of an overflow. Since all instances are
1533 one of these two forms, it is believed that there can be no overflows in
1534 internal code due to this problem. It should be noted that
1535 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1536 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1537 of these calls have also been analysed too and it is believed there are no
1538 instances in internal usage where an overflow could occur.
1540 This issue was reported by Guido Vranken.
1544 *) Prevent ASN.1 BIO excessive memory allocation
1546 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1547 a short invalid encoding can cause allocation of large amounts of memory
1548 potentially consuming excessive resources or exhausting memory.
1550 Any application parsing untrusted data through d2i BIO functions is
1551 affected. The memory based functions such as d2i_X509() are *not* affected.
1552 Since the memory based functions are used by the TLS library, TLS
1553 applications are not affected.
1555 This issue was reported by Brian Carpenter.
1561 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1562 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1563 in arbitrary stack data being returned in the buffer.
1565 This issue was reported by Guido Vranken.
1569 *) Modify behavior of ALPN to invoke callback after SNI/servername
1570 callback, such that updates to the SSL_CTX affect ALPN.
1573 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1577 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1578 methods are enabled and ssl2 is disabled the methods return NULL.
1581 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1583 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1584 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1585 provide any "EXPORT" or "LOW" strength ciphers.
1588 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1589 is by default disabled at build-time. Builds that are not configured with
1590 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1591 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1592 will need to explicitly call either of:
1594 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1596 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1598 as appropriate. Even if either of those is used, or the application
1599 explicitly uses the version-specific SSLv2_method() or its client and
1600 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1601 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1602 ciphers, and SSLv2 56-bit DES are no longer available.
1606 *) Fix a double-free in DSA code
1608 A double free bug was discovered when OpenSSL parses malformed DSA private
1609 keys and could lead to a DoS attack or memory corruption for applications
1610 that receive DSA private keys from untrusted sources. This scenario is
1613 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1618 *) Disable SRP fake user seed to address a server memory leak.
1620 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1622 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1623 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1624 was changed to ignore the "fake user" SRP seed, even if the seed
1627 Users should use SRP_VBASE_get1_by_user instead. Note that in
1628 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1629 also that even though configuring the SRP seed attempts to hide
1630 invalid usernames by continuing the handshake with fake
1631 credentials, this behaviour is not constant time and no strong
1632 guarantees are made that the handshake is indistinguishable from
1633 that of a valid user.
1637 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1639 In the BN_hex2bn function the number of hex digits is calculated using an
1640 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1641 large values of |i| this can result in |bn_expand| not allocating any
1642 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1643 field as NULL leading to a subsequent NULL ptr deref. For very large values
1644 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1645 In this case memory is allocated to the internal BIGNUM data field, but it
1646 is insufficiently sized leading to heap corruption. A similar issue exists
1647 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1648 is ever called by user applications with very large untrusted hex/dec data.
1649 This is anticipated to be a rare occurrence.
1651 All OpenSSL internal usage of these functions use data that is not expected
1652 to be untrusted, e.g. config file data or application command line
1653 arguments. If user developed applications generate config file data based
1654 on untrusted data then it is possible that this could also lead to security
1655 consequences. This is also anticipated to be rare.
1657 This issue was reported to OpenSSL by Guido Vranken.
1661 *) Fix memory issues in BIO_*printf functions
1663 The internal |fmtstr| function used in processing a "%s" format string in
1664 the BIO_*printf functions could overflow while calculating the length of a
1665 string and cause an OOB read when printing very long strings.
1667 Additionally the internal |doapr_outch| function can attempt to write to an
1668 OOB memory location (at an offset from the NULL pointer) in the event of a
1669 memory allocation failure. In 1.0.2 and below this could be caused where
1670 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1671 could be in processing a very long "%s" format string. Memory leaks can
1674 The first issue may mask the second issue dependent on compiler behaviour.
1675 These problems could enable attacks where large amounts of untrusted data
1676 is passed to the BIO_*printf functions. If applications use these functions
1677 in this way then they could be vulnerable. OpenSSL itself uses these
1678 functions when printing out human-readable dumps of ASN.1 data. Therefore
1679 applications that print this data could be vulnerable if the data is from
1680 untrusted sources. OpenSSL command line applications could also be
1681 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1682 as command line arguments.
1684 Libssl is not considered directly vulnerable. Additionally certificates etc
1685 received via remote connections via libssl are also unlikely to be able to
1686 trigger these issues because of message size limits enforced within libssl.
1688 This issue was reported to OpenSSL Guido Vranken.
1692 *) Side channel attack on modular exponentiation
1694 A side-channel attack was found which makes use of cache-bank conflicts on
1695 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1696 of RSA keys. The ability to exploit this issue is limited as it relies on
1697 an attacker who has control of code in a thread running on the same
1698 hyper-threaded core as the victim thread which is performing decryptions.
1700 This issue was reported to OpenSSL by Yuval Yarom, The University of
1701 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1702 Nadia Heninger, University of Pennsylvania with more information at
1703 http://cachebleed.info.
1707 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1708 if no keysize is specified with default_bits. This fixes an
1709 omission in an earlier change that changed all RSA/DSA key generation
1710 apps to use 2048 bits by default.
1713 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1714 *) DH small subgroups
1716 Historically OpenSSL only ever generated DH parameters based on "safe"
1717 primes. More recently (in version 1.0.2) support was provided for
1718 generating X9.42 style parameter files such as those required for RFC 5114
1719 support. The primes used in such files may not be "safe". Where an
1720 application is using DH configured with parameters based on primes that are
1721 not "safe" then an attacker could use this fact to find a peer's private
1722 DH exponent. This attack requires that the attacker complete multiple
1723 handshakes in which the peer uses the same private DH exponent. For example
1724 this could be used to discover a TLS server's private DH exponent if it's
1725 reusing the private DH exponent or it's using a static DH ciphersuite.
1727 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1728 TLS. It is not on by default. If the option is not set then the server
1729 reuses the same private DH exponent for the life of the server process and
1730 would be vulnerable to this attack. It is believed that many popular
1731 applications do set this option and would therefore not be at risk.
1733 The fix for this issue adds an additional check where a "q" parameter is
1734 available (as is the case in X9.42 based parameters). This detects the
1735 only known attack, and is the only possible defense for static DH
1736 ciphersuites. This could have some performance impact.
1738 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1739 default and cannot be disabled. This could have some performance impact.
1741 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1745 *) SSLv2 doesn't block disabled ciphers
1747 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1748 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1749 been disabled, provided that the SSLv2 protocol was not also disabled via
1752 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1753 and Sebastian Schinzel.
1757 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1759 *) BN_mod_exp may produce incorrect results on x86_64
1761 There is a carry propagating bug in the x86_64 Montgomery squaring
1762 procedure. No EC algorithms are affected. Analysis suggests that attacks
1763 against RSA and DSA as a result of this defect would be very difficult to
1764 perform and are not believed likely. Attacks against DH are considered just
1765 feasible (although very difficult) because most of the work necessary to
1766 deduce information about a private key may be performed offline. The amount
1767 of resources required for such an attack would be very significant and
1768 likely only accessible to a limited number of attackers. An attacker would
1769 additionally need online access to an unpatched system using the target
1770 private key in a scenario with persistent DH parameters and a private
1771 key that is shared between multiple clients. For example this can occur by
1772 default in OpenSSL DHE based SSL/TLS ciphersuites.
1774 This issue was reported to OpenSSL by Hanno Böck.
1778 *) Certificate verify crash with missing PSS parameter
1780 The signature verification routines will crash with a NULL pointer
1781 dereference if presented with an ASN.1 signature using the RSA PSS
1782 algorithm and absent mask generation function parameter. Since these
1783 routines are used to verify certificate signature algorithms this can be
1784 used to crash any certificate verification operation and exploited in a
1785 DoS attack. Any application which performs certificate verification is
1786 vulnerable including OpenSSL clients and servers which enable client
1789 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1793 *) X509_ATTRIBUTE memory leak
1795 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1796 memory. This structure is used by the PKCS#7 and CMS routines so any
1797 application which reads PKCS#7 or CMS data from untrusted sources is
1798 affected. SSL/TLS is not affected.
1800 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1805 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1806 This changes the decoding behaviour for some invalid messages,
1807 though the change is mostly in the more lenient direction, and
1808 legacy behaviour is preserved as much as possible.
1811 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1813 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1815 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1817 *) Alternate chains certificate forgery
1819 During certificate verification, OpenSSL will attempt to find an
1820 alternative certificate chain if the first attempt to build such a chain
1821 fails. An error in the implementation of this logic can mean that an
1822 attacker could cause certain checks on untrusted certificates to be
1823 bypassed, such as the CA flag, enabling them to use a valid leaf
1824 certificate to act as a CA and "issue" an invalid certificate.
1826 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1830 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1832 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1833 incompatibility in the handling of HMAC. The previous ABI has now been
1837 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1839 *) Malformed ECParameters causes infinite loop
1841 When processing an ECParameters structure OpenSSL enters an infinite loop
1842 if the curve specified is over a specially malformed binary polynomial
1845 This can be used to perform denial of service against any
1846 system which processes public keys, certificate requests or
1847 certificates. This includes TLS clients and TLS servers with
1848 client authentication enabled.
1850 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1854 *) Exploitable out-of-bounds read in X509_cmp_time
1856 X509_cmp_time does not properly check the length of the ASN1_TIME
1857 string and can read a few bytes out of bounds. In addition,
1858 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1861 An attacker can use this to craft malformed certificates and CRLs of
1862 various sizes and potentially cause a segmentation fault, resulting in
1863 a DoS on applications that verify certificates or CRLs. TLS clients
1864 that verify CRLs are affected. TLS clients and servers with client
1865 authentication enabled may be affected if they use custom verification
1868 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1869 independently by Hanno Böck.
1873 *) PKCS7 crash with missing EnvelopedContent
1875 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1876 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1877 with missing content and trigger a NULL pointer dereference on parsing.
1879 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1880 structures from untrusted sources are affected. OpenSSL clients and
1881 servers are not affected.
1883 This issue was reported to OpenSSL by Michal Zalewski (Google).
1887 *) CMS verify infinite loop with unknown hash function
1889 When verifying a signedData message the CMS code can enter an infinite loop
1890 if presented with an unknown hash function OID. This can be used to perform
1891 denial of service against any system which verifies signedData messages using
1893 This issue was reported to OpenSSL by Johannes Bauer.
1897 *) Race condition handling NewSessionTicket
1899 If a NewSessionTicket is received by a multi-threaded client when attempting to
1900 reuse a previous ticket then a race condition can occur potentially leading to
1901 a double free of the ticket data.
1905 *) Only support 256-bit or stronger elliptic curves with the
1906 'ecdh_auto' setting (server) or by default (client). Of supported
1907 curves, prefer P-256 (both).
1910 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1912 *) ClientHello sigalgs DoS fix
1914 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1915 invalid signature algorithms extension a NULL pointer dereference will
1916 occur. This can be exploited in a DoS attack against the server.
1918 This issue was was reported to OpenSSL by David Ramos of Stanford
1921 [Stephen Henson and Matt Caswell]
1923 *) Multiblock corrupted pointer fix
1925 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1926 feature only applies on 64 bit x86 architecture platforms that support AES
1927 NI instructions. A defect in the implementation of "multiblock" can cause
1928 OpenSSL's internal write buffer to become incorrectly set to NULL when
1929 using non-blocking IO. Typically, when the user application is using a
1930 socket BIO for writing, this will only result in a failed connection.
1931 However if some other BIO is used then it is likely that a segmentation
1932 fault will be triggered, thus enabling a potential DoS attack.
1934 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1938 *) Segmentation fault in DTLSv1_listen fix
1940 The DTLSv1_listen function is intended to be stateless and processes the
1941 initial ClientHello from many peers. It is common for user code to loop
1942 over the call to DTLSv1_listen until a valid ClientHello is received with
1943 an associated cookie. A defect in the implementation of DTLSv1_listen means
1944 that state is preserved in the SSL object from one invocation to the next
1945 that can lead to a segmentation fault. Errors processing the initial
1946 ClientHello can trigger this scenario. An example of such an error could be
1947 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1950 This issue was reported to OpenSSL by Per Allansson.
1954 *) Segmentation fault in ASN1_TYPE_cmp fix
1956 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1957 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1958 certificate signature algorithm consistency this can be used to crash any
1959 certificate verification operation and exploited in a DoS attack. Any
1960 application which performs certificate verification is vulnerable including
1961 OpenSSL clients and servers which enable client authentication.
1965 *) Segmentation fault for invalid PSS parameters fix
1967 The signature verification routines will crash with a NULL pointer
1968 dereference if presented with an ASN.1 signature using the RSA PSS
1969 algorithm and invalid parameters. Since these routines are used to verify
1970 certificate signature algorithms this can be used to crash any
1971 certificate verification operation and exploited in a DoS attack. Any
1972 application which performs certificate verification is vulnerable including
1973 OpenSSL clients and servers which enable client authentication.
1975 This issue was was reported to OpenSSL by Brian Carpenter.
1979 *) ASN.1 structure reuse memory corruption fix
1981 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1982 memory corruption via an invalid write. Such reuse is and has been
1983 strongly discouraged and is believed to be rare.
1985 Applications that parse structures containing CHOICE or ANY DEFINED BY
1986 components may be affected. Certificate parsing (d2i_X509 and related
1987 functions) are however not affected. OpenSSL clients and servers are
1992 *) PKCS7 NULL pointer dereferences fix
1994 The PKCS#7 parsing code does not handle missing outer ContentInfo
1995 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1996 missing content and trigger a NULL pointer dereference on parsing.
1998 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1999 otherwise parse PKCS#7 structures from untrusted sources are
2000 affected. OpenSSL clients and servers are not affected.
2002 This issue was reported to OpenSSL by Michal Zalewski (Google).
2006 *) DoS via reachable assert in SSLv2 servers fix
2008 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2009 servers that both support SSLv2 and enable export cipher suites by sending
2010 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2012 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2013 (OpenSSL development team).
2017 *) Empty CKE with client auth and DHE fix
2019 If client auth is used then a server can seg fault in the event of a DHE
2020 ciphersuite being selected and a zero length ClientKeyExchange message
2021 being sent by the client. This could be exploited in a DoS attack.
2025 *) Handshake with unseeded PRNG fix
2027 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2028 with an unseeded PRNG. The conditions are:
2029 - The client is on a platform where the PRNG has not been seeded
2030 automatically, and the user has not seeded manually
2031 - A protocol specific client method version has been used (i.e. not
2032 SSL_client_methodv23)
2033 - A ciphersuite is used that does not require additional random data from
2034 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2036 If the handshake succeeds then the client random that has been used will
2037 have been generated from a PRNG with insufficient entropy and therefore the
2038 output may be predictable.
2040 For example using the following command with an unseeded openssl will
2041 succeed on an unpatched platform:
2043 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2047 *) Use After Free following d2i_ECPrivatekey error fix
2049 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2050 could cause a use after free condition. This, in turn, could cause a double
2051 free in several private key parsing functions (such as d2i_PrivateKey
2052 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2053 for applications that receive EC private keys from untrusted
2054 sources. This scenario is considered rare.
2056 This issue was discovered by the BoringSSL project and fixed in their
2061 *) X509_to_X509_REQ NULL pointer deref fix
2063 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2064 the certificate key is invalid. This function is rarely used in practice.
2066 This issue was discovered by Brian Carpenter.
2070 *) Removed the export ciphers from the DEFAULT ciphers
2073 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2075 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2076 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2077 So far those who have to target multiple platforms would compromise
2078 and argue that binary targeting say ARMv5 would still execute on
2079 ARMv8. "Universal" build resolves this compromise by providing
2080 near-optimal performance even on newer platforms.
2083 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2084 (other platforms pending).
2085 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2087 *) Add support for the SignedCertificateTimestampList certificate and
2088 OCSP response extensions from RFC6962.
2091 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2092 for corner cases. (Certain input points at infinity could lead to
2093 bogus results, with non-infinity inputs mapped to infinity too.)
2096 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2097 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2098 common cases are optimized and there still is room for further
2099 improvements. Vector Permutation AES for Altivec is also added.
2102 *) Add support for little-endian ppc64 Linux target.
2103 [Marcelo Cerri (IBM)]
2105 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2106 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2107 are optimized and there still is room for further improvements.
2108 Both 32- and 64-bit modes are supported.
2109 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2111 *) Improved ARMv7 NEON support.
2114 *) Support for SPARC Architecture 2011 crypto extensions, first
2115 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2116 SHA256/512, MD5, GHASH and modular exponentiation.
2117 [Andy Polyakov, David Miller]
2119 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2121 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2123 *) Support for new and upcoming Intel processors, including AVX2,
2124 BMI and SHA ISA extensions. This includes additional "stitched"
2125 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2128 This work was sponsored by Intel Corp.
2131 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2132 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2133 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2136 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2137 this fixes a limitation in previous versions of OpenSSL.
2140 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2141 MGF1 digest and OAEP label.
2144 *) Add EVP support for key wrapping algorithms, to avoid problems with
2145 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2146 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2147 algorithms and include tests cases.
2150 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2152 [Douglas E. Engert, Steve Henson]
2154 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2155 difference in days and seconds between two tm or ASN1_TIME structures.
2158 *) Add -rev test option to s_server to just reverse order of characters
2159 received by client and send back to server. Also prints an abbreviated
2160 summary of the connection parameters.
2163 *) New option -brief for s_client and s_server to print out a brief summary
2164 of connection parameters.
2167 *) Add callbacks for arbitrary TLS extensions.
2168 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2170 *) New option -crl_download in several openssl utilities to download CRLs
2171 from CRLDP extension in certificates.
2174 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2177 *) New function X509_CRL_diff to generate a delta CRL from the difference
2178 of two full CRLs. Add support to "crl" utility.
2181 *) New functions to set lookup_crls function and to retrieve
2182 X509_STORE from X509_STORE_CTX.
2185 *) Print out deprecated issuer and subject unique ID fields in
2189 *) Extend OCSP I/O functions so they can be used for simple general purpose
2190 HTTP as well as OCSP. New wrapper function which can be used to download
2191 CRLs using the OCSP API.
2194 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2197 *) SSL_CONF* functions. These provide a common framework for application
2198 configuration using configuration files or command lines.
2201 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2202 message callback and prints the results. Needs compile time option
2203 "enable-ssl-trace". New options to s_client and s_server to enable
2207 *) New ctrl and macro to retrieve supported points extensions.
2208 Print out extension in s_server and s_client.
2211 *) New functions to retrieve certificate signature and signature
2215 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2219 *) New Suite B modes for TLS code. These use and enforce the requirements
2220 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2221 only use Suite B curves. The Suite B modes can be set by using the
2222 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2225 *) New chain verification flags for Suite B levels of security. Check
2226 algorithms are acceptable when flags are set in X509_verify_cert.
2229 *) Make tls1_check_chain return a set of flags indicating checks passed
2230 by a certificate chain. Add additional tests to handle client
2231 certificates: checks for matching certificate type and issuer name
2235 *) If an attempt is made to use a signature algorithm not in the peer
2236 preference list abort the handshake. If client has no suitable
2237 signature algorithms in response to a certificate request do not
2238 use the certificate.
2241 *) If server EC tmp key is not in client preference list abort handshake.
2244 *) Add support for certificate stores in CERT structure. This makes it
2245 possible to have different stores per SSL structure or one store in
2246 the parent SSL_CTX. Include distinct stores for certificate chain
2247 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2248 to build and store a certificate chain in CERT structure: returning
2249 an error if the chain cannot be built: this will allow applications
2250 to test if a chain is correctly configured.
2252 Note: if the CERT based stores are not set then the parent SSL_CTX
2253 store is used to retain compatibility with existing behaviour.
2257 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2258 mask based on the current session, check mask when sending client
2259 hello and checking the requested ciphersuite.
2262 *) New ctrls to retrieve and set certificate types in a certificate
2263 request message. Print out received values in s_client. If certificate
2264 types is not set with custom values set sensible values based on
2265 supported signature algorithms.
2268 *) Support for distinct client and server supported signature algorithms.
2271 *) Add certificate callback. If set this is called whenever a certificate
2272 is required by client or server. An application can decide which
2273 certificate chain to present based on arbitrary criteria: for example
2274 supported signature algorithms. Add very simple example to s_server.
2275 This fixes many of the problems and restrictions of the existing client
2276 certificate callback: for example you can now clear an existing
2277 certificate and specify the whole chain.
2280 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2281 the certificate can be used for (if anything). Set valid_flags field
2282 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2283 to have similar checks in it.
2285 Add new "cert_flags" field to CERT structure and include a "strict mode".
2286 This enforces some TLS certificate requirements (such as only permitting
2287 certificate signature algorithms contained in the supported algorithms
2288 extension) which some implementations ignore: this option should be used
2289 with caution as it could cause interoperability issues.
2292 *) Update and tidy signature algorithm extension processing. Work out
2293 shared signature algorithms based on preferences and peer algorithms
2294 and print them out in s_client and s_server. Abort handshake if no
2295 shared signature algorithms.
2298 *) Add new functions to allow customised supported signature algorithms
2299 for SSL and SSL_CTX structures. Add options to s_client and s_server
2303 *) New function SSL_certs_clear() to delete all references to certificates
2304 from an SSL structure. Before this once a certificate had been added
2305 it couldn't be removed.
2308 *) Integrate hostname, email address and IP address checking with certificate
2309 verification. New verify options supporting checking in openssl utility.
2312 *) Fixes and wildcard matching support to hostname and email checking
2313 functions. Add manual page.
2314 [Florian Weimer (Red Hat Product Security Team)]
2316 *) New functions to check a hostname email or IP address against a
2317 certificate. Add options x509 utility to print results of checks against
2321 *) Fix OCSP checking.
2322 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2324 *) Initial experimental support for explicitly trusted non-root CAs.
2325 OpenSSL still tries to build a complete chain to a root but if an
2326 intermediate CA has a trust setting included that is used. The first
2327 setting is used: whether to trust (e.g., -addtrust option to the x509
2331 *) Add -trusted_first option which attempts to find certificates in the
2332 trusted store even if an untrusted chain is also supplied.
2335 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2336 platform support for Linux and Android.
2339 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2342 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2343 When in FIPS mode the approved implementations are used as normal,
2344 when not in FIPS mode the internal unapproved versions are used instead.
2345 This means that the FIPS capable OpenSSL isn't forced to use the
2346 (often lower performance) FIPS implementations outside FIPS mode.
2349 *) Transparently support X9.42 DH parameters when calling
2350 PEM_read_bio_DHparameters. This means existing applications can handle
2351 the new parameter format automatically.
2354 *) Initial experimental support for X9.42 DH parameter format: mainly
2355 to support use of 'q' parameter for RFC5114 parameters.
2358 *) Add DH parameters from RFC5114 including test data to dhtest.
2361 *) Support for automatic EC temporary key parameter selection. If enabled
2362 the most preferred EC parameters are automatically used instead of
2363 hardcoded fixed parameters. Now a server just has to call:
2364 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2365 support ECDH and use the most appropriate parameters.
2368 *) Enhance and tidy EC curve and point format TLS extension code. Use
2369 static structures instead of allocation if default values are used.
2370 New ctrls to set curves we wish to support and to retrieve shared curves.
2371 Print out shared curves in s_server. New options to s_server and s_client
2372 to set list of supported curves.
2375 *) New ctrls to retrieve supported signature algorithms and
2376 supported curve values as an array of NIDs. Extend openssl utility
2377 to print out received values.
2380 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2381 between NIDs and the more common NIST names such as "P-256". Enhance
2382 ecparam utility and ECC method to recognise the NIST names for curves.
2385 *) Enhance SSL/TLS certificate chain handling to support different
2386 chains for each certificate instead of one chain in the parent SSL_CTX.
2389 *) Support for fixed DH ciphersuite client authentication: where both
2390 server and client use DH certificates with common parameters.
2393 *) Support for fixed DH ciphersuites: those requiring DH server
2397 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2399 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2400 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2401 X509_CINF_get_signature were reverted post internal team review.
2403 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2405 *) Build fixes for the Windows and OpenVMS platforms
2406 [Matt Caswell and Richard Levitte]
2408 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2410 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2411 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2412 dereference. This could lead to a Denial Of Service attack. Thanks to
2413 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2417 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2418 dtls1_buffer_record function under certain conditions. In particular this
2419 could occur if an attacker sent repeated DTLS records with the same
2420 sequence number but for the next epoch. The memory leak could be exploited
2421 by an attacker in a Denial of Service attack through memory exhaustion.
2422 Thanks to Chris Mueller for reporting this issue.
2426 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2427 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2428 method would be set to NULL which could later result in a NULL pointer
2429 dereference. Thanks to Frank Schmirler for reporting this issue.
2433 *) Abort handshake if server key exchange message is omitted for ephemeral
2436 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2437 reporting this issue.
2441 *) Remove non-export ephemeral RSA code on client and server. This code
2442 violated the TLS standard by allowing the use of temporary RSA keys in
2443 non-export ciphersuites and could be used by a server to effectively
2444 downgrade the RSA key length used to a value smaller than the server
2445 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2446 INRIA or reporting this issue.
2450 *) Fixed issue where DH client certificates are accepted without verification.
2451 An OpenSSL server will accept a DH certificate for client authentication
2452 without the certificate verify message. This effectively allows a client to
2453 authenticate without the use of a private key. This only affects servers
2454 which trust a client certificate authority which issues certificates
2455 containing DH keys: these are extremely rare and hardly ever encountered.
2456 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2461 *) Ensure that the session ID context of an SSL is updated when its
2462 SSL_CTX is updated via SSL_set_SSL_CTX.
2464 The session ID context is typically set from the parent SSL_CTX,
2465 and can vary with the CTX.
2468 *) Fix various certificate fingerprint issues.
2470 By using non-DER or invalid encodings outside the signed portion of a
2471 certificate the fingerprint can be changed without breaking the signature.
2472 Although no details of the signed portion of the certificate can be changed
2473 this can cause problems with some applications: e.g. those using the
2474 certificate fingerprint for blacklists.
2476 1. Reject signatures with non zero unused bits.
2478 If the BIT STRING containing the signature has non zero unused bits reject
2479 the signature. All current signature algorithms require zero unused bits.
2481 2. Check certificate algorithm consistency.
2483 Check the AlgorithmIdentifier inside TBS matches the one in the
2484 certificate signature. NB: this will result in signature failure
2485 errors for some broken certificates.
2487 Thanks to Konrad Kraszewski from Google for reporting this issue.
2489 3. Check DSA/ECDSA signatures use DER.
2491 Re-encode DSA/ECDSA signatures and compare with the original received
2492 signature. Return an error if there is a mismatch.
2494 This will reject various cases including garbage after signature
2495 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2496 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2497 (negative or with leading zeroes).
2499 Further analysis was conducted and fixes were developed by Stephen Henson
2500 of the OpenSSL core team.
2505 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2506 results on some platforms, including x86_64. This bug occurs at random
2507 with a very low probability, and is not known to be exploitable in any
2508 way, though its exact impact is difficult to determine. Thanks to Pieter
2509 Wuille (Blockstream) who reported this issue and also suggested an initial
2510 fix. Further analysis was conducted by the OpenSSL development team and
2511 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2512 the OpenSSL core team.
2516 *) Do not resume sessions on the server if the negotiated protocol
2517 version does not match the session's version. Resuming with a different
2518 version, while not strictly forbidden by the RFC, is of questionable
2519 sanity and breaks all known clients.
2520 [David Benjamin, Emilia Käsper]
2522 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2523 early CCS messages during renegotiation. (Note that because
2524 renegotiation is encrypted, this early CCS was not exploitable.)
2527 *) Tighten client-side session ticket handling during renegotiation:
2528 ensure that the client only accepts a session ticket if the server sends
2529 the extension anew in the ServerHello. Previously, a TLS client would
2530 reuse the old extension state and thus accept a session ticket if one was
2531 announced in the initial ServerHello.
2533 Similarly, ensure that the client requires a session ticket if one
2534 was advertised in the ServerHello. Previously, a TLS client would
2535 ignore a missing NewSessionTicket message.
2538 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2540 *) SRTP Memory Leak.
2542 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2543 sends a carefully crafted handshake message, to cause OpenSSL to fail
2544 to free up to 64k of memory causing a memory leak. This could be
2545 exploited in a Denial Of Service attack. This issue affects OpenSSL
2546 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2547 whether SRTP is used or configured. Implementations of OpenSSL that
2548 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2550 The fix was developed by the OpenSSL team.
2554 *) Session Ticket Memory Leak.
2556 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2557 integrity of that ticket is first verified. In the event of a session
2558 ticket integrity check failing, OpenSSL will fail to free memory
2559 causing a memory leak. By sending a large number of invalid session
2560 tickets an attacker could exploit this issue in a Denial Of Service
2565 *) Build option no-ssl3 is incomplete.
2567 When OpenSSL is configured with "no-ssl3" as a build option, servers
2568 could accept and complete a SSL 3.0 handshake, and clients could be
2569 configured to send them.
2571 [Akamai and the OpenSSL team]
2573 *) Add support for TLS_FALLBACK_SCSV.
2574 Client applications doing fallback retries should call
2575 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2577 [Adam Langley, Bodo Moeller]
2579 *) Add additional DigestInfo checks.
2581 Re-encode DigestInto in DER and check against the original when
2582 verifying RSA signature: this will reject any improperly encoded
2583 DigestInfo structures.
2585 Note: this is a precautionary measure and no attacks are currently known.
2589 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2591 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2592 SRP code can be overrun an internal buffer. Add sanity check that
2593 g, A, B < N to SRP code.
2595 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2596 Group for discovering this issue.
2600 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2601 TLS 1.0 instead of higher protocol versions when the ClientHello message
2602 is badly fragmented. This allows a man-in-the-middle attacker to force a
2603 downgrade to TLS 1.0 even if both the server and the client support a
2604 higher protocol version, by modifying the client's TLS records.
2606 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2607 researching this issue.
2611 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2612 to a denial of service attack. A malicious server can crash the client
2613 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2614 ciphersuite and sending carefully crafted handshake messages.
2616 Thanks to Felix Gröbert (Google) for discovering and researching this
2621 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2622 to leak memory. This can be exploited through a Denial of Service attack.
2623 Thanks to Adam Langley for discovering and researching this issue.
2627 *) An attacker can force openssl to consume large amounts of memory whilst
2628 processing DTLS handshake messages. This can be exploited through a
2629 Denial of Service attack.
2630 Thanks to Adam Langley for discovering and researching this issue.
2634 *) An attacker can force an error condition which causes openssl to crash
2635 whilst processing DTLS packets due to memory being freed twice. This
2636 can be exploited through a Denial of Service attack.
2637 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2642 *) If a multithreaded client connects to a malicious server using a resumed
2643 session and the server sends an ec point format extension it could write
2644 up to 255 bytes to freed memory.
2646 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2651 *) A malicious server can crash an OpenSSL client with a null pointer
2652 dereference (read) by specifying an SRP ciphersuite even though it was not
2653 properly negotiated with the client. This can be exploited through a
2654 Denial of Service attack.
2656 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2657 discovering and researching this issue.
2661 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2662 X509_name_oneline, X509_name_print_ex et al. to leak some information
2663 from the stack. Applications may be affected if they echo pretty printing
2664 output to the attacker.
2666 Thanks to Ivan Fratric (Google) for discovering this issue.
2668 [Emilia Käsper, and Steve Henson]
2670 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2671 for corner cases. (Certain input points at infinity could lead to
2672 bogus results, with non-infinity inputs mapped to infinity too.)
2675 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2677 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2678 handshake can force the use of weak keying material in OpenSSL
2679 SSL/TLS clients and servers.
2681 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2682 researching this issue. (CVE-2014-0224)
2683 [KIKUCHI Masashi, Steve Henson]
2685 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2686 OpenSSL DTLS client the code can be made to recurse eventually crashing
2689 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2691 [Imre Rad, Steve Henson]
2693 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2694 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2695 client or server. This is potentially exploitable to run arbitrary
2696 code on a vulnerable client or server.
2698 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2699 [Jüri Aedla, Steve Henson]
2701 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2702 are subject to a denial of service attack.
2704 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2705 this issue. (CVE-2014-3470)
2706 [Felix Gröbert, Ivan Fratric, Steve Henson]
2708 *) Harmonize version and its documentation. -f flag is used to display
2710 [mancha <mancha1@zoho.com>]
2712 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2713 in i2d_ECPrivateKey.
2714 [mancha <mancha1@zoho.com>]
2716 *) Fix some double frees. These are not thought to be exploitable.
2717 [mancha <mancha1@zoho.com>]
2719 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2721 *) A missing bounds check in the handling of the TLS heartbeat extension
2722 can be used to reveal up to 64k of memory to a connected client or
2725 Thanks for Neel Mehta of Google Security for discovering this bug and to
2726 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2727 preparing the fix (CVE-2014-0160)
2728 [Adam Langley, Bodo Moeller]
2730 *) Fix for the attack described in the paper "Recovering OpenSSL
2731 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2732 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2733 http://eprint.iacr.org/2014/140
2735 Thanks to Yuval Yarom and Naomi Benger for discovering this
2736 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2737 [Yuval Yarom and Naomi Benger]
2739 *) TLS pad extension: draft-agl-tls-padding-03
2741 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2742 TLS client Hello record length value would otherwise be > 255 and
2743 less that 512 pad with a dummy extension containing zeroes so it
2744 is at least 512 bytes long.
2746 [Adam Langley, Steve Henson]
2748 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2750 *) Fix for TLS record tampering bug. A carefully crafted invalid
2751 handshake could crash OpenSSL with a NULL pointer exception.
2752 Thanks to Anton Johansson for reporting this issues.
2755 *) Keep original DTLS digest and encryption contexts in retransmission
2756 structures so we can use the previous session parameters if they need
2757 to be resent. (CVE-2013-6450)
2760 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2761 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2762 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2763 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2764 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2765 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2766 [Rob Stradling, Adam Langley]
2768 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2770 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2771 supporting platforms or when small records were transferred.
2772 [Andy Polyakov, Steve Henson]
2774 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2776 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2778 This addresses the flaw in CBC record processing discovered by
2779 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2780 at: http://www.isg.rhul.ac.uk/tls/
2782 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2783 Security Group at Royal Holloway, University of London
2784 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2785 Emilia Käsper for the initial patch.
2787 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2789 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2790 ciphersuites which can be exploited in a denial of service attack.
2791 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2792 and detecting this bug and to Wolfgang Ettlinger
2793 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2797 *) Return an error when checking OCSP signatures when key is NULL.
2798 This fixes a DoS attack. (CVE-2013-0166)
2801 *) Make openssl verify return errors.
2802 [Chris Palmer <palmer@google.com> and Ben Laurie]
2804 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2805 the right response is stapled. Also change SSL_get_certificate()
2806 so it returns the certificate actually sent.
2807 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2808 [Rob Stradling <rob.stradling@comodo.com>]
2810 *) Fix possible deadlock when decoding public keys.
2813 *) Don't use TLS 1.0 record version number in initial client hello
2817 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2819 *) Sanity check record length before skipping explicit IV in TLS
2820 1.2, 1.1 and DTLS to fix DoS attack.
2822 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2823 fuzzing as a service testing platform.
2827 *) Initialise tkeylen properly when encrypting CMS messages.
2828 Thanks to Solar Designer of Openwall for reporting this issue.
2831 *) In FIPS mode don't try to use composite ciphers as they are not
2835 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2837 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2838 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2839 mean any application compiled against OpenSSL 1.0.0 headers setting
2840 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2841 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2842 0x10000000L Any application which was previously compiled against
2843 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2844 will need to be recompiled as a result. Letting be results in
2845 inability to disable specifically TLS 1.1 and in client context,
2846 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2849 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2850 disable just protocol X, but all protocols above X *if* there are
2851 protocols *below* X still enabled. In more practical terms it means
2852 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2853 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2854 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2858 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2860 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2861 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2862 in CRYPTO_realloc_clean.
2864 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2865 issue and to Adam Langley <agl@chromium.org> for fixing it.
2867 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2869 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2872 *) Workarounds for some broken servers that "hang" if a client hello
2873 record length exceeds 255 bytes.
2875 1. Do not use record version number > TLS 1.0 in initial client
2876 hello: some (but not all) hanging servers will now work.
2877 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2878 the number of ciphers sent in the client hello. This should be
2879 set to an even number, such as 50, for example by passing:
2880 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2881 Most broken servers should now work.
2882 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2883 TLS 1.2 client support entirely.
2886 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2889 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2891 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2892 STRING form instead of a DigestInfo.
2895 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2896 and the RSA_sign/RSA_verify functions. This was made more apparent when
2897 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2898 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2899 the correct format in RSA_verify so both forms transparently work.
2902 *) Some servers which support TLS 1.0 can choke if we initially indicate
2903 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2904 encrypted premaster secret. As a workaround use the maximum permitted
2905 client version in client hello, this should keep such servers happy
2906 and still work with previous versions of OpenSSL.
2909 *) Add support for TLS/DTLS heartbeats.
2910 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2912 *) Add support for SCTP.
2913 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2915 *) Improved PRNG seeding for VOS.
2916 [Paul Green <Paul.Green@stratus.com>]
2918 *) Extensive assembler packs updates, most notably:
2920 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2921 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2922 - x86_64: bit-sliced AES implementation;
2923 - ARM: NEON support, contemporary platforms optimizations;
2924 - s390x: z196 support;
2925 - *: GHASH and GF(2^m) multiplication implementations;
2929 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2930 (removal of unnecessary code)
2931 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2933 *) Add TLS key material exporter from RFC 5705.
2936 *) Add DTLS-SRTP negotiation from RFC 5764.
2939 *) Add Next Protocol Negotiation,
2940 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2941 disabled with a no-npn flag to config or Configure. Code donated
2943 [Adam Langley <agl@google.com> and Ben Laurie]
2945 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2946 NIST-P256, NIST-P521, with constant-time single point multiplication on
2947 typical inputs. Compiler support for the nonstandard type __uint128_t is
2948 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2949 Code made available under Apache License version 2.0.
2951 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2952 line to include this in your build of OpenSSL, and run "make depend" (or
2953 "make update"). This enables the following EC_METHODs:
2955 EC_GFp_nistp224_method()
2956 EC_GFp_nistp256_method()
2957 EC_GFp_nistp521_method()
2959 EC_GROUP_new_by_curve_name() will automatically use these (while
2960 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2962 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2964 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2965 all platforms. Move ssize_t definition from e_os.h to the public
2966 header file e_os2.h as it now appears in public header file cms.h
2969 *) New -sigopt option to the ca, req and x509 utilities. Additional
2970 signature parameters can be passed using this option and in
2974 *) Add RSA PSS signing function. This will generate and set the
2975 appropriate AlgorithmIdentifiers for PSS based on those in the
2976 corresponding EVP_MD_CTX structure. No application support yet.
2979 *) Support for companion algorithm specific ASN1 signing routines.
2980 New function ASN1_item_sign_ctx() signs a pre-initialised
2981 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2982 the appropriate parameters.
2985 *) Add new algorithm specific ASN1 verification initialisation function
2986 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2987 handling will be the same no matter what EVP_PKEY_METHOD is used.
2988 Add a PSS handler to support verification of PSS signatures: checked
2989 against a number of sample certificates.
2992 *) Add signature printing for PSS. Add PSS OIDs.
2993 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2995 *) Add algorithm specific signature printing. An individual ASN1 method
2996 can now print out signatures instead of the standard hex dump.
2998 More complex signatures (e.g. PSS) can print out more meaningful
2999 information. Include DSA version that prints out the signature
3003 *) Password based recipient info support for CMS library: implementing
3007 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3008 neatly separates the code into cipher and PBE sections and is required
3009 for some algorithms that split PBES2 into separate pieces (such as
3010 password based CMS).
3013 *) Session-handling fixes:
3014 - Fix handling of connections that are resuming with a session ID,
3015 but also support Session Tickets.
3016 - Fix a bug that suppressed issuing of a new ticket if the client
3017 presented a ticket with an expired session.
3018 - Try to set the ticket lifetime hint to something reasonable.
3019 - Make tickets shorter by excluding irrelevant information.
3020 - On the client side, don't ignore renewed tickets.
3021 [Adam Langley, Bodo Moeller (Google)]
3023 *) Fix PSK session representation.
3026 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3028 This work was sponsored by Intel.
3031 *) Add GCM support to TLS library. Some custom code is needed to split
3032 the IV between the fixed (from PRF) and explicit (from TLS record)
3033 portions. This adds all GCM ciphersuites supported by RFC5288 and
3034 RFC5289. Generalise some AES* cipherstrings to include GCM and
3035 add a special AESGCM string for GCM only.
3038 *) Expand range of ctrls for AES GCM. Permit setting invocation
3039 field on decrypt and retrieval of invocation field only on encrypt.
3042 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3043 As required by RFC5289 these ciphersuites cannot be used if for
3044 versions of TLS earlier than 1.2.
3047 *) For FIPS capable OpenSSL interpret a NULL default public key method
3048 as unset and return the appropriate default but do *not* set the default.
3049 This means we can return the appropriate method in applications that
3050 switch between FIPS and non-FIPS modes.
3053 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3054 ENGINE is used then we cannot handle that in the FIPS module so we
3055 keep original code iff non-FIPS operations are allowed.
3058 *) Add -attime option to openssl utilities.
3059 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3061 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3064 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3065 FIPS EC methods unconditionally for now.
3068 *) New build option no-ec2m to disable characteristic 2 code.
3071 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3072 all cases can be covered as some introduce binary incompatibilities.
3075 *) Redirect RSA operations to FIPS module including keygen,
3076 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3079 *) Add similar low level API blocking to ciphers.
3082 *) Low level digest APIs are not approved in FIPS mode: any attempt
3083 to use these will cause a fatal error. Applications that *really* want
3084 to use them can use the private_* version instead.
3087 *) Redirect cipher operations to FIPS module for FIPS builds.
3090 *) Redirect digest operations to FIPS module for FIPS builds.
3093 *) Update build system to add "fips" flag which will link in fipscanister.o
3094 for static and shared library builds embedding a signature if needed.
3097 *) Output TLS supported curves in preference order instead of numerical
3098 order. This is currently hardcoded for the highest order curves first.
3099 This should be configurable so applications can judge speed vs strength.
3102 *) Add TLS v1.2 server support for client authentication.
3105 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3109 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3110 FIPS modules versions.
3113 *) Add TLS v1.2 client side support for client authentication. Keep cache
3114 of handshake records longer as we don't know the hash algorithm to use
3115 until after the certificate request message is received.
3118 *) Initial TLS v1.2 client support. Add a default signature algorithms
3119 extension including all the algorithms we support. Parse new signature
3120 format in client key exchange. Relax some ECC signing restrictions for
3121 TLS v1.2 as indicated in RFC5246.
3124 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3125 to new signature format when needed using client digest preference.
3126 All server ciphersuites should now work correctly in TLS v1.2. No client
3127 support yet and no support for client certificates.
3130 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3131 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3132 ciphersuites. At present only RSA key exchange ciphersuites work with
3133 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3134 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3135 and version checking.
3138 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3139 with this defined it will not be affected by any changes to ssl internal
3140 structures. Add several utility functions to allow openssl application
3141 to work with OPENSSL_NO_SSL_INTERN defined.
3144 *) A long standing patch to add support for SRP from EdelWeb (Peter
3145 Sylvester and Christophe Renou) was integrated.
3146 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3147 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3150 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3153 *) Permit abbreviated handshakes when renegotiating using the function
3154 SSL_renegotiate_abbreviated().
3155 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3157 *) Add call to ENGINE_register_all_complete() to
3158 ENGINE_load_builtin_engines(), so some implementations get used
3159 automatically instead of needing explicit application support.
3162 *) Add support for TLS key exporter as described in RFC5705.
3163 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3165 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3166 a few changes are required:
3168 Add SSL_OP_NO_TLSv1_1 flag.
3169 Add TLSv1_1 methods.
3170 Update version checking logic to handle version 1.1.
3171 Add explicit IV handling (ported from DTLS code).
3172 Add command line options to s_client/s_server.
3175 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3177 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3178 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3179 content decryption and always return the same error. Note: this attack
3180 needs on average 2^20 messages so it only affects automated senders. The
3181 old behaviour can be re-enabled in the CMS code by setting the
3182 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3183 an MMA defence is not necessary.
3184 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3185 this issue. (CVE-2012-0884)
3188 *) Fix CVE-2011-4619: make sure we really are receiving a
3189 client hello before rejecting multiple SGC restarts. Thanks to
3190 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3193 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3195 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3196 Thanks to Antonio Martin, Enterprise Secure Access Research and
3197 Development, Cisco Systems, Inc. for discovering this bug and
3198 preparing a fix. (CVE-2012-0050)
3201 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3203 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3204 of the Vaudenay padding oracle attack on CBC mode encryption
3205 which enables an efficient plaintext recovery attack against
3206 the OpenSSL implementation of DTLS. Their attack exploits timing
3207 differences arising during decryption processing. A research
3208 paper describing this attack can be found at:
3209 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3210 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3211 Security Group at Royal Holloway, University of London
3212 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3213 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3214 for preparing the fix. (CVE-2011-4108)
3215 [Robin Seggelmann, Michael Tuexen]
3217 *) Clear bytes used for block padding of SSL 3.0 records.
3219 [Adam Langley (Google)]
3221 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3222 Kadianakis <desnacked@gmail.com> for discovering this issue and
3223 Adam Langley for preparing the fix. (CVE-2011-4619)
3224 [Adam Langley (Google)]
3226 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3227 [Andrey Kulikov <amdeich@gmail.com>]
3229 *) Prevent malformed RFC3779 data triggering an assertion failure.
3230 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3231 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3232 [Rob Austein <sra@hactrn.net>]
3234 *) Improved PRNG seeding for VOS.
3235 [Paul Green <Paul.Green@stratus.com>]
3237 *) Fix ssl_ciph.c set-up race.
3238 [Adam Langley (Google)]
3240 *) Fix spurious failures in ecdsatest.c.
3241 [Emilia Käsper (Google)]
3243 *) Fix the BIO_f_buffer() implementation (which was mixing different
3244 interpretations of the '..._len' fields).
3245 [Adam Langley (Google)]
3247 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3248 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3249 threads won't reuse the same blinding coefficients.
3251 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3252 lock to call BN_BLINDING_invert_ex, and avoids one use of
3253 BN_BLINDING_update for each BN_BLINDING structure (previously,
3254 the last update always remained unused).
3255 [Emilia Käsper (Google)]
3257 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3258 [Bob Buckholz (Google)]
3260 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3262 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3263 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3264 [Kaspar Brand <ossl@velox.ch>]
3266 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3267 for multi-threaded use of ECDH. (CVE-2011-3210)
3268 [Adam Langley (Google)]
3270 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3273 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3274 signature public key algorithm by using OID xref utilities instead.
3275 Before this you could only use some ECC ciphersuites with SHA1 only.
3278 *) Add protection against ECDSA timing attacks as mentioned in the paper
3279 by Billy Bob Brumley and Nicola Tuveri, see:
3281 http://eprint.iacr.org/2011/232.pdf
3283 [Billy Bob Brumley and Nicola Tuveri]
3285 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3287 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3288 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3290 *) Fix bug in string printing code: if *any* escaping is enabled we must
3291 escape the escape character (backslash) or the resulting string is
3295 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3297 *) Disable code workaround for ancient and obsolete Netscape browsers
3298 and servers: an attacker can use it in a ciphersuite downgrade attack.
3299 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3302 *) Fixed J-PAKE implementation error, originally discovered by
3303 Sebastien Martini, further info and confirmation from Stefan
3304 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3307 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3309 *) Fix extension code to avoid race conditions which can result in a buffer
3310 overrun vulnerability: resumed sessions must not be modified as they can
3311 be shared by multiple threads. CVE-2010-3864
3314 *) Fix WIN32 build system to correctly link an ENGINE directory into
3318 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3320 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3322 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3324 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3326 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3327 context. The operation can be customised via the ctrl mechanism in
3328 case ENGINEs want to include additional functionality.
3331 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3334 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3335 output hashes compatible with older versions of OpenSSL.
3336 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3338 *) Fix compression algorithm handling: if resuming a session use the
3339 compression algorithm of the resumed session instead of determining
3340 it from client hello again. Don't allow server to change algorithm.
3343 *) Add load_crls() function to apps tidying load_certs() too. Add option
3344 to verify utility to allow additional CRLs to be included.
3347 *) Update OCSP request code to permit adding custom headers to the request:
3348 some responders need this.
3351 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3353 [Julia Lawall <julia@diku.dk>]
3355 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3356 needlessly dereferenced structures, used obsolete functions and
3357 didn't handle all updated verify codes correctly.
3360 *) Disable MD2 in the default configuration.
3363 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3364 indicate the initial BIO being pushed or popped. This makes it possible
3365 to determine whether the BIO is the one explicitly called or as a result
3366 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3367 it handles reference counts correctly and doesn't zero out the I/O bio
3368 when it is not being explicitly popped. WARNING: applications which
3369 included workarounds for the old buggy behaviour will need to be modified
3370 or they could free up already freed BIOs.
3373 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3374 renaming to all platforms (within the 0.9.8 branch, this was
3375 done conditionally on Netware platforms to avoid a name clash).
3376 [Guenter <lists@gknw.net>]
3378 *) Add ECDHE and PSK support to DTLS.
3379 [Michael Tuexen <tuexen@fh-muenster.de>]
3381 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3385 *) Add "missing" function EVP_MD_flags() (without this the only way to
3386 retrieve a digest flags is by accessing the structure directly. Update
3387 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3388 or cipher is registered as in the "from" argument. Print out all
3389 registered digests in the dgst usage message instead of manually
3390 attempting to work them out.
3393 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3394 this allows the use of compression and extensions. Change default cipher
3395 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3396 by default unless an application cipher string requests it.
3399 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3400 key ids to find matching certificates and keys but some PKCS#12 files
3401 don't follow the (somewhat unwritten) rules and this strategy fails.
3402 Now just gather all certificates together and the first private key
3403 then look for the first certificate that matches the key.
3406 *) Support use of registered digest and cipher names for dgst and cipher
3407 commands instead of having to add each one as a special case. So now
3414 openssl dgst -sha256 foo
3416 and this works for ENGINE based algorithms too.
3420 *) Update Gost ENGINE to support parameter files.
3421 [Victor B. Wagner <vitus@cryptocom.ru>]
3423 *) Support GeneralizedTime in ca utility.
3424 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3426 *) Enhance the hash format used for certificate directory links. The new
3427 form uses the canonical encoding (meaning equivalent names will work
3428 even if they aren't identical) and uses SHA1 instead of MD5. This form
3429 is incompatible with the older format and as a result c_rehash should
3430 be used to rebuild symbolic links.
3433 *) Make PKCS#8 the default write format for private keys, replacing the
3434 traditional format. This form is standardised, more secure and doesn't
3435 include an implicit MD5 dependency.
3438 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3439 committed to OpenSSL should pass this lot as a minimum.
3442 *) Add session ticket override functionality for use by EAP-FAST.
3443 [Jouni Malinen <j@w1.fi>]
3445 *) Modify HMAC functions to return a value. Since these can be implemented
3446 in an ENGINE errors can occur.
3449 *) Type-checked OBJ_bsearch_ex.
3452 *) Type-checked OBJ_bsearch. Also some constification necessitated
3453 by type-checking. Still to come: TXT_DB, bsearch(?),
3454 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3458 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3459 seconds to a tm structure directly, instead of going through OS
3460 specific date routines. This avoids any issues with OS routines such
3461 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3462 and X509_time_adj_ex() to cover the extended range. The existing
3463 X509_time_adj() is still usable and will no longer have any date issues.
3466 *) Delta CRL support. New use deltas option which will attempt to locate
3467 and search any appropriate delta CRLs available.
3469 This work was sponsored by Google.
3472 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3473 code and add additional score elements. Validate alternate CRL paths
3474 as part of the CRL checking and indicate a new error "CRL path validation
3475 error" in this case. Applications wanting additional details can use
3476 the verify callback and check the new "parent" field. If this is not
3477 NULL CRL path validation is taking place. Existing applications won't
3478 see this because it requires extended CRL support which is off by
3481 This work was sponsored by Google.
3484 *) Support for freshest CRL extension.
3486 This work was sponsored by Google.
3489 *) Initial indirect CRL support. Currently only supported in the CRLs
3490 passed directly and not via lookup. Process certificate issuer
3491 CRL entry extension and lookup CRL entries by bother issuer name
3492 and serial number. Check and process CRL issuer entry in IDP extension.
3494 This work was sponsored by Google.
3497 *) Add support for distinct certificate and CRL paths. The CRL issuer
3498 certificate is validated separately in this case. Only enabled if
3499 an extended CRL support flag is set: this flag will enable additional
3500 CRL functionality in future.
3502 This work was sponsored by Google.
3505 *) Add support for policy mappings extension.
3507 This work was sponsored by Google.
3510 *) Fixes to pathlength constraint, self issued certificate handling,
3511 policy processing to align with RFC3280 and PKITS tests.
3513 This work was sponsored by Google.
3516 *) Support for name constraints certificate extension. DN, email, DNS
3517 and URI types are currently supported.
3519 This work was sponsored by Google.
3522 *) To cater for systems that provide a pointer-based thread ID rather
3523 than numeric, deprecate the current numeric thread ID mechanism and
3524 replace it with a structure and associated callback type. This
3525 mechanism allows a numeric "hash" to be extracted from a thread ID in
3526 either case, and on platforms where pointers are larger than 'long',
3527 mixing is done to help ensure the numeric 'hash' is usable even if it
3528 can't be guaranteed unique. The default mechanism is to use "&errno"
3529 as a pointer-based thread ID to distinguish between threads.
3531 Applications that want to provide their own thread IDs should now use
3532 CRYPTO_THREADID_set_callback() to register a callback that will call
3533 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3535 Note that ERR_remove_state() is now deprecated, because it is tied
3536 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3537 to free the current thread's error state should be replaced by
3538 ERR_remove_thread_state(NULL).
3540 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3541 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3542 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3543 application was previously providing a numeric thread callback that
3544 was inappropriate for distinguishing threads, then uniqueness might
3545 have been obtained with &errno that happened immediately in the
3546 intermediate development versions of OpenSSL; this is no longer the
3547 case, the numeric thread callback will now override the automatic use
3549 [Geoff Thorpe, with help from Bodo Moeller]
3551 *) Initial support for different CRL issuing certificates. This covers a
3552 simple case where the self issued certificates in the chain exist and
3553 the real CRL issuer is higher in the existing chain.
3555 This work was sponsored by Google.
3558 *) Removed effectively defunct crypto/store from the build.
3561 *) Revamp of STACK to provide stronger type-checking. Still to come:
3562 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3563 ASN1_STRING, CONF_VALUE.
3566 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3567 RAM on SSL connections. This option can save about 34k per idle SSL.
3570 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3571 STACK, TXT_DB, bsearch, qsort.
3574 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3575 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3576 support for data, signedData, compressedData, digestedData and
3577 encryptedData, envelopedData types included. Scripts to check against
3578 RFC4134 examples draft and interop and consistency checks of many
3579 content types and variants.
3582 *) Add options to enc utility to support use of zlib compression BIO.
3585 *) Extend mk1mf to support importing of options and assembly language
3586 files from Configure script, currently only included in VC-WIN32.
3587 The assembly language rules can now optionally generate the source
3588 files from the associated perl scripts.
3591 *) Implement remaining functionality needed to support GOST ciphersuites.
3592 Interop testing has been performed using CryptoPro implementations.
3593 [Victor B. Wagner <vitus@cryptocom.ru>]
3595 *) s390x assembler pack.
3598 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3602 *) Implement Opaque PRF Input TLS extension as specified in
3603 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3604 official specification yet and no extension type assignment by
3605 IANA exists, this extension (for now) will have to be explicitly
3606 enabled when building OpenSSL by providing the extension number
3607 to use. For example, specify an option
3609 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3611 to the "config" or "Configure" script to enable the extension,
3612 assuming extension number 0x9527 (which is a completely arbitrary
3613 and unofficial assignment based on the MD5 hash of the Internet
3614 Draft). Note that by doing so, you potentially lose
3615 interoperability with other TLS implementations since these might
3616 be using the same extension number for other purposes.
3618 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3619 opaque PRF input value to use in the handshake. This will create
3620 an interal copy of the length-'len' string at 'src', and will
3621 return non-zero for success.
3623 To get more control and flexibility, provide a callback function
3626 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3627 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3631 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3634 Callback function 'cb' will be called in handshakes, and is
3635 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3636 Argument 'arg' is for application purposes (the value as given to
3637 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3638 be provided to the callback function). The callback function
3639 has to return non-zero to report success: usually 1 to use opaque
3640 PRF input just if possible, or 2 to enforce use of the opaque PRF
3641 input. In the latter case, the library will abort the handshake
3642 if opaque PRF input is not successfully negotiated.
3644 Arguments 'peerinput' and 'len' given to the callback function
3645 will always be NULL and 0 in the case of a client. A server will
3646 see the client's opaque PRF input through these variables if
3647 available (NULL and 0 otherwise). Note that if the server
3648 provides an opaque PRF input, the length must be the same as the
3649 length of the client's opaque PRF input.
3651 Note that the callback function will only be called when creating
3652 a new session (session resumption can resume whatever was
3653 previously negotiated), and will not be called in SSL 2.0
3654 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3655 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3656 for applications that need to enforce opaque PRF input.
3660 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3663 [Victor B. Wagner <vitus@cryptocom.ru>]
3665 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3666 RFC4507bis. The encrypted ticket format is an encrypted encoded
3667 SSL_SESSION structure, that way new session features are automatically
3670 If a client application caches session in an SSL_SESSION structure
3671 support is transparent because tickets are now stored in the encoded
3674 The SSL_CTX structure automatically generates keys for ticket
3675 protection in servers so again support should be possible
3676 with no application modification.
3678 If a client or server wishes to disable RFC4507 support then the option
3679 SSL_OP_NO_TICKET can be set.
3681 Add a TLS extension debugging callback to allow the contents of any client
3682 or server extensions to be examined.
3684 This work was sponsored by Google.
3687 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3688 OpenSSL should now compile cleanly on gcc 4.2
3689 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3691 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3692 support including streaming MAC support: this is required for GOST
3693 ciphersuite support.
3694 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3696 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3697 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3698 to output in BER and PEM format.
3701 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3702 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3703 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3704 ENGINE support for HMAC keys which are unextractable. New -mac and
3705 -macopt options to dgst utility.
3708 *) New option -sigopt to dgst utility. Update dgst to use
3709 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3710 alternative signing parameters such as X9.31 or PSS in the dgst
3714 *) Change ssl_cipher_apply_rule(), the internal function that does
3715 the work each time a ciphersuite string requests enabling
3716 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3717 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3718 the order of disabled ciphersuites such that those ciphersuites
3719 that most recently went from enabled to disabled not only stay
3720 in order with respect to each other, but also have higher priority
3721 than other disabled ciphersuites the next time ciphersuites are
3724 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3725 the same ciphersuites as with "HIGH" alone, but in a specific
3726 order where the PSK ciphersuites come first (since they are the
3727 most recently disabled ciphersuites when "HIGH" is parsed).
3729 Also, change ssl_create_cipher_list() (using this new
3730 funcionality) such that between otherwise identical
3731 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3735 *) Change ssl_create_cipher_list() so that it automatically
3736 arranges the ciphersuites in reasonable order before starting
3737 to process the rule string. Thus, the definition for "DEFAULT"
3738 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3739 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3740 This makes it much easier to arrive at a reasonable default order
3741 in applications for which anonymous ciphers are OK (meaning
3742 that you can't actually use DEFAULT).
3743 [Bodo Moeller; suggested by Victor Duchovni]
3745 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3746 processing) into multiple integers instead of setting
3747 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3748 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3749 (These masks as well as the individual bit definitions are hidden
3750 away into the non-exported interface ssl/ssl_locl.h, so this
3751 change to the definition of the SSL_CIPHER structure shouldn't
3752 affect applications.) This give us more bits for each of these
3753 categories, so there is no longer a need to coagulate AES128 and
3754 AES256 into a single algorithm bit, and to coagulate Camellia128
3755 and Camellia256 into a single algorithm bit, which has led to all
3758 Thus, among other things, the kludge introduced in 0.9.7m and
3759 0.9.8e for masking out AES256 independently of AES128 or masking
3760 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3762 With the change, we also introduce new ciphersuite aliases that
3763 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3767 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3768 Use the leftmost N bytes of the signature input if the input is
3769 larger than the prime q (with N being the size in bytes of q).
3772 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3773 it yet and it is largely untested.
3776 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3779 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3780 some compilers (gcc 4.2 and later) reject their use. Safestack is
3781 reimplemented. Update ASN1 to avoid use of legacy functions.
3784 *) Win32/64 targets are linked with Winsock2.
3787 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3788 to external functions. This can be used to increase CRL handling
3789 efficiency especially when CRLs are very large by (for example) storing
3790 the CRL revoked certificates in a database.
3793 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3794 new CRLs added to a directory can be used. New command line option
3795 -verify_return_error to s_client and s_server. This causes real errors
3796 to be returned by the verify callback instead of carrying on no matter
3797 what. This reflects the way a "real world" verify callback would behave.
3800 *) GOST engine, supporting several GOST algorithms and public key formats.
3801 Kindly donated by Cryptocom.
3804 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3805 partitioned by DP are handled but no indirect CRL or reason partitioning
3806 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3807 selected via a scoring technique which handles IDP and AKID in CRLs.
3810 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3811 will ultimately be used for all verify operations: this will remove the
3812 X509_STORE dependency on certificate verification and allow alternative
3813 lookup methods. X509_STORE based implementations of these two callbacks.
3816 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3817 Modify get_crl() to find a valid (unexpired) CRL if possible.
3820 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3821 this would be called X509_CRL_cmp() but that name is already used by
3822 a function that just compares CRL issuer names. Cache several CRL
3823 extensions in X509_CRL structure and cache CRLDP in X509.
3826 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3827 this maps equivalent X509_NAME structures into a consistent structure.
3828 Name comparison can then be performed rapidly using memcmp().
3831 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3835 *) Allow digests to supply their own micalg string for S/MIME type using
3836 the ctrl EVP_MD_CTRL_MICALG.
3839 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3840 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3841 ctrl. It can then customise the structure before and/or after signing
3845 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3846 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3847 to free up any added signature OIDs.
3850 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3851 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3852 digest and cipher tables. New options added to openssl utility:
3853 list-message-digest-algorithms and list-cipher-algorithms.
3856 *) Change the array representation of binary polynomials: the list
3857 of degrees of non-zero coefficients is now terminated with -1.
3858 Previously it was terminated with 0, which was also part of the
3859 value; thus, the array representation was not applicable to
3860 polynomials where t^0 has coefficient zero. This change makes
3861 the array representation useful in a more general context.
3864 *) Various modifications and fixes to SSL/TLS cipher string
3865 handling. For ECC, the code now distinguishes between fixed ECDH
3866 with RSA certificates on the one hand and with ECDSA certificates
3867 on the other hand, since these are separate ciphersuites. The
3868 unused code for Fortezza ciphersuites has been removed.
3870 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3871 (not "ECDHE"). For consistency with the code for DH
3872 certificates, use of ECDH certificates is now considered ECDH
3873 authentication, not RSA or ECDSA authentication (the latter is
3874 merely the CA's signing algorithm and not actively used in the
3877 The temporary ciphersuite alias "ECCdraft" is no longer
3878 available, and ECC ciphersuites are no longer excluded from "ALL"
3879 and "DEFAULT". The following aliases now exist for RFC 4492
3880 ciphersuites, most of these by analogy with the DH case:
3882 kECDHr - ECDH cert, signed with RSA
3883 kECDHe - ECDH cert, signed with ECDSA
3884 kECDH - ECDH cert (signed with either RSA or ECDSA)
3885 kEECDH - ephemeral ECDH
3886 ECDH - ECDH cert or ephemeral ECDH
3892 AECDH - anonymous ECDH
3893 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3897 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3898 Use correct micalg parameters depending on digest(s) in signed message.
3901 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3902 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3905 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3906 an engine to register a method. Add ENGINE lookups for methods and
3907 functional reference processing.
3910 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3911 EVP_{Sign,Verify}* which allow an application to customise the signature
3915 *) New -resign option to smime utility. This adds one or more signers
3916 to an existing PKCS#7 signedData structure. Also -md option to use an
3917 alternative message digest algorithm for signing.
3920 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3921 create PKCS7 structures containing multiple signers. Update smime
3922 application to support multiple signers.
3925 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3929 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3930 Reorganize PBE internals to lookup from a static table using NIDs,
3931 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3932 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3933 PRF which will be automatically used with PBES2.
3936 *) Replace the algorithm specific calls to generate keys in "req" with the
3940 *) Update PKCS#7 enveloped data routines to use new API. This is now
3941 supported by any public key method supporting the encrypt operation. A
3942 ctrl is added to allow the public key algorithm to examine or modify
3943 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3947 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3948 a default digest type to use. In most cases this will be SHA1 but some
3949 algorithms (such as GOST) need to specify an alternative digest. The
3950 return value indicates how strong the preference is 1 means optional and
3951 2 is mandatory (that is it is the only supported type). Modify
3952 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3953 use the default md. Update openssl utilities to use the default digest
3954 type for signing if it is not explicitly indicated.
3957 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3958 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3959 signing method from the key type. This effectively removes the link
3960 between digests and public key types.
3963 *) Add an OID cross reference table and utility functions. Its purpose is to
3964 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3965 rsaEncryption. This will allow some of the algorithm specific hackery
3966 needed to use the correct OID to be removed.
3969 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3970 structures for PKCS7_sign(). They are now set up by the relevant public
3974 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3977 *) Add support for key derivation (agreement) in the API, DH method and
3981 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3982 public and private key formats. As a side effect these add additional
3983 command line functionality not previously available: DSA signatures can be
3984 generated and verified using pkeyutl and DH key support and generation in
3989 [Oliver Tappe <zooey@hirschkaefer.de>]
3991 *) New make target "install_html_docs" installs HTML renditions of the
3993 [Oliver Tappe <zooey@hirschkaefer.de>]
3995 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3996 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3997 support key and parameter generation and add initial key generation
3998 functionality for RSA.
4001 *) Add functions for main EVP_PKEY_method operations. The undocumented
4002 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4003 EVP_PKEY_{encrypt,decrypt}_old.
4006 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4007 key API, doesn't do much yet.
4010 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4011 public key algorithms. New option to openssl utility:
4012 "list-public-key-algorithms" to print out info.
4015 *) Implement the Supported Elliptic Curves Extension for
4016 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4019 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4020 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4023 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4024 utilities such as rsa, dsa, dsaparam etc except they process any key
4028 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4029 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4030 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4034 *) Initial support for pluggable public key ASN1.
4035 De-spaghettify the public key ASN1 handling. Move public and private
4036 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4037 algorithm specific handling to a single module within the relevant
4038 algorithm directory. Add functions to allow (near) opaque processing
4039 of public and private key structures.
4042 *) Implement the Supported Point Formats Extension for
4043 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4046 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4047 for the psk identity [hint] and the psk callback functions to the
4048 SSL_SESSION, SSL and SSL_CTX structure.
4051 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4055 SSL_CTX_use_psk_identity_hint
4056 SSL_get_psk_identity_hint
4057 SSL_get_psk_identity
4058 SSL_use_psk_identity_hint
4060 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4062 *) Add RFC 3161 compliant time stamp request creation, response generation
4063 and response verification functionality.
4064 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4066 *) Add initial support for TLS extensions, specifically for the server_name
4067 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4068 have new members for a host name. The SSL data structure has an
4069 additional member SSL_CTX *initial_ctx so that new sessions can be
4070 stored in that context to allow for session resumption, even after the
4071 SSL has been switched to a new SSL_CTX in reaction to a client's
4072 server_name extension.
4074 New functions (subject to change):
4076 SSL_get_servername()
4077 SSL_get_servername_type()
4080 New CTRL codes and macros (subject to change):
4082 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4083 - SSL_CTX_set_tlsext_servername_callback()
4084 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4085 - SSL_CTX_set_tlsext_servername_arg()
4086 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4088 openssl s_client has a new '-servername ...' option.
4090 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4091 '-key2 ...', '-servername_fatal' (subject to change). This allows
4092 testing the HostName extension for a specific single host name ('-cert'
4093 and '-key' remain fallbacks for handshakes without HostName
4094 negotiation). If the unrecognized_name alert has to be sent, this by
4095 default is a warning; it becomes fatal with the '-servername_fatal'
4098 [Peter Sylvester, Remy Allais, Christophe Renou]
4100 *) Whirlpool hash implementation is added.
4103 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4104 bn(64,32). Because of instruction set limitations it doesn't have
4105 any negative impact on performance. This was done mostly in order
4106 to make it possible to share assembler modules, such as bn_mul_mont
4107 implementations, between 32- and 64-bit builds without hassle.
4110 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4111 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4115 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4116 dedicated Montgomery multiplication procedure, is introduced.
4117 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4118 "64-bit" performance on certain 32-bit targets.
4121 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4122 in SSL structures. New SSL ctrl to set maximum send fragment size.
4123 Save memory by seeting the I/O buffer sizes dynamically instead of
4124 using the maximum available value.
4127 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4128 in addition to the text details.
4131 *) Very, very preliminary EXPERIMENTAL support for printing of general
4132 ASN1 structures. This currently produces rather ugly output and doesn't
4133 handle several customised structures at all.
4136 *) Integrated support for PVK file format and some related formats such
4137 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4138 these in the 'rsa' and 'dsa' utilities.
4141 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4144 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4145 place for the (very old) "NETSCAPE" format certificates which are now
4146 handled using new ASN1 code equivalents.
4149 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4150 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4151 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4154 *) Modify CRL distribution points extension code to print out previously
4155 unsupported fields. Enhance extension setting code to allow setting of
4159 *) Add print and set support for Issuing Distribution Point CRL extension.
4162 *) Change 'Configure' script to enable Camellia by default.
4165 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4167 *) When rejecting SSL/TLS records due to an incorrect version number, never
4168 update s->server with a new major version number. As of
4169 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4170 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4171 the previous behavior could result in a read attempt at NULL when
4172 receiving specific incorrect SSL/TLS records once record payload
4173 protection is active. (CVE-2010-0740)
4174 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4176 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4177 could be crashed if the relevant tables were not present (e.g. chrooted).
4178 [Tomas Hoger <thoger@redhat.com>]
4180 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4182 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
4183 [Martin Olsson, Neel Mehta]
4185 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4186 accommodate for stack sorting, always a write lock!).
4189 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4190 excessive delays in the RAND_poll(): over a minute. As a workaround
4191 include a time check in the inner Heap32Next loop too.
4194 *) The code that handled flushing of data in SSL/TLS originally used the
4195 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4196 the problem outlined in PR#1949. The fix suggested there however can
4197 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4198 of Apache). So instead simplify the code to flush unconditionally.
4199 This should be fine since flushing with no data to flush is a no op.
4202 *) Handle TLS versions 2.0 and later properly and correctly use the
4203 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4204 off ancient servers have a habit of sticking around for a while...
4207 *) Modify compression code so it frees up structures without using the
4208 ex_data callbacks. This works around a problem where some applications
4209 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4210 restarting) then use compression (e.g. SSL with compression) later.
4211 This results in significant per-connection memory leaks and
4212 has caused some security issues including CVE-2008-1678 and
4216 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4217 change when encrypting or decrypting.
4220 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4221 connect and renegotiate with servers which do not support RI.
4222 Until RI is more widely deployed this option is enabled by default.
4225 *) Add "missing" ssl ctrls to clear options and mode.
4228 *) If client attempts to renegotiate and doesn't support RI respond with
4229 a no_renegotiation alert as required by RFC5746. Some renegotiating
4230 TLS clients will continue a connection gracefully when they receive
4231 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4232 waiting for a server hello which it will never receive. Now we treat a
4233 received no_renegotiation alert as a fatal error. This is because
4234 applications requesting a renegotiation might well expect it to succeed
4235 and would have no code in place to handle the server denying it so the
4236 only safe thing to do is to terminate the connection.
4239 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4240 peer supports secure renegotiation and 0 otherwise. Print out peer
4241 renegotiation support in s_client/s_server.
4244 *) Replace the highly broken and deprecated SPKAC certification method with
4245 the updated NID creation version. This should correctly handle UTF8.
4248 *) Implement RFC5746. Re-enable renegotiation but require the extension
4249 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4250 turns out to be a bad idea. It has been replaced by
4251 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4252 SSL_CTX_set_options(). This is really not recommended unless you
4253 know what you are doing.
4254 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4256 *) Fixes to stateless session resumption handling. Use initial_ctx when
4257 issuing and attempting to decrypt tickets in case it has changed during
4258 servername handling. Use a non-zero length session ID when attempting
4259 stateless session resumption: this makes it possible to determine if
4260 a resumption has occurred immediately after receiving server hello
4261 (several places in OpenSSL subtly assume this) instead of later in
4265 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4266 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4267 fixes for a few places where the return code is not checked
4269 [Julia Lawall <julia@diku.dk>]
4271 *) Add --strict-warnings option to Configure script to include devteam
4272 warnings in other configurations.
4275 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4276 makes it possible to install openssl libraries in locations which
4277 have names other than "lib", for example "/usr/lib64" which some
4279 [Steve Henson, based on patch from Jeremy Utley]
4281 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4282 X690 8.9.12 and can produce some misleading textual output of OIDs.
4283 [Steve Henson, reported by Dan Kaminsky]
4285 *) Delete MD2 from algorithm tables. This follows the recommendation in
4286 several standards that it is not used in new applications due to
4287 several cryptographic weaknesses. For binary compatibility reasons
4288 the MD2 API is still compiled in by default.
4291 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4295 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4296 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4298 [Guenter <lists@gknw.net>]
4300 *) Fix the server certificate chain building code to use X509_verify_cert(),
4301 it used to have an ad-hoc builder which was unable to cope with anything
4302 other than a simple chain.
4303 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4305 *) Don't check self signed certificate signatures in X509_verify_cert()
4306 by default (a flag can override this): it just wastes time without
4307 adding any security. As a useful side effect self signed root CAs
4308 with non-FIPS digests are now usable in FIPS mode.
4311 *) In dtls1_process_out_of_seq_message() the check if the current message
4312 is already buffered was missing. For every new message was memory
4313 allocated, allowing an attacker to perform an denial of service attack
4314 with sending out of seq handshake messages until there is no memory
4315 left. Additionally every future messege was buffered, even if the
4316 sequence number made no sense and would be part of another handshake.
4317 So only messages with sequence numbers less than 10 in advance will be
4318 buffered. (CVE-2009-1378)
4319 [Robin Seggelmann, discovered by Daniel Mentz]
4321 *) Records are buffered if they arrive with a future epoch to be
4322 processed after finishing the corresponding handshake. There is
4323 currently no limitation to this buffer allowing an attacker to perform
4324 a DOS attack with sending records with future epochs until there is no
4325 memory left. This patch adds the pqueue_size() function to determine
4326 the size of a buffer and limits the record buffer to 100 entries.
4328 [Robin Seggelmann, discovered by Daniel Mentz]
4330 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4331 parent structure is freed. (CVE-2009-1379)
4334 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4335 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4338 [Ilya O. <vrghost@gmail.com>]
4340 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4342 *) Disable renegotiation completely - this fixes a severe security
4343 problem (CVE-2009-3555) at the cost of breaking all
4344 renegotiation. Renegotiation can be re-enabled by setting
4345 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4346 run-time. This is really not recommended unless you know what
4350 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4352 *) Don't set val to NULL when freeing up structures, it is freed up by
4353 underlying code. If sizeof(void *) > sizeof(long) this can result in
4354 zeroing past the valid field. (CVE-2009-0789)
4355 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4357 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4358 checked correctly. This would allow some invalid signed attributes to
4359 appear to verify correctly. (CVE-2009-0591)
4360 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4362 *) Reject UniversalString and BMPString types with invalid lengths. This
4363 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4364 a legal length. (CVE-2009-0590)
4367 *) Set S/MIME signing as the default purpose rather than setting it
4368 unconditionally. This allows applications to override it at the store
4372 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4373 to handle some structures.
4376 *) Improve efficiency of mem_gets: don't search whole buffer each time
4378 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4380 *) New -hex option for openssl rand.
4383 *) Print out UTF8String and NumericString when parsing ASN1.
4386 *) Support NumericString type for name components.
4389 *) Allow CC in the environment to override the automatically chosen
4390 compiler. Note that nothing is done to ensure flags work with the
4394 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4396 *) Properly check EVP_VerifyFinal() and similar return values
4398 [Ben Laurie, Bodo Moeller, Google Security Team]
4400 *) Enable TLS extensions by default.
4403 *) Allow the CHIL engine to be loaded, whether the application is
4404 multithreaded or not. (This does not release the developer from the
4405 obligation to set up the dynamic locking callbacks.)
4406 [Sander Temme <sander@temme.net>]
4408 *) Use correct exit code if there is an error in dgst command.
4409 [Steve Henson; problem pointed out by Roland Dirlewanger]
4411 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4412 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4415 *) Add experimental JPAKE support, including demo authentication in
4416 s_client and s_server.
4419 *) Set the comparison function in v3_addr_canonize().
4420 [Rob Austein <sra@hactrn.net>]
4422 *) Add support for XMPP STARTTLS in s_client.
4423 [Philip Paeps <philip@freebsd.org>]
4425 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4426 to ensure that even with this option, only ciphersuites in the
4427 server's preference list will be accepted. (Note that the option
4428 applies only when resuming a session, so the earlier behavior was
4429 just about the algorithm choice for symmetric cryptography.)
4432 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4434 *) Fix NULL pointer dereference if a DTLS server received
4435 ChangeCipherSpec as first record (CVE-2009-1386).
4438 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4439 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4442 *) The fix in 0.9.8c that supposedly got rid of unsafe
4443 double-checked locking was incomplete for RSA blinding,
4444 addressing just one layer of what turns out to have been
4445 doubly unsafe triple-checked locking.
4447 So now fix this for real by retiring the MONT_HELPER macro
4448 in crypto/rsa/rsa_eay.c.
4450 [Bodo Moeller; problem pointed out by Marius Schilder]
4452 *) Various precautionary measures:
4454 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4456 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4457 (NB: This would require knowledge of the secret session ticket key
4458 to exploit, in which case you'd be SOL either way.)
4460 - Change bn_nist.c so that it will properly handle input BIGNUMs
4461 outside the expected range.
4463 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4466 [Neel Mehta, Bodo Moeller]
4468 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4469 the load fails. Useful for distros.
4470 [Ben Laurie and the FreeBSD team]
4472 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4475 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4478 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4480 This work was sponsored by Logica.
4483 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4484 keystores. Support for SSL/TLS client authentication too.
4485 Not compiled unless enable-capieng specified to Configure.
4487 This work was sponsored by Logica.
4490 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4491 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4492 attribute creation routines such as certificate requests and PKCS#12
4496 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4498 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4499 handshake which could lead to a cilent crash as found using the
4500 Codenomicon TLS test suite (CVE-2008-1672)
4501 [Steve Henson, Mark Cox]
4503 *) Fix double free in TLS server name extensions which could lead to
4504 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4507 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4509 Clear the error queue to ensure that error entries left from
4510 older function calls do not interfere with the correct operation.
4511 [Lutz Jaenicke, Erik de Castro Lopo]
4513 *) Remove root CA certificates of commercial CAs:
4515 The OpenSSL project does not recommend any specific CA and does not
4516 have any policy with respect to including or excluding any CA.
4517 Therefore it does not make any sense to ship an arbitrary selection
4518 of root CA certificates with the OpenSSL software.
4521 *) RSA OAEP patches to fix two separate invalid memory reads.
4522 The first one involves inputs when 'lzero' is greater than
4523 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4524 before the beginning of from). The second one involves inputs where
4525 the 'db' section contains nothing but zeroes (there is a one-byte
4526 invalid read after the end of 'db').
4527 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4529 *) Partial backport from 0.9.9-dev:
4531 Introduce bn_mul_mont (dedicated Montgomery multiplication
4532 procedure) as a candidate for BIGNUM assembler implementation.
4533 While 0.9.9-dev uses assembler for various architectures, only
4534 x86_64 is available by default here in the 0.9.8 branch, and
4535 32-bit x86 is available through a compile-time setting.
4537 To try the 32-bit x86 assembler implementation, use Configure
4538 option "enable-montasm" (which exists only for this backport).
4540 As "enable-montasm" for 32-bit x86 disclaims code stability
4541 anyway, in this constellation we activate additional code
4542 backported from 0.9.9-dev for further performance improvements,
4543 namely BN_from_montgomery_word. (To enable this otherwise,
4544 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4546 [Andy Polyakov (backport partially by Bodo Moeller)]
4548 *) Add TLS session ticket callback. This allows an application to set
4549 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4550 values. This is useful for key rollover for example where several key
4551 sets may exist with different names.
4554 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4555 This was broken until now in 0.9.8 releases, such that the only way
4556 a registered ENGINE could be used (assuming it initialises
4557 successfully on the host) was to explicitly set it as the default
4558 for the relevant algorithms. This is in contradiction with 0.9.7
4559 behaviour and the documentation. With this fix, when an ENGINE is
4560 registered into a given algorithm's table of implementations, the
4561 'uptodate' flag is reset so that auto-discovery will be used next
4562 time a new context for that algorithm attempts to select an
4564 [Ian Lister (tweaked by Geoff Thorpe)]
4566 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4567 implementation in the following ways:
4569 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4572 Lack of BER streaming support means one pass streaming processing is
4573 only supported if data is detached: setting the streaming flag is
4574 ignored for embedded content.
4576 CMS support is disabled by default and must be explicitly enabled
4577 with the enable-cms configuration option.
4580 *) Update the GMP engine glue to do direct copies between BIGNUM and
4581 mpz_t when openssl and GMP use the same limb size. Otherwise the
4582 existing "conversion via a text string export" trick is still used.
4583 [Paul Sheer <paulsheer@gmail.com>]
4585 *) Zlib compression BIO. This is a filter BIO which compressed and
4586 uncompresses any data passed through it.
4589 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4590 RFC3394 compatible AES key wrapping.
4593 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4594 sets string data without copying. X509_ALGOR_set0() and
4595 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4596 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4597 from an X509_ATTRIBUTE structure optionally checking it occurs only
4598 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4602 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4603 to get the expected BN_FLG_CONSTTIME behavior.
4604 [Bodo Moeller (Google)]
4608 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4609 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4610 - added some more tests to do_tests.pl
4611 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4612 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4613 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4614 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4615 - various changes to netware.pl to enable gcc-cross builds on Win32
4617 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4618 - various changes to fix missing prototype warnings
4619 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4620 - added AES, WHIRLPOOL and CPUID assembler code to build files
4621 - added missing AES assembler make rules to mk1mf.pl
4622 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4623 [Guenter Knauf <eflash@gmx.net>]
4625 *) Implement certificate status request TLS extension defined in RFC3546.
4626 A client can set the appropriate parameters and receive the encoded
4627 OCSP response via a callback. A server can query the supplied parameters
4628 and set the encoded OCSP response in the callback. Add simplified examples
4629 to s_client and s_server.
4632 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4634 *) Fix various bugs:
4635 + Binary incompatibility of ssl_ctx_st structure
4636 + DTLS interoperation with non-compliant servers
4637 + Don't call get_session_cb() without proposed session
4638 + Fix ia64 assembler code
4639 [Andy Polyakov, Steve Henson]
4641 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4643 *) DTLS Handshake overhaul. There were longstanding issues with
4644 OpenSSL DTLS implementation, which were making it impossible for
4645 RFC 4347 compliant client to communicate with OpenSSL server.
4646 Unfortunately just fixing these incompatibilities would "cut off"
4647 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4648 server keeps tolerating non RFC compliant syntax. The opposite is
4649 not true, 0.9.8f client can not communicate with earlier server.
4650 This update even addresses CVE-2007-4995.
4653 *) Changes to avoid need for function casts in OpenSSL: some compilers
4654 (gcc 4.2 and later) reject their use.
4655 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4658 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4659 RFC4507bis. The encrypted ticket format is an encrypted encoded
4660 SSL_SESSION structure, that way new session features are automatically
4663 If a client application caches session in an SSL_SESSION structure
4664 support is transparent because tickets are now stored in the encoded
4667 The SSL_CTX structure automatically generates keys for ticket
4668 protection in servers so again support should be possible
4669 with no application modification.
4671 If a client or server wishes to disable RFC4507 support then the option
4672 SSL_OP_NO_TICKET can be set.
4674 Add a TLS extension debugging callback to allow the contents of any client
4675 or server extensions to be examined.
4677 This work was sponsored by Google.
4680 *) Add initial support for TLS extensions, specifically for the server_name
4681 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4682 have new members for a host name. The SSL data structure has an
4683 additional member SSL_CTX *initial_ctx so that new sessions can be
4684 stored in that context to allow for session resumption, even after the
4685 SSL has been switched to a new SSL_CTX in reaction to a client's
4686 server_name extension.
4688 New functions (subject to change):
4690 SSL_get_servername()
4691 SSL_get_servername_type()
4694 New CTRL codes and macros (subject to change):
4696 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4697 - SSL_CTX_set_tlsext_servername_callback()
4698 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4699 - SSL_CTX_set_tlsext_servername_arg()
4700 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4702 openssl s_client has a new '-servername ...' option.
4704 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4705 '-key2 ...', '-servername_fatal' (subject to change). This allows
4706 testing the HostName extension for a specific single host name ('-cert'
4707 and '-key' remain fallbacks for handshakes without HostName
4708 negotiation). If the unrecognized_name alert has to be sent, this by
4709 default is a warning; it becomes fatal with the '-servername_fatal'
4712 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4714 *) Add AES and SSE2 assembly language support to VC++ build.
4717 *) Mitigate attack on final subtraction in Montgomery reduction.
4720 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4721 (which previously caused an internal error).
4724 *) Squeeze another 10% out of IGE mode when in != out.
4727 *) AES IGE mode speedup.
4728 [Dean Gaudet (Google)]
4730 *) Add the Korean symmetric 128-bit cipher SEED (see
4731 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4732 add SEED ciphersuites from RFC 4162:
4734 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4735 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4736 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4737 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4739 To minimize changes between patchlevels in the OpenSSL 0.9.8
4740 series, SEED remains excluded from compilation unless OpenSSL
4741 is configured with 'enable-seed'.
4742 [KISA, Bodo Moeller]
4744 *) Mitigate branch prediction attacks, which can be practical if a
4745 single processor is shared, allowing a spy process to extract
4746 information. For detailed background information, see
4747 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4748 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4749 and Necessary Software Countermeasures"). The core of the change
4750 are new versions BN_div_no_branch() and
4751 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4752 respectively, which are slower, but avoid the security-relevant
4753 conditional branches. These are automatically called by BN_div()
4754 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4755 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4756 remove a conditional branch.
4758 BN_FLG_CONSTTIME is the new name for the previous
4759 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4760 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4761 in the exponent causes BN_mod_exp_mont() to use the alternative
4762 implementation in BN_mod_exp_mont_consttime().) The old name
4763 remains as a deprecated alias.
4765 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4766 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4767 constant-time implementations for more than just exponentiation.
4768 Here too the old name is kept as a deprecated alias.
4770 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4771 the BN_BLINDING structure gets an independent copy of the
4772 modulus. This means that the previous "BIGNUM *m" argument to
4773 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4774 essentially becomes "const BIGNUM *m", although we can't actually
4775 change this in the header file before 0.9.9. It allows
4776 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4777 enable BN_FLG_CONSTTIME.
4779 [Matthew D Wood (Intel Corp)]
4781 *) In the SSL/TLS server implementation, be strict about session ID
4782 context matching (which matters if an application uses a single
4783 external cache for different purposes). Previously,
4784 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4785 set. This did ensure strict client verification, but meant that,
4786 with applications using a single external cache for quite
4787 different requirements, clients could circumvent ciphersuite
4788 restrictions for a given session ID context by starting a session
4789 in a different context.
4792 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4793 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4794 authentication-only ciphersuites.
4797 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4798 not complete and could lead to a possible single byte overflow
4799 (CVE-2007-5135) [Ben Laurie]
4801 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4803 *) Since AES128 and AES256 (and similarly Camellia128 and
4804 Camellia256) share a single mask bit in the logic of
4805 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4806 kludge to work properly if AES128 is available and AES256 isn't
4807 (or if Camellia128 is available and Camellia256 isn't).
4810 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4811 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4812 When a point or a seed is encoded in a BIT STRING, we need to
4813 prevent the removal of trailing zero bits to get the proper DER
4814 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4815 of a NamedBitList, for which trailing 0 bits need to be removed.)
4818 *) Have SSL/TLS server implementation tolerate "mismatched" record
4819 protocol version while receiving ClientHello even if the
4820 ClientHello is fragmented. (The server can't insist on the
4821 particular protocol version it has chosen before the ServerHello
4822 message has informed the client about his choice.)
4825 *) Add RFC 3779 support.
4826 [Rob Austein for ARIN, Ben Laurie]
4828 *) Load error codes if they are not already present instead of using a
4829 static variable. This allows them to be cleanly unloaded and reloaded.
4830 Improve header file function name parsing.
4833 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4834 or CAPABILITY handshake as required by RFCs.
4837 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4839 *) Introduce limits to prevent malicious keys being able to
4840 cause a denial of service. (CVE-2006-2940)
4841 [Steve Henson, Bodo Moeller]
4843 *) Fix ASN.1 parsing of certain invalid structures that can result
4844 in a denial of service. (CVE-2006-2937) [Steve Henson]
4846 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4847 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4849 *) Fix SSL client code which could crash if connecting to a
4850 malicious SSLv2 server. (CVE-2006-4343)
4851 [Tavis Ormandy and Will Drewry, Google Security Team]
4853 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4854 match only those. Before that, "AES256-SHA" would be interpreted
4855 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4856 the same strength classification in 0.9.7h) as we currently only
4857 have a single AES bit in the ciphersuite description bitmap.
4858 That change, however, also applied to ciphersuite strings such as
4859 "RC4-MD5" that intentionally matched multiple ciphersuites --
4860 namely, SSL 2.0 ciphersuites in addition to the more common ones
4861 from SSL 3.0/TLS 1.0.
4863 So we change the selection algorithm again: Naming an explicit
4864 ciphersuite selects this one ciphersuite, and any other similar
4865 ciphersuite (same bitmap) from *other* protocol versions.
4866 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4867 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4869 Since SSL 2.0 does not have any ciphersuites for which the
4870 128/256 bit distinction would be relevant, this works for now.
4871 The proper fix will be to use different bits for AES128 and
4872 AES256, which would have avoided the problems from the beginning;
4873 however, bits are scarce, so we can only do this in a new release
4874 (not just a patchlevel) when we can change the SSL_CIPHER
4875 definition to split the single 'unsigned long mask' bitmap into
4876 multiple values to extend the available space.
4880 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4882 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4883 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4885 *) Add AES IGE and biIGE modes.
4888 *) Change the Unix randomness entropy gathering to use poll() when
4889 possible instead of select(), since the latter has some
4890 undesirable limitations.
4891 [Darryl Miles via Richard Levitte and Bodo Moeller]
4893 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4894 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4895 cannot be implicitly activated as part of, e.g., the "AES" alias.
4896 However, please upgrade to OpenSSL 0.9.9[-dev] for
4897 non-experimental use of the ECC ciphersuites to get TLS extension
4898 support, which is required for curve and point format negotiation
4899 to avoid potential handshake problems.
4902 *) Disable rogue ciphersuites:
4904 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4905 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4906 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4908 The latter two were purportedly from
4909 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4912 Also deactivate the remaining ciphersuites from
4913 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4914 unofficial, and the ID has long expired.
4917 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4918 dual-core machines) and other potential thread-safety issues.
4921 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4922 versions), which is now available for royalty-free use
4923 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4924 Also, add Camellia TLS ciphersuites from RFC 4132.
4926 To minimize changes between patchlevels in the OpenSSL 0.9.8
4927 series, Camellia remains excluded from compilation unless OpenSSL
4928 is configured with 'enable-camellia'.
4931 *) Disable the padding bug check when compression is in use. The padding
4932 bug check assumes the first packet is of even length, this is not
4933 necessarily true if compresssion is enabled and can result in false
4934 positives causing handshake failure. The actual bug test is ancient
4935 code so it is hoped that implementations will either have fixed it by
4936 now or any which still have the bug do not support compression.
4939 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4941 *) When applying a cipher rule check to see if string match is an explicit
4942 cipher suite and only match that one cipher suite if it is.
4945 *) Link in manifests for VC++ if needed.
4946 [Austin Ziegler <halostatue@gmail.com>]
4948 *) Update support for ECC-based TLS ciphersuites according to
4949 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4950 TLS extensions, which are supported starting with the 0.9.9
4951 branch, not in the OpenSSL 0.9.8 branch).
4954 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4955 opaque EVP_CIPHER_CTX handling.
4958 *) Fixes and enhancements to zlib compression code. We now only use
4959 "zlib1.dll" and use the default __cdecl calling convention on Win32
4960 to conform with the standards mentioned here:
4961 http://www.zlib.net/DLL_FAQ.txt
4962 Static zlib linking now works on Windows and the new --with-zlib-include
4963 --with-zlib-lib options to Configure can be used to supply the location
4964 of the headers and library. Gracefully handle case where zlib library
4968 *) Several fixes and enhancements to the OID generation code. The old code
4969 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4970 handle numbers larger than ULONG_MAX, truncated printing and had a
4971 non standard OBJ_obj2txt() behaviour.
4974 *) Add support for building of engines under engine/ as shared libraries
4975 under VC++ build system.
4978 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4979 Hopefully, we will not see any false combination of paths any more.
4982 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4984 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4985 (part of SSL_OP_ALL). This option used to disable the
4986 countermeasure against man-in-the-middle protocol-version
4987 rollback in the SSL 2.0 server implementation, which is a bad
4988 idea. (CVE-2005-2969)
4990 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4991 for Information Security, National Institute of Advanced Industrial
4992 Science and Technology [AIST], Japan)]
4994 *) Add two function to clear and return the verify parameter flags.
4997 *) Keep cipherlists sorted in the source instead of sorting them at
4998 runtime, thus removing the need for a lock.
5001 *) Avoid some small subgroup attacks in Diffie-Hellman.
5002 [Nick Mathewson and Ben Laurie]
5004 *) Add functions for well-known primes.
5007 *) Extended Windows CE support.
5008 [Satoshi Nakamura and Andy Polyakov]
5010 *) Initialize SSL_METHOD structures at compile time instead of during
5011 runtime, thus removing the need for a lock.
5014 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5015 attempting to decrypt each encrypted key in turn. Add support to
5019 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5021 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5024 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5027 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5028 key into the same file any more.
5031 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5034 *) Add -utf8 command line and config file option to 'ca'.
5035 [Stefan <stf@udoma.org]
5037 *) Removed the macro des_crypt(), as it seems to conflict with some
5038 libraries. Use DES_crypt().
5041 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5042 involves renaming the source and generated shared-libs for
5043 both. The engines will accept the corrected or legacy ids
5044 ('ncipher' and '4758_cca' respectively) when binding. NB,
5045 this only applies when building 'shared'.
5046 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5048 *) Add attribute functions to EVP_PKEY structure. Modify
5049 PKCS12_create() to recognize a CSP name attribute and
5050 use it. Make -CSP option work again in pkcs12 utility.
5053 *) Add new functionality to the bn blinding code:
5054 - automatic re-creation of the BN_BLINDING parameters after
5055 a fixed number of uses (currently 32)
5056 - add new function for parameter creation
5057 - introduce flags to control the update behaviour of the
5058 BN_BLINDING parameters
5059 - hide BN_BLINDING structure
5060 Add a second BN_BLINDING slot to the RSA structure to improve
5061 performance when a single RSA object is shared among several
5065 *) Add support for DTLS.
5066 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5068 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5069 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5072 *) Remove buggy and incomplete DH cert support from
5073 ssl/ssl_rsa.c and ssl/s3_both.c
5076 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5077 the apps/openssl applications.
5080 *) Compile clean with "-Wall -Wmissing-prototypes
5081 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5082 DEBUG_SAFESTACK must also be set.
5085 *) Change ./Configure so that certain algorithms can be disabled by default.
5086 The new counterpiece to "no-xxx" is "enable-xxx".
5088 The patented RC5 and MDC2 algorithms will now be disabled unless
5089 "enable-rc5" and "enable-mdc2", respectively, are specified.
5091 (IDEA remains enabled despite being patented. This is because IDEA
5092 is frequently required for interoperability, and there is no license
5093 fee for non-commercial use. As before, "no-idea" can be used to
5094 avoid this algorithm.)
5098 *) Add processing of proxy certificates (see RFC 3820). This work was
5099 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5100 EGEE (Enabling Grids for E-science in Europe).
5103 *) RC4 performance overhaul on modern architectures/implementations, such
5104 as Intel P4, IA-64 and AMD64.
5107 *) New utility extract-section.pl. This can be used specify an alternative
5108 section number in a pod file instead of having to treat each file as
5109 a separate case in Makefile. This can be done by adding two lines to the
5112 =for comment openssl_section:XXX
5114 The blank line is mandatory.
5118 *) New arguments -certform, -keyform and -pass for s_client and s_server
5119 to allow alternative format key and certificate files and passphrase
5123 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5124 update associated structures and add various utility functions.
5126 Add new policy related verify parameters, include policy checking in
5127 standard verify code. Enhance 'smime' application with extra parameters
5128 to support policy checking and print out.
5131 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5132 Nehemiah processors. These extensions support AES encryption in hardware
5133 as well as RNG (though RNG support is currently disabled).
5134 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5136 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5139 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5140 [Andy Polyakov and a number of other people]
5142 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5143 implementation contributed by IBM.
5144 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5146 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5147 exponent rather than 'unsigned long'. There is a corresponding change to
5148 the new 'rsa_keygen' element of the RSA_METHOD structure.
5149 [Jelte Jansen, Geoff Thorpe]
5151 *) Functionality for creating the initial serial number file is now
5152 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5154 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5155 number file to 1, which is bound to cause problems. To avoid
5156 the problems while respecting compatibility between different 0.9.7
5157 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5158 CA.pl for serial number initialization. With the new release 0.9.8,
5159 we can fix the problem directly in the 'ca' utility.)
5162 *) Reduced header interdepencies by declaring more opaque objects in
5163 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5164 give fewer recursive includes, which could break lazy source code - so
5165 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5166 developers should define this symbol when building and using openssl to
5167 ensure they track the recommended behaviour, interfaces, [etc], but
5168 backwards-compatible behaviour prevails when this isn't defined.
5171 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5174 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5175 This will generate a random key of the appropriate length based on the
5176 cipher context. The EVP_CIPHER can provide its own random key generation
5177 routine to support keys of a specific form. This is used in the des and
5178 3des routines to generate a key of the correct parity. Update S/MIME
5179 code to use new functions and hence generate correct parity DES keys.
5180 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5181 valid (weak or incorrect parity).
5184 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5185 as looking them up. This is useful when the verified structure may contain
5186 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5187 present unless the new PKCS7_NO_CRL flag is asserted.
5190 *) Extend ASN1 oid configuration module. It now additionally accepts the
5193 shortName = some long name, 1.2.3.4
5196 *) Reimplemented the BN_CTX implementation. There is now no more static
5197 limitation on the number of variables it can handle nor the depth of the
5198 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5199 information can now expand as required, and rather than having a single
5200 static array of bignums, BN_CTX now uses a linked-list of such arrays
5201 allowing it to expand on demand whilst maintaining the usefulness of
5202 BN_CTX's "bundling".
5205 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5206 to allow all RSA operations to function using a single BN_CTX.
5209 *) Preliminary support for certificate policy evaluation and checking. This
5210 is initially intended to pass the tests outlined in "Conformance Testing
5211 of Relying Party Client Certificate Path Processing Logic" v1.07.
5214 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5215 remained unused and not that useful. A variety of other little bignum
5216 tweaks and fixes have also been made continuing on from the audit (see
5220 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5221 associated ASN1, EVP and SSL functions and old ASN1 macros.
5224 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5225 and this should never fail. So the return value from the use of
5226 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5227 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5230 *) BN_CTX_get() should return zero-valued bignums, providing the same
5231 initialised value as BN_new().
5232 [Geoff Thorpe, suggested by Ulf Möller]
5234 *) Support for inhibitAnyPolicy certificate extension.
5237 *) An audit of the BIGNUM code is underway, for which debugging code is
5238 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5239 is considered valid when processing BIGNUMs, and causes execution to
5240 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5241 further steps are taken to deliberately pollute unused data in BIGNUM
5242 structures to try and expose faulty code further on. For now, openssl will
5243 (in its default mode of operation) continue to tolerate the inconsistent
5244 forms that it has tolerated in the past, but authors and packagers should
5245 consider trying openssl and their own applications when compiled with
5246 these debugging symbols defined. It will help highlight potential bugs in
5247 their own code, and will improve the test coverage for OpenSSL itself. At
5248 some point, these tighter rules will become openssl's default to improve
5249 maintainability, though the assert()s and other overheads will remain only
5250 in debugging configurations. See bn.h for more details.
5251 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5253 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5254 that can only be obtained through BN_CTX_new() (which implicitly
5255 initialises it). The presence of this function only made it possible
5256 to overwrite an existing structure (and cause memory leaks).
5259 *) Because of the callback-based approach for implementing LHASH as a
5260 template type, lh_insert() adds opaque objects to hash-tables and
5261 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5262 to clean up those corresponding objects before destroying the hash table
5263 (and losing the object pointers). So some over-zealous constifications in
5264 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5265 objects as "const" and the lh_doall[_arg] callback wrappers are not
5266 prototyped to have "const" restrictions on the object pointers they are
5267 given (and so aren't required to cast them away any more).
5270 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5271 (speed) prefers to use its own implementation. The two implementations
5272 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5273 its object type properly exposed (MS_TM) instead of casting to/from "char
5274 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5275 aren't necessarily the greatest nomenclatures - but this is what was used
5276 internally to the implementation so I've used that for now.
5279 *) Ensure that deprecated functions do not get compiled when
5280 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5281 the self-tests were still using deprecated key-generation functions so
5282 these have been updated also.
5285 *) Reorganise PKCS#7 code to separate the digest location functionality
5286 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5287 New function PKCS7_set_digest() to set the digest type for PKCS#7
5288 digestedData type. Add additional code to correctly generate the
5289 digestedData type and add support for this type in PKCS7 initialization
5293 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5294 structure of type "other".
5297 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5298 sure the loop does correctly stop and breaking ("division by zero")
5299 modulus operations are not performed. The (pre-generated) prime
5300 table crypto/bn/bn_prime.h was already correct, but it could not be
5301 re-generated on some platforms because of the "division by zero"
5302 situation in the script.
5303 [Ralf S. Engelschall]
5305 *) Update support for ECC-based TLS ciphersuites according to
5306 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5307 SHA-1 now is only used for "small" curves (where the
5308 representation of a field element takes up to 24 bytes); for
5309 larger curves, the field element resulting from ECDH is directly
5310 used as premaster secret.
5311 [Douglas Stebila (Sun Microsystems Laboratories)]
5313 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5314 curve secp160r1 to the tests.
5315 [Douglas Stebila (Sun Microsystems Laboratories)]
5317 *) Add the possibility to load symbols globally with DSO.
5318 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5320 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5321 control of the error stack.
5324 *) Add support for STORE in ENGINE.
5327 *) Add the STORE type. The intention is to provide a common interface
5328 to certificate and key stores, be they simple file-based stores, or
5329 HSM-type store, or LDAP stores, or...
5330 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5333 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5334 pass a list of arguments to any function as well as provide a way
5335 for a function to pass data back to the caller.
5338 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5339 works like BUF_strdup() but can be used to duplicate a portion of
5340 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5344 *) Add the function sk_find_ex() which works like sk_find(), but will
5345 return an index to an element even if an exact match couldn't be
5346 found. The index is guaranteed to point at the element where the
5347 searched-for key would be inserted to preserve sorting order.
5350 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5351 takes an extra flags argument for optional functionality. Currently,
5352 the following flags are defined:
5354 OBJ_BSEARCH_VALUE_ON_NOMATCH
5355 This one gets OBJ_bsearch_ex() to return a pointer to the first
5356 element where the comparing function returns a negative or zero
5359 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5360 This one gets OBJ_bsearch_ex() to return a pointer to the first
5361 element where the comparing function returns zero. This is useful
5362 if there are more than one element where the comparing function
5366 *) Make it possible to create self-signed certificates with 'openssl ca'
5367 in such a way that the self-signed certificate becomes part of the
5368 CA database and uses the same mechanisms for serial number generation
5369 as all other certificate signing. The new flag '-selfsign' enables
5370 this functionality. Adapt CA.sh and CA.pl.in.
5373 *) Add functionality to check the public key of a certificate request
5374 against a given private. This is useful to check that a certificate
5375 request can be signed by that key (self-signing).
5378 *) Make it possible to have multiple active certificates with the same
5379 subject in the CA index file. This is done only if the keyword
5380 'unique_subject' is set to 'no' in the main CA section (default
5381 if 'CA_default') of the configuration file. The value is saved
5382 with the database itself in a separate index attribute file,
5383 named like the index file with '.attr' appended to the name.
5386 *) Generate muti valued AVAs using '+' notation in config files for
5390 *) Support for nameConstraints certificate extension.
5393 *) Support for policyConstraints certificate extension.
5396 *) Support for policyMappings certificate extension.
5399 *) Make sure the default DSA_METHOD implementation only uses its
5400 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5401 and change its own handlers to be NULL so as to remove unnecessary
5402 indirection. This lets alternative implementations fallback to the
5403 default implementation more easily.
5406 *) Support for directoryName in GeneralName related extensions
5410 *) Make it possible to link applications using Makefile.shared.
5411 Make that possible even when linking against static libraries!
5414 *) Support for single pass processing for S/MIME signing. This now
5415 means that S/MIME signing can be done from a pipe, in addition
5416 cleartext signing (multipart/signed type) is effectively streaming
5417 and the signed data does not need to be all held in memory.
5419 This is done with a new flag PKCS7_STREAM. When this flag is set
5420 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5421 is done after the data is output (and digests calculated) in
5422 SMIME_write_PKCS7().
5425 *) Add full support for -rpath/-R, both in shared libraries and
5426 applications, at least on the platforms where it's known how
5430 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5431 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5432 will now compute a table of multiples of the generator that
5433 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5434 faster (notably in the case of a single point multiplication,
5435 scalar * generator).
5436 [Nils Larsch, Bodo Moeller]
5438 *) IPv6 support for certificate extensions. The various extensions
5439 which use the IP:a.b.c.d can now take IPv6 addresses using the
5440 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5444 *) Added an ENGINE that implements RSA by performing private key
5445 exponentiations with the GMP library. The conversions to and from
5446 GMP's mpz_t format aren't optimised nor are any montgomery forms
5447 cached, and on x86 it appears OpenSSL's own performance has caught up.
5448 However there are likely to be other architectures where GMP could
5449 provide a boost. This ENGINE is not built in by default, but it can be
5450 specified at Configure time and should be accompanied by the necessary
5451 linker additions, eg;
5452 ./config -DOPENSSL_USE_GMP -lgmp
5455 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5456 testing availability of engines with "-t" - the old behaviour is
5457 produced by increasing the feature's verbosity with "-tt".
5460 *) ECDSA routines: under certain error conditions uninitialized BN objects
5461 could be freed. Solution: make sure initialization is performed early
5462 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5466 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5467 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5468 software implementations. For DSA and DH, parameter generation can
5469 also be overridden by providing the appropriate method callbacks.
5472 *) Change the "progress" mechanism used in key-generation and
5473 primality testing to functions that take a new BN_GENCB pointer in
5474 place of callback/argument pairs. The new API functions have "_ex"
5475 postfixes and the older functions are reimplemented as wrappers for
5476 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5477 declarations of the old functions to help (graceful) attempts to
5478 migrate to the new functions. Also, the new key-generation API
5479 functions operate on a caller-supplied key-structure and return
5480 success/failure rather than returning a key or NULL - this is to
5481 help make "keygen" another member function of RSA_METHOD etc.
5483 Example for using the new callback interface:
5485 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5489 BN_GENCB_set(&my_cb, my_callback, my_arg);
5491 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5492 /* For the meaning of a, b in calls to my_callback(), see the
5493 * documentation of the function that calls the callback.
5494 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5495 * my_callback should return 1 if it wants BN_is_prime_ex()
5496 * to continue, or 0 to stop.
5501 *) Change the ZLIB compression method to be stateful, and make it
5502 available to TLS with the number defined in
5503 draft-ietf-tls-compression-04.txt.
5506 *) Add the ASN.1 structures and functions for CertificatePair, which
5507 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5509 CertificatePair ::= SEQUENCE {
5510 forward [0] Certificate OPTIONAL,
5511 reverse [1] Certificate OPTIONAL,
5512 -- at least one of the pair shall be present -- }
5514 Also implement the PEM functions to read and write certificate
5515 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5517 This needed to be defined, mostly for the sake of the LDAP
5518 attribute crossCertificatePair, but may prove useful elsewhere as
5522 *) Make it possible to inhibit symlinking of shared libraries in
5523 Makefile.shared, for Cygwin's sake.
5526 *) Extend the BIGNUM API by creating a function
5527 void BN_set_negative(BIGNUM *a, int neg);
5528 and a macro that behave like
5529 int BN_is_negative(const BIGNUM *a);
5531 to avoid the need to access 'a->neg' directly in applications.
5534 *) Implement fast modular reduction for pseudo-Mersenne primes
5535 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5536 EC_GROUP_new_curve_GFp() will now automatically use this
5538 [Nils Larsch <nla@trustcenter.de>]
5540 *) Add new lock type (CRYPTO_LOCK_BN).
5543 *) Change the ENGINE framework to automatically load engines
5544 dynamically from specific directories unless they could be
5545 found to already be built in or loaded. Move all the
5546 current engines except for the cryptodev one to a new
5548 The engines in engines/ are built as shared libraries if
5549 the "shared" options was given to ./Configure or ./config.
5550 Otherwise, they are inserted in libcrypto.a.
5551 /usr/local/ssl/engines is the default directory for dynamic
5552 engines, but that can be overridden at configure time through
5553 the usual use of --prefix and/or --openssldir, and at run
5554 time with the environment variable OPENSSL_ENGINES.
5555 [Geoff Thorpe and Richard Levitte]
5557 *) Add Makefile.shared, a helper makefile to build shared
5558 libraries. Adapt Makefile.org.
5561 *) Add version info to Win32 DLLs.
5562 [Peter 'Luna' Runestig" <peter@runestig.com>]
5564 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5565 can be added using this API to created arbitrary PKCS#12
5566 files while avoiding the low level API.
5568 New options to PKCS12_create(), key or cert can be NULL and
5569 will then be omitted from the output file. The encryption
5570 algorithm NIDs can be set to -1 for no encryption, the mac
5571 iteration count can be set to 0 to omit the mac.
5573 Enhance pkcs12 utility by making the -nokeys and -nocerts
5574 options work when creating a PKCS#12 file. New option -nomac
5575 to omit the mac, NONE can be set for an encryption algorithm.
5576 New code is modified to use the enhanced PKCS12_create()
5577 instead of the low level API.
5580 *) Extend ASN1 encoder to support indefinite length constructed
5581 encoding. This can output sequences tags and octet strings in
5582 this form. Modify pk7_asn1.c to support indefinite length
5583 encoding. This is experimental and needs additional code to
5584 be useful, such as an ASN1 bio and some enhanced streaming
5587 Extend template encode functionality so that tagging is passed
5588 down to the template encoder.
5591 *) Let 'openssl req' fail if an argument to '-newkey' is not
5592 recognized instead of using RSA as a default.
5595 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5596 As these are not official, they are not included in "ALL";
5597 the "ECCdraft" ciphersuite group alias can be used to select them.
5598 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5600 *) Add ECDH engine support.
5601 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5603 *) Add ECDH in new directory crypto/ecdh/.
5604 [Douglas Stebila (Sun Microsystems Laboratories)]
5606 *) Let BN_rand_range() abort with an error after 100 iterations
5607 without success (which indicates a broken PRNG).
5610 *) Change BN_mod_sqrt() so that it verifies that the input value
5611 is really the square of the return value. (Previously,
5612 BN_mod_sqrt would show GIGO behaviour.)
5615 *) Add named elliptic curves over binary fields from X9.62, SECG,
5616 and WAP/WTLS; add OIDs that were still missing.
5618 [Sheueling Chang Shantz and Douglas Stebila
5619 (Sun Microsystems Laboratories)]
5621 *) Extend the EC library for elliptic curves over binary fields
5622 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5625 EC_GF2m_simple_method
5629 EC_GROUP_new_curve_GF2m
5630 EC_GROUP_set_curve_GF2m
5631 EC_GROUP_get_curve_GF2m
5632 EC_POINT_set_affine_coordinates_GF2m
5633 EC_POINT_get_affine_coordinates_GF2m
5634 EC_POINT_set_compressed_coordinates_GF2m
5636 Point compression for binary fields is disabled by default for
5637 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5640 As binary polynomials are represented as BIGNUMs, various members
5641 of the EC_GROUP and EC_POINT data structures can be shared
5642 between the implementations for prime fields and binary fields;
5643 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5644 are essentially identical to their ..._GFp counterparts.
5645 (For simplicity, the '..._GFp' prefix has been dropped from
5646 various internal method names.)
5648 An internal 'field_div' method (similar to 'field_mul' and
5649 'field_sqr') has been added; this is used only for binary fields.
5651 [Sheueling Chang Shantz and Douglas Stebila
5652 (Sun Microsystems Laboratories)]
5654 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5655 through methods ('mul', 'precompute_mult').
5657 The generic implementations (now internally called 'ec_wNAF_mul'
5658 and 'ec_wNAF_precomputed_mult') remain the default if these
5659 methods are undefined.
5661 [Sheueling Chang Shantz and Douglas Stebila
5662 (Sun Microsystems Laboratories)]
5664 *) New function EC_GROUP_get_degree, which is defined through
5665 EC_METHOD. For curves over prime fields, this returns the bit
5666 length of the modulus.
5668 [Sheueling Chang Shantz and Douglas Stebila
5669 (Sun Microsystems Laboratories)]
5671 *) New functions EC_GROUP_dup, EC_POINT_dup.
5672 (These simply call ..._new and ..._copy).
5674 [Sheueling Chang Shantz and Douglas Stebila
5675 (Sun Microsystems Laboratories)]
5677 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5678 Polynomials are represented as BIGNUMs (where the sign bit is not
5679 used) in the following functions [macros]:
5682 BN_GF2m_sub [= BN_GF2m_add]
5683 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5684 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5685 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5687 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5688 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5689 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5690 BN_GF2m_cmp [= BN_ucmp]
5692 (Note that only the 'mod' functions are actually for fields GF(2^m).
5693 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5695 For some functions, an the irreducible polynomial defining a
5696 field can be given as an 'unsigned int[]' with strictly
5697 decreasing elements giving the indices of those bits that are set;
5698 i.e., p[] represents the polynomial
5699 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5701 p[0] > p[1] > ... > p[k] = 0.
5702 This applies to the following functions:
5707 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5708 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5710 BN_GF2m_mod_sqrt_arr
5711 BN_GF2m_mod_solve_quad_arr
5715 Conversion can be performed by the following functions:
5720 bntest.c has additional tests for binary polynomial arithmetic.
5722 Two implementations for BN_GF2m_mod_div() are available.
5723 The default algorithm simply uses BN_GF2m_mod_inv() and
5724 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5725 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5726 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5728 [Sheueling Chang Shantz and Douglas Stebila
5729 (Sun Microsystems Laboratories)]
5731 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5732 functionality is disabled at compile-time.
5733 [Douglas Stebila <douglas.stebila@sun.com>]
5735 *) Change default behaviour of 'openssl asn1parse' so that more
5736 information is visible when viewing, e.g., a certificate:
5738 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5739 mode the content of non-printable OCTET STRINGs is output in a
5740 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5741 avoid the appearance of a printable string.
5742 [Nils Larsch <nla@trustcenter.de>]
5744 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5746 EC_GROUP_set_asn1_flag()
5747 EC_GROUP_get_asn1_flag()
5748 EC_GROUP_set_point_conversion_form()
5749 EC_GROUP_get_point_conversion_form()
5750 These control ASN1 encoding details:
5751 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5752 has been set to OPENSSL_EC_NAMED_CURVE.
5753 - Points are encoded in uncompressed form by default; options for
5754 asn1_for are as for point2oct, namely
5755 POINT_CONVERSION_COMPRESSED
5756 POINT_CONVERSION_UNCOMPRESSED
5757 POINT_CONVERSION_HYBRID
5759 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5762 EC_GROUP_get0_seed()
5763 EC_GROUP_get_seed_len()
5764 This is used only for ASN1 purposes (so far).
5765 [Nils Larsch <nla@trustcenter.de>]
5767 *) Add 'field_type' member to EC_METHOD, which holds the NID
5768 of the appropriate field type OID. The new function
5769 EC_METHOD_get_field_type() returns this value.
5770 [Nils Larsch <nla@trustcenter.de>]
5775 EC_POINT_point2hex()
5776 EC_POINT_hex2point()
5777 providing useful interfaces to EC_POINT_point2oct() and
5778 EC_POINT_oct2point().
5779 [Nils Larsch <nla@trustcenter.de>]
5781 *) Change internals of the EC library so that the functions
5782 EC_GROUP_set_generator()
5783 EC_GROUP_get_generator()
5784 EC_GROUP_get_order()
5785 EC_GROUP_get_cofactor()
5786 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5787 to methods, which would lead to unnecessary code duplication when
5788 adding different types of curves.
5789 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5791 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5792 arithmetic, and such that modified wNAFs are generated
5793 (which avoid length expansion in many cases).
5796 *) Add a function EC_GROUP_check_discriminant() (defined via
5797 EC_METHOD) that verifies that the curve discriminant is non-zero.
5799 Add a function EC_GROUP_check() that makes some sanity tests
5800 on a EC_GROUP, its generator and order. This includes
5801 EC_GROUP_check_discriminant().
5802 [Nils Larsch <nla@trustcenter.de>]
5804 *) Add ECDSA in new directory crypto/ecdsa/.
5806 Add applications 'openssl ecparam' and 'openssl ecdsa'
5807 (these are based on 'openssl dsaparam' and 'openssl dsa').
5809 ECDSA support is also included in various other files across the
5810 library. Most notably,
5811 - 'openssl req' now has a '-newkey ecdsa:file' option;
5812 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5813 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5814 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5815 them suitable for ECDSA where domain parameters must be
5816 extracted before the specific public key;
5817 - ECDSA engine support has been added.
5818 [Nils Larsch <nla@trustcenter.de>]
5820 *) Include some named elliptic curves, and add OIDs from X9.62,
5821 SECG, and WAP/WTLS. Each curve can be obtained from the new
5823 EC_GROUP_new_by_curve_name(),
5824 and the list of available named curves can be obtained with
5825 EC_get_builtin_curves().
5826 Also add a 'curve_name' member to EC_GROUP objects, which can be
5828 EC_GROUP_set_curve_name()
5829 EC_GROUP_get_curve_name()
5830 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5832 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5833 was actually never needed) and in BN_mul(). The removal in BN_mul()
5834 required a small change in bn_mul_part_recursive() and the addition
5835 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5836 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5837 bn_sub_words() and bn_add_words() except they take arrays with
5841 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5843 *) Cleanse PEM buffers before freeing them since they may contain
5845 [Benjamin Bennett <ben@psc.edu>]
5847 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5848 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5849 authentication-only ciphersuites.
5852 *) Since AES128 and AES256 share a single mask bit in the logic of
5853 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5854 kludge to work properly if AES128 is available and AES256 isn't.
5857 *) Expand security boundary to match 1.1.1 module.
5860 *) Remove redundant features: hash file source, editing of test vectors
5861 modify fipsld to use external fips_premain.c signature.
5864 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5865 run algorithm test programs.
5868 *) Make algorithm test programs more tolerant of whitespace.
5871 *) Have SSL/TLS server implementation tolerate "mismatched" record
5872 protocol version while receiving ClientHello even if the
5873 ClientHello is fragmented. (The server can't insist on the
5874 particular protocol version it has chosen before the ServerHello
5875 message has informed the client about his choice.)
5878 *) Load error codes if they are not already present instead of using a
5879 static variable. This allows them to be cleanly unloaded and reloaded.
5882 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5884 *) Introduce limits to prevent malicious keys being able to
5885 cause a denial of service. (CVE-2006-2940)
5886 [Steve Henson, Bodo Moeller]
5888 *) Fix ASN.1 parsing of certain invalid structures that can result
5889 in a denial of service. (CVE-2006-2937) [Steve Henson]
5891 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5892 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5894 *) Fix SSL client code which could crash if connecting to a
5895 malicious SSLv2 server. (CVE-2006-4343)
5896 [Tavis Ormandy and Will Drewry, Google Security Team]
5898 *) Change ciphersuite string processing so that an explicit
5899 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5900 will no longer include "AES128-SHA"), and any other similar
5901 ciphersuite (same bitmap) from *other* protocol versions (so that
5902 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5903 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5904 changes from 0.9.8b and 0.9.8d.
5907 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5909 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5910 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5912 *) Change the Unix randomness entropy gathering to use poll() when
5913 possible instead of select(), since the latter has some
5914 undesirable limitations.
5915 [Darryl Miles via Richard Levitte and Bodo Moeller]
5917 *) Disable rogue ciphersuites:
5919 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5920 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5921 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5923 The latter two were purportedly from
5924 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5927 Also deactive the remaining ciphersuites from
5928 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5929 unofficial, and the ID has long expired.
5932 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5933 dual-core machines) and other potential thread-safety issues.
5936 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5938 *) Adapt fipsld and the build system to link against the validated FIPS
5939 module in FIPS mode.
5942 *) Fixes for VC++ 2005 build under Windows.
5945 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5946 from a Windows bash shell such as MSYS. It is autodetected from the
5947 "config" script when run from a VC++ environment. Modify standard VC++
5948 build to use fipscanister.o from the GNU make build.
5951 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5953 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5954 The value now differs depending on if you build for FIPS or not.
5955 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5956 safely run with a non-FIPSed libcrypto, as it may crash because of
5957 the difference induced by this change.
5960 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5962 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5963 (part of SSL_OP_ALL). This option used to disable the
5964 countermeasure against man-in-the-middle protocol-version
5965 rollback in the SSL 2.0 server implementation, which is a bad
5966 idea. (CVE-2005-2969)
5968 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5969 for Information Security, National Institute of Advanced Industrial
5970 Science and Technology [AIST], Japan)]
5972 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5973 mainly for FIPS compliance and not fully integrated at this stage.
5976 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5977 the exponentiation using a fixed-length exponent. (Otherwise,
5978 the information leaked through timing could expose the secret key
5979 after many signatures; cf. Bleichenbacher's attack on DSA with
5983 *) Make a new fixed-window mod_exp implementation the default for
5984 RSA, DSA, and DH private-key operations so that the sequence of
5985 squares and multiplies and the memory access pattern are
5986 independent of the particular secret key. This will mitigate
5987 cache-timing and potential related attacks.
5989 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5990 and this is automatically used by BN_mod_exp_mont() if the new flag
5991 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5992 will use this BN flag for private exponents unless the flag
5993 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5994 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5996 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5998 *) Change the client implementation for SSLv23_method() and
5999 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6000 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6001 (Previously, the SSL 2.0 backwards compatible Client Hello
6002 message format would be used even with SSL_OP_NO_SSLv2.)
6005 *) Add support for smime-type MIME parameter in S/MIME messages which some
6009 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6010 a threadsafe manner. Modify rsa code to use new function and add calls
6011 to dsa and dh code (which had race conditions before).
6014 *) Include the fixed error library code in the C error file definitions
6015 instead of fixing them up at runtime. This keeps the error code
6016 structures constant.
6019 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6021 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6024 *) Fixes for newer kerberos headers. NB: the casts are needed because
6025 the 'length' field is signed on one version and unsigned on another
6026 with no (?) obvious way to tell the difference, without these VC++
6027 complains. Also the "definition" of FAR (blank) is no longer included
6028 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6029 some needed definitions.
6032 *) Undo Cygwin change.
6035 *) Added support for proxy certificates according to RFC 3820.
6036 Because they may be a security thread to unaware applications,
6037 they must be explicitly allowed in run-time. See
6038 docs/HOWTO/proxy_certificates.txt for further information.
6041 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6043 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6044 server and client random values. Previously
6045 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6046 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6048 This change has negligible security impact because:
6050 1. Server and client random values still have 24 bytes of pseudo random
6053 2. Server and client random values are sent in the clear in the initial
6056 3. The master secret is derived using the premaster secret (48 bytes in
6057 size for static RSA ciphersuites) as well as client server and random
6060 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6063 [Stephen Henson, reported by UK NISCC]
6065 *) Use Windows randomness collection on Cygwin.
6068 *) Fix hang in EGD/PRNGD query when communication socket is closed
6069 prematurely by EGD/PRNGD.
6070 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6072 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6075 *) Back-port of selected performance improvements from development
6076 branch, as well as improved support for PowerPC platforms.
6079 *) Add lots of checks for memory allocation failure, error codes to indicate
6080 failure and freeing up memory if a failure occurs.
6081 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6083 *) Add new -passin argument to dgst.
6086 *) Perform some character comparisons of different types in X509_NAME_cmp:
6087 this is needed for some certificates that re-encode DNs into UTF8Strings
6088 (in violation of RFC3280) and can't or won't issue name rollover
6092 *) Make an explicit check during certificate validation to see that
6093 the CA setting in each certificate on the chain is correct. As a
6094 side effect always do the following basic checks on extensions,
6095 not just when there's an associated purpose to the check:
6097 - if there is an unhandled critical extension (unless the user
6098 has chosen to ignore this fault)
6099 - if the path length has been exceeded (if one is set at all)
6100 - that certain extensions fit the associated purpose (if one has
6104 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6106 *) Avoid a race condition when CRLs are checked in a multi threaded
6107 environment. This would happen due to the reordering of the revoked
6108 entries during signature checking and serial number lookup. Now the
6109 encoding is cached and the serial number sort performed under a lock.
6110 Add new STACK function sk_is_sorted().
6113 *) Add Delta CRL to the extension code.
6116 *) Various fixes to s3_pkt.c so alerts are sent properly.
6117 [David Holmes <d.holmes@f5.com>]
6119 *) Reduce the chances of duplicate issuer name and serial numbers (in
6120 violation of RFC3280) using the OpenSSL certificate creation utilities.
6121 This is done by creating a random 64 bit value for the initial serial
6122 number when a serial number file is created or when a self signed
6123 certificate is created using 'openssl req -x509'. The initial serial
6124 number file is created using 'openssl x509 -next_serial' in CA.pl
6125 rather than being initialized to 1.
6128 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6130 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6131 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6132 [Joe Orton, Steve Henson]
6134 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6136 [Joe Orton, Steve Henson]
6138 *) Make it possible to have multiple active certificates with the same
6139 subject in the CA index file. This is done only if the keyword
6140 'unique_subject' is set to 'no' in the main CA section (default
6141 if 'CA_default') of the configuration file. The value is saved
6142 with the database itself in a separate index attribute file,
6143 named like the index file with '.attr' appended to the name.
6146 *) X509 verify fixes. Disable broken certificate workarounds when
6147 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6148 keyUsage extension present. Don't accept CRLs with unhandled critical
6149 extensions: since verify currently doesn't process CRL extensions this
6150 rejects a CRL with *any* critical extensions. Add new verify error codes
6154 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6155 A clarification of RFC2560 will require the use of OCTET STRINGs and
6156 some implementations cannot handle the current raw format. Since OpenSSL
6157 copies and compares OCSP nonces as opaque blobs without any attempt at
6158 parsing them this should not create any compatibility issues.
6161 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6162 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6163 this HMAC (and other) operations are several times slower than OpenSSL
6167 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6168 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6170 *) Use the correct content when signing type "other".
6173 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6175 *) Fix various bugs revealed by running the NISCC test suite:
6177 Stop out of bounds reads in the ASN1 code when presented with
6178 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6180 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6182 If verify callback ignores invalid public key errors don't try to check
6183 certificate signature with the NULL public key.
6187 *) New -ignore_err option in ocsp application to stop the server
6188 exiting on the first error in a request.
6191 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6192 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6196 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6197 extra data after the compression methods not only for TLS 1.0
6198 but also for SSL 3.0 (as required by the specification).
6199 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6201 *) Change X509_certificate_type() to mark the key as exported/exportable
6202 when it's 512 *bits* long, not 512 bytes.
6205 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6206 blocks during encryption.
6209 *) Various fixes to base64 BIO and non blocking I/O. On write
6210 flushes were not handled properly if the BIO retried. On read
6211 data was not being buffered properly and had various logic bugs.
6212 This also affects blocking I/O when the data being decoded is a
6216 *) Various S/MIME bugfixes and compatibility changes:
6217 output correct application/pkcs7 MIME type if
6218 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6219 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6220 of files as .eml work). Correctly handle very long lines in MIME
6224 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6226 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6227 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6228 a protocol version number mismatch like a decryption error
6229 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6232 *) Turn on RSA blinding by default in the default implementation
6233 to avoid a timing attack. Applications that don't want it can call
6234 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6235 They would be ill-advised to do so in most cases.
6236 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6238 *) Change RSA blinding code so that it works when the PRNG is not
6239 seeded (in this case, the secret RSA exponent is abused as
6240 an unpredictable seed -- if it is not unpredictable, there
6241 is no point in blinding anyway). Make RSA blinding thread-safe
6242 by remembering the creator's thread ID in rsa->blinding and
6243 having all other threads use local one-time blinding factors
6244 (this requires more computation than sharing rsa->blinding, but
6245 avoids excessive locking; and if an RSA object is not shared
6246 between threads, blinding will still be very fast).
6249 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6250 ENGINE as defaults for all supported algorithms irrespective of
6251 the 'flags' parameter. 'flags' is now honoured, so applications
6252 should make sure they are passing it correctly.
6255 *) Target "mingw" now allows native Windows code to be generated in
6256 the Cygwin environment as well as with the MinGW compiler.
6259 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6261 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6262 via timing by performing a MAC computation even if incorrect
6263 block cipher padding has been found. This is a countermeasure
6264 against active attacks where the attacker has to distinguish
6265 between bad padding and a MAC verification error. (CVE-2003-0078)
6267 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6268 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6269 Martin Vuagnoux (EPFL, Ilion)]
6271 *) Make the no-err option work as intended. The intention with no-err
6272 is not to have the whole error stack handling routines removed from
6273 libcrypto, it's only intended to remove all the function name and
6274 reason texts, thereby removing some of the footprint that may not
6275 be interesting if those errors aren't displayed anyway.
6277 NOTE: it's still possible for any application or module to have it's
6278 own set of error texts inserted. The routines are there, just not
6279 used by default when no-err is given.
6282 *) Add support for FreeBSD on IA64.
6283 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6285 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6286 Kerberos function mit_des_cbc_cksum(). Before this change,
6287 the value returned by DES_cbc_cksum() was like the one from
6288 mit_des_cbc_cksum(), except the bytes were swapped.
6289 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6291 *) Allow an application to disable the automatic SSL chain building.
6292 Before this a rather primitive chain build was always performed in
6293 ssl3_output_cert_chain(): an application had no way to send the
6294 correct chain if the automatic operation produced an incorrect result.
6296 Now the chain builder is disabled if either:
6298 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6300 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6302 The reasoning behind this is that an application would not want the
6303 auto chain building to take place if extra chain certificates are
6304 present and it might also want a means of sending no additional
6305 certificates (for example the chain has two certificates and the
6309 *) Add the possibility to build without the ENGINE framework.
6310 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6312 *) Under Win32 gmtime() can return NULL: check return value in
6313 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6316 *) DSA routines: under certain error conditions uninitialized BN objects
6317 could be freed. Solution: make sure initialization is performed early
6318 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6319 Nils Larsch <nla@trustcenter.de> via PR#459)
6322 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6323 checked on reconnect on the client side, therefore session resumption
6324 could still fail with a "ssl session id is different" error. This
6325 behaviour is masked when SSL_OP_ALL is used due to
6326 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6327 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6328 followup to PR #377.
6331 *) IA-32 assembler support enhancements: unified ELF targets, support
6332 for SCO/Caldera platforms, fix for Cygwin shared build.
6335 *) Add support for FreeBSD on sparc64. As a consequence, support for
6336 FreeBSD on non-x86 processors is separate from x86 processors on
6337 the config script, much like the NetBSD support.
6338 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6340 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6342 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6345 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6346 code (06) was taken as the first octet of the session ID and the last
6347 octet was ignored consequently. As a result SSLv2 client side session
6348 caching could not have worked due to the session ID mismatch between
6350 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6354 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6355 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6359 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6360 seems that in spite of existing for more than a year, many application
6361 author have done nothing to provide the necessary callbacks, which
6362 means that this particular engine will not work properly anywhere.
6363 This is a very unfortunate situation which forces us, in the name
6364 of usability, to give the hw_ncipher.c a static lock, which is part
6366 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6367 appear in 0.9.8 or later. We EXPECT application authors to have
6368 dealt properly with this when 0.9.8 is released (unless we actually
6369 make such changes in the libcrypto locking code that changes will
6370 have to be made anyway).
6373 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6374 octets have been read, EOF or an error occurs. Without this change
6375 some truncated ASN1 structures will not produce an error.
6378 *) Disable Heimdal support, since it hasn't been fully implemented.
6379 Still give the possibility to force the use of Heimdal, but with
6380 warnings and a request that patches get sent to openssl-dev.
6383 *) Add the VC-CE target, introduce the WINCE sysname, and add
6384 INSTALL.WCE and appropriate conditionals to make it build.
6385 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6387 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6388 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6389 edit numbers of the version.
6390 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6392 *) Introduce safe string copy and catenation functions
6393 (BUF_strlcpy() and BUF_strlcat()).
6394 [Ben Laurie (CHATS) and Richard Levitte]
6396 *) Avoid using fixed-size buffers for one-line DNs.
6397 [Ben Laurie (CHATS)]
6399 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6400 resizing buffers containing secrets, and use where appropriate.
6401 [Ben Laurie (CHATS)]
6403 *) Avoid using fixed size buffers for configuration file location.
6404 [Ben Laurie (CHATS)]
6406 *) Avoid filename truncation for various CA files.
6407 [Ben Laurie (CHATS)]
6409 *) Use sizeof in preference to magic numbers.
6410 [Ben Laurie (CHATS)]
6412 *) Avoid filename truncation in cert requests.
6413 [Ben Laurie (CHATS)]
6415 *) Add assertions to check for (supposedly impossible) buffer
6417 [Ben Laurie (CHATS)]
6419 *) Don't cache truncated DNS entries in the local cache (this could
6420 potentially lead to a spoofing attack).
6421 [Ben Laurie (CHATS)]
6423 *) Fix various buffers to be large enough for hex/decimal
6424 representations in a platform independent manner.
6425 [Ben Laurie (CHATS)]
6427 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6428 resizing buffers containing secrets, and use where appropriate.
6429 [Ben Laurie (CHATS)]
6431 *) Add BIO_indent() to avoid much slightly worrying code to do
6433 [Ben Laurie (CHATS)]
6435 *) Convert sprintf()/BIO_puts() to BIO_printf().
6436 [Ben Laurie (CHATS)]
6438 *) buffer_gets() could terminate with the buffer only half
6440 [Ben Laurie (CHATS)]
6442 *) Add assertions to prevent user-supplied crypto functions from
6443 overflowing internal buffers by having large block sizes, etc.
6444 [Ben Laurie (CHATS)]
6446 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6448 [Ben Laurie (CHATS)]
6450 *) Eliminate unused copy of key in RC4.
6451 [Ben Laurie (CHATS)]
6453 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6454 [Ben Laurie (CHATS)]
6456 *) Fix off-by-one error in EGD path.
6457 [Ben Laurie (CHATS)]
6459 *) If RANDFILE path is too long, ignore instead of truncating.
6460 [Ben Laurie (CHATS)]
6462 *) Eliminate unused and incorrectly sized X.509 structure
6464 [Ben Laurie (CHATS)]
6466 *) Eliminate unused and dangerous function knumber().
6467 [Ben Laurie (CHATS)]
6469 *) Eliminate unused and dangerous structure, KSSL_ERR.
6470 [Ben Laurie (CHATS)]
6472 *) Protect against overlong session ID context length in an encoded
6473 session object. Since these are local, this does not appear to be
6475 [Ben Laurie (CHATS)]
6477 *) Change from security patch (see 0.9.6e below) that did not affect
6478 the 0.9.6 release series:
6480 Remote buffer overflow in SSL3 protocol - an attacker could
6481 supply an oversized master key in Kerberos-enabled versions.
6483 [Ben Laurie (CHATS)]
6485 *) Change the SSL kerb5 codes to match RFC 2712.
6488 *) Make -nameopt work fully for req and add -reqopt switch.
6489 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6491 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6492 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6494 *) Make sure tests can be performed even if the corresponding algorithms
6495 have been removed entirely. This was also the last step to make
6496 OpenSSL compilable with DJGPP under all reasonable conditions.
6497 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6499 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6500 to allow version independent disabling of normally unselected ciphers,
6501 which may be activated as a side-effect of selecting a single cipher.
6503 (E.g., cipher list string "RSA" enables ciphersuites that are left
6504 out of "ALL" because they do not provide symmetric encryption.
6505 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6506 [Lutz Jaenicke, Bodo Moeller]
6508 *) Add appropriate support for separate platform-dependent build
6509 directories. The recommended way to make a platform-dependent
6510 build directory is the following (tested on Linux), maybe with
6513 # Place yourself outside of the OpenSSL source tree. In
6514 # this example, the environment variable OPENSSL_SOURCE
6515 # is assumed to contain the absolute OpenSSL source directory.
6516 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6517 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6518 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6519 mkdir -p `dirname $F`
6520 ln -s $OPENSSL_SOURCE/$F $F
6523 To be absolutely sure not to disturb the source tree, a "make clean"
6524 is a good thing. If it isn't successful, don't worry about it,
6525 it probably means the source directory is very clean.
6528 *) Make sure any ENGINE control commands make local copies of string
6529 pointers passed to them whenever necessary. Otherwise it is possible
6530 the caller may have overwritten (or deallocated) the original string
6531 data when a later ENGINE operation tries to use the stored values.
6532 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6534 *) Improve diagnostics in file reading and command-line digests.
6535 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6537 *) Add AES modes CFB and OFB to the object database. Correct an
6538 error in AES-CFB decryption.
6541 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6542 allows existing EVP_CIPHER_CTX structures to be reused after
6543 calling EVP_*Final(). This behaviour is used by encryption
6544 BIOs and some applications. This has the side effect that
6545 applications must explicitly clean up cipher contexts with
6546 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6549 *) Check the values of dna and dnb in bn_mul_recursive before calling
6550 bn_mul_comba (a non zero value means the a or b arrays do not contain
6551 n2 elements) and fallback to bn_mul_normal if either is not zero.
6554 *) Fix escaping of non-ASCII characters when using the -subj option
6555 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6558 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6559 form for "surname", serialNumber has no short form.
6560 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6561 therefore remove "mail" short name for "internet 7".
6562 The OID for unique identifiers in X509 certificates is
6563 x500UniqueIdentifier, not uniqueIdentifier.
6564 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6567 *) Add an "init" command to the ENGINE config module and auto initialize
6568 ENGINEs. Without any "init" command the ENGINE will be initialized
6569 after all ctrl commands have been executed on it. If init=1 the
6570 ENGINE is initailized at that point (ctrls before that point are run
6571 on the uninitialized ENGINE and after on the initialized one). If
6572 init=0 then the ENGINE will not be iniatialized at all.
6575 *) Fix the 'app_verify_callback' interface so that the user-defined
6576 argument is actually passed to the callback: In the
6577 SSL_CTX_set_cert_verify_callback() prototype, the callback
6578 declaration has been changed from
6581 int (*cb)(X509_STORE_CTX *,void *);
6582 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6583 i=s->ctx->app_verify_callback(&ctx)
6584 has been changed into
6585 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6587 To update applications using SSL_CTX_set_cert_verify_callback(),
6588 a dummy argument can be added to their callback functions.
6589 [D. K. Smetters <smetters@parc.xerox.com>]
6591 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6592 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6594 *) Add and OPENSSL_LOAD_CONF define which will cause
6595 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6596 This allows older applications to transparently support certain
6597 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6598 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6599 load the config file and OPENSSL_add_all_algorithms_conf() which will
6600 always load it have also been added.
6603 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6604 Adjust NIDs and EVP layer.
6605 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6607 *) Config modules support in openssl utility.
6609 Most commands now load modules from the config file,
6610 though in a few (such as version) this isn't done
6611 because it couldn't be used for anything.
6613 In the case of ca and req the config file used is
6614 the same as the utility itself: that is the -config
6615 command line option can be used to specify an
6619 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6620 use "openssl_conf" if filename is NULL use default openssl config file.
6623 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6624 config section name. Add a new flag to tolerate a missing config file
6625 and move code to CONF_modules_load_file().
6628 *) Support for crypto accelerator cards from Accelerated Encryption
6629 Processing, www.aep.ie. (Use engine 'aep')
6630 The support was copied from 0.9.6c [engine] and adapted/corrected
6631 to work with the new engine framework.
6632 [AEP Inc. and Richard Levitte]
6634 *) Support for SureWare crypto accelerator cards from Baltimore
6635 Technologies. (Use engine 'sureware')
6636 The support was copied from 0.9.6c [engine] and adapted
6637 to work with the new engine framework.
6640 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6641 make the newer ENGINE framework commands for the CHIL engine work.
6642 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6644 *) Make it possible to produce shared libraries on ReliantUNIX.
6645 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6647 *) Add the configuration target debug-linux-ppro.
6648 Make 'openssl rsa' use the general key loading routines
6649 implemented in apps.c, and make those routines able to
6650 handle the key format FORMAT_NETSCAPE and the variant
6652 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6654 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6655 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6657 *) Add -keyform to rsautl, and document -engine.
6658 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6660 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6661 BIO_R_NO_SUCH_FILE error code rather than the generic
6662 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6665 *) Add new functions
6667 ERR_peek_last_error_line
6668 ERR_peek_last_error_line_data.
6669 These are similar to
6672 ERR_peek_error_line_data,
6673 but report on the latest error recorded rather than the first one
6674 still in the error queue.
6675 [Ben Laurie, Bodo Moeller]
6677 *) default_algorithms option in ENGINE config module. This allows things
6679 default_algorithms = ALL
6680 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6683 *) Preliminary ENGINE config module.
6686 *) New experimental application configuration code.
6689 *) Change the AES code to follow the same name structure as all other
6690 symmetric ciphers, and behave the same way. Move everything to
6691 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6692 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6694 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6695 [Ben Laurie and Theo de Raadt]
6697 *) Add option to output public keys in req command.
6698 [Massimiliano Pala madwolf@openca.org]
6700 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6701 (up to about 10% better than before for P-192 and P-224).
6704 *) New functions/macros
6706 SSL_CTX_set_msg_callback(ctx, cb)
6707 SSL_CTX_set_msg_callback_arg(ctx, arg)
6708 SSL_set_msg_callback(ssl, cb)
6709 SSL_set_msg_callback_arg(ssl, arg)
6711 to request calling a callback function
6713 void cb(int write_p, int version, int content_type,
6714 const void *buf, size_t len, SSL *ssl, void *arg)
6716 whenever a protocol message has been completely received
6717 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6718 protocol version according to which the SSL library interprets
6719 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6720 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6721 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6722 specification (change_cipher_spec(20), alert(21), handshake(22)).
6723 'buf' and 'len' point to the actual message, 'ssl' to the
6724 SSL object, and 'arg' is the application-defined value set by
6725 SSL[_CTX]_set_msg_callback_arg().
6727 'openssl s_client' and 'openssl s_server' have new '-msg' options
6728 to enable a callback that displays all protocol messages.
6731 *) Change the shared library support so shared libraries are built as
6732 soon as the corresponding static library is finished, and thereby get
6733 openssl and the test programs linked against the shared library.
6734 This still only happens when the keyword "shard" has been given to
6735 the configuration scripts.
6737 NOTE: shared library support is still an experimental thing, and
6738 backward binary compatibility is still not guaranteed.
6739 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6741 *) Add support for Subject Information Access extension.
6742 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6744 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6745 additional bytes when new memory had to be allocated, not just
6746 when reusing an existing buffer.
6749 *) New command line and configuration option 'utf8' for the req command.
6750 This allows field values to be specified as UTF8 strings.
6753 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6754 runs for the former and machine-readable output for the latter.
6757 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6758 of the e-mail address in the DN (i.e., it will go into a certificate
6759 extension only). The new configuration file option 'email_in_dn = no'
6760 has the same effect.
6761 [Massimiliano Pala madwolf@openca.org]
6763 *) Change all functions with names starting with des_ to be starting
6764 with DES_ instead. Add wrappers that are compatible with libdes,
6765 but are named _ossl_old_des_*. Finally, add macros that map the
6766 des_* symbols to the corresponding _ossl_old_des_* if libdes
6767 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6768 desired, the des_* symbols will be mapped to DES_*, with one
6771 Since we provide two compatibility mappings, the user needs to
6772 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6773 compatibility is desired. The default (i.e., when that macro
6774 isn't defined) is OpenSSL 0.9.6c compatibility.
6776 There are also macros that enable and disable the support of old
6777 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6778 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6779 are defined, the default will apply: to support the old des routines.
6781 In either case, one must include openssl/des.h to get the correct
6782 definitions. Do not try to just include openssl/des_old.h, that
6785 NOTE: This is a major break of an old API into a new one. Software
6786 authors are encouraged to switch to the DES_ style functions. Some
6787 time in the future, des_old.h and the libdes compatibility functions
6788 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6789 default), and then completely removed.
6792 *) Test for certificates which contain unsupported critical extensions.
6793 If such a certificate is found during a verify operation it is
6794 rejected by default: this behaviour can be overridden by either
6795 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6796 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6797 X509_supported_extension() has also been added which returns 1 if a
6798 particular extension is supported.
6801 *) Modify the behaviour of EVP cipher functions in similar way to digests
6802 to retain compatibility with existing code.
6805 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6806 compatibility with existing code. In particular the 'ctx' parameter does
6807 not have to be to be initialized before the call to EVP_DigestInit() and
6808 it is tidied up after a call to EVP_DigestFinal(). New function
6809 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6810 EVP_MD_CTX_copy() changed to not require the destination to be
6811 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6812 requires the destination to be valid.
6814 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6815 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6818 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6819 so that complete 'Handshake' protocol structures are kept in memory
6820 instead of overwriting 'msg_type' and 'length' with 'body' data.
6823 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6824 [Massimo Santin via Richard Levitte]
6826 *) Major restructuring to the underlying ENGINE code. This includes
6827 reduction of linker bloat, separation of pure "ENGINE" manipulation
6828 (initialisation, etc) from functionality dealing with implementations
6829 of specific crypto iterfaces. This change also introduces integrated
6830 support for symmetric ciphers and digest implementations - so ENGINEs
6831 can now accelerate these by providing EVP_CIPHER and EVP_MD
6832 implementations of their own. This is detailed in crypto/engine/README
6833 as it couldn't be adequately described here. However, there are a few
6834 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6835 were changed in the original introduction of ENGINE code have now
6836 reverted back - the hooking from this code to ENGINE is now a good
6837 deal more passive and at run-time, operations deal directly with
6838 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6839 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6840 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6841 they were not being used by the framework as there is no concept of a
6842 BIGNUM_METHOD and they could not be generalised to the new
6843 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6844 ENGINE_cpy() has been removed as it cannot be consistently defined in
6848 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6851 *) Change mkdef.pl to sort symbols that get the same entry number,
6852 and make sure the automatically generated functions ERR_load_*
6853 become part of libeay.num as well.
6856 *) New function SSL_renegotiate_pending(). This returns true once
6857 renegotiation has been requested (either SSL_renegotiate() call
6858 or HelloRequest/ClientHello received from the peer) and becomes
6859 false once a handshake has been completed.
6860 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6861 sends a HelloRequest, but does not ensure that a handshake takes
6862 place. SSL_renegotiate_pending() is useful for checking if the
6863 client has followed the request.)
6866 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6867 By default, clients may request session resumption even during
6868 renegotiation (if session ID contexts permit); with this option,
6869 session resumption is possible only in the first handshake.
6871 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6872 more bits available for options that should not be part of
6873 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6876 *) Add some demos for certificate and certificate request creation.
6879 *) Make maximum certificate chain size accepted from the peer application
6880 settable (SSL*_get/set_max_cert_list()), as proposed by
6881 "Douglas E. Engert" <deengert@anl.gov>.
6884 *) Add support for shared libraries for Unixware-7
6885 (Boyd Lynn Gerber <gerberb@zenez.com>).
6888 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6889 be done prior to destruction. Use this to unload error strings from
6890 ENGINEs that load their own error strings. NB: This adds two new API
6891 functions to "get" and "set" this destroy handler in an ENGINE.
6894 *) Alter all existing ENGINE implementations (except "openssl" and
6895 "openbsd") to dynamically instantiate their own error strings. This
6896 makes them more flexible to be built both as statically-linked ENGINEs
6897 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6898 Also, add stub code to each that makes building them as self-contained
6899 shared-libraries easier (see README.ENGINE).
6902 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6903 implementations into applications that are completely implemented in
6904 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6905 commands that can be used to configure what shared-library to load and
6906 to control aspects of the way it is handled. Also, made an update to
6907 the README.ENGINE file that brings its information up-to-date and
6908 provides some information and instructions on the "dynamic" ENGINE
6909 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6912 *) Make it possible to unload ranges of ERR strings with a new
6913 "ERR_unload_strings" function.
6916 *) Add a copy() function to EVP_MD.
6919 *) Make EVP_MD routines take a context pointer instead of just the
6920 md_data void pointer.
6923 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6924 that the digest can only process a single chunk of data
6925 (typically because it is provided by a piece of
6926 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6927 is only going to provide a single chunk of data, and hence the
6928 framework needn't accumulate the data for oneshot drivers.
6931 *) As with "ERR", make it possible to replace the underlying "ex_data"
6932 functions. This change also alters the storage and management of global
6933 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6934 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6935 index counters. The API functions that use this state have been changed
6936 to take a "class_index" rather than pointers to the class's local STACK
6937 and counter, and there is now an API function to dynamically create new
6938 classes. This centralisation allows us to (a) plug a lot of the
6939 thread-safety problems that existed, and (b) makes it possible to clean
6940 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6941 such data would previously have always leaked in application code and
6942 workarounds were in place to make the memory debugging turn a blind eye
6943 to it. Application code that doesn't use this new function will still
6944 leak as before, but their memory debugging output will announce it now
6945 rather than letting it slide.
6947 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6948 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6949 has a return value to indicate success or failure.
6952 *) Make it possible to replace the underlying "ERR" functions such that the
6953 global state (2 LHASH tables and 2 locks) is only used by the "default"
6954 implementation. This change also adds two functions to "get" and "set"
6955 the implementation prior to it being automatically set the first time
6956 any other ERR function takes place. Ie. an application can call "get",
6957 pass the return value to a module it has just loaded, and that module
6958 can call its own "set" function using that value. This means the
6959 module's "ERR" operations will use (and modify) the error state in the
6960 application and not in its own statically linked copy of OpenSSL code.
6963 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6964 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6965 the operation, and provides a more encapsulated way for external code
6966 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6967 to use these functions rather than manually incrementing the counts.
6969 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6972 *) Add EVP test program.
6975 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6978 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6979 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6980 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6981 These allow a CRL to be built without having to access X509_CRL fields
6982 directly. Modify 'ca' application to use new functions.
6985 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6986 bug workarounds. Rollback attack detection is a security feature.
6987 The problem will only arise on OpenSSL servers when TLSv1 is not
6988 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6989 Software authors not wanting to support TLSv1 will have special reasons
6990 for their choice and can explicitly enable this option.
6991 [Bodo Moeller, Lutz Jaenicke]
6993 *) Rationalise EVP so it can be extended: don't include a union of
6994 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6995 (similar to those existing for EVP_CIPHER_CTX).
7000 EVP_MD_CTX_init(&md); /* new function call */
7001 EVP_DigestInit(&md, EVP_sha1());
7002 EVP_DigestUpdate(&md, in, len);
7003 EVP_DigestFinal(&md, out, NULL);
7004 EVP_MD_CTX_cleanup(&md); /* new function call */
7008 *) Make DES key schedule conform to the usual scheme, as well as
7009 correcting its structure. This means that calls to DES functions
7010 now have to pass a pointer to a des_key_schedule instead of a
7011 plain des_key_schedule (which was actually always a pointer
7014 des_key_schedule ks;
7016 des_set_key_checked(..., &ks);
7017 des_ncbc_encrypt(..., &ks, ...);
7019 (Note that a later change renames 'des_...' into 'DES_...'.)
7022 *) Initial reduction of linker bloat: the use of some functions, such as
7023 PEM causes large amounts of unused functions to be linked in due to
7024 poor organisation. For example pem_all.c contains every PEM function
7025 which has a knock on effect of linking in large amounts of (unused)
7026 ASN1 code. Grouping together similar functions and splitting unrelated
7027 functions prevents this.
7030 *) Cleanup of EVP macros.
7033 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7034 correct _ecb suffix.
7037 *) Add initial OCSP responder support to ocsp application. The
7038 revocation information is handled using the text based index
7039 use by the ca application. The responder can either handle
7040 requests generated internally, supplied in files (for example
7041 via a CGI script) or using an internal minimal server.
7044 *) Add configuration choices to get zlib compression for TLS.
7047 *) Changes to Kerberos SSL for RFC 2712 compliance:
7048 1. Implemented real KerberosWrapper, instead of just using
7049 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7050 2. Implemented optional authenticator field of KerberosWrapper.
7052 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7053 and authenticator structs; see crypto/krb5/.
7055 Generalized Kerberos calls to support multiple Kerberos libraries.
7056 [Vern Staats <staatsvr@asc.hpc.mil>,
7057 Jeffrey Altman <jaltman@columbia.edu>
7058 via Richard Levitte]
7060 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7061 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7062 values for each of the key sizes rather than having just
7063 parameters (and 'speed' generating keys each time).
7066 *) Speed up EVP routines.
7069 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7070 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7071 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7072 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7074 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7075 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7076 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7079 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7081 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7084 *) Added the OS2-EMX target.
7085 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7087 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7088 to support NCONF routines in extension code. New function CONF_set_nconf()
7089 to allow functions which take an NCONF to also handle the old LHASH
7090 structure: this means that the old CONF compatible routines can be
7091 retained (in particular wrt extensions) without having to duplicate the
7092 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7095 *) Enhance the general user interface with mechanisms for inner control
7096 and with possibilities to have yes/no kind of prompts.
7099 *) Change all calls to low level digest routines in the library and
7100 applications to use EVP. Add missing calls to HMAC_cleanup() and
7101 don't assume HMAC_CTX can be copied using memcpy().
7102 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7104 *) Add the possibility to control engines through control names but with
7105 arbitrary arguments instead of just a string.
7106 Change the key loaders to take a UI_METHOD instead of a callback
7107 function pointer. NOTE: this breaks binary compatibility with earlier
7108 versions of OpenSSL [engine].
7109 Adapt the nCipher code for these new conditions and add a card insertion
7113 *) Enhance the general user interface with mechanisms to better support
7114 dialog box interfaces, application-defined prompts, the possibility
7115 to use defaults (for example default passwords from somewhere else)
7116 and interrupts/cancellations.
7119 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7120 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7123 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7124 tidy up some unnecessarily weird code in 'sk_new()').
7125 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7127 *) Change the key loading routines for ENGINEs to use the same kind
7128 callback (pem_password_cb) as all other routines that need this
7132 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7133 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7134 than this minimum value is recommended.
7137 *) New random seeder for OpenVMS, using the system process statistics
7138 that are easily reachable.
7141 *) Windows apparently can't transparently handle global
7142 variables defined in DLLs. Initialisations such as:
7144 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7146 won't compile. This is used by the any applications that need to
7147 declare their own ASN1 modules. This was fixed by adding the option
7148 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7149 needed for static libraries under Win32.
7152 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7153 setting of purpose and trust fields. New X509_STORE trust and
7154 purpose functions and tidy up setting in other SSL functions.
7157 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7158 structure. These are inherited by X509_STORE_CTX when it is
7159 initialised. This allows various defaults to be set in the
7160 X509_STORE structure (such as flags for CRL checking and custom
7161 purpose or trust settings) for functions which only use X509_STORE_CTX
7162 internally such as S/MIME.
7164 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7165 trust settings if they are not set in X509_STORE. This allows X509_STORE
7166 purposes and trust (in S/MIME for example) to override any set by default.
7168 Add command line options for CRL checking to smime, s_client and s_server
7172 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7173 are set then the CRL is looked up in the X509_STORE structure and
7174 its validity and signature checked, then if the certificate is found
7175 in the CRL the verify fails with a revoked error.
7177 Various new CRL related callbacks added to X509_STORE_CTX structure.
7179 Command line options added to 'verify' application to support this.
7181 This needs some additional work, such as being able to handle multiple
7182 CRLs with different times, extension based lookup (rather than just
7183 by subject name) and ultimately more complete V2 CRL extension
7187 *) Add a general user interface API (crypto/ui/). This is designed
7188 to replace things like des_read_password and friends (backward
7189 compatibility functions using this new API are provided).
7190 The purpose is to remove prompting functions from the DES code
7191 section as well as provide for prompting through dialog boxes in
7192 a window system and the like.
7195 *) Add "ex_data" support to ENGINE so implementations can add state at a
7196 per-structure level rather than having to store it globally.
7199 *) Make it possible for ENGINE structures to be copied when retrieved by
7200 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7201 This causes the "original" ENGINE structure to act like a template,
7202 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7203 operational state can be localised to each ENGINE structure, despite the
7204 fact they all share the same "methods". New ENGINE structures returned in
7205 this case have no functional references and the return value is the single
7206 structural reference. This matches the single structural reference returned
7207 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7211 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7212 needs to match any other type at all we need to manually clear the
7216 *) Changes to the "openssl engine" utility to include;
7217 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7218 about an ENGINE's available control commands.
7219 - executing control commands from command line arguments using the
7220 '-pre' and '-post' switches. '-post' is only used if '-t' is
7221 specified and the ENGINE is successfully initialised. The syntax for
7222 the individual commands are colon-separated, for example;
7223 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7226 *) New dynamic control command support for ENGINEs. ENGINEs can now
7227 declare their own commands (numbers), names (strings), descriptions,
7228 and input types for run-time discovery by calling applications. A
7229 subset of these commands are implicitly classed as "executable"
7230 depending on their input type, and only these can be invoked through
7231 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7232 can be based on user input, config files, etc). The distinction is
7233 that "executable" commands cannot return anything other than a boolean
7234 result and can only support numeric or string input, whereas some
7235 discoverable commands may only be for direct use through
7236 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7237 pointers, or other custom uses. The "executable" commands are to
7238 support parameterisations of ENGINE behaviour that can be
7239 unambiguously defined by ENGINEs and used consistently across any
7240 OpenSSL-based application. Commands have been added to all the
7241 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7242 control over shared-library paths without source code alterations.
7245 *) Changed all ENGINE implementations to dynamically allocate their
7246 ENGINEs rather than declaring them statically. Apart from this being
7247 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7248 this also allows the implementations to compile without using the
7249 internal engine_int.h header.
7252 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7253 'const' value. Any code that should be able to modify a RAND_METHOD
7254 should already have non-const pointers to it (ie. they should only
7255 modify their own ones).
7258 *) Made a variety of little tweaks to the ENGINE code.
7259 - "atalla" and "ubsec" string definitions were moved from header files
7260 to C code. "nuron" string definitions were placed in variables
7261 rather than hard-coded - allowing parameterisation of these values
7262 later on via ctrl() commands.
7263 - Removed unused "#if 0"'d code.
7264 - Fixed engine list iteration code so it uses ENGINE_free() to release
7265 structural references.
7266 - Constified the RAND_METHOD element of ENGINE structures.
7267 - Constified various get/set functions as appropriate and added
7268 missing functions (including a catch-all ENGINE_cpy that duplicates
7269 all ENGINE values onto a new ENGINE except reference counts/state).
7270 - Removed NULL parameter checks in get/set functions. Setting a method
7271 or function to NULL is a way of cancelling out a previously set
7272 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7273 and doesn't justify the extra error symbols and code.
7274 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7275 flags from engine_int.h to engine.h.
7276 - Changed prototypes for ENGINE handler functions (init(), finish(),
7277 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7280 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7281 to the algorithm using long division. The binary algorithm can be
7282 used only if the modulus is odd. On 32-bit systems, it is faster
7283 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7284 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7285 up to 450 bits. In 64-bit environments, the binary algorithm
7286 appears to be advantageous for much longer moduli; here we use it
7287 for moduli up to 2048 bits.
7290 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7291 could not support the combine flag in choice fields.
7294 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7295 extensions from a certificate request to the certificate.
7298 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7299 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7300 file: this allows the display of the certificate about to be
7301 signed to be customised, to allow certain fields to be included
7302 or excluded and extension details. The old system didn't display
7303 multicharacter strings properly, omitted fields not in the policy
7304 and couldn't display additional details such as extensions.
7307 *) Function EC_POINTs_mul for multiple scalar multiplication
7308 of an arbitrary number of elliptic curve points
7309 \sum scalars[i]*points[i],
7310 optionally including the generator defined for the EC_GROUP:
7311 scalar*generator + \sum scalars[i]*points[i].
7313 EC_POINT_mul is a simple wrapper function for the typical case
7314 that the point list has just one item (besides the optional
7318 *) First EC_METHODs for curves over GF(p):
7320 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7321 operations and provides various method functions that can also
7322 operate with faster implementations of modular arithmetic.
7324 EC_GFp_mont_method() reuses most functions that are part of
7325 EC_GFp_simple_method, but uses Montgomery arithmetic.
7327 [Bodo Moeller; point addition and point doubling
7328 implementation directly derived from source code provided by
7329 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7331 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7332 crypto/ec/ec_lib.c):
7334 Curves are EC_GROUP objects (with an optional group generator)
7335 based on EC_METHODs that are built into the library.
7337 Points are EC_POINT objects based on EC_GROUP objects.
7339 Most of the framework would be able to handle curves over arbitrary
7340 finite fields, but as there are no obvious types for fields other
7341 than GF(p), some functions are limited to that for now.
7344 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7345 that the file contains a complete HTTP response.
7348 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7349 change the def and num file printf format specifier from "%-40sXXX"
7350 to "%-39s XXX". The latter will always guarantee a space after the
7351 field while the former will cause them to run together if the field
7352 is 40 of more characters long.
7355 *) Constify the cipher and digest 'method' functions and structures
7356 and modify related functions to take constant EVP_MD and EVP_CIPHER
7360 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7361 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7364 *) Modify EVP_Digest*() routines so they now return values. Although the
7365 internal software routines can never fail additional hardware versions
7369 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7371 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7372 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7375 ERR_R_NESTED_ASN1_ERROR
7377 ERR_R_MISSING_ASN1_EOS
7378 were 4 .. 9, conflicting with
7379 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7381 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7382 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7384 Add new error code 'ERR_R_INTERNAL_ERROR'.
7387 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7391 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7392 sets the subject name for a new request or supersedes the
7393 subject name in a given request. Formats that can be parsed are
7394 'CN=Some Name, OU=myOU, C=IT'
7396 'CN=Some Name/OU=myOU/C=IT'.
7398 Add options '-batch' and '-verbose' to 'openssl req'.
7399 [Massimiliano Pala <madwolf@hackmasters.net>]
7401 *) Introduce the possibility to access global variables through
7402 functions on platform were that's the best way to handle exporting
7403 global variables in shared libraries. To enable this functionality,
7404 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7405 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7406 is normally done by Configure or something similar).
7408 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7409 in the source file (foo.c) like this:
7411 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7412 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7414 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7415 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7417 OPENSSL_DECLARE_GLOBAL(int,foo);
7418 #define foo OPENSSL_GLOBAL_REF(foo)
7419 OPENSSL_DECLARE_GLOBAL(double,bar);
7420 #define bar OPENSSL_GLOBAL_REF(bar)
7422 The #defines are very important, and therefore so is including the
7423 header file everywhere where the defined globals are used.
7425 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7426 of ASN.1 items, but that structure is a bit different.
7428 The largest change is in util/mkdef.pl which has been enhanced with
7429 better and easier to understand logic to choose which symbols should
7430 go into the Windows .def files as well as a number of fixes and code
7431 cleanup (among others, algorithm keywords are now sorted
7432 lexicographically to avoid constant rewrites).
7435 *) In BN_div() keep a copy of the sign of 'num' before writing the
7436 result to 'rm' because if rm==num the value will be overwritten
7437 and produce the wrong result if 'num' is negative: this caused
7438 problems with BN_mod() and BN_nnmod().
7441 *) Function OCSP_request_verify(). This checks the signature on an
7442 OCSP request and verifies the signer certificate. The signer
7443 certificate is just checked for a generic purpose and OCSP request
7447 *) Add OCSP_check_validity() function to check the validity of OCSP
7448 responses. OCSP responses are prepared in real time and may only
7449 be a few seconds old. Simply checking that the current time lies
7450 between thisUpdate and nextUpdate max reject otherwise valid responses
7451 caused by either OCSP responder or client clock inaccuracy. Instead
7452 we allow thisUpdate and nextUpdate to fall within a certain period of
7453 the current time. The age of the response can also optionally be
7454 checked. Two new options -validity_period and -status_age added to
7458 *) If signature or public key algorithm is unrecognized print out its
7459 OID rather that just UNKNOWN.
7462 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7463 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7464 ID to be generated from the issuer certificate alone which can then be
7465 passed to OCSP_id_issuer_cmp().
7468 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7469 ASN1 modules to export functions returning ASN1_ITEM pointers
7470 instead of the ASN1_ITEM structures themselves. This adds several
7471 new macros which allow the underlying ASN1 function/structure to
7472 be accessed transparently. As a result code should not use ASN1_ITEM
7473 references directly (such as &X509_it) but instead use the relevant
7474 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7475 use of the new ASN1 code on platforms where exporting structures
7476 is problematical (for example in shared libraries) but exporting
7477 functions returning pointers to structures is not.
7480 *) Add support for overriding the generation of SSL/TLS session IDs.
7481 These callbacks can be registered either in an SSL_CTX or per SSL.
7482 The purpose of this is to allow applications to control, if they wish,
7483 the arbitrary values chosen for use as session IDs, particularly as it
7484 can be useful for session caching in multiple-server environments. A
7485 command-line switch for testing this (and any client code that wishes
7486 to use such a feature) has been added to "s_server".
7487 [Geoff Thorpe, Lutz Jaenicke]
7489 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7490 of the form '#if defined(...) || defined(...) || ...' and
7491 '#if !defined(...) && !defined(...) && ...'. This also avoids
7492 the growing number of special cases it was previously handling.
7495 *) Make all configuration macros available for application by making
7496 sure they are available in opensslconf.h, by giving them names starting
7497 with "OPENSSL_" to avoid conflicts with other packages and by making
7498 sure e_os2.h will cover all platform-specific cases together with
7500 Additionally, it is now possible to define configuration/platform-
7501 specific names (called "system identities"). In the C code, these
7502 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7503 macro with the name beginning with "OPENSSL_SYS_", which is determined
7504 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7508 *) New option -set_serial to 'req' and 'x509' this allows the serial
7509 number to use to be specified on the command line. Previously self
7510 signed certificates were hard coded with serial number 0 and the
7511 CA options of 'x509' had to use a serial number in a file which was
7515 *) New options to 'ca' utility to support V2 CRL entry extensions.
7516 Currently CRL reason, invalidity date and hold instruction are
7517 supported. Add new CRL extensions to V3 code and some new objects.
7520 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7521 disable standard block padding (aka PKCS#5 padding) in the EVP
7522 API, which was previously mandatory. This means that the data is
7523 not padded in any way and so the total length much be a multiple
7524 of the block size, otherwise an error occurs.
7527 *) Initial (incomplete) OCSP SSL support.
7530 *) New function OCSP_parse_url(). This splits up a URL into its host,
7531 port and path components: primarily to parse OCSP URLs. New -url
7532 option to ocsp utility.
7535 *) New nonce behavior. The return value of OCSP_check_nonce() now
7536 reflects the various checks performed. Applications can decide
7537 whether to tolerate certain situations such as an absent nonce
7538 in a response when one was present in a request: the ocsp application
7539 just prints out a warning. New function OCSP_add1_basic_nonce()
7540 this is to allow responders to include a nonce in a response even if
7541 the request is nonce-less.
7544 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7545 skipped when using openssl x509 multiple times on a single input file,
7546 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7549 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7550 set string type: to handle setting ASN1_TIME structures. Fix ca
7551 utility to correctly initialize revocation date of CRLs.
7554 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7555 the clients preferred ciphersuites and rather use its own preferences.
7556 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7557 Internet Explorer by ensuring unchanged hash method during stepup.
7558 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7561 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7562 to aes and add a new 'exist' option to print out symbols that don't
7566 *) Additional options to ocsp utility to allow flags to be set and
7567 additional certificates supplied.
7570 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7571 OCSP client a number of certificate to only verify the response
7575 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7576 handle the new API. Currently only ECB, CBC modes supported. Add new
7579 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7580 Encryption Standard (AES) Ciphersuites for Transport Layer
7581 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7582 not enabled by default and were not part of the "ALL" ciphersuite
7583 alias because they were not yet official; they could be
7584 explicitly requested by specifying the "AESdraft" ciphersuite
7585 group alias. In the final release of OpenSSL 0.9.7, the group
7586 alias is called "AES" and is part of "ALL".)
7587 [Ben Laurie, Steve Henson, Bodo Moeller]
7589 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7590 request to response.
7593 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7594 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7595 extract information from a certificate request. OCSP_response_create()
7596 creates a response and optionally adds a basic response structure.
7597 OCSP_basic_add1_status() adds a complete single response to a basic
7598 response and returns the OCSP_SINGLERESP structure just added (to allow
7599 extensions to be included for example). OCSP_basic_add1_cert() adds a
7600 certificate to a basic response and OCSP_basic_sign() signs a basic
7601 response with various flags. New helper functions ASN1_TIME_check()
7602 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7603 (converts ASN1_TIME to GeneralizedTime).
7606 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7607 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7608 structure from a certificate. X509_pubkey_digest() digests the public_key
7609 contents: this is used in various key identifiers.
7612 *) Make sk_sort() tolerate a NULL argument.
7613 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7615 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7616 passed by the function are trusted implicitly. If any of them signed the
7617 response then it is assumed to be valid and is not verified.
7620 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7621 to data. This was previously part of the PKCS7 ASN1 code. This
7622 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7623 [Steve Henson, reported by Kenneth R. Robinette
7624 <support@securenetterm.com>]
7626 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7627 routines: without these tracing memory leaks is very painful.
7628 Fix leaks in PKCS12 and PKCS7 routines.
7631 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7632 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7633 effectively meant GeneralizedTime would never be used. Now it
7634 is initialised to -1 but X509_time_adj() now has to check the value
7635 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7636 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7637 [Steve Henson, reported by Kenneth R. Robinette
7638 <support@securenetterm.com>]
7640 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7641 result in a zero length in the ASN1_INTEGER structure which was
7642 not consistent with the structure when d2i_ASN1_INTEGER() was used
7643 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7644 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7645 where it did not print out a minus for negative ASN1_INTEGER.
7648 *) Add summary printout to ocsp utility. The various functions which
7649 convert status values to strings have been renamed to:
7650 OCSP_response_status_str(), OCSP_cert_status_str() and
7651 OCSP_crl_reason_str() and are no longer static. New options
7652 to verify nonce values and to disable verification. OCSP response
7653 printout format cleaned up.
7656 *) Add additional OCSP certificate checks. These are those specified
7657 in RFC2560. This consists of two separate checks: the CA of the
7658 certificate being checked must either be the OCSP signer certificate
7659 or the issuer of the OCSP signer certificate. In the latter case the
7660 OCSP signer certificate must contain the OCSP signing extended key
7661 usage. This check is performed by attempting to match the OCSP
7662 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7663 in the OCSP_CERTID structures of the response.
7666 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7667 and related routines. This uses the standard OpenSSL certificate
7668 verify routines to perform initial checks (just CA validity) and
7669 to obtain the certificate chain. Then additional checks will be
7670 performed on the chain. Currently the root CA is checked to see
7671 if it is explicitly trusted for OCSP signing. This is used to set
7672 a root CA as a global signing root: that is any certificate that
7673 chains to that CA is an acceptable OCSP signing certificate.
7676 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7677 extensions from a separate configuration file.
7678 As when reading extensions from the main configuration file,
7679 the '-extensions ...' option may be used for specifying the
7681 [Massimiliano Pala <madwolf@comune.modena.it>]
7683 *) New OCSP utility. Allows OCSP requests to be generated or
7684 read. The request can be sent to a responder and the output
7685 parsed, outputed or printed in text form. Not complete yet:
7686 still needs to check the OCSP response validity.
7689 *) New subcommands for 'openssl ca':
7690 'openssl ca -status <serial>' prints the status of the cert with
7691 the given serial number (according to the index file).
7692 'openssl ca -updatedb' updates the expiry status of certificates
7694 [Massimiliano Pala <madwolf@comune.modena.it>]
7696 *) New '-newreq-nodes' command option to CA.pl. This is like
7697 '-newreq', but calls 'openssl req' with the '-nodes' option
7698 so that the resulting key is not encrypted.
7699 [Damien Miller <djm@mindrot.org>]
7701 *) New configuration for the GNU Hurd.
7702 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7704 *) Initial code to implement OCSP basic response verify. This
7705 is currently incomplete. Currently just finds the signer's
7706 certificate and verifies the signature on the response.
7709 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7710 value of OPENSSLDIR. This is available via the new '-d' option
7711 to 'openssl version', and is also included in 'openssl version -a'.
7714 *) Allowing defining memory allocation callbacks that will be given
7715 file name and line number information in additional arguments
7716 (a const char* and an int). The basic functionality remains, as
7717 well as the original possibility to just replace malloc(),
7718 realloc() and free() by functions that do not know about these
7719 additional arguments. To register and find out the current
7720 settings for extended allocation functions, the following
7721 functions are provided:
7723 CRYPTO_set_mem_ex_functions
7724 CRYPTO_set_locked_mem_ex_functions
7725 CRYPTO_get_mem_ex_functions
7726 CRYPTO_get_locked_mem_ex_functions
7728 These work the same way as CRYPTO_set_mem_functions and friends.
7729 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7730 extended allocation function is enabled.
7731 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7732 a conventional allocation function is enabled.
7733 [Richard Levitte, Bodo Moeller]
7735 *) Finish off removing the remaining LHASH function pointer casts.
7736 There should no longer be any prototype-casting required when using
7737 the LHASH abstraction, and any casts that remain are "bugs". See
7738 the callback types and macros at the head of lhash.h for details
7739 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7742 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7743 If /dev/[u]random devices are not available or do not return enough
7744 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7746 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7747 /etc/entropy will be queried once each in this sequence, querying stops
7748 when enough entropy was collected without querying more sockets.
7751 *) Change the Unix RAND_poll() variant to be able to poll several
7752 random devices, as specified by DEVRANDOM, until a sufficient amount
7753 of data has been collected. We spend at most 10 ms on each file
7754 (select timeout) and read in non-blocking mode. DEVRANDOM now
7755 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7756 (previously it was just the string "/dev/urandom"), so on typical
7757 platforms the 10 ms delay will never occur.
7758 Also separate out the Unix variant to its own file, rand_unix.c.
7759 For VMS, there's a currently-empty rand_vms.c.
7762 *) Move OCSP client related routines to ocsp_cl.c. These
7763 provide utility functions which an application needing
7764 to issue a request to an OCSP responder and analyse the
7765 response will typically need: as opposed to those which an
7766 OCSP responder itself would need which will be added later.
7768 OCSP_request_sign() signs an OCSP request with an API similar
7769 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7770 response. OCSP_response_get1_basic() extracts basic response
7771 from response. OCSP_resp_find_status(): finds and extracts status
7772 information from an OCSP_CERTID structure (which will be created
7773 when the request structure is built). These are built from lower
7774 level functions which work on OCSP_SINGLERESP structures but
7775 won't normally be used unless the application wishes to examine
7776 extensions in the OCSP response for example.
7778 Replace nonce routines with a pair of functions.
7779 OCSP_request_add1_nonce() adds a nonce value and optionally
7780 generates a random value. OCSP_check_nonce() checks the
7781 validity of the nonce in an OCSP response.
7784 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7785 This doesn't copy the supplied OCSP_CERTID and avoids the
7786 need to free up the newly created id. Change return type
7787 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7788 This can then be used to add extensions to the request.
7789 Deleted OCSP_request_new(), since most of its functionality
7790 is now in OCSP_REQUEST_new() (and the case insensitive name
7791 clash) apart from the ability to set the request name which
7792 will be added elsewhere.
7795 *) Update OCSP API. Remove obsolete extensions argument from
7796 various functions. Extensions are now handled using the new
7797 OCSP extension code. New simple OCSP HTTP function which
7798 can be used to send requests and parse the response.
7801 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7802 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7803 uses the special reorder version of SET OF to sort the attributes
7804 and reorder them to match the encoded order. This resolves a long
7805 standing problem: a verify on a PKCS7 structure just after signing
7806 it used to fail because the attribute order did not match the
7807 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7808 it uses the received order. This is necessary to tolerate some broken
7809 software that does not order SET OF. This is handled by encoding
7810 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7811 to produce the required SET OF.
7814 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7815 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7816 files to get correct declarations of the ASN.1 item variables.
7819 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7820 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7821 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7822 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7823 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7824 ASN1_ITEM and no wrapper functions.
7827 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7828 replace the old function pointer based I/O routines. Change most of
7829 the *_d2i_bio() and *_d2i_fp() functions to use these.
7832 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7833 lines, recognice more "algorithms" that can be deselected, and make
7834 it complain about algorithm deselection that isn't recognised.
7837 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7838 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7839 to use new functions. Add NO_ASN1_OLD which can be set to remove
7840 some old style ASN1 functions: this can be used to determine if old
7841 code will still work when these eventually go away.
7844 *) New extension functions for OCSP structures, these follow the
7845 same conventions as certificates and CRLs.
7848 *) New function X509V3_add1_i2d(). This automatically encodes and
7849 adds an extension. Its behaviour can be customised with various
7850 flags to append, replace or delete. Various wrappers added for
7851 certificates and CRLs.
7854 *) Fix to avoid calling the underlying ASN1 print routine when
7855 an extension cannot be parsed. Correct a typo in the
7856 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7859 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7860 entries for variables.
7863 *) Add functionality to apps/openssl.c for detecting locking
7864 problems: As the program is single-threaded, all we have
7865 to do is register a locking callback using an array for
7866 storing which locks are currently held by the program.
7869 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7870 SSL_get_ex_data_X509_STORE_idx(), which is used in
7871 ssl_verify_cert_chain() and thus can be called at any time
7872 during TLS/SSL handshakes so that thread-safety is essential.
7873 Unfortunately, the ex_data design is not at all suited
7874 for multi-threaded use, so it probably should be abolished.
7877 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7878 [Broadcom, tweaked and integrated by Geoff Thorpe]
7880 *) Move common extension printing code to new function
7881 X509V3_print_extensions(). Reorganise OCSP print routines and
7882 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7885 *) New function X509_signature_print() to remove duplication in some
7889 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7890 set (this was treated exactly the same as SET OF previously). This
7891 is used to reorder the STACK representing the structure to match the
7892 encoding. This will be used to get round a problem where a PKCS7
7893 structure which was signed could not be verified because the STACK
7894 order did not reflect the encoded order.
7897 *) Reimplement the OCSP ASN1 module using the new code.
7900 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7901 for its ASN1 operations. The old style function pointers still exist
7902 for now but they will eventually go away.
7905 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7906 completely replaces the old ASN1 functionality with a table driven
7907 encoder and decoder which interprets an ASN1_ITEM structure describing
7908 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7909 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7910 has also been converted to the new form.
7913 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7914 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7915 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7916 for negative moduli.
7919 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7920 of not touching the result's sign bit.
7923 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7927 *) Changed the LHASH code to use prototypes for callbacks, and created
7928 macros to declare and implement thin (optionally static) functions
7929 that provide type-safety and avoid function pointer casting for the
7930 type-specific callbacks.
7933 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7935 [Veers Staats <staatsvr@asc.hpc.mil>,
7936 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7938 *) Reformat the FAQ so the different questions and answers can be divided
7939 in sections depending on the subject.
7942 *) Have the zlib compression code load ZLIB.DLL dynamically under
7946 *) New function BN_mod_sqrt for computing square roots modulo a prime
7947 (using the probabilistic Tonelli-Shanks algorithm unless
7948 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7949 be handled deterministically).
7950 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7952 *) Make BN_mod_inverse faster by explicitly handling small quotients
7953 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7954 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7957 *) New function BN_kronecker.
7960 *) Fix BN_gcd so that it works on negative inputs; the result is
7961 positive unless both parameters are zero.
7962 Previously something reasonably close to an infinite loop was
7963 possible because numbers could be growing instead of shrinking
7964 in the implementation of Euclid's algorithm.
7967 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7968 sign of the number in question.
7970 Fix BN_is_word(a,w) to work correctly for w == 0.
7972 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7973 because its test if the absolute value of 'a' equals 'w'.
7974 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7975 it exists mostly for use in the implementations of BN_is_zero(),
7976 BN_is_one(), and BN_is_word().
7979 *) New function BN_swap.
7982 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7983 the exponentiation functions are more likely to produce reasonable
7984 results on negative inputs.
7987 *) Change BN_mod_mul so that the result is always non-negative.
7988 Previously, it could be negative if one of the factors was negative;
7989 I don't think anyone really wanted that behaviour.
7992 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7993 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7994 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7995 and add new functions:
8004 BN_mod_lshift1_quick
8008 These functions always generate non-negative results.
8010 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8011 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8013 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8014 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8015 be reduced modulo m.
8016 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8019 The following entry accidentally appeared in the CHANGES file
8020 distributed with OpenSSL 0.9.7. The modifications described in
8021 it do *not* apply to OpenSSL 0.9.7.
8023 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8024 was actually never needed) and in BN_mul(). The removal in BN_mul()
8025 required a small change in bn_mul_part_recursive() and the addition
8026 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8027 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8028 bn_sub_words() and bn_add_words() except they take arrays with
8033 *) In 'openssl passwd', verify passwords read from the terminal
8034 unless the '-salt' option is used (which usually means that
8035 verification would just waste user's time since the resulting
8036 hash is going to be compared with some given password hash)
8037 or the new '-noverify' option is used.
8039 This is an incompatible change, but it does not affect
8040 non-interactive use of 'openssl passwd' (passwords on the command
8041 line, '-stdin' option, '-in ...' option) and thus should not
8045 *) Remove all references to RSAref, since there's no more need for it.
8048 *) Make DSO load along a path given through an environment variable
8049 (SHLIB_PATH) with shl_load().
8052 *) Constify the ENGINE code as a result of BIGNUM constification.
8053 Also constify the RSA code and most things related to it. In a
8054 few places, most notable in the depth of the ASN.1 code, ugly
8055 casts back to non-const were required (to be solved at a later
8059 *) Make it so the openssl application has all engines loaded by default.
8062 *) Constify the BIGNUM routines a little more.
8065 *) Add the following functions:
8067 ENGINE_load_cswift()
8069 ENGINE_load_atalla()
8071 ENGINE_load_builtin_engines()
8073 That way, an application can itself choose if external engines that
8074 are built-in in OpenSSL shall ever be used or not. The benefit is
8075 that applications won't have to be linked with libdl or other dso
8076 libraries unless it's really needed.
8078 Changed 'openssl engine' to load all engines on demand.
8079 Changed the engine header files to avoid the duplication of some
8080 declarations (they differed!).
8083 *) 'openssl engine' can now list capabilities.
8086 *) Better error reporting in 'openssl engine'.
8089 *) Never call load_dh_param(NULL) in s_server.
8092 *) Add engine application. It can currently list engines by name and
8093 identity, and test if they are actually available.
8096 *) Improve RPM specification file by forcing symbolic linking and making
8097 sure the installed documentation is also owned by root.root.
8098 [Damien Miller <djm@mindrot.org>]
8100 *) Give the OpenSSL applications more possibilities to make use of
8101 keys (public as well as private) handled by engines.
8104 *) Add OCSP code that comes from CertCo.
8107 *) Add VMS support for the Rijndael code.
8110 *) Added untested support for Nuron crypto accelerator.
8113 *) Add support for external cryptographic devices. This code was
8114 previously distributed separately as the "engine" branch.
8115 [Geoff Thorpe, Richard Levitte]
8117 *) Rework the filename-translation in the DSO code. It is now possible to
8118 have far greater control over how a "name" is turned into a filename
8119 depending on the operating environment and any oddities about the
8120 different shared library filenames on each system.
8123 *) Support threads on FreeBSD-elf in Configure.
8126 *) Fix for SHA1 assembly problem with MASM: it produces
8127 warnings about corrupt line number information when assembling
8128 with debugging information. This is caused by the overlapping
8130 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8133 NCONF_get_number() has no error checking at all. As a replacement,
8134 NCONF_get_number_e() is defined (_e for "error checking") and is
8135 promoted strongly. The old NCONF_get_number is kept around for
8136 binary backward compatibility.
8137 Make it possible for methods to load from something other than a BIO,
8138 by providing a function pointer that is given a name instead of a BIO.
8139 For example, this could be used to load configuration data from an
8143 *) Fix for non blocking accept BIOs. Added new I/O special reason
8144 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8145 with non blocking I/O was not possible because no retry code was
8146 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8150 *) Added the beginnings of Rijndael support.
8153 *) Fix for bug in DirectoryString mask setting. Add support for
8154 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8155 to allow certificate printing to more controllable, additional
8156 'certopt' option to 'x509' to allow new printing options to be
8160 *) Clean old EAY MD5 hack from e_os.h.
8163 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8165 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8166 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8167 [Joe Orton, Steve Henson]
8169 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8171 *) Fix additional bug revealed by the NISCC test suite:
8173 Stop bug triggering large recursion when presented with
8174 certain ASN.1 tags (CVE-2003-0851)
8177 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8179 *) Fix various bugs revealed by running the NISCC test suite:
8181 Stop out of bounds reads in the ASN1 code when presented with
8182 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8184 If verify callback ignores invalid public key errors don't try to check
8185 certificate signature with the NULL public key.
8189 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8190 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8194 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8195 extra data after the compression methods not only for TLS 1.0
8196 but also for SSL 3.0 (as required by the specification).
8197 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8199 *) Change X509_certificate_type() to mark the key as exported/exportable
8200 when it's 512 *bits* long, not 512 bytes.
8203 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8205 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8206 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8207 a protocol version number mismatch like a decryption error
8208 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8211 *) Turn on RSA blinding by default in the default implementation
8212 to avoid a timing attack. Applications that don't want it can call
8213 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8214 They would be ill-advised to do so in most cases.
8215 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8217 *) Change RSA blinding code so that it works when the PRNG is not
8218 seeded (in this case, the secret RSA exponent is abused as
8219 an unpredictable seed -- if it is not unpredictable, there
8220 is no point in blinding anyway). Make RSA blinding thread-safe
8221 by remembering the creator's thread ID in rsa->blinding and
8222 having all other threads use local one-time blinding factors
8223 (this requires more computation than sharing rsa->blinding, but
8224 avoids excessive locking; and if an RSA object is not shared
8225 between threads, blinding will still be very fast).
8228 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8230 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8231 via timing by performing a MAC computation even if incorrrect
8232 block cipher padding has been found. This is a countermeasure
8233 against active attacks where the attacker has to distinguish
8234 between bad padding and a MAC verification error. (CVE-2003-0078)
8236 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8237 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8238 Martin Vuagnoux (EPFL, Ilion)]
8240 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8242 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8243 memory from it's contents. This is done with a counter that will
8244 place alternating values in each byte. This can be used to solve
8245 two issues: 1) the removal of calls to memset() by highly optimizing
8246 compilers, and 2) cleansing with other values than 0, since those can
8247 be read through on certain media, for example a swap space on disk.
8250 *) Bugfix: client side session caching did not work with external caching,
8251 because the session->cipher setting was not restored when reloading
8252 from the external cache. This problem was masked, when
8253 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8254 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8257 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8258 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8259 [Zeev Lieber <zeev-l@yahoo.com>]
8261 *) Undo an undocumented change introduced in 0.9.6e which caused
8262 repeated calls to OpenSSL_add_all_ciphers() and
8263 OpenSSL_add_all_digests() to be ignored, even after calling
8267 *) Change the default configuration reader to deal with last line not
8268 being properly terminated.
8271 *) Change X509_NAME_cmp() so it applies the special rules on handling
8272 DN values that are of type PrintableString, as well as RDNs of type
8273 emailAddress where the value has the type ia5String.
8274 [stefank@valicert.com via Richard Levitte]
8276 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8277 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8278 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8279 the bitwise-OR of the two for use by the majority of applications
8280 wanting this behaviour, and update the docs. The documented
8281 behaviour and actual behaviour were inconsistent and had been
8282 changing anyway, so this is more a bug-fix than a behavioural
8284 [Geoff Thorpe, diagnosed by Nadav Har'El]
8286 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8287 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8290 *) Fix initialization code race conditions in
8291 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8292 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8293 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8294 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8295 ssl2_get_cipher_by_char(),
8296 ssl3_get_cipher_by_char().
8297 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8299 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8300 the cached sessions are flushed, as the remove_cb() might use ex_data
8301 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8302 (see [openssl.org #212]).
8303 [Geoff Thorpe, Lutz Jaenicke]
8305 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8306 length, instead of the encoding length to d2i_ASN1_OBJECT.
8309 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8311 *) [In 0.9.6g-engine release:]
8312 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8313 [Lynn Gazis <lgazis@rainbow.com>]
8315 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8317 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8318 and get fix the header length calculation.
8319 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8320 Alon Kantor <alonk@checkpoint.com> (and others),
8323 *) Use proper error handling instead of 'assertions' in buffer
8324 overflow checks added in 0.9.6e. This prevents DoS (the
8325 assertions could call abort()).
8326 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8328 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8330 *) Add various sanity checks to asn1_get_length() to reject
8331 the ASN1 length bytes if they exceed sizeof(long), will appear
8332 negative or the content length exceeds the length of the
8334 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8336 *) Fix cipher selection routines: ciphers without encryption had no flags
8337 for the cipher strength set and where therefore not handled correctly
8338 by the selection routines (PR #130).
8341 *) Fix EVP_dsa_sha macro.
8345 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8346 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8347 that was added in OpenSSL 0.9.6d.
8349 As the countermeasure turned out to be incompatible with some
8350 broken SSL implementations, the new option is part of SSL_OP_ALL.
8351 SSL_OP_ALL is usually employed when compatibility with weird SSL
8352 implementations is desired (e.g. '-bugs' option to 's_client' and
8353 's_server'), so the new option is automatically set in many
8357 *) Changes in security patch:
8359 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8360 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8361 Air Force Materiel Command, USAF, under agreement number
8364 *) Add various sanity checks to asn1_get_length() to reject
8365 the ASN1 length bytes if they exceed sizeof(long), will appear
8366 negative or the content length exceeds the length of the
8367 supplied buffer. (CVE-2002-0659)
8368 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8370 *) Assertions for various potential buffer overflows, not known to
8372 [Ben Laurie (CHATS)]
8374 *) Various temporary buffers to hold ASCII versions of integers were
8375 too small for 64 bit platforms. (CVE-2002-0655)
8376 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8378 *) Remote buffer overflow in SSL3 protocol - an attacker could
8379 supply an oversized session ID to a client. (CVE-2002-0656)
8380 [Ben Laurie (CHATS)]
8382 *) Remote buffer overflow in SSL2 protocol - an attacker could
8383 supply an oversized client master key. (CVE-2002-0656)
8384 [Ben Laurie (CHATS)]
8386 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8388 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8389 encoded as NULL) with id-dsa-with-sha1.
8390 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8392 *) Check various X509_...() return values in apps/req.c.
8393 [Nils Larsch <nla@trustcenter.de>]
8395 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8396 an end-of-file condition would erroneously be flagged, when the CRLF
8397 was just at the end of a processed block. The bug was discovered when
8398 processing data through a buffering memory BIO handing the data to a
8399 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8400 <ptsekov@syntrex.com> and Nedelcho Stanev.
8403 *) Implement a countermeasure against a vulnerability recently found
8404 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8405 before application data chunks to avoid the use of known IVs
8406 with data potentially chosen by the attacker.
8409 *) Fix length checks in ssl3_get_client_hello().
8412 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8413 to prevent ssl3_read_internal() from incorrectly assuming that
8414 ssl3_read_bytes() found application data while handshake
8415 processing was enabled when in fact s->s3->in_read_app_data was
8416 merely automatically cleared during the initial handshake.
8417 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8419 *) Fix object definitions for Private and Enterprise: they were not
8420 recognized in their shortname (=lowercase) representation. Extend
8421 obj_dat.pl to issue an error when using undefined keywords instead
8422 of silently ignoring the problem (Svenning Sorensen
8423 <sss@sss.dnsalias.net>).
8426 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8427 generators, i.e. generators other than 2 and 5. (Previously, the
8428 code did not properly initialise the 'add' and 'rem' values to
8429 BN_generate_prime().)
8431 In the new general case, we do not insist that 'generator' is
8432 actually a primitive root: This requirement is rather pointless;
8433 a generator of the order-q subgroup is just as good, if not
8437 *) Map new X509 verification errors to alerts. Discovered and submitted by
8438 Tom Wu <tom@arcot.com>.
8441 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8442 returning non-zero before the data has been completely received
8443 when using non-blocking I/O.
8444 [Bodo Moeller; problem pointed out by John Hughes]
8446 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8447 [Ben Laurie, Lutz Jaenicke]
8449 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8450 Yoram Zahavi <YoramZ@gilian.com>).
8453 *) Add information about CygWin 1.3 and on, and preserve proper
8454 configuration for the versions before that.
8455 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8457 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8458 check whether we deal with a copy of a session and do not delete from
8459 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8460 <izhar@checkpoint.com>.
8463 *) Do not store session data into the internal session cache, if it
8464 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8465 flag is set). Proposed by Aslam <aslam@funk.com>.
8468 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8472 *) [In 0.9.6d-engine release:]
8473 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8474 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8476 *) Add the configuration target linux-s390x.
8477 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8479 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8480 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8481 variable as an indication that a ClientHello message has been
8482 received. As the flag value will be lost between multiple
8483 invocations of ssl3_accept when using non-blocking I/O, the
8484 function may not be aware that a handshake has actually taken
8485 place, thus preventing a new session from being added to the
8488 To avoid this problem, we now set s->new_session to 2 instead of
8489 using a local variable.
8490 [Lutz Jaenicke, Bodo Moeller]
8492 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8493 if the SSL_R_LENGTH_MISMATCH error is detected.
8494 [Geoff Thorpe, Bodo Moeller]
8496 *) New 'shared_ldflag' column in Configure platform table.
8499 *) Fix EVP_CIPHER_mode macro.
8500 ["Dan S. Camper" <dan@bti.net>]
8502 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8503 type, we must throw them away by setting rr->length to 0.
8504 [D P Chang <dpc@qualys.com>]
8506 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8508 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8509 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8510 worked incorrectly for those cases where range = 10..._2 and
8511 3*range is two bits longer than range.)
8514 *) Only add signing time to PKCS7 structures if it is not already
8518 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8519 OBJ_ld_ce should be OBJ_id_ce.
8520 Also some ip-pda OIDs in crypto/objects/objects.txt were
8521 incorrect (cf. RFC 3039).
8522 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8524 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8525 returns early because it has nothing to do.
8526 [Andy Schneider <andy.schneider@bjss.co.uk>]
8528 *) [In 0.9.6c-engine release:]
8529 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8530 [Andy Schneider <andy.schneider@bjss.co.uk>]
8532 *) [In 0.9.6c-engine release:]
8533 Add support for Cryptographic Appliance's keyserver technology.
8534 (Use engine 'keyclient')
8535 [Cryptographic Appliances and Geoff Thorpe]
8537 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8538 is called via tools/c89.sh because arguments have to be
8539 rearranged (all '-L' options must appear before the first object
8541 [Richard Shapiro <rshapiro@abinitio.com>]
8543 *) [In 0.9.6c-engine release:]
8544 Add support for Broadcom crypto accelerator cards, backported
8546 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8548 *) [In 0.9.6c-engine release:]
8549 Add support for SureWare crypto accelerator cards from
8550 Baltimore Technologies. (Use engine 'sureware')
8551 [Baltimore Technologies and Mark Cox]
8553 *) [In 0.9.6c-engine release:]
8554 Add support for crypto accelerator cards from Accelerated
8555 Encryption Processing, www.aep.ie. (Use engine 'aep')
8556 [AEP Inc. and Mark Cox]
8558 *) Add a configuration entry for gcc on UnixWare.
8559 [Gary Benson <gbenson@redhat.com>]
8561 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8562 messages are stored in a single piece (fixed-length part and
8563 variable-length part combined) and fix various bugs found on the way.
8566 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8567 instead. BIO_gethostbyname() does not know what timeouts are
8568 appropriate, so entries would stay in cache even when they have
8570 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8572 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8573 faced with a pathologically small ClientHello fragment that does
8574 not contain client_version: Instead of aborting with an error,
8575 simply choose the highest available protocol version (i.e.,
8576 TLS 1.0 unless it is disabled). In practice, ClientHello
8577 messages are never sent like this, but this change gives us
8578 strictly correct behaviour at least for TLS.
8581 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8582 never resets s->method to s->ctx->method when called from within
8583 one of the SSL handshake functions.
8584 [Bodo Moeller; problem pointed out by Niko Baric]
8586 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8587 (sent using the client's version number) if client_version is
8588 smaller than the protocol version in use. Also change
8589 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8590 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8591 the client will at least see that alert.
8594 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8598 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8599 client receives HelloRequest while in a handshake.
8600 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8602 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8603 should end in 'break', not 'goto end' which circumvents various
8604 cleanups done in state SSL_ST_OK. But session related stuff
8605 must be disabled for SSL_ST_OK in the case that we just sent a
8608 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8609 before just sending a HelloRequest.
8610 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8612 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8613 reveal whether illegal block cipher padding was found or a MAC
8614 verification error occurred. (Neither SSLerr() codes nor alerts
8615 are directly visible to potential attackers, but the information
8616 may leak via logfiles.)
8618 Similar changes are not required for the SSL 2.0 implementation
8619 because the number of padding bytes is sent in clear for SSL 2.0,
8620 and the extra bytes are just ignored. However ssl/s2_pkt.c
8621 failed to verify that the purported number of padding bytes is in
8625 *) Add OpenUNIX-8 support including shared libraries
8626 (Boyd Lynn Gerber <gerberb@zenez.com>).
8629 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8630 'wristwatch attack' using huge encoding parameters (cf.
8631 James H. Manger's CRYPTO 2001 paper). Note that the
8632 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8633 encoding parameters and hence was not vulnerable.
8636 *) BN_sqr() bug fix.
8637 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8639 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8640 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8641 followed by modular reduction.
8642 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8644 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8645 equivalent based on BN_pseudo_rand() instead of BN_rand().
8648 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8649 This function was broken, as the check for a new client hello message
8650 to handle SGC did not allow these large messages.
8651 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8654 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8657 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8658 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8661 *) Rework the configuration and shared library support for Tru64 Unix.
8662 The configuration part makes use of modern compiler features and
8663 still retains old compiler behavior for those that run older versions
8664 of the OS. The shared library support part includes a variant that
8665 uses the RPATH feature, and is available through the special
8666 configuration target "alpha-cc-rpath", which will never be selected
8668 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8670 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8671 with the same message size as in ssl3_get_certificate_request().
8672 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8673 messages might inadvertently be reject as too long.
8674 [Petr Lampa <lampa@fee.vutbr.cz>]
8676 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8679 *) Modified SSL library such that the verify_callback that has been set
8680 specificly for an SSL object with SSL_set_verify() is actually being
8681 used. Before the change, a verify_callback set with this function was
8682 ignored and the verify_callback() set in the SSL_CTX at the time of
8683 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8684 to allow the necessary settings.
8687 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8688 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8689 done automatically (in contradiction to the requirements of the C
8690 standard). This made problems when used from OpenSSH.
8693 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8694 dh->length and always used
8696 BN_rand_range(priv_key, dh->p).
8698 BN_rand_range() is not necessary for Diffie-Hellman, and this
8699 specific range makes Diffie-Hellman unnecessarily inefficient if
8700 dh->length (recommended exponent length) is much smaller than the
8701 length of dh->p. We could use BN_rand_range() if the order of
8702 the subgroup was stored in the DH structure, but we only have
8707 BN_rand(priv_key, l, ...)
8709 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8715 RSA_eay_public_encrypt
8716 RSA_eay_private_decrypt
8717 RSA_eay_private_encrypt (signing)
8718 RSA_eay_public_decrypt (signature verification)
8720 (default implementations for RSA_public_encrypt,
8721 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8722 always reject numbers >= n.
8725 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8726 to synchronize access to 'locking_thread'. This is necessary on
8727 systems where access to 'locking_thread' (an 'unsigned long'
8728 variable) is not atomic.
8731 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8732 *before* setting the 'crypto_lock_rand' flag. The previous code had
8733 a race condition if 0 is a valid thread ID.
8734 [Travis Vitek <vitek@roguewave.com>]
8736 *) Add support for shared libraries under Irix.
8737 [Albert Chin-A-Young <china@thewrittenword.com>]
8739 *) Add configuration option to build on Linux on both big-endian and
8741 [Ralf Baechle <ralf@uni-koblenz.de>]
8743 *) Add the possibility to create shared libraries on HP-UX.
8746 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8748 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8749 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8750 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8751 PRNG state recovery was possible based on the output of
8752 one PRNG request appropriately sized to gain knowledge on
8753 'md' followed by enough consecutive 1-byte PRNG requests
8754 to traverse all of 'state'.
8756 1. When updating 'md_local' (the current thread's copy of 'md')
8757 during PRNG output generation, hash all of the previous
8758 'md_local' value, not just the half used for PRNG output.
8760 2. Make the number of bytes from 'state' included into the hash
8761 independent from the number of PRNG bytes requested.
8763 The first measure alone would be sufficient to avoid
8764 Markku-Juhani's attack. (Actually it had never occurred
8765 to me that the half of 'md_local' used for chaining was the
8766 half from which PRNG output bytes were taken -- I had always
8767 assumed that the secret half would be used.) The second
8768 measure makes sure that additional data from 'state' is never
8769 mixed into 'md_local' in small portions; this heuristically
8770 further strengthens the PRNG.
8773 *) Fix crypto/bn/asm/mips3.s.
8776 *) When only the key is given to "enc", the IV is undefined. Print out
8777 an error message in this case.
8780 *) Handle special case when X509_NAME is empty in X509 printing routines.
8783 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8784 positive and less than q.
8787 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8788 used: it isn't thread safe and the add_lock_callback should handle
8790 [Paul Rose <Paul.Rose@bridge.com>]
8792 *) Verify that incoming data obeys the block size in
8793 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8797 [Ulf Möller, Bodo Möller]
8799 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8800 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8801 when fixing the server behaviour for backwards-compatible 'client
8802 hello' messages. (Note that the attack is impractical against
8803 SSL 3.0 and TLS 1.0 anyway because length and version checking
8804 means that the probability of guessing a valid ciphertext is
8805 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8808 Before 0.9.5, the countermeasure (hide the error by generating a
8809 random 'decryption result') did not work properly because
8810 ERR_clear_error() was missing, meaning that SSL_get_error() would
8811 detect the supposedly ignored error.
8813 Both problems are now fixed.
8816 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8817 (previously it was 1024).
8820 *) Fix for compatibility mode trust settings: ignore trust settings
8821 unless some valid trust or reject settings are present.
8824 *) Fix for blowfish EVP: its a variable length cipher.
8827 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8828 parameters in DSA public key structures and return an error in the
8829 DSA routines if parameters are absent.
8832 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8833 in the current directory if neither $RANDFILE nor $HOME was set.
8834 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8835 caused some confusion to Windows users who haven't defined $HOME.
8836 Thus RAND_file_name() is changed again: e_os.h can define a
8837 DEFAULT_HOME, which will be used if $HOME is not set.
8838 For Windows, we use "C:"; on other platforms, we still require
8839 environment variables.
8841 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8842 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8843 having multiple threads call RAND_poll() concurrently.
8846 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8847 combination of a flag and a thread ID variable.
8848 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8849 flag), *other* threads can enter ssleay_add_bytes without obeying
8850 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8851 that they do not hold after the first thread unsets add_do_not_lock).
8854 *) Change bctest again: '-x' expressions are not available in all
8858 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8860 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8861 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8863 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8864 the default extension for executables, if any. Also, make the perl
8865 scripts that use symlink() to test if it really exists and use "cp"
8866 if it doesn't. All this made OpenSSL compilable and installable in
8870 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8871 If SEQUENCE is length is indefinite just set c->slen to the total
8872 amount of data available.
8873 [Steve Henson, reported by shige@FreeBSD.org]
8874 [This change does not apply to 0.9.7.]
8876 *) Change bctest to avoid here-documents inside command substitution
8877 (workaround for FreeBSD /bin/sh bug).
8878 For compatibility with Ultrix, avoid shell functions (introduced
8879 in the bctest version that searches along $PATH).
8882 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8883 with des_encrypt() defined on some operating systems, like Solaris
8887 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8888 On the Importance of Eliminating Errors in Cryptographic
8889 Computations, J. Cryptology 14 (2001) 2, 101-119,
8890 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8893 *) MIPS assembler BIGNUM division bug fix.
8896 *) Disabled incorrect Alpha assembler code.
8899 *) Fix PKCS#7 decode routines so they correctly update the length
8900 after reading an EOC for the EXPLICIT tag.
8902 [This change does not apply to 0.9.7.]
8904 *) Fix bug in PKCS#12 key generation routines. This was triggered
8905 if a 3DES key was generated with a 0 initial byte. Include
8906 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8907 (but broken) behaviour.
8910 *) Enhance bctest to search for a working bc along $PATH and print
8912 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8914 *) Fix memory leaks in err.c: free err_data string if necessary;
8915 don't write to the wrong index in ERR_set_error_data.
8918 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8922 *) Replace rdtsc with _emit statements for VC++ version 5.
8923 [Jeremy Cooper <jeremy@baymoo.org>]
8925 *) Make it possible to reuse SSLv2 sessions.
8928 *) In copy_email() check for >= 0 as a return value for
8929 X509_NAME_get_index_by_NID() since 0 is a valid index.
8930 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8932 *) Avoid coredump with unsupported or invalid public keys by checking if
8933 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8934 PKCS7_verify() fails with non detached data.
8937 *) Don't use getenv in library functions when run as setuid/setgid.
8938 New function OPENSSL_issetugid().
8941 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8942 due to incorrect handling of multi-threading:
8944 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8946 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8948 3. Count how many times MemCheck_off() has been called so that
8949 nested use can be treated correctly. This also avoids
8950 inband-signalling in the previous code (which relied on the
8951 assumption that thread ID 0 is impossible).
8954 *) Add "-rand" option also to s_client and s_server.
8957 *) Fix CPU detection on Irix 6.x.
8958 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8959 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8961 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8964 [This change does not apply to 0.9.7.]
8966 *) Use the cached encoding of an X509_NAME structure rather than
8967 copying it. This is apparently the reason for the libsafe "errors"
8968 but the code is actually correct.
8971 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8972 Bleichenbacher's DSA attack.
8973 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8974 to be set and top=0 forces the highest bit to be set; top=-1 is new
8975 and leaves the highest bit random.
8976 [Ulf Moeller, Bodo Moeller]
8978 *) In the NCONF_...-based implementations for CONF_... queries
8979 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8980 a temporary CONF structure with the data component set to NULL
8981 (which gives segmentation faults in lh_retrieve).
8982 Instead, use NULL for the CONF pointer in CONF_get_string and
8983 CONF_get_number (which may use environment variables) and directly
8984 return NULL from CONF_get_section.
8987 *) Fix potential buffer overrun for EBCDIC.
8990 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8991 keyUsage if basicConstraints absent for a CA.
8994 *) Make SMIME_write_PKCS7() write mail header values with a format that
8995 is more generally accepted (no spaces before the semicolon), since
8996 some programs can't parse those values properly otherwise. Also make
8997 sure BIO's that break lines after each write do not create invalid
9001 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9002 macros previously used would not encode an empty SEQUENCE OF
9003 and break the signature.
9005 [This change does not apply to 0.9.7.]
9007 *) Zero the premaster secret after deriving the master secret in
9011 *) Add some EVP_add_digest_alias registrations (as found in
9012 OpenSSL_add_all_digests()) to SSL_library_init()
9013 aka OpenSSL_add_ssl_algorithms(). This provides improved
9014 compatibility with peers using X.509 certificates
9015 with unconventional AlgorithmIdentifier OIDs.
9018 *) Fix for Irix with NO_ASM.
9019 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9021 *) ./config script fixes.
9022 [Ulf Moeller, Richard Levitte]
9024 *) Fix 'openssl passwd -1'.
9027 *) Change PKCS12_key_gen_asc() so it can cope with non null
9028 terminated strings whose length is passed in the passlen
9029 parameter, for example from PEM callbacks. This was done
9030 by adding an extra length parameter to asc2uni().
9031 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9033 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9034 call failed, free the DSA structure.
9037 *) Fix to uni2asc() to cope with zero length Unicode strings.
9038 These are present in some PKCS#12 files.
9041 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9042 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9043 when writing a 32767 byte record.
9044 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9046 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9047 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9049 (RSA objects have a reference count access to which is protected
9050 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9051 so they are meant to be shared between threads.)
9052 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9053 "Reddie, Steven" <Steven.Reddie@ca.com>]
9055 *) Fix a deadlock in CRYPTO_mem_leaks().
9058 *) Use better test patterns in bntest.
9061 *) rand_win.c fix for Borland C.
9064 *) BN_rshift bugfix for n == 0.
9067 *) Add a 'bctest' script that checks for some known 'bc' bugs
9068 so that 'make test' does not abort just because 'bc' is broken.
9071 *) Store verify_result within SSL_SESSION also for client side to
9072 avoid potential security hole. (Re-used sessions on the client side
9073 always resulted in verify_result==X509_V_OK, not using the original
9074 result of the server certificate verification.)
9077 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9078 SSL3_RT_APPLICATION_DATA, return 0.
9079 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9083 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9084 releases, have been re-implemented by renaming the previous
9085 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9086 and ssl3_read_internal, respectively, and adding 'peek' parameters
9087 to them. The new ssl[23]_{read,peek} functions are calls to
9088 ssl[23]_read_internal with the 'peek' flag set appropriately.
9089 A 'peek' parameter has also been added to ssl3_read_bytes, which
9090 does the actual work for ssl3_read_internal.
9093 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9094 the method-specific "init()" handler. Also clean up ex_data after
9095 calling the method-specific "finish()" handler. Previously, this was
9096 happening the other way round.
9099 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9100 The previous value, 12, was not always sufficient for BN_mod_exp().
9103 *) Make sure that shared libraries get the internal name engine with
9104 the full version number and not just 0. This should mark the
9105 shared libraries as not backward compatible. Of course, this should
9106 be changed again when we can guarantee backward binary compatibility.
9109 *) Fix typo in get_cert_by_subject() in by_dir.c
9110 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9112 *) Rework the system to generate shared libraries:
9114 - Make note of the expected extension for the shared libraries and
9115 if there is a need for symbolic links from for example libcrypto.so.0
9116 to libcrypto.so.0.9.7. There is extended info in Configure for
9119 - Make as few rebuilds of the shared libraries as possible.
9121 - Still avoid linking the OpenSSL programs with the shared libraries.
9123 - When installing, install the shared libraries separately from the
9127 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9129 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9130 and not in SSL_clear because the latter is also used by the
9131 accept/connect functions; previously, the settings made by
9132 SSL_set_read_ahead would be lost during the handshake.
9133 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9135 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9136 Previously, it would create entries for disabled algorithms no
9140 *) Added several new manual pages for SSL_* function.
9143 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9145 *) In ssl23_get_client_hello, generate an error message when faced
9146 with an initial SSL 3.0/TLS record that is too small to contain the
9147 first two bytes of the ClientHello message, i.e. client_version.
9148 (Note that this is a pathologic case that probably has never happened
9149 in real life.) The previous approach was to use the version number
9150 from the record header as a substitute; but our protocol choice
9151 should not depend on that one because it is not authenticated
9152 by the Finished messages.
9155 *) More robust randomness gathering functions for Windows.
9156 [Jeffrey Altman <jaltman@columbia.edu>]
9158 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9159 not set then we don't setup the error code for issuer check errors
9160 to avoid possibly overwriting other errors which the callback does
9161 handle. If an application does set the flag then we assume it knows
9162 what it is doing and can handle the new informational codes
9166 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9167 a general "ANY" type, as such it should be able to decode anything
9168 including tagged types. However it didn't check the class so it would
9169 wrongly interpret tagged types in the same way as their universal
9170 counterpart and unknown types were just rejected. Changed so that the
9171 tagged and unknown types are handled in the same way as a SEQUENCE:
9172 that is the encoding is stored intact. There is also a new type
9173 "V_ASN1_OTHER" which is used when the class is not universal, in this
9174 case we have no idea what the actual type is so we just lump them all
9178 *) On VMS, stdout may very well lead to a file that is written to
9179 in a record-oriented fashion. That means that every write() will
9180 write a separate record, which will be read separately by the
9181 programs trying to read from it. This can be very confusing.
9183 The solution is to put a BIO filter in the way that will buffer
9184 text until a linefeed is reached, and then write everything a
9185 line at a time, so every record written will be an actual line,
9186 not chunks of lines and not (usually doesn't happen, but I've
9187 seen it once) several lines in one record. BIO_f_linebuffer() is
9190 Currently, it's a VMS-only method, because that's where it has
9191 been tested well enough.
9194 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9195 it can return incorrect results.
9196 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9197 but it was in 0.9.6-beta[12].)
9200 *) Disable the check for content being present when verifying detached
9201 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9202 include zero length content when signing messages.
9205 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9206 BIO_ctrl (for BIO pairs).
9209 *) Add DSO method for VMS.
9212 *) Bug fix: Montgomery multiplication could produce results with the
9216 *) Add RPM specification openssl.spec and modify it to build three
9217 packages. The default package contains applications, application
9218 documentation and run-time libraries. The devel package contains
9219 include files, static libraries and function documentation. The
9220 doc package contains the contents of the doc directory. The original
9221 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9224 *) Add a large number of documentation files for many SSL routines.
9225 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9227 *) Add a configuration entry for Sony News 4.
9228 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9230 *) Don't set the two most significant bits to one when generating a
9231 random number < q in the DSA library.
9234 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9235 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9236 the underlying transport is blocking) if a handshake took place.
9237 (The default behaviour is needed by applications such as s_client
9238 and s_server that use select() to determine when to use SSL_read;
9239 but for applications that know in advance when to expect data, it
9240 just makes things more complicated.)
9243 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9247 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9248 work better on such systems.
9249 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9251 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9252 Update PKCS12_parse() so it copies the friendlyName and the
9253 keyid to the certificates aux info.
9256 *) Fix bug in PKCS7_verify() which caused an infinite loop
9257 if there was more than one signature.
9258 [Sven Uszpelkat <su@celocom.de>]
9260 *) Major change in util/mkdef.pl to include extra information
9261 about each symbol, as well as presenting variables as well
9262 as functions. This change means that there's n more need
9263 to rebuild the .num files when some algorithms are excluded.
9266 *) Allow the verify time to be set by an application,
9267 rather than always using the current time.
9270 *) Phase 2 verify code reorganisation. The certificate
9271 verify code now looks up an issuer certificate by a
9272 number of criteria: subject name, authority key id
9273 and key usage. It also verifies self signed certificates
9274 by the same criteria. The main comparison function is
9275 X509_check_issued() which performs these checks.
9277 Lot of changes were necessary in order to support this
9278 without completely rewriting the lookup code.
9280 Authority and subject key identifier are now cached.
9282 The LHASH 'certs' is X509_STORE has now been replaced
9283 by a STACK_OF(X509_OBJECT). This is mainly because an
9284 LHASH can't store or retrieve multiple objects with
9285 the same hash value.
9287 As a result various functions (which were all internal
9288 use only) have changed to handle the new X509_STORE
9289 structure. This will break anything that messed round
9290 with X509_STORE internally.
9292 The functions X509_STORE_add_cert() now checks for an
9293 exact match, rather than just subject name.
9295 The X509_STORE API doesn't directly support the retrieval
9296 of multiple certificates matching a given criteria, however
9297 this can be worked round by performing a lookup first
9298 (which will fill the cache with candidate certificates)
9299 and then examining the cache for matches. This is probably
9300 the best we can do without throwing out X509_LOOKUP
9301 entirely (maybe later...).
9303 The X509_VERIFY_CTX structure has been enhanced considerably.
9305 All certificate lookup operations now go via a get_issuer()
9306 callback. Although this currently uses an X509_STORE it
9307 can be replaced by custom lookups. This is a simple way
9308 to bypass the X509_STORE hackery necessary to make this
9309 work and makes it possible to use more efficient techniques
9310 in future. A very simple version which uses a simple
9311 STACK for its trusted certificate store is also provided
9312 using X509_STORE_CTX_trusted_stack().
9314 The verify_cb() and verify() callbacks now have equivalents
9315 in the X509_STORE_CTX structure.
9317 X509_STORE_CTX also has a 'flags' field which can be used
9318 to customise the verify behaviour.
9321 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9322 excludes S/MIME capabilities.
9325 *) When a certificate request is read in keep a copy of the
9326 original encoding of the signed data and use it when outputting
9327 again. Signatures then use the original encoding rather than
9328 a decoded, encoded version which may cause problems if the
9329 request is improperly encoded.
9332 *) For consistency with other BIO_puts implementations, call
9333 buffer_write(b, ...) directly in buffer_puts instead of calling
9336 In BIO_puts, increment b->num_write as in BIO_write.
9337 [Peter.Sylvester@EdelWeb.fr]
9339 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9340 BN_zero, we may not return a BIGNUM with an array consisting of
9344 *) Avoid calling abort() from within the library when problems are
9345 detected, except if preprocessor symbols have been defined
9346 (such as REF_CHECK, BN_DEBUG etc.).
9349 *) New openssl application 'rsautl'. This utility can be
9350 used for low level RSA operations. DER public key
9351 BIO/fp routines also added.
9354 *) New Configure entry and patches for compiling on QNX 4.
9355 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9357 *) A demo state-machine implementation was sponsored by
9358 Nuron (http://www.nuron.com/) and is now available in
9359 demos/state_machine.
9362 *) New options added to the 'dgst' utility for signature
9363 generation and verification.
9366 *) Unrecognized PKCS#7 content types are now handled via a
9367 catch all ASN1_TYPE structure. This allows unsupported
9368 types to be stored as a "blob" and an application can
9369 encode and decode it manually.
9372 *) Fix various signed/unsigned issues to make a_strex.c
9374 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9376 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9377 length if passed a buffer. ASN1_INTEGER_to_BN failed
9378 if passed a NULL BN and its argument was negative.
9379 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9381 *) Modification to PKCS#7 encoding routines to output definite
9382 length encoding. Since currently the whole structures are in
9383 memory there's not real point in using indefinite length
9384 constructed encoding. However if OpenSSL is compiled with
9385 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9388 *) Added BIO_vprintf() and BIO_vsnprintf().
9391 *) Added more prefixes to parse for in the the strings written
9392 through a logging bio, to cover all the levels that are available
9393 through syslog. The prefixes are now:
9395 PANIC, EMERG, EMR => LOG_EMERG
9396 ALERT, ALR => LOG_ALERT
9397 CRIT, CRI => LOG_CRIT
9398 ERROR, ERR => LOG_ERR
9399 WARNING, WARN, WAR => LOG_WARNING
9400 NOTICE, NOTE, NOT => LOG_NOTICE
9401 INFO, INF => LOG_INFO
9402 DEBUG, DBG => LOG_DEBUG
9404 and as before, if none of those prefixes are present at the
9405 beginning of the string, LOG_ERR is chosen.
9407 On Win32, the LOG_* levels are mapped according to this:
9409 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9410 LOG_WARNING => EVENTLOG_WARNING_TYPE
9411 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9415 *) Made it possible to reconfigure with just the configuration
9416 argument "reconf" or "reconfigure". The command line arguments
9417 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9418 and are retrieved from there when reconfiguring.
9422 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9424 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9427 *) The obj_dat.pl script was messing up the sorting of object
9428 names. The reason was that it compared the quoted version
9429 of strings as a result "OCSP" > "OCSP Signing" because
9430 " > SPACE. Changed script to store unquoted versions of
9431 names and add quotes on output. It was also omitting some
9432 names from the lookup table if they were given a default
9433 value (that is if SN is missing it is given the same
9434 value as LN and vice versa), these are now added on the
9435 grounds that if an object has a name we should be able to
9436 look it up. Finally added warning output when duplicate
9437 short or long names are found.
9440 *) Changes needed for Tandem NSK.
9441 [Scott Uroff <scott@xypro.com>]
9443 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9444 RSA_padding_check_SSLv23(), special padding was never detected
9445 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9446 version rollback attacks was not effective.
9448 In s23_clnt.c, don't use special rollback-attack detection padding
9449 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9450 client; similarly, in s23_srvr.c, don't do the rollback check if
9451 SSL 2.0 is the only protocol enabled in the server.
9454 *) Make it possible to get hexdumps of unprintable data with 'openssl
9455 asn1parse'. By implication, the functions ASN1_parse_dump() and
9456 BIO_dump_indent() are added.
9459 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9460 these print out strings and name structures based on various
9461 flags including RFC2253 support and proper handling of
9462 multibyte characters. Added options to the 'x509' utility
9463 to allow the various flags to be set.
9466 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9467 Also change the functions X509_cmp_current_time() and
9468 X509_gmtime_adj() work with an ASN1_TIME structure,
9469 this will enable certificates using GeneralizedTime in validity
9470 dates to be checked.
9473 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9474 negative public key encodings) on by default,
9475 NO_NEG_PUBKEY_BUG can be set to disable it.
9478 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9479 content octets. An i2c_ASN1_OBJECT is unnecessary because
9480 the encoding can be trivially obtained from the structure.
9483 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9484 not read locks (CRYPTO_r_[un]lock).
9487 *) A first attempt at creating official support for shared
9488 libraries through configuration. I've kept it so the
9489 default is static libraries only, and the OpenSSL programs
9490 are always statically linked for now, but there are
9491 preparations for dynamic linking in place.
9492 This has been tested on Linux and Tru64.
9495 *) Randomness polling function for Win9x, as described in:
9496 Peter Gutmann, Software Generation of Practically Strong
9500 *) Fix so PRNG is seeded in req if using an already existing
9504 *) New options to smime application. -inform and -outform
9505 allow alternative formats for the S/MIME message including
9506 PEM and DER. The -content option allows the content to be
9507 specified separately. This should allow things like Netscape
9508 form signing output easier to verify.
9511 *) Fix the ASN1 encoding of tags using the 'long form'.
9514 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9515 STRING types. These convert content octets to and from the
9516 underlying type. The actual tag and length octets are
9517 already assumed to have been read in and checked. These
9518 are needed because all other string types have virtually
9519 identical handling apart from the tag. By having versions
9520 of the ASN1 functions that just operate on content octets
9521 IMPLICIT tagging can be handled properly. It also allows
9522 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9523 and ASN1_INTEGER are identical apart from the tag.
9526 *) Change the handling of OID objects as follows:
9528 - New object identifiers are inserted in objects.txt, following
9529 the syntax given in objects.README.
9530 - objects.pl is used to process obj_mac.num and create a new
9532 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9535 This is currently kind of a hack, and the perl code in objects.pl
9536 isn't very elegant, but it works as I intended. The simplest way
9537 to check that it worked correctly is to look in obj_dat.h and
9538 check the array nid_objs and make sure the objects haven't moved
9539 around (this is important!). Additions are OK, as well as
9540 consistent name changes.
9543 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9546 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9547 The given file adds to whatever has already been seeded into the
9548 random pool through the RANDFILE configuration file option or
9549 environment variable, or the default random state file.
9552 *) mkstack.pl now sorts each macro group into lexical order.
9553 Previously the output order depended on the order the files
9554 appeared in the directory, resulting in needless rewriting
9558 *) Patches to make OpenSSL compile under Win32 again. Mostly
9559 work arounds for the VC++ problem that it treats func() as
9560 func(void). Also stripped out the parts of mkdef.pl that
9561 added extra typesafe functions: these no longer exist.
9564 *) Reorganisation of the stack code. The macros are now all
9565 collected in safestack.h . Each macro is defined in terms of
9566 a "stack macro" of the form SKM_<name>(type, a, b). The
9567 DEBUG_SAFESTACK is now handled in terms of function casts,
9568 this has the advantage of retaining type safety without the
9569 use of additional functions. If DEBUG_SAFESTACK is not defined
9570 then the non typesafe macros are used instead. Also modified the
9571 mkstack.pl script to handle the new form. Needs testing to see
9572 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9573 the default if no major problems. Similar behaviour for ASN1_SET_OF
9574 and PKCS12_STACK_OF.
9577 *) When some versions of IIS use the 'NET' form of private key the
9578 key derivation algorithm is different. Normally MD5(password) is
9579 used as a 128 bit RC4 key. In the modified case
9580 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9581 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9582 as the old Netscape_RSA functions except they have an additional
9583 'sgckey' parameter which uses the modified algorithm. Also added
9584 an -sgckey command line option to the rsa utility. Thanks to
9585 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9586 algorithm to openssl-dev.
9589 *) The evp_local.h macros were using 'c.##kname' which resulted in
9590 invalid expansion on some systems (SCO 5.0.5 for example).
9591 Corrected to 'c.kname'.
9592 [Phillip Porch <root@theporch.com>]
9594 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9595 a STACK of email addresses from a certificate or request, these look
9596 in the subject name and the subject alternative name extensions and
9597 omit any duplicate addresses.
9600 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9601 This makes DSA verification about 2 % faster.
9604 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9605 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9606 plus overhead for 1024 bit moduli).
9607 This makes exponentiations about 0.5 % faster for 1024 bit
9608 exponents (as measured by "openssl speed rsa2048").
9611 *) Rename memory handling macros to avoid conflicts with other
9613 Malloc => OPENSSL_malloc
9614 Malloc_locked => OPENSSL_malloc_locked
9615 Realloc => OPENSSL_realloc
9616 Free => OPENSSL_free
9619 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9620 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9623 *) CygWin32 support.
9624 [John Jarvie <jjarvie@newsguy.com>]
9626 *) The type-safe stack code has been rejigged. It is now only compiled
9627 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9628 by default all type-specific stack functions are "#define"d back to
9629 standard stack functions. This results in more streamlined output
9630 but retains the type-safety checking possibilities of the original
9634 *) The STACK code has been cleaned up, and certain type declarations
9635 that didn't make a lot of sense have been brought in line. This has
9636 also involved a cleanup of sorts in safestack.h to more correctly
9637 map type-safe stack functions onto their plain stack counterparts.
9638 This work has also resulted in a variety of "const"ifications of
9639 lots of the code, especially "_cmp" operations which should normally
9640 be prototyped with "const" parameters anyway.
9643 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9644 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9645 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9646 where all of 'md' is used each time the PRNG is used, but 'state'
9647 is used only indexed by a cyclic counter. As entropy may not be
9648 well distributed from the beginning, 'md' is important as a
9649 chaining variable. However, the output function chains only half
9650 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9651 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9652 in all of 'state' being rewritten, with the new values depending
9653 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9656 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9657 the handshake is continued after ssl_verify_cert_chain();
9658 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9659 can lead to 'unexplainable' connection aborts later.
9660 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9662 *) Major EVP API cipher revision.
9663 Add hooks for extra EVP features. This allows various cipher
9664 parameters to be set in the EVP interface. Support added for variable
9665 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9666 setting of RC2 and RC5 parameters.
9668 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9671 Remove lots of duplicated code from the EVP library. For example *every*
9672 cipher init() function handles the 'iv' in the same way according to the
9673 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9674 for CFB and OFB modes they zero ctx->num.
9676 New functionality allows removal of S/MIME code RC2 hack.
9678 Most of the routines have the same form and so can be declared in terms
9681 By shifting this to the top level EVP_CipherInit() it can be removed from
9682 all individual ciphers. If the cipher wants to handle IVs or keys
9683 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9686 Change lots of functions like EVP_EncryptUpdate() to now return a
9687 value: although software versions of the algorithms cannot fail
9688 any installed hardware versions can.
9691 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9692 this option is set, tolerate broken clients that send the negotiated
9693 protocol version number instead of the requested protocol version
9697 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9698 i.e. non-zero for export ciphersuites, zero otherwise.
9699 Previous versions had this flag inverted, inconsistent with
9700 rsa_tmp_cb (..._TMP_RSA_CB).
9701 [Bodo Moeller; problem reported by Amit Chopra]
9703 *) Add missing DSA library text string. Work around for some IIS
9704 key files with invalid SEQUENCE encoding.
9707 *) Add a document (doc/standards.txt) that list all kinds of standards
9708 and so on that are implemented in OpenSSL.
9711 *) Enhance c_rehash script. Old version would mishandle certificates
9712 with the same subject name hash and wouldn't handle CRLs at all.
9713 Added -fingerprint option to crl utility, to support new c_rehash
9717 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9720 *) Fix for SSL server purpose checking. Server checking was
9721 rejecting certificates which had extended key usage present
9722 but no ssl client purpose.
9723 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9725 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9726 is a little unclear about how a blank password is handled.
9727 Since the password in encoded as a BMPString with terminating
9728 double NULL a zero length password would end up as just the
9729 double NULL. However no password at all is different and is
9730 handled differently in the PKCS#12 key generation code. NS
9731 treats a blank password as zero length. MSIE treats it as no
9732 password on export: but it will try both on import. We now do
9733 the same: PKCS12_parse() tries zero length and no password if
9734 the password is set to "" or NULL (NULL is now a valid password:
9735 it wasn't before) as does the pkcs12 application.
9738 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9739 perror when PEM_read_bio_X509_REQ fails, the error message must
9740 be obtained from the error queue.
9743 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9744 it in ERR_remove_state if appropriate, and change ERR_get_state
9745 accordingly to avoid race conditions (this is necessary because
9746 thread_hash is no longer constant once set).
9749 *) Bugfix for linux-elf makefile.one.
9752 *) RSA_get_default_method() will now cause a default
9753 RSA_METHOD to be chosen if one doesn't exist already.
9754 Previously this was only set during a call to RSA_new()
9755 or RSA_new_method(NULL) meaning it was possible for
9756 RSA_get_default_method() to return NULL.
9759 *) Added native name translation to the existing DSO code
9760 that will convert (if the flag to do so is set) filenames
9761 that are sufficiently small and have no path information
9762 into a canonical native form. Eg. "blah" converted to
9763 "libblah.so" or "blah.dll" etc.
9766 *) New function ERR_error_string_n(e, buf, len) which is like
9767 ERR_error_string(e, buf), but writes at most 'len' bytes
9768 including the 0 terminator. For ERR_error_string_n, 'buf'
9770 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9772 *) CONF library reworked to become more general. A new CONF
9773 configuration file reader "class" is implemented as well as a
9774 new functions (NCONF_*, for "New CONF") to handle it. The now
9775 old CONF_* functions are still there, but are reimplemented to
9776 work in terms of the new functions. Also, a set of functions
9777 to handle the internal storage of the configuration data is
9778 provided to make it easier to write new configuration file
9779 reader "classes" (I can definitely see something reading a
9780 configuration file in XML format, for example), called _CONF_*,
9781 or "the configuration storage API"...
9783 The new configuration file reading functions are:
9785 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9786 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9788 NCONF_default, NCONF_WIN32
9790 NCONF_dump_fp, NCONF_dump_bio
9792 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9793 NCONF_new creates a new CONF object. This works in the same way
9794 as other interfaces in OpenSSL, like the BIO interface.
9795 NCONF_dump_* dump the internal storage of the configuration file,
9796 which is useful for debugging. All other functions take the same
9797 arguments as the old CONF_* functions wth the exception of the
9798 first that must be a `CONF *' instead of a `LHASH *'.
9800 To make it easer to use the new classes with the old CONF_* functions,
9801 the function CONF_set_default_method is provided.
9804 *) Add '-tls1' option to 'openssl ciphers', which was already
9805 mentioned in the documentation but had not been implemented.
9806 (This option is not yet really useful because even the additional
9807 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9810 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9811 OpenSSL-based applications) load shared libraries and bind to
9812 them in a portable way.
9813 [Geoff Thorpe, with contributions from Richard Levitte]
9815 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9817 *) Make sure _lrotl and _lrotr are only used with MSVC.
9819 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9820 (the default implementation of RAND_status).
9822 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9823 to '-clrext' (= clear extensions), as intended and documented.
9824 [Bodo Moeller; inconsistency pointed out by Michael Attili
9827 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9828 was larger than the MD block size.
9829 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9831 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9832 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9833 using the passed key: if the passed key was a private key the result
9834 of X509_print(), for example, would be to print out all the private key
9838 *) des_quad_cksum() byte order bug fix.
9839 [Ulf Möller, using the problem description in krb4-0.9.7, where
9840 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9842 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9844 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9846 *) For easily testing in shell scripts whether some command
9847 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9848 returns with exit code 0 iff no command of the given name is available.
9849 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9850 the output goes to stdout and nothing is printed to stderr.
9851 Additional arguments are always ignored.
9853 Since for each cipher there is a command of the same name,
9854 the 'no-cipher' compilation switches can be tested this way.
9856 ('openssl no-XXX' is not able to detect pseudo-commands such
9857 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9860 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9863 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9864 is set; it will be thrown away anyway because each handshake creates
9866 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9867 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9868 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9869 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9872 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9873 'Q' and 'R' lose their special meanings (quit/renegotiate).
9874 This is part of what -quiet does; unlike -quiet, -ign_eof
9875 does not suppress any output.
9878 *) Add compatibility options to the purpose and trust code. The
9879 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9880 accepts a certificate or CA, this was the previous behaviour,
9881 with all the associated security issues.
9883 X509_TRUST_COMPAT is the old trust behaviour: only and
9884 automatically trust self signed roots in certificate store. A
9885 new trust setting X509_TRUST_DEFAULT is used to specify that
9886 a purpose has no associated trust setting and it should instead
9887 use the value in the default purpose.
9890 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9891 and fix a memory leak.
9894 *) In util/mkerr.pl (which implements 'make errors'), preserve
9895 reason strings from the previous version of the .c file, as
9896 the default to have only downcase letters (and digits) in
9897 automatically generated reasons codes is not always appropriate.
9900 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9901 using strerror. Previously, ERR_reason_error_string() returned
9902 library names as reason strings for SYSerr; but SYSerr is a special
9903 case where small numbers are errno values, not library numbers.
9906 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9907 converts DSA parameters into DH parameters. (When creating parameters,
9908 DSA_generate_parameters is used.)
9911 *) Include 'length' (recommended exponent length) in C code generated
9912 by 'openssl dhparam -C'.
9915 *) The second argument to set_label in perlasm was already being used
9916 so couldn't be used as a "file scope" flag. Moved to third argument
9920 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9921 instead of RAND_bytes for encryption IVs and salts.
9924 *) Include RAND_status() into RAND_METHOD instead of implementing
9925 it only for md_rand.c Otherwise replacing the PRNG by calling
9926 RAND_set_rand_method would be impossible.
9929 *) Don't let DSA_generate_key() enter an infinite loop if the random
9930 number generation fails.
9933 *) New 'rand' application for creating pseudo-random output.
9936 *) Added configuration support for Linux/IA64
9937 [Rolf Haberrecker <rolf@suse.de>]
9939 *) Assembler module support for Mingw32.
9942 *) Shared library support for HPUX (in shlib/).
9943 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9945 *) Shared library support for Solaris gcc.
9946 [Lutz Behnke <behnke@trustcenter.de>]
9948 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9950 *) PKCS7_encrypt() was adding text MIME headers twice because they
9951 were added manually and by SMIME_crlf_copy().
9954 *) In bntest.c don't call BN_rand with zero bits argument.
9955 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9957 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9958 case was implemented. This caused BN_div_recp() to fail occasionally.
9961 *) Add an optional second argument to the set_label() in the perl
9962 assembly language builder. If this argument exists and is set
9963 to 1 it signals that the assembler should use a symbol whose
9964 scope is the entire file, not just the current function. This
9965 is needed with MASM which uses the format label:: for this scope.
9966 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9968 *) Change the ASN1 types so they are typedefs by default. Before
9969 almost all types were #define'd to ASN1_STRING which was causing
9970 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9974 *) Change names of new functions to the new get1/get0 naming
9975 convention: After 'get1', the caller owns a reference count
9976 and has to call ..._free; 'get0' returns a pointer to some
9977 data structure without incrementing reference counters.
9978 (Some of the existing 'get' functions increment a reference
9979 counter, some don't.)
9980 Similarly, 'set1' and 'add1' functions increase reference
9981 counters or duplicate objects.
9984 *) Allow for the possibility of temp RSA key generation failure:
9985 the code used to assume it always worked and crashed on failure.
9988 *) Fix potential buffer overrun problem in BIO_printf().
9989 [Ulf Möller, using public domain code by Patrick Powell; problem
9990 pointed out by David Sacerdote <das33@cornell.edu>]
9992 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9993 RAND_egd() and RAND_status(). In the command line application,
9994 the EGD socket can be specified like a seed file using RANDFILE
9998 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9999 Some CAs (e.g. Verisign) distribute certificates in this form.
10002 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10003 list to exclude them. This means that no special compilation option
10004 is needed to use anonymous DH: it just needs to be included in the
10008 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10009 EVP_MD_type. The old functionality is available in a new macro called
10010 EVP_MD_md(). Change code that uses it and update docs.
10013 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10014 where the 'void *' argument is replaced by a function pointer argument.
10015 Previously 'void *' was abused to point to functions, which works on
10016 many platforms, but is not correct. As these functions are usually
10017 called by macros defined in OpenSSL header files, most source code
10018 should work without changes.
10021 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10022 sections with information on -D... compiler switches used for
10023 compiling the library so that applications can see them. To enable
10024 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10025 must be defined. E.g.,
10026 #define OPENSSL_ALGORITHM_DEFINES
10027 #include <openssl/opensslconf.h>
10028 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10029 [Richard Levitte, Ulf and Bodo Möller]
10031 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10035 *) Change the 'other' type in certificate aux info to a STACK_OF
10036 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10037 the required ASN1 format: arbitrary types determined by an OID.
10040 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10041 argument to 'req'. This is not because the function is newer or
10042 better than others it just uses the work 'NEW' in the certificate
10043 request header lines. Some software needs this.
10046 *) Reorganise password command line arguments: now passwords can be
10047 obtained from various sources. Delete the PEM_cb function and make
10048 it the default behaviour: i.e. if the callback is NULL and the
10049 usrdata argument is not NULL interpret it as a null terminated pass
10050 phrase. If usrdata and the callback are NULL then the pass phrase
10051 is prompted for as usual.
10054 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10055 the support is automatically enabled. The resulting binaries will
10056 autodetect the card and use it if present.
10057 [Ben Laurie and Compaq Inc.]
10059 *) Work around for Netscape hang bug. This sends certificate request
10060 and server done in one record. Since this is perfectly legal in the
10061 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10062 the bugs/SSLv3 entry for more info.
10065 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10068 *) Add -rand argument to smime and pkcs12 applications and read/write
10072 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10075 *) Add command line password options to the remaining applications.
10078 *) Bug fix for BN_div_recp() for numerators with an even number of
10082 *) More tests in bntest.c, and changed test_bn output.
10085 *) ./config recognizes MacOS X now.
10088 *) Bug fix for BN_div() when the first words of num and divsor are
10089 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10092 *) Add support for various broken PKCS#8 formats, and command line
10093 options to produce them.
10096 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10097 get temporary BIGNUMs from a BN_CTX.
10100 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10104 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10105 include a #define from the old name to the new. The original intent
10106 was that statically linked binaries could for example just call
10107 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10108 link with digests. This never worked because SSLeay_add_all_digests()
10109 and SSLeay_add_all_ciphers() were in the same source file so calling
10110 one would link with the other. They are now in separate source files.
10113 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10116 *) Use a less unusual form of the Miller-Rabin primality test (it used
10117 a binary algorithm for exponentiation integrated into the Miller-Rabin
10118 loop, our standard modexp algorithms are faster).
10121 *) Support for the EBCDIC character set completed.
10122 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10124 *) Source code cleanups: use const where appropriate, eliminate casts,
10125 use void * instead of char * in lhash.
10128 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10129 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10130 this the server could overwrite ephemeral keys that the client
10134 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10135 using 50 iterations of the Rabin-Miller test.
10137 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10138 iterations of the Rabin-Miller test as required by the appendix
10139 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10140 As BN_is_prime_fasttest includes trial division, DSA parameter
10141 generation becomes much faster.
10143 This implies a change for the callback functions in DSA_is_prime
10144 and DSA_generate_parameters: The callback function is called once
10145 for each positive witness in the Rabin-Miller test, not just
10146 occasionally in the inner loop; and the parameters to the
10147 callback function now provide an iteration count for the outer
10148 loop rather than for the current invocation of the inner loop.
10149 DSA_generate_parameters additionally can call the callback
10150 function with an 'iteration count' of -1, meaning that a
10151 candidate has passed the trial division test (when q is generated
10152 from an application-provided seed, trial division is skipped).
10155 *) New function BN_is_prime_fasttest that optionally does trial
10156 division before starting the Rabin-Miller test and has
10157 an additional BN_CTX * argument (whereas BN_is_prime always
10158 has to allocate at least one BN_CTX).
10159 'callback(1, -1, cb_arg)' is called when a number has passed the
10160 trial division stage.
10163 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10167 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10170 *) New function BN_pseudo_rand().
10173 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10174 bignum version of BN_from_montgomery() with the working code from
10175 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10179 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10180 made it impossible to use the same SSL_SESSION data structure in
10181 SSL2 clients in multiple threads.
10184 *) The return value of RAND_load_file() no longer counts bytes obtained
10185 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10186 to seed the PRNG (previously an explicit byte count was required).
10187 [Ulf Möller, Bodo Möller]
10189 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10190 used (char *) instead of (void *) and had casts all over the place.
10193 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10196 *) Retain source code compatibility for BN_prime_checks macro:
10197 BN_is_prime(..., BN_prime_checks, ...) now uses
10198 BN_prime_checks_for_size to determine the appropriate number of
10199 Rabin-Miller iterations.
10202 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10203 DH_CHECK_P_NOT_SAFE_PRIME.
10204 (Check if this is true? OpenPGP calls them "strong".)
10207 *) Merge the functionality of "dh" and "gendh" programs into a new program
10208 "dhparam". The old programs are retained for now but will handle DH keys
10209 (instead of parameters) in future.
10212 *) Make the ciphers, s_server and s_client programs check the return values
10213 when a new cipher list is set.
10216 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10217 ciphers. Before when the 56bit ciphers were enabled the sorting was
10220 The syntax for the cipher sorting has been extended to support sorting by
10221 cipher-strength (using the strength_bits hard coded in the tables).
10222 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10224 Fix a bug in the cipher-command parser: when supplying a cipher command
10225 string with an "undefined" symbol (neither command nor alphanumeric
10226 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10227 an error is flagged.
10229 Due to the strength-sorting extension, the code of the
10230 ssl_create_cipher_list() function was completely rearranged. I hope that
10231 the readability was also increased :-)
10232 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10234 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10235 for the first serial number and places 2 in the serial number file. This
10236 avoids problems when the root CA is created with serial number zero and
10237 the first user certificate has the same issuer name and serial number
10241 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10242 the new code. Add documentation for this stuff.
10245 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10246 X509_*() to X509at_*() on the grounds that they don't handle X509
10247 structures and behave in an analogous way to the X509v3 functions:
10248 they shouldn't be called directly but wrapper functions should be used
10251 So we also now have some wrapper functions that call the X509at functions
10252 when passed certificate requests. (TO DO: similar things can be done with
10253 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10254 things. Some of these need some d2i or i2d and print functionality
10255 because they handle more complex structures.)
10258 *) Add missing #ifndefs that caused missing symbols when building libssl
10259 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10260 NO_RSA in ssl/s2*.c.
10261 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10263 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10264 has a return value which indicates the quality of the random data
10265 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10266 error queue. New function RAND_pseudo_bytes() generates output that is
10267 guaranteed to be unique but not unpredictable. RAND_add is like
10268 RAND_seed, but takes an extra argument for an entropy estimate
10269 (RAND_seed always assumes full entropy).
10272 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10273 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10274 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10275 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10276 false-positive rate of at most 2^-80 for random input.
10279 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10282 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10283 in the 0.9.5 release), this returns the chain
10284 from an X509_CTX structure with a dup of the stack and all
10285 the X509 reference counts upped: so the stack will exist
10286 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10289 Also make SSL_SESSION_print() print out the verify return
10293 *) Add manpage for the pkcs12 command. Also change the default
10294 behaviour so MAC iteration counts are used unless the new
10295 -nomaciter option is used. This improves file security and
10296 only older versions of MSIE (4.0 for example) need it.
10299 *) Honor the no-xxx Configure options when creating .DEF files.
10302 *) Add PKCS#10 attributes to field table: challengePassword,
10303 unstructuredName and unstructuredAddress. These are taken from
10304 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10305 international characters are used.
10307 More changes to X509_ATTRIBUTE code: allow the setting of types
10308 based on strings. Remove the 'loc' parameter when adding
10309 attributes because these will be a SET OF encoding which is sorted
10313 *) Initial changes to the 'req' utility to allow request generation
10314 automation. This will allow an application to just generate a template
10315 file containing all the field values and have req construct the
10318 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10319 used all over the place including certificate requests and PKCS#7
10320 structures. They are currently handled manually where necessary with
10321 some primitive wrappers for PKCS#7. The new functions behave in a
10322 manner analogous to the X509 extension functions: they allow
10323 attributes to be looked up by NID and added.
10325 Later something similar to the X509V3 code would be desirable to
10326 automatically handle the encoding, decoding and printing of the
10327 more complex types. The string types like challengePassword can
10328 be handled by the string table functions.
10330 Also modified the multi byte string table handling. Now there is
10331 a 'global mask' which masks out certain types. The table itself
10332 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10333 is useful when for example there is only one permissible type
10334 (as in countryName) and using the mask might result in no valid
10338 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10339 SSL_get_peer_finished to allow applications to obtain the latest
10340 Finished messages sent to the peer or expected from the peer,
10341 respectively. (SSL_get_peer_finished is usually the Finished message
10342 actually received from the peer, otherwise the protocol will be aborted.)
10344 As the Finished message are message digests of the complete handshake
10345 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10346 be used for external authentication procedures when the authentication
10347 provided by SSL/TLS is not desired or is not enough.
10350 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10351 the host supports BWX extension and if Compaq C is present on the
10352 $PATH. Just exploiting of the BWX extension results in 20-30%
10353 performance kick for some algorithms, e.g. DES and RC4 to mention
10354 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10358 *) Add support for MS "fast SGC". This is arguably a violation of the
10359 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10360 weak crypto and after checking the certificate is SGC a second one
10361 with strong crypto. MS SGC stops the first handshake after receiving
10362 the server certificate message and sends a second client hello. Since
10363 a server will typically do all the time consuming operations before
10364 expecting any further messages from the client (server key exchange
10365 is the most expensive) there is little difference between the two.
10367 To get OpenSSL to support MS SGC we have to permit a second client
10368 hello message after we have sent server done. In addition we have to
10369 reset the MAC if we do get this second client hello.
10372 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10373 if a DER encoded private key is RSA or DSA traditional format. Changed
10374 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10375 format DER encoded private key. Newer code should use PKCS#8 format which
10376 has the key type encoded in the ASN1 structure. Added DER private key
10377 support to pkcs8 application.
10380 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10381 ciphersuites has been selected (as required by the SSL 3/TLS 1
10382 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10383 is set, we interpret this as a request to violate the specification
10384 (the worst that can happen is a handshake failure, and 'correct'
10385 behaviour would result in a handshake failure anyway).
10388 *) In SSL_CTX_add_session, take into account that there might be multiple
10389 SSL_SESSION structures with the same session ID (e.g. when two threads
10390 concurrently obtain them from an external cache).
10391 The internal cache can handle only one SSL_SESSION with a given ID,
10392 so if there's a conflict, we now throw out the old one to achieve
10396 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10397 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10398 some routines that use cipher OIDs: some ciphers do not have OIDs
10399 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10403 *) Simplify the trust setting structure and code. Now we just have
10404 two sequences of OIDs for trusted and rejected settings. These will
10405 typically have values the same as the extended key usage extension
10406 and any application specific purposes.
10408 The trust checking code now has a default behaviour: it will just
10409 check for an object with the same NID as the passed id. Functions can
10410 be provided to override either the default behaviour or the behaviour
10411 for a given id. SSL client, server and email already have functions
10412 in place for compatibility: they check the NID and also return "trusted"
10413 if the certificate is self signed.
10416 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10417 traditional format into an EVP_PKEY structure.
10420 *) Add a password callback function PEM_cb() which either prompts for
10421 a password if usr_data is NULL or otherwise assumes it is a null
10422 terminated password. Allow passwords to be passed on command line
10423 environment or config files in a few more utilities.
10426 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10427 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10428 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10429 Update documentation.
10432 *) Support for ASN1 "NULL" type. This could be handled before by using
10433 ASN1_TYPE but there wasn't any function that would try to read a NULL
10434 and produce an error if it couldn't. For compatibility we also have
10435 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10436 don't allocate anything because they don't need to.
10439 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10441 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10443 *) Rebuild of the memory allocation routines used by OpenSSL code and
10444 possibly others as well. The purpose is to make an interface that
10445 provide hooks so anyone can build a separate set of allocation and
10446 deallocation routines to be used by OpenSSL, for example memory
10447 pool implementations, or something else, which was previously hard
10448 since Malloc(), Realloc() and Free() were defined as macros having
10449 the values malloc, realloc and free, respectively (except for Win32
10450 compilations). The same is provided for memory debugging code.
10451 OpenSSL already comes with functionality to find memory leaks, but
10452 this gives people a chance to debug other memory problems.
10454 With these changes, a new set of functions and macros have appeared:
10456 CRYPTO_set_mem_debug_functions() [F]
10457 CRYPTO_get_mem_debug_functions() [F]
10458 CRYPTO_dbg_set_options() [F]
10459 CRYPTO_dbg_get_options() [F]
10460 CRYPTO_malloc_debug_init() [M]
10462 The memory debug functions are NULL by default, unless the library
10463 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10464 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10465 gives the standard debugging functions that come with OpenSSL) or
10466 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10467 provided by the library user) must be used. When the standard
10468 debugging functions are used, CRYPTO_dbg_set_options can be used to
10469 request additional information:
10470 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10471 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10473 Also, things like CRYPTO_set_mem_functions will always give the
10474 expected result (the new set of functions is used for allocation
10475 and deallocation) at all times, regardless of platform and compiler
10478 To finish it up, some functions that were never use in any other
10479 way than through macros have a new API and new semantic:
10481 CRYPTO_dbg_malloc()
10482 CRYPTO_dbg_realloc()
10485 All macros of value have retained their old syntax.
10486 [Richard Levitte and Bodo Moeller]
10488 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10489 ordering of SMIMECapabilities wasn't in "strength order" and there
10490 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10494 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10495 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10496 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10498 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10499 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10500 functionality to handle multipart/signed properly) and a utility
10501 called 'smime' to call all this stuff. This is based on code I
10502 originally wrote for Celo who have kindly allowed it to be
10503 included in OpenSSL.
10506 *) Add variants des_set_key_checked and des_set_key_unchecked of
10507 des_set_key (aka des_key_sched). Global variable des_check_key
10508 decides which of these is called by des_set_key; this way
10509 des_check_key behaves as it always did, but applications and
10510 the library itself, which was buggy for des_check_key == 1,
10511 have a cleaner way to pick the version they need.
10514 *) New function PKCS12_newpass() which changes the password of a
10518 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10519 dynamic mix. In both cases the ids can be used as an index into the
10520 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10521 functions so they accept a list of the field values and the
10522 application doesn't need to directly manipulate the X509_TRUST
10526 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10530 *) Modify the way the V3 extension code looks up extensions. This now
10531 works in a similar way to the object code: we have some "standard"
10532 extensions in a static table which is searched with OBJ_bsearch()
10533 and the application can add dynamic ones if needed. The file
10534 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10535 updated whenever a new extension is added to the core code and kept
10536 in ext_nid order. There is a simple program 'tabtest.c' which checks
10537 this. New extensions are not added too often so this file can readily
10538 be maintained manually.
10540 There are two big advantages in doing things this way. The extensions
10541 can be looked up immediately and no longer need to be "added" using
10542 X509V3_add_standard_extensions(): this function now does nothing.
10543 [Side note: I get *lots* of email saying the extension code doesn't
10544 work because people forget to call this function]
10545 Also no dynamic allocation is done unless new extensions are added:
10546 so if we don't add custom extensions there is no need to call
10547 X509V3_EXT_cleanup().
10550 *) Modify enc utility's salting as follows: make salting the default. Add a
10551 magic header, so unsalted files fail gracefully instead of just decrypting
10552 to garbage. This is because not salting is a big security hole, so people
10553 should be discouraged from doing it.
10556 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10557 digest to be passed on the command line but it only used this
10558 parameter when signing a certificate. Modified so all relevant
10559 operations are affected by the digest parameter including the
10560 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10561 DSA key was used because it didn't fix the digest.
10564 *) Initial certificate chain verify code. Currently tests the untrusted
10565 certificates for consistency with the verify purpose (which is set
10566 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10568 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10569 this is because it will reject chains with invalid extensions whereas
10570 every previous version of OpenSSL and SSLeay made no checks at all.
10572 Trust code: checks the root CA for the relevant trust settings. Trust
10573 settings have an initial value consistent with the verify purpose: e.g.
10574 if the verify purpose is for SSL client use it expects the CA to be
10575 trusted for SSL client use. However the default value can be changed to
10576 permit custom trust settings: one example of this would be to only trust
10577 certificates from a specific "secure" set of CAs.
10579 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10580 which should be used for version portability: especially since the
10581 verify structure is likely to change more often now.
10583 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10584 to set them. If not set then assume SSL clients will verify SSL servers
10587 Two new options to the verify program: -untrusted allows a set of
10588 untrusted certificates to be passed in and -purpose which sets the
10589 intended purpose of the certificate. If a purpose is set then the
10590 new chain verify code is used to check extension consistency.
10593 *) Support for the authority information access extension.
10596 *) Modify RSA and DSA PEM read routines to transparently handle
10597 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10598 public keys in a format compatible with certificate
10599 SubjectPublicKeyInfo structures. Unfortunately there were already
10600 functions called *_PublicKey_* which used various odd formats so
10601 these are retained for compatibility: however the DSA variants were
10602 never in a public release so they have been deleted. Changed dsa/rsa
10603 utilities to handle the new format: note no releases ever handled public
10604 keys so we should be OK.
10606 The primary motivation for this change is to avoid the same fiasco
10607 that dogs private keys: there are several incompatible private key
10608 formats some of which are standard and some OpenSSL specific and
10609 require various evil hacks to allow partial transparent handling and
10610 even then it doesn't work with DER formats. Given the option anything
10611 other than PKCS#8 should be dumped: but the other formats have to
10612 stay in the name of compatibility.
10614 With public keys and the benefit of hindsight one standard format
10615 is used which works with EVP_PKEY, RSA or DSA structures: though
10616 it clearly returns an error if you try to read the wrong kind of key.
10618 Added a -pubkey option to the 'x509' utility to output the public key.
10619 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10620 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10621 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10622 that do the same as the EVP_PKEY_assign_*() except they up the
10623 reference count of the added key (they don't "swallow" the
10627 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10628 CRLs would fail if the file contained no certificates or no CRLs:
10629 added a new function to read in both types and return the number
10630 read: this means that if none are read it will be an error. The
10631 DER versions of the certificate and CRL reader would always fail
10632 because it isn't possible to mix certificates and CRLs in DER format
10633 without choking one or the other routine. Changed this to just read
10634 a certificate: this is the best we can do. Also modified the code
10635 in apps/verify.c to take notice of return codes: it was previously
10636 attempting to read in certificates from NULL pointers and ignoring
10637 any errors: this is one reason why the cert and CRL reader seemed
10638 to work. It doesn't check return codes from the default certificate
10639 routines: these may well fail if the certificates aren't installed.
10642 *) Code to support otherName option in GeneralName.
10645 *) First update to verify code. Change the verify utility
10646 so it warns if it is passed a self signed certificate:
10647 for consistency with the normal behaviour. X509_verify
10648 has been modified to it will now verify a self signed
10649 certificate if *exactly* the same certificate appears
10650 in the store: it was previously impossible to trust a
10651 single self signed certificate. This means that:
10652 openssl verify ss.pem
10653 now gives a warning about a self signed certificate but
10654 openssl verify -CAfile ss.pem ss.pem
10658 *) For servers, store verify_result in SSL_SESSION data structure
10659 (and add it to external session representation).
10660 This is needed when client certificate verifications fails,
10661 but an application-provided verification callback (set by
10662 SSL_CTX_set_cert_verify_callback) allows accepting the session
10663 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10664 but returns 1): When the session is reused, we have to set
10665 ssl->verify_result to the appropriate error code to avoid
10667 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10669 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10670 case in PKCS7_dataInit() where the signed PKCS7 structure
10671 didn't contain any existing data because it was being created.
10672 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10674 *) Add a salt to the key derivation routines in enc.c. This
10675 forms the first 8 bytes of the encrypted file. Also add a
10676 -S option to allow a salt to be input on the command line.
10679 *) New function X509_cmp(). Oddly enough there wasn't a function
10680 to compare two certificates. We do this by working out the SHA1
10681 hash and comparing that. X509_cmp() will be needed by the trust
10685 *) SSL_get1_session() is like SSL_get_session(), but increments
10686 the reference count in the SSL_SESSION returned.
10687 [Geoff Thorpe <geoff@eu.c2.net>]
10689 *) Fix for 'req': it was adding a null to request attributes.
10690 Also change the X509_LOOKUP and X509_INFO code to handle
10691 certificate auxiliary information.
10694 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10698 *) Add the possibility to add extra information to the memory leak
10699 detecting output, to form tracebacks, showing from where each
10700 allocation was originated: CRYPTO_push_info("constant string") adds
10701 the string plus current file name and line number to a per-thread
10702 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10703 is like calling CYRPTO_pop_info() until the stack is empty.
10704 Also updated memory leak detection code to be multi-thread-safe.
10707 *) Add options -text and -noout to pkcs7 utility and delete the
10708 encryption options which never did anything. Update docs.
10711 *) Add options to some of the utilities to allow the pass phrase
10712 to be included on either the command line (not recommended on
10713 OSes like Unix) or read from the environment. Update the
10714 manpages and fix a few bugs.
10717 *) Add a few manpages for some of the openssl commands.
10720 *) Fix the -revoke option in ca. It was freeing up memory twice,
10721 leaking and not finding already revoked certificates.
10724 *) Extensive changes to support certificate auxiliary information.
10725 This involves the use of X509_CERT_AUX structure and X509_AUX
10726 functions. An X509_AUX function such as PEM_read_X509_AUX()
10727 can still read in a certificate file in the usual way but it
10728 will also read in any additional "auxiliary information". By
10729 doing things this way a fair degree of compatibility can be
10730 retained: existing certificates can have this information added
10731 using the new 'x509' options.
10733 Current auxiliary information includes an "alias" and some trust
10734 settings. The trust settings will ultimately be used in enhanced
10735 certificate chain verification routines: currently a certificate
10736 can only be trusted if it is self signed and then it is trusted
10740 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10741 The problem was that one of the replacement routines had not been working
10742 since SSLeay releases. For now the offending routine has been replaced
10743 with non-optimised assembler. Even so, this now gives around 95%
10744 performance improvement for 1024 bit RSA signs.
10747 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10748 handling. Most clients have the effective key size in bits equal to
10749 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10750 A few however don't do this and instead use the size of the decrypted key
10751 to determine the RC2 key length and the AlgorithmIdentifier to determine
10752 the effective key length. In this case the effective key length can still
10753 be 40 bits but the key length can be 168 bits for example. This is fixed
10754 by manually forcing an RC2 key into the EVP_PKEY structure because the
10755 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10756 the key length and effective key length are equal.
10759 *) Add a bunch of functions that should simplify the creation of
10760 X509_NAME structures. Now you should be able to do:
10761 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10762 and have it automatically work out the correct field type and fill in
10763 the structures. The more adventurous can try:
10764 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10765 and it will (hopefully) work out the correct multibyte encoding.
10768 *) Change the 'req' utility to use the new field handling and multibyte
10769 copy routines. Before the DN field creation was handled in an ad hoc
10770 way in req, ca, and x509 which was rather broken and didn't support
10771 BMPStrings or UTF8Strings. Since some software doesn't implement
10772 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10773 using the dirstring_type option. See the new comment in the default
10774 openssl.cnf for more info.
10777 *) Make crypto/rand/md_rand.c more robust:
10778 - Assure unique random numbers after fork().
10779 - Make sure that concurrent threads access the global counter and
10780 md serializably so that we never lose entropy in them
10781 or use exactly the same state in multiple threads.
10782 Access to the large state is not always serializable because
10783 the additional locking could be a performance killer, and
10784 md should be large enough anyway.
10787 *) New file apps/app_rand.c with commonly needed functionality
10788 for handling the random seed file.
10790 Use the random seed file in some applications that previously did not:
10792 dsaparam -genkey (which also ignored its '-rand' option),
10795 x509 (when signing).
10796 Except on systems with /dev/urandom, it is crucial to have a random
10797 seed file at least for key creation, DSA signing, and for DH exchanges;
10798 for RSA signatures we could do without one.
10800 gendh and gendsa (unlike genrsa) used to read only the first byte
10801 of each file listed in the '-rand' option. The function as previously
10802 found in genrsa is now in app_rand.c and is used by all programs
10803 that support '-rand'.
10806 *) In RAND_write_file, use mode 0600 for creating files;
10807 don't just chmod when it may be too late.
10810 *) Report an error from X509_STORE_load_locations
10811 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10814 *) New function ASN1_mbstring_copy() this copies a string in either
10815 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10816 into an ASN1_STRING type. A mask of permissible types is passed
10817 and it chooses the "minimal" type to use or an error if not type
10821 *) Add function equivalents to the various macros in asn1.h. The old
10822 macros are retained with an M_ prefix. Code inside the library can
10823 use the M_ macros. External code (including the openssl utility)
10824 should *NOT* in order to be "shared library friendly".
10827 *) Add various functions that can check a certificate's extensions
10828 to see if it usable for various purposes such as SSL client,
10829 server or S/MIME and CAs of these types. This is currently
10830 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10831 verification. Also added a -purpose flag to x509 utility to
10832 print out all the purposes.
10835 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10839 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10840 for, obtain and decode and extension and obtain its critical flag.
10841 This allows all the necessary extension code to be handled in a
10842 single function call.
10845 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10846 platforms. See crypto/rc4/rc4_enc.c for further details.
10849 *) New -noout option to asn1parse. This causes no output to be produced
10850 its main use is when combined with -strparse and -out to extract data
10851 from a file (which may not be in ASN.1 format).
10854 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10855 when producing the local key id.
10856 [Richard Levitte <levitte@stacken.kth.se>]
10858 *) New option -dhparam in s_server. This allows a DH parameter file to be
10859 stated explicitly. If it is not stated then it tries the first server
10860 certificate file. The previous behaviour hard coded the filename
10864 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10865 a public key to be input or output. For example:
10866 openssl rsa -in key.pem -pubout -out pubkey.pem
10867 Also added necessary DSA public key functions to handle this.
10870 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10871 in the message. This was handled by allowing
10872 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10873 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10875 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10876 to the end of the strings whereas this didn't. This would cause problems
10877 if strings read with d2i_ASN1_bytes() were later modified.
10878 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10880 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10881 data and it contains EOF it will end up returning an error. This is
10882 caused by input 46 bytes long. The cause is due to the way base64
10883 BIOs find the start of base64 encoded data. They do this by trying a
10884 trial decode on each line until they find one that works. When they
10885 do a flag is set and it starts again knowing it can pass all the
10886 data directly through the decoder. Unfortunately it doesn't reset
10887 the context it uses. This means that if EOF is reached an attempt
10888 is made to pass two EOFs through the context and this causes the
10889 resulting error. This can also cause other problems as well. As is
10890 usual with these problems it takes *ages* to find and the fix is
10891 trivial: move one line.
10892 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10894 *) Ugly workaround to get s_client and s_server working under Windows. The
10895 old code wouldn't work because it needed to select() on sockets and the
10896 tty (for keypresses and to see if data could be written). Win32 only
10897 supports select() on sockets so we select() with a 1s timeout on the
10898 sockets and then see if any characters are waiting to be read, if none
10899 are present then we retry, we also assume we can always write data to
10900 the tty. This isn't nice because the code then blocks until we've
10901 received a complete line of data and it is effectively polling the
10902 keyboard at 1s intervals: however it's quite a bit better than not
10903 working at all :-) A dedicated Windows application might handle this
10904 with an event loop for example.
10907 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10908 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10909 will be called when RSA_sign() and RSA_verify() are used. This is useful
10910 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10911 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10912 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10913 This necessitated the support of an extra signature type NID_md5_sha1
10914 for SSL signatures and modifications to the SSL library to use it instead
10915 of calling RSA_public_decrypt() and RSA_private_encrypt().
10918 *) Add new -verify -CAfile and -CApath options to the crl program, these
10919 will lookup a CRL issuers certificate and verify the signature in a
10920 similar way to the verify program. Tidy up the crl program so it
10921 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10922 less strict. It will now permit CRL extensions even if it is not
10923 a V2 CRL: this will allow it to tolerate some broken CRLs.
10926 *) Initialize all non-automatic variables each time one of the openssl
10927 sub-programs is started (this is necessary as they may be started
10928 multiple times from the "OpenSSL>" prompt).
10929 [Lennart Bang, Bodo Moeller]
10931 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10932 removing all other RSA functionality (this is what NO_RSA does). This
10933 is so (for example) those in the US can disable those operations covered
10934 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10938 *) Non-copying interface to BIO pairs.
10939 (still largely untested)
10942 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10943 ASCII string. This was handled independently in various places before.
10946 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10947 UTF8 strings a character at a time.
10950 *) Use client_version from client hello to select the protocol
10951 (s23_srvr.c) and for RSA client key exchange verification
10952 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10955 *) Add various utility functions to handle SPKACs, these were previously
10956 handled by poking round in the structure internals. Added new function
10957 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10958 print, verify and generate SPKACs. Based on an original idea from
10959 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10962 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10965 *) Allow the config file extension section to be overwritten on the
10966 command line. Based on an original idea from Massimiliano Pala
10967 <madwolf@comune.modena.it>. The new option is called -extensions
10968 and can be applied to ca, req and x509. Also -reqexts to override
10969 the request extensions in req and -crlexts to override the crl extensions
10973 *) Add new feature to the SPKAC handling in ca. Now you can include
10974 the same field multiple times by preceding it by "XXXX." for example:
10977 this is the same syntax as used in the req config file.
10980 *) Allow certificate extensions to be added to certificate requests. These
10981 are specified in a 'req_extensions' option of the req section of the
10982 config file. They can be printed out with the -text option to req but
10983 are otherwise ignored at present.
10986 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10987 data read consists of only the final block it would not decrypted because
10988 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10989 A misplaced 'break' also meant the decrypted final block might not be
10990 copied until the next read.
10993 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10994 a few extra parameters to the DH structure: these will be useful if
10995 for example we want the value of 'q' or implement X9.42 DH.
10998 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10999 provides hooks that allow the default DSA functions or functions on a
11000 "per key" basis to be replaced. This allows hardware acceleration and
11001 hardware key storage to be handled without major modification to the
11002 library. Also added low level modexp hooks and CRYPTO_EX structure and
11003 associated functions.
11006 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11007 as "read only": it can't be written to and the buffer it points to will
11008 not be freed. Reading from a read only BIO is much more efficient than
11009 a normal memory BIO. This was added because there are several times when
11010 an area of memory needs to be read from a BIO. The previous method was
11011 to create a memory BIO and write the data to it, this results in two
11012 copies of the data and an O(n^2) reading algorithm. There is a new
11013 function BIO_new_mem_buf() which creates a read only memory BIO from
11014 an area of memory. Also modified the PKCS#7 routines to use read only
11018 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11019 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11020 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11021 but a retry condition occurred while trying to read the rest.
11024 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11025 NID_pkcs7_encrypted by default: this was wrong since this should almost
11026 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11027 the encrypted data type: this is a more sensible place to put it and it
11028 allows the PKCS#12 code to be tidied up that duplicated this
11032 *) Changed obj_dat.pl script so it takes its input and output files on
11033 the command line. This should avoid shell escape redirection problems
11037 *) Initial support for certificate extension requests, these are included
11038 in things like Xenroll certificate requests. Included functions to allow
11039 extensions to be obtained and added.
11042 *) -crlf option to s_client and s_server for sending newlines as
11043 CRLF (as required by many protocols).
11046 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11048 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11049 [Ralf S. Engelschall]
11051 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11052 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11054 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11058 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11059 DH parameters/keys (q is lost during that conversion, but the resulting
11060 DH parameters contain its length).
11062 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11063 much faster than DH_generate_parameters (which creates parameters
11064 where p = 2*q + 1), and also the smaller q makes DH computations
11065 much more efficient (160-bit exponentiation instead of 1024-bit
11066 exponentiation); so this provides a convenient way to support DHE
11067 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11068 utter importance to use
11069 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11071 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11072 when such DH parameters are used, because otherwise small subgroup
11073 attacks may become possible!
11076 *) Avoid memory leak in i2d_DHparams.
11079 *) Allow the -k option to be used more than once in the enc program:
11080 this allows the same encrypted message to be read by multiple recipients.
11083 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11084 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11085 it will always use the numerical form of the OID, even if it has a short
11089 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11090 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11091 otherwise bn_mod_exp was called. In the case of hardware keys for example
11092 no private key components need be present and it might store extra data
11093 in the RSA structure, which cannot be accessed from bn_mod_exp.
11094 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11095 private key operations.
11098 *) Added support for SPARC Linux.
11101 *) pem_password_cb function type incompatibly changed from
11102 typedef int pem_password_cb(char *buf, int size, int rwflag);
11104 ....(char *buf, int size, int rwflag, void *userdata);
11105 so that applications can pass data to their callbacks:
11106 The PEM[_ASN1]_{read,write}... functions and macros now take an
11107 additional void * argument, which is just handed through whenever
11108 the password callback is called.
11109 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11111 New function SSL_CTX_set_default_passwd_cb_userdata.
11113 Compatibility note: As many C implementations push function arguments
11114 onto the stack in reverse order, the new library version is likely to
11115 interoperate with programs that have been compiled with the old
11116 pem_password_cb definition (PEM_whatever takes some data that
11117 happens to be on the stack as its last argument, and the callback
11118 just ignores this garbage); but there is no guarantee whatsoever that
11121 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11122 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11123 problems not only on Windows, but also on some Unix platforms.
11124 To avoid problematic command lines, these definitions are now in an
11125 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11126 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11129 *) MIPS III/IV assembler module is reimplemented.
11132 *) More DES library cleanups: remove references to srand/rand and
11133 delete an unused file.
11136 *) Add support for the the free Netwide assembler (NASM) under Win32,
11137 since not many people have MASM (ml) and it can be hard to obtain.
11138 This is currently experimental but it seems to work OK and pass all
11139 the tests. Check out INSTALL.W32 for info.
11142 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11143 without temporary keys kept an extra copy of the server key,
11144 and connections with temporary keys did not free everything in case
11148 *) New function RSA_check_key and new openssl rsa option -check
11149 for verifying the consistency of RSA keys.
11150 [Ulf Moeller, Bodo Moeller]
11152 *) Various changes to make Win32 compile work:
11153 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11154 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11155 comparison" warnings.
11156 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11159 *) Add a debugging option to PKCS#5 v2 key generation function: when
11160 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11161 derived keys are printed to stderr.
11164 *) Copy the flags in ASN1_STRING_dup().
11165 [Roman E. Pavlov <pre@mo.msk.ru>]
11167 *) The x509 application mishandled signing requests containing DSA
11168 keys when the signing key was also DSA and the parameters didn't match.
11170 It was supposed to omit the parameters when they matched the signing key:
11171 the verifying software was then supposed to automatically use the CA's
11172 parameters if they were absent from the end user certificate.
11174 Omitting parameters is no longer recommended. The test was also
11175 the wrong way round! This was probably due to unusual behaviour in
11176 EVP_cmp_parameters() which returns 1 if the parameters match.
11177 This meant that parameters were omitted when they *didn't* match and
11178 the certificate was useless. Certificates signed with 'ca' didn't have
11180 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11182 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11183 The interface is as follows:
11184 Applications can use
11185 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11186 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11187 "off" is now the default.
11188 The library internally uses
11189 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11190 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11191 to disable memory-checking temporarily.
11193 Some inconsistent states that previously were possible (and were
11194 even the default) are now avoided.
11196 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11197 with each memory chunk allocated; this is occasionally more helpful
11198 than just having a counter.
11200 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11202 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11206 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11207 which largely parallels "options", but is for changing API behaviour,
11208 whereas "options" are about protocol behaviour.
11209 Initial "mode" flags are:
11211 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11212 a single record has been written.
11213 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11214 retries use the same buffer location.
11215 (But all of the contents must be
11219 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11222 *) Fix problems with no-hmac etc.
11223 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11225 *) New functions RSA_get_default_method(), RSA_set_method() and
11226 RSA_get_method(). These allows replacement of RSA_METHODs without having
11227 to mess around with the internals of an RSA structure.
11230 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11231 Also really enable memory leak checks in openssl.c and in some
11233 [Chad C. Mulligan, Bodo Moeller]
11235 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11236 up the length of negative integers. This has now been simplified to just
11237 store the length when it is first determined and use it later, rather
11238 than trying to keep track of where data is copied and updating it to
11240 [Steve Henson, reported by Brien Wheeler
11241 <bwheeler@authentica-security.com>]
11243 *) Add a new function PKCS7_signatureVerify. This allows the verification
11244 of a PKCS#7 signature but with the signing certificate passed to the
11245 function itself. This contrasts with PKCS7_dataVerify which assumes the
11246 certificate is present in the PKCS#7 structure. This isn't always the
11247 case: certificates can be omitted from a PKCS#7 structure and be
11248 distributed by "out of band" means (such as a certificate database).
11251 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11252 function prototypes in pem.h, also change util/mkdef.pl to add the
11253 necessary function names.
11256 *) mk1mf.pl (used by Windows builds) did not properly read the
11257 options set by Configure in the top level Makefile, and Configure
11258 was not even able to write more than one option correctly.
11259 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11262 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11263 file to be loaded from a BIO or FILE pointer. The BIO version will
11264 for example allow memory BIOs to contain config info.
11267 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11268 Whoever hopes to achieve shared-library compatibility across versions
11269 must use this, not the compile-time macro.
11270 (Exercise 0.9.4: Which is the minimum library version required by
11272 Note: All this applies only to multi-threaded programs, others don't
11276 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11277 through a BIO pair triggered the default case, i.e.
11278 SSLerr(...,SSL_R_UNKNOWN_STATE).
11281 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11282 can use the SSL library even if none of the specific BIOs is
11286 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11287 for the encoded length.
11288 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11290 *) Add initial documentation of the X509V3 functions.
11293 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11294 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11295 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11296 secure PKCS#8 private key format with a high iteration count.
11299 *) Fix determination of Perl interpreter: A perl or perl5
11300 _directory_ in $PATH was also accepted as the interpreter.
11301 [Ralf S. Engelschall]
11303 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11304 wrong with it but it was very old and did things like calling
11305 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11306 unusual formatting.
11309 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11310 to use the new extension code.
11313 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11314 with macros. This should make it easier to change their form, add extra
11315 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11319 *) Add to configuration table a new entry that can specify an alternative
11320 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11321 according to Mark Crispin <MRC@Panda.COM>.
11325 *) DES CBC did not update the IV. Weird.
11328 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11329 Changing the behaviour of the former might break existing programs --
11330 where IV updating is needed, des_ncbc_encrypt can be used.
11333 *) When bntest is run from "make test" it drives bc to check its
11334 calculations, as well as internally checking them. If an internal check
11335 fails, it needs to cause bc to give a non-zero result or make test carries
11336 on without noticing the failure. Fixed.
11339 *) DES library cleanups.
11342 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11343 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11344 ciphers. NOTE: although the key derivation function has been verified
11345 against some published test vectors it has not been extensively tested
11346 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11350 *) Instead of "mkdir -p", which is not fully portable, use new
11351 Perl script "util/mkdir-p.pl".
11354 *) Rewrite the way password based encryption (PBE) is handled. It used to
11355 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11356 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11357 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11358 the 'parameter' field of the AlgorithmIdentifier is passed to the
11359 underlying key generation function so it must do its own ASN1 parsing.
11360 This has also changed the EVP_PBE_CipherInit() function which now has a
11361 'parameter' argument instead of literal salt and iteration count values
11362 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11365 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11366 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11367 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11368 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11369 value was just used as a "magic string" and not used directly its
11370 value doesn't matter.
11373 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11377 *) "linux-sparc64" configuration (ultrapenguin).
11378 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11379 "linux-sparc" configuration.
11380 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11382 *) config now generates no-xxx options for missing ciphers.
11385 *) Support the EBCDIC character set (work in progress).
11386 File ebcdic.c not yet included because it has a different license.
11387 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11389 *) Support BS2000/OSD-POSIX.
11390 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11392 *) Make callbacks for key generation use void * instead of char *.
11395 *) Make S/MIME samples compile (not yet tested).
11398 *) Additional typesafe stacks.
11401 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11405 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11407 *) New configuration variant "sco5-gcc".
11409 *) Updated some demos.
11410 [Sean O Riordain, Wade Scholine]
11412 *) Add missing BIO_free at exit of pkcs12 application.
11415 *) Fix memory leak in conf.c.
11418 *) Updates for Win32 to assembler version of MD5.
11421 *) Set #! path to perl in apps/der_chop to where we found it
11422 instead of using a fixed path.
11425 *) SHA library changes for irix64-mips4-cc.
11428 *) Improvements for VMS support.
11432 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11434 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11435 This also avoids the problems with SC4.2 and unpatched SC5.
11436 [Andy Polyakov <appro@fy.chalmers.se>]
11438 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11439 These are required because of the typesafe stack would otherwise break
11440 existing code. If old code used a structure member which used to be STACK
11441 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11442 sk_num or sk_value it would produce an error because the num, data members
11443 are not present in STACK_OF. Now it just produces a warning. sk_set
11444 replaces the old method of assigning a value to sk_value
11445 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11446 that does this will no longer work (and should use sk_set instead) but
11447 this could be regarded as a "questionable" behaviour anyway.
11450 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11451 correctly handle encrypted S/MIME data.
11454 *) Change type of various DES function arguments from des_cblock
11455 (which means, in function argument declarations, pointer to char)
11456 to des_cblock * (meaning pointer to array with 8 char elements),
11457 which allows the compiler to do more typechecking; it was like
11458 that back in SSLeay, but with lots of ugly casts.
11460 Introduce new type const_des_cblock.
11463 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11464 problems: find RecipientInfo structure that matches recipient certificate
11465 and initialise the ASN1 structures properly based on passed cipher.
11468 *) Belatedly make the BN tests actually check the results.
11471 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11472 to and from BNs: it was completely broken. New compilation option
11473 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11474 key elements as negative integers.
11477 *) Reorganize and speed up MD5.
11478 [Andy Polyakov <appro@fy.chalmers.se>]
11481 [Richard Levitte <richard@levitte.org>]
11483 *) New option -out to asn1parse to allow the parsed structure to be
11484 output to a file. This is most useful when combined with the -strparse
11485 option to examine the output of things like OCTET STRINGS.
11488 *) Make SSL library a little more fool-proof by not requiring any longer
11489 that SSL_set_{accept,connect}_state be called before
11490 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11491 in many applications because usually everything *appeared* to work as
11492 intended anyway -- now it really works as intended).
11495 *) Move openssl.cnf out of lib/.
11498 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11499 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11500 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11501 [Ralf S. Engelschall]
11503 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11504 handle PKCS#7 enveloped data properly.
11505 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11507 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11508 copying pointers. The cert_st handling is changed by this in
11509 various ways (and thus what used to be known as ctx->default_cert
11510 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11511 any longer when s->cert does not give us what we need).
11512 ssl_cert_instantiate becomes obsolete by this change.
11513 As soon as we've got the new code right (possibly it already is?),
11514 we have solved a couple of bugs of the earlier code where s->cert
11515 was used as if it could not have been shared with other SSL structures.
11517 Note that using the SSL API in certain dirty ways now will result
11518 in different behaviour than observed with earlier library versions:
11519 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11520 does not influence s as it used to.
11522 In order to clean up things more thoroughly, inside SSL_SESSION
11523 we don't use CERT any longer, but a new structure SESS_CERT
11524 that holds per-session data (if available); currently, this is
11525 the peer's certificate chain and, for clients, the server's certificate
11526 and temporary key. CERT holds only those values that can have
11527 meaningful defaults in an SSL_CTX.
11530 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11531 from the internal representation. Various PKCS#7 fixes: remove some
11532 evil casts and set the enc_dig_alg field properly based on the signing
11536 *) Allow PKCS#12 password to be set from the command line or the
11537 environment. Let 'ca' get its config file name from the environment
11538 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11542 *) Allow certificate policies extension to use an IA5STRING for the
11543 organization field. This is contrary to the PKIX definition but
11544 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11548 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11549 without disallowing inline assembler and the like for non-pedantic builds.
11552 *) Support Borland C++ builder.
11553 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11555 *) Support Mingw32.
11558 *) SHA-1 cleanups and performance enhancements.
11559 [Andy Polyakov <appro@fy.chalmers.se>]
11561 *) Sparc v8plus assembler for the bignum library.
11562 [Andy Polyakov <appro@fy.chalmers.se>]
11564 *) Accept any -xxx and +xxx compiler options in Configure.
11567 *) Update HPUX configuration.
11570 *) Add missing sk_<type>_unshift() function to safestack.h
11571 [Ralf S. Engelschall]
11573 *) New function SSL_CTX_use_certificate_chain_file that sets the
11574 "extra_cert"s in addition to the certificate. (This makes sense
11575 only for "PEM" format files, as chains as a whole are not
11579 *) Support verify_depth from the SSL API.
11580 x509_vfy.c had what can be considered an off-by-one-error:
11581 Its depth (which was not part of the external interface)
11582 was actually counting the number of certificates in a chain;
11583 now it really counts the depth.
11586 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11587 instead of X509err, which often resulted in confusing error
11588 messages since the error codes are not globally unique
11589 (e.g. an alleged error in ssl3_accept when a certificate
11590 didn't match the private key).
11592 *) New function SSL_CTX_set_session_id_context that allows to set a default
11593 value (so that you don't need SSL_set_session_id_context for each
11594 connection using the SSL_CTX).
11597 *) OAEP decoding bug fix.
11600 *) Support INSTALL_PREFIX for package builders, as proposed by
11604 *) New Configure options "threads" and "no-threads". For systems
11605 where the proper compiler options are known (currently Solaris
11606 and Linux), "threads" is the default.
11609 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11612 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11613 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11614 such as /usr/local/bin.
11617 *) "make linux-shared" to build shared libraries.
11618 [Niels Poppe <niels@netbox.org>]
11620 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11623 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11624 extension adding in x509 utility.
11627 *) Remove NOPROTO sections and error code comments.
11630 *) Partial rewrite of the DEF file generator to now parse the ANSI
11634 *) New Configure options --prefix=DIR and --openssldir=DIR.
11637 *) Complete rewrite of the error code script(s). It is all now handled
11638 by one script at the top level which handles error code gathering,
11639 header rewriting and C source file generation. It should be much better
11640 than the old method: it now uses a modified version of Ulf's parser to
11641 read the ANSI prototypes in all header files (thus the old K&R definitions
11642 aren't needed for error creation any more) and do a better job of
11643 translating function codes into names. The old 'ASN1 error code imbedded
11644 in a comment' is no longer necessary and it doesn't use .err files which
11645 have now been deleted. Also the error code call doesn't have to appear all
11646 on one line (which resulted in some large lines...).
11649 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11652 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11653 0 (which usually indicates a closed connection), but continue reading.
11656 *) Fix some race conditions.
11659 *) Add support for CRL distribution points extension. Add Certificate
11660 Policies and CRL distribution points documentation.
11663 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11666 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11667 8 of keying material. Merlin has also confirmed interop with this fix
11668 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11669 [Merlin Hughes <merlin@baltimore.ie>]
11671 *) Fix lots of warnings.
11672 [Richard Levitte <levitte@stacken.kth.se>]
11674 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11675 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11676 [Richard Levitte <levitte@stacken.kth.se>]
11678 *) Fix problems with sizeof(long) == 8.
11679 [Andy Polyakov <appro@fy.chalmers.se>]
11681 *) Change functions to ANSI C.
11684 *) Fix typos in error codes.
11685 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11687 *) Remove defunct assembler files from Configure.
11690 *) SPARC v8 assembler BIGNUM implementation.
11691 [Andy Polyakov <appro@fy.chalmers.se>]
11693 *) Support for Certificate Policies extension: both print and set.
11694 Various additions to support the r2i method this uses.
11697 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11698 return a const string when you are expecting an allocated buffer.
11701 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11702 types DirectoryString and DisplayText.
11705 *) Add code to allow r2i extensions to access the configuration database,
11706 add an LHASH database driver and add several ctx helper functions.
11709 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11710 fail when they extended the size of a BIGNUM.
11713 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11714 support typesafe stack.
11717 *) Fix typo in SSL_[gs]et_options().
11718 [Nils Frostberg <nils@medcom.se>]
11720 *) Delete various functions and files that belonged to the (now obsolete)
11721 old X509V3 handling code.
11724 *) New Configure option "rsaref".
11727 *) Don't auto-generate pem.h.
11730 *) Introduce type-safe ASN.1 SETs.
11733 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11734 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11736 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11737 that links with OpenSSL (well at least cause lots of warnings), but fear
11738 not: the conversion is trivial, and it eliminates loads of evil casts. A
11739 few STACKed things have been converted already. Feel free to convert more.
11740 In the fullness of time, I'll do away with the STACK type altogether.
11743 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11744 specified in <certfile> by updating the entry in the index.txt file.
11745 This way one no longer has to edit the index.txt file manually for
11746 revoking a certificate. The -revoke option does the gory details now.
11747 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11749 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11750 `-text' option at all and this way the `-noout -text' combination was
11751 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11752 [Ralf S. Engelschall]
11754 *) Make sure a corresponding plain text error message exists for the
11755 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11756 verify callback function determined that a certificate was revoked.
11757 [Ralf S. Engelschall]
11759 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11760 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11761 all available cipers including rc5, which was forgotten until now.
11762 In order to let the testing shell script know which algorithms
11763 are available, a new (up to now undocumented) command
11764 "openssl list-cipher-commands" is used.
11767 *) Bugfix: s_client occasionally would sleep in select() when
11768 it should have checked SSL_pending() first.
11771 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11772 the raw DSA values prior to ASN.1 encoding.
11775 *) Tweaks to Configure
11776 [Niels Poppe <niels@netbox.org>]
11778 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11782 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11785 *) New config option to avoid instructions that are illegal on the 80386.
11786 The default code is faster, but requires at least a 486.
11789 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11790 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11791 same as SSL2_VERSION anyway.
11794 *) New "-showcerts" option for s_client.
11797 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11798 application. Various cleanups and fixes.
11801 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11802 modify error routines to work internally. Add error codes and PBE init
11803 to library startup routines.
11806 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11807 packing functions to asn1 and evp. Changed function names and error
11808 codes along the way.
11811 *) PKCS12 integration: and so it begins... First of several patches to
11812 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11813 objects to objects.h
11816 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11817 and display support for Thawte strong extranet extension.
11820 *) Add LinuxPPC support.
11821 [Jeff Dubrule <igor@pobox.org>]
11823 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11824 bn_div_words in alpha.s.
11825 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11827 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11828 OAEP isn't supported when OpenSSL is built with RSAref.
11829 [Ulf Moeller <ulf@fitug.de>]
11831 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11832 so they no longer are missing under -DNOPROTO.
11833 [Soren S. Jorvang <soren@t.dk>]
11836 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11838 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11839 doesn't work when the session is reused. Coming soon!
11842 *) Fix a security hole, that allows sessions to be reused in the wrong
11843 context thus bypassing client cert protection! All software that uses
11844 client certs and session caches in multiple contexts NEEDS PATCHING to
11845 allow session reuse! A fuller solution is in the works.
11846 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11848 *) Some more source tree cleanups (removed obsolete files
11849 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11850 permission on "config" script to be executable) and a fix for the INSTALL
11852 [Ulf Moeller <ulf@fitug.de>]
11854 *) Remove some legacy and erroneous uses of malloc, free instead of
11856 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11858 *) Make rsa_oaep_test return non-zero on error.
11859 [Ulf Moeller <ulf@fitug.de>]
11861 *) Add support for native Solaris shared libraries. Configure
11862 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11863 if someone would make that last step automatic.
11864 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11866 *) ctx_size was not built with the right compiler during "make links". Fixed.
11869 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11870 except NULL ciphers". This means the default cipher list will no longer
11871 enable NULL ciphers. They need to be specifically enabled e.g. with
11872 the string "DEFAULT:eNULL".
11875 *) Fix to RSA private encryption routines: if p < q then it would
11876 occasionally produce an invalid result. This will only happen with
11877 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11880 *) Be less restrictive and allow also `perl util/perlpath.pl
11881 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11882 because this way one can also use an interpreter named `perl5' (which is
11883 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11884 installed as `perl').
11885 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11887 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11888 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11890 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11891 advapi32.lib to Win32 build and change the pem test comparison
11892 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11893 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11894 and crypto/des/ede_cbcm_enc.c.
11897 *) DES quad checksum was broken on big-endian architectures. Fixed.
11900 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11901 Win32 test batch file so it (might) work again. The Win32 test batch file
11902 is horrible: I feel ill....
11905 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11906 in e_os.h. Audit of header files to check ANSI and non ANSI
11907 sections: 10 functions were absent from non ANSI section and not exported
11908 from Windows DLLs. Fixed up libeay.num for new functions.
11911 *) Make `openssl version' output lines consistent.
11912 [Ralf S. Engelschall]
11914 *) Fix Win32 symbol export lists for BIO functions: Added
11915 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11916 to ms/libeay{16,32}.def.
11917 [Ralf S. Engelschall]
11919 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11920 fine under Unix and passes some trivial tests I've now added. But the
11921 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11922 added to make sure no one expects that this stuff really works in the
11923 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11924 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11926 [Ralf S. Engelschall]
11928 *) Fix the generation of two part addresses in perl.
11929 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11931 *) Add config entry for Linux on MIPS.
11932 [John Tobey <jtobey@channel1.com>]
11934 *) Make links whenever Configure is run, unless we are on Windoze.
11937 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11938 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11942 *) Add a useful kludge to allow package maintainers to specify compiler and
11943 other platforms details on the command line without having to patch the
11944 Configure script everytime: One now can use ``perl Configure
11945 <id>:<details>'', i.e. platform ids are allowed to have details appended
11946 to them (separated by colons). This is treated as there would be a static
11947 pre-configured entry in Configure's %table under key <id> with value
11948 <details> and ``perl Configure <id>'' is called. So, when you want to
11949 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11950 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11951 now, which overrides the FreeBSD-elf entry on-the-fly.
11952 [Ralf S. Engelschall]
11954 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11957 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11958 on the `perl Configure ...' command line. This way one can compile
11959 OpenSSL libraries with Position Independent Code (PIC) which is needed
11960 for linking it into DSOs.
11961 [Ralf S. Engelschall]
11963 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11967 *) Cleaned up the LICENSE document: The official contact for any license
11968 questions now is the OpenSSL core team under openssl-core@openssl.org.
11969 And add a paragraph about the dual-license situation to make sure people
11970 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11971 to the OpenSSL toolkit.
11972 [Ralf S. Engelschall]
11974 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11975 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11976 Additionally cleaned up the `make links' target: Remove unnecessary
11977 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11978 to speed processing and no longer clutter the display with confusing
11979 stuff. Instead only the actually done links are displayed.
11980 [Ralf S. Engelschall]
11982 *) Permit null encryption ciphersuites, used for authentication only. It used
11983 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11984 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11988 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11989 signed attributes when verifying signatures (this would break them),
11990 the detached data encoding was wrong and public keys obtained using
11991 X509_get_pubkey() weren't freed.
11994 *) Add text documentation for the BUFFER functions. Also added a work around
11995 to a Win95 console bug. This was triggered by the password read stuff: the
11996 last character typed gets carried over to the next fread(). If you were
11997 generating a new cert request using 'req' for example then the last
11998 character of the passphrase would be CR which would then enter the first
12002 *) Added the new `Includes OpenSSL Cryptography Software' button as
12003 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12004 button and can be used by applications based on OpenSSL to show the
12005 relationship to the OpenSSL project.
12006 [Ralf S. Engelschall]
12008 *) Remove confusing variables in function signatures in files
12009 ssl/ssl_lib.c and ssl/ssl.h.
12010 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12012 *) Don't install bss_file.c under PREFIX/include/
12013 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12015 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12016 functions that return function pointers and has support for NT specific
12017 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12018 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12019 unsigned to signed types: this was killing the Win32 compile.
12022 *) Add new certificate file to stack functions,
12023 SSL_add_dir_cert_subjects_to_stack() and
12024 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12025 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12026 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12027 This means that Apache-SSL and similar packages don't have to mess around
12028 to add as many CAs as they want to the preferred list.
12031 *) Experiment with doxygen documentation. Currently only partially applied to
12033 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12034 openssl.doxy as the configuration file.
12037 *) Get rid of remaining C++-style comments which strict C compilers hate.
12038 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12040 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12041 compiled in by default: it has problems with large keys.
12044 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12045 DH private keys and/or callback functions which directly correspond to
12046 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12047 is needed for applications which have to configure certificates on a
12048 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12050 For the RSA certificate situation is makes no difference, but
12051 for the DSA certificate situation this fixes the "no shared cipher"
12052 problem where the OpenSSL cipher selection procedure failed because the
12053 temporary keys were not overtaken from the context and the API provided
12054 no way to reconfigure them.
12055 The new functions now let applications reconfigure the stuff and they
12056 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12057 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12058 non-public-API function ssl_cert_instantiate() is used as a helper
12059 function and also to reduce code redundancy inside ssl_rsa.c.
12060 [Ralf S. Engelschall]
12062 *) Move s_server -dcert and -dkey options out of the undocumented feature
12063 area because they are useful for the DSA situation and should be
12064 recognized by the users.
12065 [Ralf S. Engelschall]
12067 *) Fix the cipher decision scheme for export ciphers: the export bits are
12068 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12069 SSL_EXP_MASK. So, the original variable has to be used instead of the
12070 already masked variable.
12071 [Richard Levitte <levitte@stacken.kth.se>]
12073 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12074 [Richard Levitte <levitte@stacken.kth.se>]
12076 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12077 from `int' to `unsigned int' because it's a length and initialized by
12078 EVP_DigestFinal() which expects an `unsigned int *'.
12079 [Richard Levitte <levitte@stacken.kth.se>]
12081 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12082 script. Instead use the usual Shell->Perl transition trick.
12083 [Ralf S. Engelschall]
12085 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12086 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12087 -noout -modulus' as it's already the case for `openssl rsa -noout
12088 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12089 currently the public key is printed (a decision which was already done by
12090 `openssl dsa -modulus' in the past) which serves a similar purpose.
12091 Additionally the NO_RSA no longer completely removes the whole -modulus
12092 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12094 [Ralf S. Engelschall]
12096 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12097 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12098 [Arne Ansper <arne@ats.cyber.ee>]
12100 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12101 to be added. Now both 'req' and 'ca' can use new objects defined in the
12105 *) Add cool BIO that does syslog (or event log on NT).
12106 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12108 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12109 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12110 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12111 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12114 *) Add preliminary config info for new extension code.
12117 *) Make RSA_NO_PADDING really use no padding.
12118 [Ulf Moeller <ulf@fitug.de>]
12120 *) Generate errors when private/public key check is done.
12123 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12124 for some CRL extensions and new objects added.
12127 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12128 key usage extension and fuller support for authority key id.
12131 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12132 padding method for RSA, which is recommended for new applications in PKCS
12133 #1 v2.0 (RFC 2437, October 1998).
12134 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12135 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12136 against Bleichbacher's attack on RSA.
12137 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12140 *) Updates to the new SSL compression code
12141 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12143 *) Fix so that the version number in the master secret, when passed
12144 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12145 (because the server will not accept higher), that the version number
12146 is 0x03,0x01, not 0x03,0x00
12147 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12149 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12150 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12151 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12154 *) Support for RAW extensions where an arbitrary extension can be
12155 created by including its DER encoding. See apps/openssl.cnf for
12159 *) Make sure latest Perl versions don't interpret some generated C array
12160 code as Perl array code in the crypto/err/err_genc.pl script.
12161 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12163 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12164 not many people have the assembler. Various Win32 compilation fixes and
12165 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12166 build instructions.
12169 *) Modify configure script 'Configure' to automatically create crypto/date.h
12170 file under Win32 and also build pem.h from pem.org. New script
12171 util/mkfiles.pl to create the MINFO file on environments that can't do a
12172 'make files': perl util/mkfiles.pl >MINFO should work.
12175 *) Major rework of DES function declarations, in the pursuit of correctness
12176 and purity. As a result, many evil casts evaporated, and some weirdness,
12177 too. You may find this causes warnings in your code. Zapping your evil
12178 casts will probably fix them. Mostly.
12181 *) Fix for a typo in asn1.h. Bug fix to object creation script
12182 obj_dat.pl. It considered a zero in an object definition to mean
12183 "end of object": none of the objects in objects.h have any zeros
12184 so it wasn't spotted.
12185 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12187 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12188 Masking (CBCM). In the absence of test vectors, the best I have been able
12189 to do is check that the decrypt undoes the encrypt, so far. Send me test
12190 vectors if you have them.
12193 *) Correct calculation of key length for export ciphers (too much space was
12194 allocated for null ciphers). This has not been tested!
12197 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12198 message is now correct (it understands "crypto" and "ssl" on its
12199 command line). There is also now an "update" option. This will update
12200 the util/ssleay.num and util/libeay.num files with any new functions.
12202 perl util/mkdef.pl crypto ssl update
12203 it will update them.
12206 *) Overhauled the Perl interface (perl/*):
12207 - ported BN stuff to OpenSSL's different BN library
12208 - made the perl/ source tree CVS-aware
12209 - renamed the package from SSLeay to OpenSSL (the files still contain
12210 their history because I've copied them in the repository)
12211 - removed obsolete files (the test scripts will be replaced
12212 by better Test::Harness variants in the future)
12213 [Ralf S. Engelschall]
12215 *) First cut for a very conservative source tree cleanup:
12216 1. merge various obsolete readme texts into doc/ssleay.txt
12217 where we collect the old documents and readme texts.
12218 2. remove the first part of files where I'm already sure that we no
12219 longer need them because of three reasons: either they are just temporary
12220 files which were left by Eric or they are preserved original files where
12221 I've verified that the diff is also available in the CVS via "cvs diff
12222 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12223 the crypto/md/ stuff).
12224 [Ralf S. Engelschall]
12226 *) More extension code. Incomplete support for subject and issuer alt
12227 name, issuer and authority key id. Change the i2v function parameters
12228 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12229 what that's for :-) Fix to ASN1 macro which messed up
12230 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12233 *) Preliminary support for ENUMERATED type. This is largely copied from the
12237 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12238 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12240 *) Make sure `make rehash' target really finds the `openssl' program.
12241 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12243 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12244 like to hear about it if this slows down other processors.
12247 *) Add CygWin32 platform information to Configure script.
12248 [Alan Batie <batie@aahz.jf.intel.com>]
12250 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12251 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12253 *) New program nseq to manipulate netscape certificate sequences
12256 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12260 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12261 but the BN code had some problems that would cause failures when
12262 doing certificate verification and some other functions.
12263 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12265 *) Add ASN1 and PEM code to support netscape certificate sequences.
12268 *) Add ASN1 and PEM code to support netscape certificate sequences.
12271 *) Add several PKIX and private extended key usage OIDs.
12274 *) Modify the 'ca' program to handle the new extension code. Modify
12275 openssl.cnf for new extension format, add comments.
12278 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12279 and add a sample to openssl.cnf so req -x509 now adds appropriate
12283 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12284 error code, add initial support to X509_print() and x509 application.
12287 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12288 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12289 stuff is currently isolated and isn't even compiled yet.
12292 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12293 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12294 Removed the versions check from X509 routines when loading extensions:
12295 this allows certain broken certificates that don't set the version
12296 properly to be processed.
12299 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12300 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12301 can still be regenerated with "make depend".
12304 *) Spelling mistake in C version of CAST-128.
12305 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12307 *) Changes to the error generation code. The perl script err-code.pl
12308 now reads in the old error codes and retains the old numbers, only
12309 adding new ones if necessary. It also only changes the .err files if new
12310 codes are added. The makefiles have been modified to only insert errors
12311 when needed (to avoid needlessly modifying header files). This is done
12312 by only inserting errors if the .err file is newer than the auto generated
12313 C file. To rebuild all the error codes from scratch (the old behaviour)
12314 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12315 or delete all the .err files.
12318 *) CAST-128 was incorrectly implemented for short keys. The C version has
12319 been fixed, but is untested. The assembler versions are also fixed, but
12320 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12321 to regenerate it if needed.
12322 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12323 Hagino <itojun@kame.net>]
12325 *) File was opened incorrectly in randfile.c.
12326 [Ulf Möller <ulf@fitug.de>]
12328 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12329 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12330 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12331 al: it's just almost always a UTCTime. Note this patch adds new error
12332 codes so do a "make errors" if there are problems.
12335 *) Correct Linux 1 recognition in config.
12336 [Ulf Möller <ulf@fitug.de>]
12338 *) Remove pointless MD5 hash when using DSA keys in ca.
12339 [Anonymous <nobody@replay.com>]
12341 *) Generate an error if given an empty string as a cert directory. Also
12342 generate an error if handed NULL (previously returned 0 to indicate an
12343 error, but didn't set one).
12344 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12346 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12349 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12350 parameters. This was causing a warning which killed off the Win32 compile.
12353 *) Remove C++ style comments from crypto/bn/bn_local.h.
12354 [Neil Costigan <neil.costigan@celocom.com>]
12356 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12357 based on a text string, looking up short and long names and finally
12358 "dot" format. The "dot" format stuff didn't work. Added new function
12359 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12360 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12361 OID is not part of the table.
12364 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12365 X509_LOOKUP_by_alias().
12368 *) Sort openssl functions by name.
12371 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12372 encryption from sample DSA keys (in case anyone is interested the password
12376 *) Make _all_ *_free functions accept a NULL pointer.
12377 [Frans Heymans <fheymans@isaserver.be>]
12379 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12381 [Anonymous <nobody@replay.com>]
12383 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12384 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12386 *) Don't blow it for numeric -newkey arguments to apps/req.
12387 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12389 *) Temp key "for export" tests were wrong in s3_srvr.c.
12390 [Anonymous <nobody@replay.com>]
12392 *) Add prototype for temp key callback functions
12393 SSL_CTX_set_tmp_{rsa,dh}_callback().
12396 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12397 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12400 *) X509_name_add_entry() freed the wrong thing after an error.
12401 [Arne Ansper <arne@ats.cyber.ee>]
12403 *) rsa_eay.c would attempt to free a NULL context.
12404 [Arne Ansper <arne@ats.cyber.ee>]
12406 *) BIO_s_socket() had a broken should_retry() on Windoze.
12407 [Arne Ansper <arne@ats.cyber.ee>]
12409 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12410 [Arne Ansper <arne@ats.cyber.ee>]
12412 *) Make sure the already existing X509_STORE->depth variable is initialized
12413 in X509_STORE_new(), but document the fact that this variable is still
12414 unused in the certificate verification process.
12415 [Ralf S. Engelschall]
12417 *) Fix the various library and apps files to free up pkeys obtained from
12418 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12421 *) Fix reference counting in X509_PUBKEY_get(). This makes
12422 demos/maurice/example2.c work, amongst others, probably.
12423 [Steve Henson and Ben Laurie]
12425 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12426 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12427 are no longer created. This way we have a single and consistent command
12428 line interface `openssl <command>', similar to `cvs <command>'.
12429 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12431 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12432 BIT STRING wrapper always have zero unused bits.
12435 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12438 *) Make the top-level INSTALL documentation easier to understand.
12441 *) Makefiles updated to exit if an error occurs in a sub-directory
12442 make (including if user presses ^C) [Paul Sutton]
12444 *) Make Montgomery context stuff explicit in RSA data structure.
12447 *) Fix build order of pem and err to allow for generated pem.h.
12450 *) Fix renumbering bug in X509_NAME_delete_entry().
12453 *) Enhanced the err-ins.pl script so it makes the error library number
12454 global and can add a library name. This is needed for external ASN1 and
12455 other error libraries.
12458 *) Fixed sk_insert which never worked properly.
12461 *) Fix ASN1 macros so they can handle indefinite length constructed
12462 EXPLICIT tags. Some non standard certificates use these: they can now
12466 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12467 into a single doc/ssleay.txt bundle. This way the information is still
12468 preserved but no longer messes up this directory. Now it's new room for
12469 the new set of documentation files.
12470 [Ralf S. Engelschall]
12472 *) SETs were incorrectly DER encoded. This was a major pain, because they
12473 shared code with SEQUENCEs, which aren't coded the same. This means that
12474 almost everything to do with SETs or SEQUENCEs has either changed name or
12475 number of arguments.
12476 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12478 *) Fix test data to work with the above.
12481 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12482 was already fixed by Eric for 0.9.1 it seems.
12483 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12485 *) Autodetect FreeBSD3.
12488 *) Fix various bugs in Configure. This affects the following platforms:
12492 unixware-2.0-pentium
12496 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12497 before they are needed.
12500 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12504 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12506 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12507 changed SSLeay to OpenSSL in version strings.
12508 [Ralf S. Engelschall]
12510 *) Some fixups to the top-level documents.
12513 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12514 because the symlink to include/ was missing.
12515 [Ralf S. Engelschall]
12517 *) Incorporated the popular no-RSA/DSA-only patches
12518 which allow to compile a RSA-free SSLeay.
12519 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12521 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12522 when "ssleay" is still not found.
12523 [Ralf S. Engelschall]
12525 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12526 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12528 *) Updated the README file.
12529 [Ralf S. Engelschall]
12531 *) Added various .cvsignore files in the CVS repository subdirs
12532 to make a "cvs update" really silent.
12533 [Ralf S. Engelschall]
12535 *) Recompiled the error-definition header files and added
12536 missing symbols to the Win32 linker tables.
12537 [Ralf S. Engelschall]
12539 *) Cleaned up the top-level documents;
12540 o new files: CHANGES and LICENSE
12541 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12542 o merged COPYRIGHT into LICENSE
12543 o removed obsolete TODO file
12544 o renamed MICROSOFT to INSTALL.W32
12545 [Ralf S. Engelschall]
12547 *) Removed dummy files from the 0.9.1b source tree:
12548 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12549 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12550 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12551 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12552 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12553 [Ralf S. Engelschall]
12555 *) Added various platform portability fixes.
12558 *) The Genesis of the OpenSSL rpject:
12559 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12560 Young and Tim J. Hudson created while they were working for C2Net until
12562 [The OpenSSL Project]
12565 Changes between 0.9.0b and 0.9.1b [not released]
12567 *) Updated a few CA certificates under certs/
12570 *) Changed some BIGNUM api stuff.
12573 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12574 DGUX x86, Linux Alpha, etc.
12577 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12578 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12582 *) Add -strparse option to asn1pars program which parses nested
12584 [Dr Stephen Henson <shenson@bigfoot.com>]
12586 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12589 *) DSA fix for "ca" program.
12592 *) Added "-genkey" option to "dsaparam" program.
12595 *) Added RIPE MD160 (rmd160) message digest.
12598 *) Added -a (all) option to "ssleay version" command.
12601 *) Added PLATFORM define which is the id given to Configure.
12604 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12607 *) Extended the ASN.1 parser routines.
12610 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12613 *) Added a BN_CTX to the BN library.
12616 *) Fixed the weak key values in DES library
12619 *) Changed API in EVP library for cipher aliases.
12622 *) Added support for RC2/64bit cipher.
12625 *) Converted the lhash library to the crypto/mem.c functions.
12628 *) Added more recognized ASN.1 object ids.
12631 *) Added more RSA padding checks for SSL/TLS.
12634 *) Added BIO proxy/filter functionality.
12637 *) Added extra_certs to SSL_CTX which can be used
12638 send extra CA certificates to the client in the CA cert chain sending
12639 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12642 *) Now Fortezza is denied in the authentication phase because
12643 this is key exchange mechanism is not supported by SSLeay at all.
12646 *) Additional PKCS1 checks.
12649 *) Support the string "TLSv1" for all TLS v1 ciphers.
12652 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12653 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12656 *) Fixed a few memory leaks.
12659 *) Fixed various code and comment typos.
12662 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12663 bytes sent in the client random.
12664 [Edward Bishop <ebishop@spyglass.com>]