5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the [git commit log][log] and
7 pick the appropriate release branch.
9 [log]: https://github.com/openssl/openssl/commits/
11 ### Changes between 1.1.1 and 3.0.0 [xx XXX xxxx] ###
13 * Removed include/openssl/opensslconf.h.in and replaced it with
14 include/openssl/configuration.h.in, which differs in not including
15 <openssl/macros.h>. A short header include/openssl/opensslconf.h
16 was added to include both.
18 This allows internal hacks where one might need to modify the set
19 of configured macros, for example this if deprecated symbols are
20 still supposed to be available internally:
22 #include <openssl/configuration.h>
24 #undef OPENSSL_NO_DEPRECATED
25 #define OPENSSL_SUPPRESS_DEPRECATED
27 #include <openssl/macros.h>
29 This should not be used by applications that use the exported
30 symbols, as that will lead to linking errors.
33 * Fixed an an overflow bug in the x64_64 Montgomery squaring procedure
34 used in exponentiation with 512-bit moduli. No EC algorithms are
35 affected. Analysis suggests that attacks against 2-prime RSA1024,
36 3-prime RSA1536, and DSA1024 as a result of this defect would be very
37 difficult to perform and are not believed likely. Attacks against DH512
38 are considered just feasible. However, for an attack the target would
39 have to re-use the DH512 private key, which is not recommended anyway.
40 Also applications directly using the low level API BN_mod_exp may be
41 affected if they use BN_FLG_CONSTTIME.
45 * Most memory-debug features have been deprecated, and the functionality
49 * Most common options (such as -rand/-writerand, TLS version control, etc)
50 were refactored and point to newly-enhanced descriptions in openssl.pod
53 * Introduced a new method type and API, OSSL_SERIALIZER, to
54 represent generic serializers. An implementation is expected to
55 be able to serialize an object associated with a given name (such
56 as an algorithm name for an asymmetric key) into forms given by
57 implementation properties.
59 Serializers are primarily used from inside libcrypto, through
60 calls to functions like EVP_PKEY_print_private(),
61 PEM_write_bio_PrivateKey() and similar.
63 Serializers are specified in such a way that they can be made to
64 directly handle the provider side portion of an object, if this
65 provider side part comes from the same provider as the serializer
66 itself, but can also be made to handle objects in parametrized
67 form (as an OSSL_PARAM array of data). This allows a provider to
68 offer generic serializers as a service for any other provider.
72 * Added a .pragma directive to the syntax of configuration files, to
73 allow varying behavior in a supported and predictable manner.
74 Currently added pragma:
78 This allows dollar signs to be a keyword character unless it's
79 followed by a opening brace or parenthesis. This is useful for
80 platforms where dollar signs are commonly used in names, such as
81 volume names and system directory names on VMS.
85 * Added functionality to create an EVP_PKEY from user data. This
86 is effectively the same as creating a RSA, DH or DSA object and
87 then assigning them to an EVP_PKEY, but directly using algorithm
88 agnostic EVP functions. A benefit is that this should be future
89 proof for public key algorithms to come.
93 * Change the interpretation of the '--api' configuration option to
94 mean that this is a desired API compatibility level with no
95 further meaning. The previous interpretation, that this would
96 also mean to remove all deprecated symbols up to and including
97 the given version, no requires that 'no-deprecated' is also used
100 When building applications, the desired API compatibility level
101 can be set with the OPENSSL_API_COMPAT macro like before. For
102 API compatibility version below 3.0, the old style numerical
103 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
104 For version 3.0 and on, the value is expected to be the decimal
105 value calculated from the major and minor version like this:
107 MAJOR * 10000 + MINOR * 100
111 -DOPENSSL_API_COMPAT=30000 For 3.0
112 -DOPENSSL_API_COMPAT=30200 For 3.2
114 To hide declarations that are deprecated up to and including the
115 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
116 given when building the application as well.
120 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
121 access to certificate and CRL stores via URIs and OSSL_STORE
124 This adds the following functions:
126 - X509_LOOKUP_store()
127 - X509_STORE_load_file()
128 - X509_STORE_load_path()
129 - X509_STORE_load_store()
130 - SSL_add_store_cert_subjects_to_stack()
131 - SSL_CTX_set_default_verify_store()
132 - SSL_CTX_load_verify_file()
133 - SSL_CTX_load_verify_dir()
134 - SSL_CTX_load_verify_store()
136 Also, the following functions are now deprecated:
138 - X509_STORE_load_locations() (use X509_STORE_load_file(),
139 X509_STORE_load_path() or X509_STORE_load_store() instead)
140 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
141 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
145 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
146 The presence of this system service is determined at run-time.
150 * Added functionality to create an EVP_PKEY context based on data
151 for methods from providers. This takes an algorithm name and a
152 property query string and simply stores them, with the intent
153 that any operation that uses this context will use those strings
154 to fetch the needed methods implicitly, thereby making the port
155 of application written for pre-3.0 OpenSSL easier.
159 * The undocumented function NCONF_WIN32() has been deprecated; for
160 conversion details see the HISTORY section of doc/man5/config.pod
164 * Introduced the new functions EVP_DigestSignInit_ex() and
165 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
166 EVP_DigestVerifyUpdate() have been converted to functions. See the man
167 pages for further details.
171 * Over two thousand fixes were made to the documentation, including:
172 adding missing command flags, better style conformance, documentation
175 *Rich Salz, Richard Levitte*
177 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
178 X25519, X448, Ed25519 and Ed448.
182 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
187 * Deprecated the public definition of ERR_STATE as well as the function
188 ERR_get_state(). This is done in preparation of making ERR_STATE an
193 * Added ERR functionality to give callers access to the stored function
194 names that have replaced the older function code based functions.
196 New functions are ERR_get_error_func(), ERR_peek_error_func(),
197 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
198 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
199 and ERR_peek_last_error_all().
201 These functions have become deprecated: ERR_get_error_line_data(),
202 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
203 ERR_func_error_string().
207 * Extended testing to be verbose for failing tests only. The make variables
208 VERBOSE_FAILURE or VF can be used to enable this:
210 $ make VF=1 test # Unix
211 $ mms /macro=(VF=1) test ! OpenVMS
212 $ nmake VF=1 test # Windows
217 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
218 used even when parsing explicit parameters, when loading a serialized key
219 or calling `EC_GROUP_new_from_ecpkparameters()`/
220 `EC_GROUP_new_from_ecparameters()`.
221 This prevents bypass of security hardening and performance gains,
222 especially for curves with specialized EC_METHODs.
223 By default, if a key encoded with explicit parameters is loaded and later
224 serialized, the output is still encoded with explicit parameters, even if
225 internally a "named" EC_GROUP is used for computation.
229 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
230 this change, EC_GROUP_set_generator would accept order and/or cofactor as
231 NULL. After this change, only the cofactor parameter can be NULL. It also
232 does some minimal sanity checks on the passed order.
237 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
238 An attack is simple, if the first CMS_recipientInfo is valid but the
239 second CMS_recipientInfo is chosen ciphertext. If the second
240 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
241 encryption key will be replaced by garbage, and the message cannot be
242 decoded, but if the RSA decryption fails, the correct encryption key is
243 used and the recipient will not notice the attack.
244 As a work around for this potential attack the length of the decrypted
245 key must be equal to the cipher default key length, in case the
246 certifiate is not given and all recipientInfo are tried out.
247 The old behaviour can be re-enabled in the CMS code by setting the
248 CMS_DEBUG_DECRYPT flag.
252 * Early start up entropy quality from the DEVRANDOM seed source has been
253 improved for older Linux systems. The RAND subsystem will wait for
254 /dev/random to be producing output before seeding from /dev/urandom.
255 The seeded state is stored for future library initialisations using
256 a system global shared memory segment. The shared memory identifier
257 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
258 the desired value. The default identifier is 114.
262 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
263 when primes for RSA keys are computed.
264 Since we previously always generated primes == 2 (mod 3) for RSA keys,
265 the 2-prime and 3-prime RSA modules were easy to distinguish, since
266 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
267 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
268 This avoids possible fingerprinting of newly generated RSA modules.
272 * Correct the extended master secret constant on EBCDIC systems. Without this
273 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
274 negotiate EMS will fail. Unfortunately this also means that TLS connections
275 between EBCDIC systems with this fix, and EBCDIC systems without this
276 fix will fail if they negotiate EMS.
280 * Changed the library initialisation so that the config file is now loaded
281 by default. This was already the case for libssl. It now occurs for both
282 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
283 OPENSSL_init_crypto() to suppress automatic loading of a config file.
287 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
288 where the former acts as a replacement for ERR_put_error(), and the
289 latter replaces the combination ERR_put_error()+ERR_add_error_data().
290 ERR_raise_data() adds more flexibility by taking a format string and
291 an arbitrary number of arguments following it, to be processed with
296 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
297 to check if a named provider is loaded and available. When called, it
298 will also activate all fallback providers if such are still present.
302 * Enforce a minimum DH modulus size of 512 bits.
306 * Changed DH parameters to generate the order q subgroup instead of 2q.
307 Previously generated DH parameters are still accepted by DH_check
308 but DH_generate_key works around that by clearing bit 0 of the
309 private key for those. This avoids leaking bit 0 of the private key.
313 * Significantly reduce secure memory usage by the randomness pools.
317 * {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
322 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
323 algorithms. An implementation of a key exchange algorithm can be obtained
324 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
325 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
326 the older EVP_PKEY_derive_init() function. See the man pages for the new
327 functions for further details.
331 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
335 * Removed the function names from error messages and deprecated the
338 * Removed NextStep support and the macro OPENSSL_UNISTD
342 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
343 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
344 Also removed "export var as function" capability; we do not export
345 variables, only functions.
349 * RC5_32_set_key has been changed to return an int type, with 0 indicating
350 an error and 1 indicating success. In previous versions of OpenSSL this
351 was a void type. If a key was set longer than the maximum possible this
356 * Support SM2 signing and verification schemes with X509 certificate.
360 * Use SHA256 as the default digest for TS query in the ts app.
364 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
365 This checks that the salt length is at least 128 bits, the derived key
366 length is at least 112 bits, and that the iteration count is at least 1000.
367 For backwards compatibility these checks are disabled by default in the
368 default provider, but are enabled by default in the fips provider.
369 To enable or disable these checks use the control
370 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
374 * Default cipher lists/suites are now available via a function, the
375 #defines are deprecated.
379 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
380 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
381 for Windows Store apps easier. Also, the "no-uplink" option has been added.
385 * Join the directories crypto/x509 and crypto/x509v3
389 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
390 This changes the size when using the genpkey app when no size is given. It
391 fixes an omission in earlier changes that changed all RSA, DSA and DH
392 generation apps to use 2048 bits by default.
396 * Added command 'openssl kdf' that uses the EVP_KDF API.
400 * Added command 'openssl mac' that uses the EVP_MAC API.
404 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
405 as default directories. Also added the command 'openssl info'
406 for scripting purposes.
410 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
411 deprecated. These undocumented functions were never integrated into the EVP
412 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
413 Bi-directional IGE mode. These modes were never formally standardised and
414 usage of these functions is believed to be very small. In particular
415 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
416 is ever used. The security implications are believed to be minimal, but
417 this issue was never fixed for backwards compatibility reasons. New code
418 should not use these modes.
422 * Add prediction resistance to the DRBG reseeding process.
426 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
427 mandated by IEEE Std 1619-2018.
431 * Added newline escaping functionality to a filename when using openssl dgst.
432 This output format is to replicate the output format found in the '*sum'
433 checksum programs. This aims to preserve backward compatibility.
435 *Matt Eaton, Richard Levitte, and Paul Dale*
437 * Removed the heartbeat message in DTLS feature, as it has very
438 little usage and doesn't seem to fulfill a valuable purpose.
439 The configuration option is now deprecated.
443 * Changed the output of 'openssl {digestname} < file' to display the
444 digest name in its output.
448 * Added a new generic trace API which provides support for enabling
449 instrumentation through trace output. This feature is mainly intended
450 as an aid for developers and is disabled by default. To utilize it,
451 OpenSSL needs to be configured with the `enable-trace` option.
453 If the tracing API is enabled, the application can activate trace output
454 by registering BIOs as trace channels for a number of tracing and debugging
457 The 'openssl' application has been expanded to enable any of the types
458 available via environment variables defined by the user, and serves as
459 one possible example on how to use this functionality.
461 *Richard Levitte & Matthias St. Pierre*
463 * Added build tests for C++. These are generated files that only do one
464 thing, to include one public OpenSSL head file each. This tests that
465 the public header files can be usefully included in a C++ application.
467 This test isn't enabled by default. It can be enabled with the option
468 'enable-buildtest-c++'.
472 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
476 * Add KMAC to EVP_MAC.
480 * Added property based algorithm implementation selection framework to
485 * Added SCA hardening for modular field inversion in EC_GROUP through
486 a new dedicated field_inv() pointer in EC_METHOD.
487 This also addresses a leakage affecting conversions from projective
488 to affine coordinates.
490 *Billy Bob Brumley, Nicola Tuveri*
492 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
493 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
494 those algorithms that were already supported through the EVP_PKEY API
495 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
496 and scrypt are now wrappers that call EVP_KDF.
500 * Build devcrypto engine as a dynamic engine.
504 * Add keyed BLAKE2 to EVP_MAC.
508 * Fix a bug in the computation of the endpoint-pair shared secret used
509 by DTLS over SCTP. This breaks interoperability with older versions
510 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
511 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
512 interoperability with such broken implementations. However, enabling
513 this switch breaks interoperability with correct implementations.
515 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
516 re-used X509_PUBKEY object if the second PUBKEY is malformed.
520 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
524 * Change the license to the Apache License v2.0.
528 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
530 - Major releases (indicated by incrementing the MAJOR release number)
531 may introduce incompatible API/ABI changes.
532 - Minor releases (indicated by incrementing the MINOR release number)
533 may introduce new features but retain API/ABI compatibility.
534 - Patch releases (indicated by incrementing the PATCH number)
535 are intended for bug fixes and other improvements of existing
536 features only (like improving performance or adding documentation)
537 and retain API/ABI compatibility.
541 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
545 * Remove the 'dist' target and add a tarball building script. The
546 'dist' target has fallen out of use, and it shouldn't be
547 necessary to configure just to create a source distribution.
551 * Recreate the OS390-Unix config target. It no longer relies on a
552 special script like it did for OpenSSL pre-1.1.0.
556 * Instead of having the source directories listed in Configure, add
557 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
562 * Add GMAC to EVP_MAC.
566 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
570 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
571 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
572 to facilitate the continued use of MACs through raw private keys in
573 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
577 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
578 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
582 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
583 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
584 are retained for backwards compatibility.
588 * AES-XTS mode now enforces that its two keys are different to mitigate
589 the attacked described in "Efficient Instantiations of Tweakable
590 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
591 Details of this attack can be obtained from:
592 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
596 * Rename the object files, i.e. give them other names than in previous
597 versions. Their names now include the name of the final product, as
598 well as its type mnemonic (bin, lib, shlib).
602 * Added new option for 'openssl list', '-objects', which will display the
603 list of built in objects, i.e. OIDs with names.
607 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
608 improves application performance by removing data copies and providing
609 applications with zero-copy system calls such as sendfile and splice.
613 ### Changes between 1.1.1a and 1.1.1b [xx XXX xxxx] ###
615 * Change the info callback signals for the start and end of a post-handshake
616 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
617 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
618 confused by this and assume that a TLSv1.2 renegotiation has started. This
619 can break KeyUpdate handling. Instead we no longer signal the start and end
620 of a post handshake message exchange (although the messages themselves are
621 still signalled). This could break some applications that were expecting
622 the old signals. However without this KeyUpdate is not usable for many
627 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018] ###
629 * Timing vulnerability in DSA signature generation
631 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
632 timing side channel attack. An attacker could use variations in the signing
633 algorithm to recover the private key.
635 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
640 * Timing vulnerability in ECDSA signature generation
642 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
643 timing side channel attack. An attacker could use variations in the signing
644 algorithm to recover the private key.
646 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
651 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
652 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
653 of two gigabytes and the error handling improved.
655 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
656 categorized as a normal bug, not a security issue, because the DRBG reseeds
657 automatically and is fully functional even without additional randomness
658 provided by the application.
660 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018] ###
662 * Add a new ClientHello callback. Provides a callback interface that gives
663 the application the ability to adjust the nascent SSL object at the
664 earliest stage of ClientHello processing, immediately after extensions have
665 been collected but before they have been processed. In particular, this
666 callback can adjust the supported TLS versions in response to the contents
671 * Add SM2 base algorithm support.
675 * s390x assembly pack: add (improved) hardware-support for the following
676 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
677 aes-cfb/cfb8, aes-ecb.
681 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
682 parameter is no longer accepted, as it leads to a corrupt table. NULL
683 pem_str is reserved for alias entries only.
687 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
688 step for prime curves. The new implementation is based on formulae from
689 differential addition-and-doubling in homogeneous projective coordinates
690 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
691 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
692 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
693 to work in projective coordinates.
695 *Billy Bob Brumley, Nicola Tuveri*
697 * Change generating and checking of primes so that the error rate of not
698 being prime depends on the intended use based on the size of the input.
699 For larger primes this will result in more rounds of Miller-Rabin.
700 The maximal error rate for primes with more than 1080 bits is lowered
703 *Kurt Roeckx, Annie Yousar*
705 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
709 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
710 moving between systems, and to avoid confusion when a Windows build is
711 done with mingw vs with MSVC. For POSIX installs, there's still a
712 symlink or copy named 'tsget' to avoid that confusion as well.
716 * Revert blinding in ECDSA sign and instead make problematic addition
717 length-invariant. Switch even to fixed-length Montgomery multiplication.
721 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
722 step for binary curves. The new implementation is based on formulae from
723 differential addition-and-doubling in mixed Lopez-Dahab projective
724 coordinates, modified to independently blind the operands.
726 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
728 * Add a scaffold to optionally enhance the Montgomery ladder implementation
729 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
730 EC_METHODs to implement their own specialized "ladder step", to take
731 advantage of more favorable coordinate systems or more efficient
732 differential addition-and-doubling algorithms.
734 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
736 * Modified the random device based seed sources to keep the relevant
737 file descriptors open rather than reopening them on each access.
738 This allows such sources to operate in a chroot() jail without
739 the associated device nodes being available. This behaviour can be
740 controlled using RAND_keep_random_devices_open().
744 * Numerous side-channel attack mitigations have been applied. This may have
745 performance impacts for some algorithms for the benefit of improved
746 security. Specific changes are noted in this change log by their respective
751 * AIX shared library support overhaul. Switch to AIX "natural" way of
752 handling shared libraries, which means collecting shared objects of
753 different versions and bitnesses in one common archive. This allows to
754 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
755 doesn't affect the way 3rd party applications are linked, only how
756 multi-version installation is managed.
760 * Make ec_group_do_inverse_ord() more robust and available to other
761 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
762 mitigations are applied to the fallback BN_mod_inverse().
763 When using this function rather than BN_mod_inverse() directly, new
764 EC cryptosystem implementations are then safer-by-default.
768 * Add coordinate blinding for EC_POINT and implement projective
769 coordinate blinding for generic prime curves as a countermeasure to
770 chosen point SCA attacks.
772 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
774 * Add blinding to ECDSA and DSA signatures to protect against side channel
775 attacks discovered by Keegan Ryan (NCC Group).
779 * Enforce checking in the pkeyutl command line app to ensure that the input
780 length does not exceed the maximum supported digest length when performing
781 a sign, verify or verifyrecover operation.
785 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
786 I/O in combination with something like select() or poll() will hang. This
787 can be turned off again using SSL_CTX_clear_mode().
788 Many applications do not properly handle non-application data records, and
789 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
790 around the problems in those applications, but can also break some.
791 It's recommended to read the manpages about SSL_read(), SSL_write(),
792 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
793 SSL_CTX_set_read_ahead() again.
797 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
798 now allow empty (zero character) pass phrases.
802 * Apply blinding to binary field modular inversion and remove patent
803 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
807 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
808 binary and prime elliptic curves.
812 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
813 constant time fixed point multiplication.
817 * Revise elliptic curve scalar multiplication with timing attack
818 defenses: ec_wNAF_mul redirects to a constant time implementation
819 when computing fixed point and variable point multiplication (which
820 in OpenSSL are mostly used with secret scalars in keygen, sign,
821 ECDH derive operations).
822 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
825 * Updated CONTRIBUTING
829 * Updated DRBG / RAND to request nonce and additional low entropy
830 randomness from the system.
832 *Matthias St. Pierre*
834 * Updated 'openssl rehash' to use OpenSSL consistent default.
838 * Moved the load of the ssl_conf module to libcrypto, which helps
839 loading engines that libssl uses before libssl is initialised.
843 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
847 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
849 *Ingo Schwarze, Rich Salz*
851 * Added output of accepting IP address and port for 'openssl s_server'
855 * Added a new API for TLSv1.3 ciphersuites:
856 SSL_CTX_set_ciphersuites()
857 SSL_set_ciphersuites()
861 * Memory allocation failures consistently add an error to the error
866 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
867 in libcrypto when run as setuid/setgid.
871 * Load any config file by default when libssl is used.
875 * Added new public header file <openssl/rand_drbg.h> and documentation
876 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
878 *Matthias St. Pierre*
880 * QNX support removed (cannot find contributors to get their approval
881 for the license change).
885 * TLSv1.3 replay protection for early data has been implemented. See the
886 SSL_read_early_data() man page for further details.
890 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
891 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
892 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
893 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
894 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
895 configuration has been separated out. See the ciphers man page or the
896 SSL_CTX_set_ciphersuites() man page for more information.
900 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
901 in responder mode now supports the new "-multi" option, which
902 spawns the specified number of child processes to handle OCSP
903 requests. The "-timeout" option now also limits the OCSP
904 responder's patience to wait to receive the full client request
905 on a newly accepted connection. Child processes are respawned
906 as needed, and the CA index file is automatically reloaded
907 when changed. This makes it possible to run the "ocsp" responder
908 as a long-running service, making the OpenSSL CA somewhat more
909 feature-complete. In this mode, most diagnostic messages logged
910 after entering the event loop are logged via syslog(3) rather than
915 * Added support for X448 and Ed448. Heavily based on original work by
920 * Extend OSSL_STORE with capabilities to search and to narrow the set of
921 objects loaded. This adds the functions OSSL_STORE_expect() and
922 OSSL_STORE_find() as well as needed tools to construct searches and
923 get the search data out of them.
927 * Support for TLSv1.3 added. Note that users upgrading from an earlier
928 version of OpenSSL should review their configuration settings to ensure
929 that they are still appropriate for TLSv1.3. For further information see:
930 https://wiki.openssl.org/index.php/TLS1.3
934 * Grand redesign of the OpenSSL random generator
936 The default RAND method now utilizes an AES-CTR DRBG according to
937 NIST standard SP 800-90Ar1. The new random generator is essentially
938 a port of the default random generator from the OpenSSL FIPS 2.0
939 object module. It is a hybrid deterministic random bit generator
940 using an AES-CTR bit stream and which seeds and reseeds itself
941 automatically using trusted system entropy sources.
943 Some of its new features are:
944 - Support for multiple DRBG instances with seed chaining.
945 - The default RAND method makes use of a DRBG.
946 - There is a public and private DRBG instance.
947 - The DRBG instances are fork-safe.
948 - Keep all global DRBG instances on the secure heap if it is enabled.
949 - The public and private DRBG instance are per thread for lock free
952 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
954 * Changed Configure so it only says what it does and doesn't dump
955 so much data. Instead, ./configdata.pm should be used as a script
956 to display all sorts of configuration data.
960 * Added processing of "make variables" to Configure.
964 * Added SHA512/224 and SHA512/256 algorithm support.
968 * The last traces of Netware support, first removed in 1.1.0, have
973 * Get rid of Makefile.shared, and in the process, make the processing
974 of certain files (rc.obj, or the .def/.map/.opt files produced from
975 the ordinal files) more visible and hopefully easier to trace and
976 debug (or make silent).
980 * Make it possible to have environment variable assignments as
981 arguments to config / Configure.
985 * Add multi-prime RSA (RFC 8017) support.
989 * Add SM3 implemented according to GB/T 32905-2016
990 * Jack Lloyd <jack.lloyd@ribose.com>,
991 Ronald Tse <ronald.tse@ribose.com>,
992 Erick Borsboom <erick.borsboom@ribose.com> *
994 * Add 'Maximum Fragment Length' TLS extension negotiation and support
995 as documented in RFC6066.
996 Based on a patch from Tomasz Moń
998 *Filipe Raimundo da Silva*
1000 * Add SM4 implemented according to GB/T 32907-2016.
1001 * Jack Lloyd <jack.lloyd@ribose.com>,
1002 Ronald Tse <ronald.tse@ribose.com>,
1003 Erick Borsboom <erick.borsboom@ribose.com> *
1005 * Reimplement -newreq-nodes and ERR_error_string_n; the
1006 original author does not agree with the license change.
1010 * Add ARIA AEAD TLS support.
1014 * Some macro definitions to support VS6 have been removed. Visual
1015 Studio 6 has not worked since 1.1.0
1019 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1020 without clearing the errors.
1024 * Add "atfork" functions. If building on a system that without
1025 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1026 requirements. The RAND facility now uses/requires this.
1034 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1035 not possible to disable entirely. However, it's still possible to
1036 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1039 To disable, configure with 'no-ui-console'. 'no-ui' is still
1040 possible to use as an alias. Check at compile time with the
1041 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1042 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1046 * Add a STORE module, which implements a uniform and URI based reader of
1047 stores that can contain keys, certificates, CRLs and numerous other
1048 objects. The main API is loosely based on a few stdio functions,
1049 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1050 OSSL_STORE_error and OSSL_STORE_close.
1051 The implementation uses backends called "loaders" to implement arbitrary
1052 URI schemes. There is one built in "loader" for the 'file' scheme.
1056 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1057 then adjusted to work on FreeBSD 8.4 as well.
1058 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1059 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1063 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1064 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1065 error code calls like this:
1067 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1069 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1070 that can be encoded in C. For the foreseeable future, this will only
1073 *Richard Levitte and Tim Hudson*
1075 * Removed BSD cryptodev engine.
1079 * Add a build target 'build_all_generated', to build all generated files
1080 and only that. This can be used to prepare everything that requires
1081 things like perl for a system that lacks perl and then move everything
1082 to that system and do the rest of the build there.
1086 * In the UI interface, make it possible to duplicate the user data. This
1087 can be used by engines that need to retain the data for a longer time
1088 than just the call where this user data is passed.
1092 * Ignore the '-named_curve auto' value for compatibility of applications
1095 *Tomas Mraz <tmraz@fedoraproject.org>*
1097 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1098 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1099 alerts across multiple records (some of which could be empty). In practice
1100 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1101 prohibts this altogether and other libraries (BoringSSL, NSS) do not
1102 support this at all. Supporting it adds significant complexity to the
1103 record layer, and its removal is unlikely to cause inter-operability
1108 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1109 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1110 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1115 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1116 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1118 *Richard Levitte, Andy Polyakov*
1120 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1125 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1126 platform rather than 'mingw'.
1130 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1131 success if they are asked to add an object which already exists
1132 in the store. This change cascades to other functions which load
1133 certificates and CRLs.
1137 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1138 facilitate stack unwinding even from assembly subroutines.
1142 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1143 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1147 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1148 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1149 which is the minimum version we support.
1153 * Certificate time validation (X509_cmp_time) enforces stricter
1154 compliance with RFC 5280. Fractional seconds and timezone offsets
1155 are no longer allowed.
1159 * Add support for ARIA
1163 * s_client will now send the Server Name Indication (SNI) extension by
1164 default unless the new "-noservername" option is used. The server name is
1165 based on the host provided to the "-connect" option unless overridden by
1166 using "-servername".
1170 * Add support for SipHash
1174 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1175 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1176 prevent issues where no progress is being made and the peer continually
1177 sends unrecognised record types, using up resources processing them.
1181 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1182 using the algorithm defined in
1183 https://www.akkadia.org/drepper/SHA-crypt.txt
1187 * Heartbeat support has been removed; the ABI is changed for now.
1189 *Richard Levitte, Rich Salz*
1191 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1195 * The RSA "null" method, which was partially supported to avoid patent
1196 issues, has been replaced to always returns NULL.
1201 ### Changes between 1.1.0h and 1.1.0i [xx XXX xxxx] ###
1203 * Client DoS due to large DH parameter
1205 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1206 malicious server can send a very large prime value to the client. This will
1207 cause the client to spend an unreasonably long period of time generating a
1208 key for this prime resulting in a hang until the client has finished. This
1209 could be exploited in a Denial Of Service attack.
1211 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1216 * Cache timing vulnerability in RSA Key Generation
1218 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1219 a cache timing side channel attack. An attacker with sufficient access to
1220 mount cache timing attacks during the RSA key generation process could
1221 recover the private key.
1223 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1224 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1229 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1230 parameter is no longer accepted, as it leads to a corrupt table. NULL
1231 pem_str is reserved for alias entries only.
1235 * Revert blinding in ECDSA sign and instead make problematic addition
1236 length-invariant. Switch even to fixed-length Montgomery multiplication.
1240 * Change generating and checking of primes so that the error rate of not
1241 being prime depends on the intended use based on the size of the input.
1242 For larger primes this will result in more rounds of Miller-Rabin.
1243 The maximal error rate for primes with more than 1080 bits is lowered
1246 *Kurt Roeckx, Annie Yousar*
1248 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1252 * Add blinding to ECDSA and DSA signatures to protect against side channel
1253 attacks discovered by Keegan Ryan (NCC Group).
1257 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1258 now allow empty (zero character) pass phrases.
1262 * Certificate time validation (X509_cmp_time) enforces stricter
1263 compliance with RFC 5280. Fractional seconds and timezone offsets
1264 are no longer allowed.
1268 * Fixed a text canonicalisation bug in CMS
1270 Where a CMS detached signature is used with text content the text goes
1271 through a canonicalisation process first prior to signing or verifying a
1272 signature. This process strips trailing space at the end of lines, converts
1273 line terminators to CRLF and removes additional trailing line terminators
1274 at the end of a file. A bug in the canonicalisation process meant that
1275 some characters, such as form-feed, were incorrectly treated as whitespace
1276 and removed. This is contrary to the specification (RFC5485). This fix
1277 could mean that detached text data signed with an earlier version of
1278 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1279 signed with a fixed OpenSSL may fail to verify with an earlier version of
1280 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1281 and use the "-binary" flag (for the "cms" command line application) or set
1282 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1286 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018] ###
1288 * Constructed ASN.1 types with a recursive definition could exceed the stack
1290 Constructed ASN.1 types with a recursive definition (such as can be found
1291 in PKCS7) could eventually exceed the stack given malicious input with
1292 excessive recursion. This could result in a Denial Of Service attack. There
1293 are no such structures used within SSL/TLS that come from untrusted sources
1294 so this is considered safe.
1296 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1302 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1304 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1305 effectively reduced to only comparing the least significant bit of each
1306 byte. This allows an attacker to forge messages that would be considered as
1307 authenticated in an amount of tries lower than that guaranteed by the
1308 security claims of the scheme. The module can only be compiled by the
1309 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
1311 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1317 * Add a build target 'build_all_generated', to build all generated files
1318 and only that. This can be used to prepare everything that requires
1319 things like perl for a system that lacks perl and then move everything
1320 to that system and do the rest of the build there.
1324 * Backport SSL_OP_NO_RENGOTIATION
1326 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1327 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1328 changes this is no longer possible in 1.1.0. Therefore the new
1329 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1330 1.1.0 to provide equivalent functionality.
1332 Note that if an application built against 1.1.0h headers (or above) is run
1333 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1334 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1338 * Removed the OS390-Unix config target. It relied on a script that doesn't
1343 * rsaz_1024_mul_avx2 overflow bug on x86_64
1345 There is an overflow bug in the AVX2 Montgomery multiplication procedure
1346 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
1347 Analysis suggests that attacks against RSA and DSA as a result of this
1348 defect would be very difficult to perform and are not believed likely.
1349 Attacks against DH1024 are considered just feasible, because most of the
1350 work necessary to deduce information about a private key may be performed
1351 offline. The amount of resources required for such an attack would be
1352 significant. However, for an attack on TLS to be meaningful, the server
1353 would have to share the DH1024 private key among multiple clients, which is
1354 no longer an option since CVE-2016-0701.
1356 This only affects processors that support the AVX2 but not ADX extensions
1357 like Intel Haswell (4th generation).
1359 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1360 was originally found via the OSS-Fuzz project.
1365 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017] ###
1367 * bn_sqrx8x_internal carry bug on x86_64
1369 There is a carry propagating bug in the x86_64 Montgomery squaring
1370 procedure. No EC algorithms are affected. Analysis suggests that attacks
1371 against RSA and DSA as a result of this defect would be very difficult to
1372 perform and are not believed likely. Attacks against DH are considered just
1373 feasible (although very difficult) because most of the work necessary to
1374 deduce information about a private key may be performed offline. The amount
1375 of resources required for such an attack would be very significant and
1376 likely only accessible to a limited number of attackers. An attacker would
1377 additionally need online access to an unpatched system using the target
1378 private key in a scenario with persistent DH parameters and a private
1379 key that is shared between multiple clients.
1381 This only affects processors that support the BMI1, BMI2 and ADX extensions
1382 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1384 This issue was reported to OpenSSL by the OSS-Fuzz project.
1389 * Malformed X.509 IPAddressFamily could cause OOB read
1391 If an X.509 certificate has a malformed IPAddressFamily extension,
1392 OpenSSL could do a one-byte buffer overread. The most likely result
1393 would be an erroneous display of the certificate in text format.
1395 This issue was reported to OpenSSL by the OSS-Fuzz project.
1400 ### Changes between 1.1.0e and 1.1.0f [25 May 2017] ###
1402 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1403 platform rather than 'mingw'.
1407 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1408 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1409 which is the minimum version we support.
1413 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017] ###
1415 * Encrypt-Then-Mac renegotiation crash
1417 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1418 negotiated where it was not in the original handshake (or vice-versa) then
1419 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1420 and servers are affected.
1422 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1427 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017] ###
1429 * Truncated packet could crash via OOB read
1431 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1432 cipher is being used, then a truncated packet can cause that host to
1433 perform an out-of-bounds read, usually resulting in a crash.
1435 This issue was reported to OpenSSL by Robert Święcki of Google.
1440 * Bad (EC)DHE parameters cause a client crash
1442 If a malicious server supplies bad parameters for a DHE or ECDHE key
1443 exchange then this can result in the client attempting to dereference a
1444 NULL pointer leading to a client crash. This could be exploited in a Denial
1447 This issue was reported to OpenSSL by Guido Vranken.
1452 * BN_mod_exp may produce incorrect results on x86_64
1454 There is a carry propagating bug in the x86_64 Montgomery squaring
1455 procedure. No EC algorithms are affected. Analysis suggests that attacks
1456 against RSA and DSA as a result of this defect would be very difficult to
1457 perform and are not believed likely. Attacks against DH are considered just
1458 feasible (although very difficult) because most of the work necessary to
1459 deduce information about a private key may be performed offline. The amount
1460 of resources required for such an attack would be very significant and
1461 likely only accessible to a limited number of attackers. An attacker would
1462 additionally need online access to an unpatched system using the target
1463 private key in a scenario with persistent DH parameters and a private
1464 key that is shared between multiple clients. For example this can occur by
1465 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1466 similar to CVE-2015-3193 but must be treated as a separate problem.
1468 This issue was reported to OpenSSL by the OSS-Fuzz project.
1473 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016] ###
1475 * ChaCha20/Poly1305 heap-buffer-overflow
1477 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1478 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1479 crash. This issue is not considered to be exploitable beyond a DoS.
1481 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1486 * CMS Null dereference
1488 Applications parsing invalid CMS structures can crash with a NULL pointer
1489 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1490 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1491 structure callback if an attempt is made to free certain invalid encodings.
1492 Only CHOICE structures using a callback which do not handle NULL value are
1495 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1500 * Montgomery multiplication may produce incorrect results
1502 There is a carry propagating bug in the Broadwell-specific Montgomery
1503 multiplication procedure that handles input lengths divisible by, but
1504 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1505 and DH private keys are impossible. This is because the subroutine in
1506 question is not used in operations with the private key itself and an input
1507 of the attacker's direct choice. Otherwise the bug can manifest itself as
1508 transient authentication and key negotiation failures or reproducible
1509 erroneous outcome of public-key operations with specially crafted input.
1510 Among EC algorithms only Brainpool P-512 curves are affected and one
1511 presumably can attack ECDH key negotiation. Impact was not analyzed in
1512 detail, because pre-requisites for attack are considered unlikely. Namely
1513 multiple clients have to choose the curve in question and the server has to
1514 share the private key among them, neither of which is default behaviour.
1515 Even then only clients that chose the curve will be affected.
1517 This issue was publicly reported as transient failures and was not
1518 initially recognized as a security issue. Thanks to Richard Morgan for
1519 providing reproducible case.
1524 * Removed automatic addition of RPATH in shared libraries and executables,
1525 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1529 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016] ###
1531 * Fix Use After Free for large message sizes
1533 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1534 message larger than approx 16k is received then the underlying buffer to
1535 store the incoming message is reallocated and moved. Unfortunately a
1536 dangling pointer to the old location is left which results in an attempt to
1537 write to the previously freed location. This is likely to result in a
1538 crash, however it could potentially lead to execution of arbitrary code.
1540 This issue only affects OpenSSL 1.1.0a.
1542 This issue was reported to OpenSSL by Robert Święcki.
1547 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016] ###
1549 * OCSP Status Request extension unbounded memory growth
1551 A malicious client can send an excessively large OCSP Status Request
1552 extension. If that client continually requests renegotiation, sending a
1553 large OCSP Status Request extension each time, then there will be unbounded
1554 memory growth on the server. This will eventually lead to a Denial Of
1555 Service attack through memory exhaustion. Servers with a default
1556 configuration are vulnerable even if they do not support OCSP. Builds using
1557 the "no-ocsp" build time option are not affected.
1559 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1564 * SSL_peek() hang on empty record
1566 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1567 sends an empty record. This could be exploited by a malicious peer in a
1568 Denial Of Service attack.
1570 This issue was reported to OpenSSL by Alex Gaynor.
1575 * Excessive allocation of memory in tls_get_message_header() and
1576 dtls1_preprocess_fragment()
1578 A (D)TLS message includes 3 bytes for its length in the header for the
1579 message. This would allow for messages up to 16Mb in length. Messages of
1580 this length are excessive and OpenSSL includes a check to ensure that a
1581 peer is sending reasonably sized messages in order to avoid too much memory
1582 being consumed to service a connection. A flaw in the logic of version
1583 1.1.0 means that memory for the message is allocated too early, prior to
1584 the excessive message length check. Due to way memory is allocated in
1585 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1586 to service a connection. This could lead to a Denial of Service through
1587 memory exhaustion. However, the excessive message length check still takes
1588 place, and this would cause the connection to immediately fail. Assuming
1589 that the application calls SSL_free() on the failed connection in a timely
1590 manner then the 21Mb of allocated memory will then be immediately freed
1591 again. Therefore the excessive memory allocation will be transitory in
1592 nature. This then means that there is only a security impact if:
1594 1) The application does not call SSL_free() in a timely manner in the event
1595 that the connection fails
1597 2) The application is working in a constrained environment where there is
1598 very little free memory
1600 3) The attacker initiates multiple connection attempts such that there are
1601 multiple connections in a state where memory has been allocated for the
1602 connection; SSL_free() has not yet been called; and there is insufficient
1603 memory to service the multiple requests.
1605 Except in the instance of (1) above any Denial Of Service is likely to be
1606 transitory because as soon as the connection fails the memory is
1607 subsequently freed again in the SSL_free() call. However there is an
1608 increased risk during this period of application crashes due to the lack of
1609 memory - which would then mean a more serious Denial of Service.
1611 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1612 (CVE-2016-6307 and CVE-2016-6308)
1616 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1617 had to be removed. Primary reason is that vendor assembler can't
1618 assemble our modules with -KPIC flag. As result it, assembly
1619 support, was not even available as option. But its lack means
1620 lack of side-channel resistant code, which is incompatible with
1621 security by todays standards. Fortunately gcc is readily available
1622 prepackaged option, which we firmly point at...
1626 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016] ###
1628 * Windows command-line tool supports UTF-8 opt-in option for arguments
1629 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1630 (to any value) allows Windows user to access PKCS#12 file generated
1631 with Windows CryptoAPI and protected with non-ASCII password, as well
1632 as files generated under UTF-8 locale on Linux also protected with
1637 * To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1638 have been disabled by default and removed from DEFAULT, just like RC4.
1639 See the RC4 item below to re-enable both.
1643 * The method for finding the storage location for the Windows RAND seed file
1644 has changed. First we check %RANDFILE%. If that is not set then we check
1645 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1646 all else fails we fall back to C:\.
1650 * The EVP_EncryptUpdate() function has had its return type changed from void
1651 to int. A return of 0 indicates and error while a return of 1 indicates
1656 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1657 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1658 off the constant time implementation for RSA, DSA and DH have been made
1659 no-ops and deprecated.
1663 * Windows RAND implementation was simplified to only get entropy by
1664 calling CryptGenRandom(). Various other RAND-related tickets
1667 *Joseph Wylie Yandle, Rich Salz*
1669 * The stack and lhash API's were renamed to start with OPENSSL_SK_
1670 and OPENSSL_LH_, respectively. The old names are available
1671 with API compatibility. They new names are now completely documented.
1675 * Unify TYPE_up_ref(obj) methods signature.
1676 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1677 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1678 int (instead of void) like all others TYPE_up_ref() methods.
1679 So now these methods also check the return value of CRYPTO_atomic_add(),
1680 and the validity of object reference counter.
1682 *fdasilvayy@gmail.com*
1684 * With Windows Visual Studio builds, the .pdb files are installed
1685 alongside the installed libraries and executables. For a static
1686 library installation, ossl_static.pdb is the associate compiler
1687 generated .pdb file to be used when linking programs.
1691 * Remove openssl.spec. Packaging files belong with the packagers.
1695 * Automatic Darwin/OSX configuration has had a refresh, it will now
1696 recognise x86_64 architectures automatically. You can still decide
1697 to build for a different bitness with the environment variable
1698 KERNEL_BITS (can be 32 or 64), for example:
1700 KERNEL_BITS=32 ./config
1704 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1705 256 bit AES and HMAC with SHA256.
1709 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1713 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
1717 * To enable users to have their own config files and build file templates,
1718 Configure looks in the directory indicated by the environment variable
1719 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1720 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1721 name and is used as is.
1725 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1726 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1727 X509_CERT_FILE_CTX was removed.
1731 * "shared" builds are now the default. To create only static libraries use
1732 the "no-shared" Configure option.
1736 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1737 All of these option have not worked for some while and are fundamental
1742 * Make various cleanup routines no-ops and mark them as deprecated. Most
1743 global cleanup functions are no longer required because they are handled
1744 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1745 Explicitly de-initing can cause problems (e.g. where a library that uses
1746 OpenSSL de-inits, but an application is still using it). The affected
1747 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1748 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1749 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1750 COMP_zlib_cleanup().
1754 * --strict-warnings no longer enables runtime debugging options
1755 such as REF_DEBUG. Instead, debug options are automatically
1756 enabled with '--debug' builds.
1758 *Andy Polyakov, Emilia Käsper*
1760 * Made DH and DH_METHOD opaque. The structures for managing DH objects
1761 have been moved out of the public header files. New functions for managing
1762 these have been added.
1766 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
1767 objects have been moved out of the public header files. New
1768 functions for managing these have been added.
1772 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1773 have been moved out of the public header files. New functions for managing
1774 these have been added.
1778 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1779 moved out of the public header files. New functions for managing these
1784 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
1788 * Removed the mk1mf build scripts.
1792 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1793 it is always safe to #include a header now.
1797 * Removed the aged BC-32 config and all its supporting scripts
1801 * Removed support for Ultrix, Netware, and OS/2.
1805 * Add support for HKDF.
1807 *Alessandro Ghedini*
1809 * Add support for blake2b and blake2s
1813 * Added support for "pipelining". Ciphers that have the
1814 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1815 encryptions/decryptions simultaneously. There are currently no built-in
1816 ciphers with this property but the expectation is that engines will be able
1817 to offer it to significantly improve throughput. Support has been extended
1818 into libssl so that multiple records for a single connection can be
1819 processed in one go (for >=TLS 1.1).
1823 * Added the AFALG engine. This is an async capable engine which is able to
1824 offload work to the Linux kernel. In this initial version it only supports
1825 AES128-CBC. The kernel must be version 4.1.0 or greater.
1829 * OpenSSL now uses a new threading API. It is no longer necessary to
1830 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1831 are two supported threading models: pthreads and windows threads. It is
1832 also possible to configure OpenSSL at compile time for "no-threads". The
1833 old threading API should no longer be used. The functions have been
1834 replaced with "no-op" compatibility macros.
1836 *Alessandro Ghedini, Matt Caswell*
1838 * Modify behavior of ALPN to invoke callback after SNI/servername
1839 callback, such that updates to the SSL_CTX affect ALPN.
1843 * Add SSL_CIPHER queries for authentication and key-exchange.
1847 * Changes to the DEFAULT cipherlist:
1848 - Prefer (EC)DHE handshakes over plain RSA.
1849 - Prefer AEAD ciphers over legacy ciphers.
1850 - Prefer ECDSA over RSA when both certificates are available.
1851 - Prefer TLSv1.2 ciphers/PRF.
1852 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1857 * Change the ECC default curve list to be this, in order: x25519,
1858 secp256r1, secp521r1, secp384r1.
1862 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1863 disabled by default. They can be re-enabled using the
1864 enable-weak-ssl-ciphers option to Configure.
1868 * If the server has ALPN configured, but supports no protocols that the
1869 client advertises, send a fatal "no_application_protocol" alert.
1870 This behaviour is SHALL in RFC 7301, though it isn't universally
1871 implemented by other servers.
1875 * Add X25519 support.
1876 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1877 for public and private key encoding using the format documented in
1878 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1879 key generation and key derivation.
1881 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1886 * Deprecate SRP_VBASE_get_by_user.
1887 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1888 In order to fix an unavoidable memory leak (CVE-2016-0798),
1889 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1890 seed, even if the seed is configured.
1892 Users should use SRP_VBASE_get1_by_user instead. Note that in
1893 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1894 also that even though configuring the SRP seed attempts to hide
1895 invalid usernames by continuing the handshake with fake
1896 credentials, this behaviour is not constant time and no strong
1897 guarantees are made that the handshake is indistinguishable from
1898 that of a valid user.
1902 * Configuration change; it's now possible to build dynamic engines
1903 without having to build shared libraries and vice versa. This
1904 only applies to the engines in engines/, those in crypto/engine/
1905 will always be built into libcrypto (i.e. "static").
1907 Building dynamic engines is enabled by default; to disable, use
1908 the configuration option "disable-dynamic-engine".
1910 The only requirements for building dynamic engines are the
1911 presence of the DSO module and building with position independent
1912 code, so they will also automatically be disabled if configuring
1913 with "disable-dso" or "disable-pic".
1915 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1916 are also taken away from openssl/opensslconf.h, as they are
1921 * Configuration change; if there is a known flag to compile
1922 position independent code, it will always be applied on the
1923 libcrypto and libssl object files, and never on the application
1924 object files. This means other libraries that use routines from
1925 libcrypto / libssl can be made into shared libraries regardless
1926 of how OpenSSL was configured.
1928 If this isn't desirable, the configuration options "disable-pic"
1929 or "no-pic" can be used to disable the use of PIC. This will
1930 also disable building shared libraries and dynamic engines.
1934 * Removed JPAKE code. It was experimental and has no wide use.
1938 * The INSTALL_PREFIX Makefile variable has been renamed to
1939 DESTDIR. That makes for less confusion on what this variable
1940 is for. Also, the configuration option --install_prefix is
1945 * Heartbeat for TLS has been removed and is disabled by default
1946 for DTLS; configure with enable-heartbeats. Code that uses the
1947 old #define's might need to be updated.
1949 *Emilia Käsper, Rich Salz*
1951 * Rename REF_CHECK to REF_DEBUG.
1955 * New "unified" build system
1957 The "unified" build system is aimed to be a common system for all
1958 platforms we support. With it comes new support for VMS.
1960 This system builds supports building in a different directory tree
1961 than the source tree. It produces one Makefile (for unix family
1962 or lookalikes), or one descrip.mms (for VMS).
1964 The source of information to make the Makefile / descrip.mms is
1965 small files called 'build.info', holding the necessary
1966 information for each directory with source to compile, and a
1967 template in Configurations, like unix-Makefile.tmpl or
1970 With this change, the library names were also renamed on Windows
1971 and on VMS. They now have names that are closer to the standard
1972 on Unix, and include the major version number, and in certain
1973 cases, the architecture they are built for. See "Notes on shared
1974 libraries" in INSTALL.
1976 We rely heavily on the perl module Text::Template.
1980 * Added support for auto-initialisation and de-initialisation of the library.
1981 OpenSSL no longer requires explicit init or deinit routines to be called,
1982 except in certain circumstances. See the OPENSSL_init_crypto() and
1983 OPENSSL_init_ssl() man pages for further information.
1987 * The arguments to the DTLSv1_listen function have changed. Specifically the
1988 "peer" argument is now expected to be a BIO_ADDR object.
1990 * Rewrite of BIO networking library. The BIO library lacked consistent
1991 support of IPv6, and adding it required some more extensive
1992 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1993 which hold all types of addresses and chains of address information.
1994 It also introduces a new API, with functions like BIO_socket,
1995 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1996 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1997 have been adapted accordingly.
2001 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2006 * CRIME protection: disable compression by default, even if OpenSSL is
2007 compiled with zlib enabled. Applications can still enable compression
2008 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2009 using the SSL_CONF library to configure compression.
2013 * The signature of the session callback configured with
2014 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2015 was explicitly marked as 'const unsigned char*' instead of
2020 * Always DPURIFY. Remove the use of uninitialized memory in the
2021 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2025 * Removed many obsolete configuration items, including
2026 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2027 MD2_CHAR, MD2_INT, MD2_LONG
2029 IDEA_SHORT, IDEA_LONG
2030 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2032 *Rich Salz, with advice from Andy Polyakov*
2034 * Many BN internals have been moved to an internal header file.
2036 *Rich Salz with help from Andy Polyakov*
2038 * Configuration and writing out the results from it has changed.
2039 Files such as Makefile include/openssl/opensslconf.h and are now
2040 produced through general templates, such as Makefile.in and
2041 crypto/opensslconf.h.in and some help from the perl module
2044 Also, the center of configuration information is no longer
2045 Makefile. Instead, Configure produces a perl module in
2046 configdata.pm which holds most of the config data (in the hash
2047 table %config), the target data that comes from the target
2048 configuration in one of the Configurations/*.conf files (in
2053 * To clarify their intended purposes, the Configure options
2054 --prefix and --openssldir change their semantics, and become more
2055 straightforward and less interdependent.
2057 --prefix shall be used exclusively to give the location INSTALLTOP
2058 where programs, scripts, libraries, include files and manuals are
2059 going to be installed. The default is now /usr/local.
2061 --openssldir shall be used exclusively to give the default
2062 location OPENSSLDIR where certificates, private keys, CRLs are
2063 managed. This is also where the default openssl.cnf gets
2065 If the directory given with this option is a relative path, the
2066 values of both the --prefix value and the --openssldir value will
2067 be combined to become OPENSSLDIR.
2068 The default for --openssldir is INSTALLTOP/ssl.
2070 Anyone who uses --openssldir to specify where OpenSSL is to be
2071 installed MUST change to use --prefix instead.
2075 * The GOST engine was out of date and therefore it has been removed. An up
2076 to date GOST engine is now being maintained in an external repository.
2077 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
2078 support for GOST ciphersuites (these are only activated if a GOST engine
2083 * EGD is no longer supported by default; use enable-egd when
2086 *Ben Kaduk and Rich Salz*
2088 * The distribution now has Makefile.in files, which are used to
2089 create Makefile's when Configure is run. *Configure must be run
2090 before trying to build now.*
2094 * The return value for SSL_CIPHER_description() for error conditions
2099 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2101 Obtaining and performing DNSSEC validation of TLSA records is
2102 the application's responsibility. The application provides
2103 the TLSA records of its choice to OpenSSL, and these are then
2104 used to authenticate the peer.
2106 The TLSA records need not even come from DNS. They can, for
2107 example, be used to implement local end-entity certificate or
2108 trust-anchor "pinning", where the "pin" data takes the form
2109 of TLSA records, which can augment or replace verification
2110 based on the usual WebPKI public certification authorities.
2114 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2115 continues to support deprecated interfaces in default builds.
2116 However, applications are strongly advised to compile their
2117 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2118 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2119 or the 1.1.0 releases.
2121 In environments in which all applications have been ported to
2122 not use any deprecated interfaces OpenSSL's Configure script
2123 should be used with the --api=1.1.0 option to entirely remove
2124 support for the deprecated features from the library and
2125 unconditionally disable them in the installed headers.
2126 Essentially the same effect can be achieved with the "no-deprecated"
2127 argument to Configure, except that this will always restrict
2128 the build to just the latest API, rather than a fixed API
2131 As applications are ported to future revisions of the API,
2132 they should update their compile-time OPENSSL_API_COMPAT define
2133 accordingly, but in most cases should be able to continue to
2134 compile with later releases.
2136 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2137 0x10000000L and 0x00908000L, respectively. However those
2138 versions did not support the OPENSSL_API_COMPAT feature, and
2139 so applications are not typically tested for explicit support
2140 of just the undeprecated features of either release.
2144 * Add support for setting the minimum and maximum supported protocol.
2145 It can bet set via the SSL_set_min_proto_version() and
2146 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2147 MaxProtocol. It's recommended to use the new APIs to disable
2148 protocols instead of disabling individual protocols using
2149 SSL_set_options() or SSL_CONF's Protocol. This change also
2150 removes support for disabling TLS 1.2 in the OpenSSL TLS
2151 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2155 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2159 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2160 and integrates ECDSA and ECDH functionality into EC. Implementations can
2161 now redirect key generation and no longer need to convert to or from
2164 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2165 include the ec.h header file instead.
2169 * Remove support for all 40 and 56 bit ciphers. This includes all the export
2170 ciphers who are no longer supported and drops support the ephemeral RSA key
2171 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2175 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2176 opaque. For HMAC_CTX, the following constructors and destructors
2179 HMAC_CTX *HMAC_CTX_new(void);
2180 void HMAC_CTX_free(HMAC_CTX *ctx);
2182 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2183 destroy such methods has been added. See EVP_MD_meth_new(3) and
2184 EVP_CIPHER_meth_new(3) for documentation.
2187 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2188 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2189 EVP_MD_CTX_reset() should be called instead to reinitialise
2190 an already created structure.
2191 2) For consistency with the majority of our object creators and
2192 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2193 EVP_MD_CTX_(new|free). The old names are retained as macros
2194 for deprecated builds.
2198 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
2199 cryptographic operations to be performed asynchronously as long as an
2200 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2201 further details. Libssl has also had this capability integrated with the
2202 introduction of the new mode SSL_MODE_ASYNC and associated error
2203 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2204 pages. This work was developed in partnership with Intel Corp.
2208 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2209 always enabled now. If you want to disable the support you should
2210 exclude it using the list of supported ciphers. This also means that the
2211 "-no_ecdhe" option has been removed from s_server.
2215 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2216 SSL_{CTX_}set1_curves() which can set a list.
2220 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2221 curve you want to support using SSL_{CTX_}set1_curves().
2225 * State machine rewrite. The state machine code has been significantly
2226 refactored in order to remove much duplication of code and solve issues
2227 with the old code (see ssl/statem/README for further details). This change
2228 does have some associated API changes. Notably the SSL_state() function
2229 has been removed and replaced by SSL_get_state which now returns an
2230 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2231 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2236 * All instances of the string "ssleay" in the public API were replaced
2237 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2238 Some error codes related to internal RSA_eay API's were renamed.
2242 * The demo files in crypto/threads were moved to demo/threads.
2246 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2249 *Matt Caswell, Rich Salz*
2251 * New ASN.1 embed macro.
2253 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2254 structure is not allocated: it is part of the parent. That is instead of
2262 This reduces memory fragmentation and make it impossible to accidentally
2263 set a mandatory field to NULL.
2265 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2266 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2267 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
2272 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
2276 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
2277 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
2278 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
2279 DES and RC4 ciphersuites.
2283 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2284 This changes the decoding behaviour for some invalid messages,
2285 though the change is mostly in the more lenient direction, and
2286 legacy behaviour is preserved as much as possible.
2290 * Fix no-stdio build.
2291 * David Woodhouse <David.Woodhouse@intel.com> and also
2292 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
2294 * New testing framework
2295 The testing framework has been largely rewritten and is now using
2296 perl and the perl modules Test::Harness and an extended variant of
2297 Test::More called OpenSSL::Test to do its work. All test scripts in
2298 test/ have been rewritten into test recipes, and all direct calls to
2299 executables in test/Makefile have become individual recipes using the
2300 simplified testing OpenSSL::Test::Simple.
2302 For documentation on our testing modules, do:
2304 perldoc test/testlib/OpenSSL/Test/Simple.pm
2305 perldoc test/testlib/OpenSSL/Test.pm
2309 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
2310 are used; the latter aborts on memory leaks (usually checked on exit).
2311 Some undocumented "set malloc, etc., hooks" functions were removed
2312 and others were changed. All are now documented.
2316 * In DSA_generate_parameters_ex, if the provided seed is too short,
2319 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
2321 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
2322 from RFC4279, RFC4785, RFC5487, RFC5489.
2324 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
2325 original RSA_PSK patch.
2329 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
2330 era flag was never set throughout the codebase (only read). Also removed
2331 SSL3_FLAGS_POP_BUFFER which was only used if
2332 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
2336 * Changed the default name options in the "ca", "crl", "req" and "x509"
2337 to be "oneline" instead of "compat".
2341 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
2342 not aware of clients that still exhibit this bug, and the workaround
2343 hasn't been working properly for a while.
2347 * The return type of BIO_number_read() and BIO_number_written() as well as
2348 the corresponding num_read and num_write members in the BIO structure has
2349 changed from unsigned long to uint64_t. On platforms where an unsigned
2350 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
2355 * Given the pervasive nature of TLS extensions it is inadvisable to run
2356 OpenSSL without support for them. It also means that maintaining
2357 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
2358 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
2362 * Removed support for the two export grade static DH ciphersuites
2363 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
2364 were newly added (along with a number of other static DH ciphersuites) to
2365 1.0.2. However the two export ones have *never* worked since they were
2366 introduced. It seems strange in any case to be adding new export
2367 ciphersuites, and given "logjam" it also does not seem correct to fix them.
2371 * Version negotiation has been rewritten. In particular SSLv23_method(),
2372 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
2373 and turned into macros which simply call the new preferred function names
2374 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
2375 should use the new names instead. Also as part of this change the ssl23.h
2376 header file has been removed.
2380 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
2381 code and the associated standard is no longer considered fit-for-purpose.
2385 * RT2547 was closed. When generating a private key, try to make the
2386 output file readable only by the owner. This behavior change might
2387 be noticeable when interacting with other software.
2389 * Documented all exdata functions. Added CRYPTO_free_ex_index.
2394 * Added HTTP GET support to the ocsp command.
2398 * Changed default digest for the dgst and enc commands from MD5 to
2403 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
2407 * Added support for TLS extended master secret from
2408 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
2409 initial patch which was a great help during development.
2413 * All libssl internal structures have been removed from the public header
2414 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
2415 now redundant). Users should not attempt to access internal structures
2416 directly. Instead they should use the provided API functions.
2420 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
2421 Access to deprecated functions can be re-enabled by running config with
2422 "enable-deprecated". In addition applications wishing to use deprecated
2423 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
2424 will, by default, disable some transitive includes that previously existed
2425 in the header files (e.g. ec.h will no longer, by default, include bn.h)
2429 * Added support for OCB mode. OpenSSL has been granted a patent license
2430 compatible with the OpenSSL license for use of OCB. Details are available
2431 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
2432 for OCB can be removed by calling config with no-ocb.
2436 * SSLv2 support has been removed. It still supports receiving a SSLv2
2437 compatible client hello.
2441 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
2442 done while fixing the error code for the key-too-small case.
2444 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
2446 * CA.sh has been removed; use CA.pl instead.
2450 * Removed old DES API.
2454 * Remove various unsupported platforms:
2460 Sinix/ReliantUNIX RM400
2465 16-bit platforms such as WIN16
2469 * Clean up OPENSSL_NO_xxx #define's
2470 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
2471 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
2472 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
2473 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
2474 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
2475 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
2476 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
2477 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
2478 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
2479 - Remove MS_STATIC; it's a relic from platforms <32 bits.
2483 * Cleaned up dead code
2484 Remove all but one '#ifdef undef' which is to be looked at.
2488 * Clean up calling of xxx_free routines.
2489 Just like free(), fix most of the xxx_free routines to accept
2490 NULL. Remove the non-null checks from callers. Save much code.
2494 * Add secure heap for storage of private keys (when possible).
2495 Add BIO_s_secmem(), CBIGNUM, etc.
2496 Contributed by Akamai Technologies under our Corporate CLA.
2500 * Experimental support for a new, fast, unbiased prime candidate generator,
2501 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2503 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
2505 * New output format NSS in the sess_id command line tool. This allows
2506 exporting the session id and the master key in NSS keylog format.
2508 *Martin Kaiser <martin@kaiser.cx>*
2510 * Harmonize version and its documentation. -f flag is used to display
2513 *mancha <mancha1@zoho.com>*
2515 * Fix eckey_priv_encode so it immediately returns an error upon a failure
2516 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2518 *mancha <mancha1@zoho.com>*
2520 * Fix some double frees. These are not thought to be exploitable.
2522 *mancha <mancha1@zoho.com>*
2524 * A missing bounds check in the handling of the TLS heartbeat extension
2525 can be used to reveal up to 64k of memory to a connected client or
2528 Thanks for Neel Mehta of Google Security for discovering this bug and to
2529 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2530 preparing the fix (CVE-2014-0160)
2532 *Adam Langley, Bodo Moeller*
2534 * Fix for the attack described in the paper "Recovering OpenSSL
2535 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2536 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2537 http://eprint.iacr.org/2014/140
2539 Thanks to Yuval Yarom and Naomi Benger for discovering this
2540 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2542 *Yuval Yarom and Naomi Benger*
2544 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2545 this fixes a limitation in previous versions of OpenSSL.
2549 * Experimental encrypt-then-mac support.
2551 Experimental support for encrypt then mac from
2552 draft-gutmann-tls-encrypt-then-mac-02.txt
2554 To enable it set the appropriate extension number (0x42 for the test
2555 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2557 For non-compliant peers (i.e. just about everything) this should have no
2560 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2565 * Add EVP support for key wrapping algorithms, to avoid problems with
2566 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2567 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2568 algorithms and include tests cases.
2572 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2577 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2578 MGF1 digest and OAEP label.
2582 * Make openssl verify return errors.
2584 *Chris Palmer <palmer@google.com> and Ben Laurie*
2586 * New function ASN1_TIME_diff to calculate the difference between two
2587 ASN1_TIME structures or one structure and the current time.
2591 * Update fips_test_suite to support multiple command line options. New
2592 test to induce all self test errors in sequence and check expected
2597 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2598 sign or verify all in one operation.
2602 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
2603 test programs and fips_test_suite. Includes functionality to parse
2604 the minimal script output of fipsalgest.pl directly.
2608 * Add authorisation parameter to FIPS_module_mode_set().
2612 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2616 * Use separate DRBG fields for internal and external flags. New function
2617 FIPS_drbg_health_check() to perform on demand health checking. Add
2618 generation tests to fips_test_suite with reduced health check interval to
2619 demonstrate periodic health checking. Add "nodh" option to
2620 fips_test_suite to skip very slow DH test.
2624 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2629 * More extensive health check for DRBG checking many more failure modes.
2630 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2631 combination: call this in fips_test_suite.
2635 * Add support for canonical generation of DSA parameter 'g'. See
2638 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2639 POST to handle HMAC cases.
2643 * Add functions FIPS_module_version() and FIPS_module_version_text()
2644 to return numerical and string versions of the FIPS module number.
2648 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2649 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2650 outside the validated module in the FIPS capable OpenSSL.
2654 * Minor change to DRBG entropy callback semantics. In some cases
2655 there is no multiple of the block length between min_len and
2656 max_len. Allow the callback to return more than max_len bytes
2657 of entropy but discard any extra: it is the callback's responsibility
2658 to ensure that the extra data discarded does not impact the
2659 requested amount of entropy.
2663 * Add PRNG security strength checks to RSA, DSA and ECDSA using
2664 information in FIPS186-3, SP800-57 and SP800-131A.
2668 * CCM support via EVP. Interface is very similar to GCM case except we
2669 must supply all data in one chunk (i.e. no update, final) and the
2670 message length must be supplied if AAD is used. Add algorithm test
2675 * Initial version of POST overhaul. Add POST callback to allow the status
2676 of POST to be monitored and/or failures induced. Modify fips_test_suite
2677 to use callback. Always run all selftests even if one fails.
2681 * XTS support including algorithm test driver in the fips_gcmtest program.
2682 Note: this does increase the maximum key length from 32 to 64 bytes but
2683 there should be no binary compatibility issues as existing applications
2684 will never use XTS mode.
2688 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2689 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2690 performs algorithm blocking for unapproved PRNG types. Also do not
2691 set PRNG type in FIPS_mode_set(): leave this to the application.
2692 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2693 the standard OpenSSL PRNG: set additional data to a date time vector.
2697 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2698 This shouldn't present any incompatibility problems because applications
2699 shouldn't be using these directly and any that are will need to rethink
2700 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2704 * Extensive self tests and health checking required by SP800-90 DRBG.
2705 Remove strength parameter from FIPS_drbg_instantiate and always
2706 instantiate at maximum supported strength.
2710 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2714 * New algorithm test program fips_dhvs to handle DH primitives only testing.
2718 * New function DH_compute_key_padded() to compute a DH key and pad with
2719 leading zeroes if needed: this complies with SP800-56A et al.
2723 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2724 anything, incomplete, subject to change and largely untested at present.
2728 * Modify fipscanisteronly build option to only build the necessary object
2729 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2733 * Add experimental option FIPSSYMS to give all symbols in
2734 fipscanister.o and FIPS or fips prefix. This will avoid
2735 conflicts with future versions of OpenSSL. Add perl script
2736 util/fipsas.pl to preprocess assembly language source files
2737 and rename any affected symbols.
2741 * Add selftest checks and algorithm block of non-fips algorithms in
2742 FIPS mode. Remove DES2 from selftests.
2746 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2747 return internal method without any ENGINE dependencies. Add new
2748 tiny fips sign and verify functions.
2752 * New build option no-ec2m to disable characteristic 2 code.
2756 * New build option "fipscanisteronly". This only builds fipscanister.o
2757 and (currently) associated fips utilities. Uses the file Makefile.fips
2758 instead of Makefile.org as the prototype.
2762 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
2763 Update fips_gcmtest to use IV generator.
2767 * Initial, experimental EVP support for AES-GCM. AAD can be input by
2768 setting output buffer to NULL. The *Final function must be
2769 called although it will not retrieve any additional data. The tag
2770 can be set or retrieved with a ctrl. The IV length is by default 12
2771 bytes (96 bits) but can be set to an alternative value. If the IV
2772 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2777 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2778 underlying do_cipher function handles all cipher semantics itself
2779 including padding and finalisation. This is useful if (for example)
2780 an ENGINE cipher handles block padding itself. The behaviour of
2781 do_cipher is subtly changed if this flag is set: the return value
2782 is the number of characters written to the output buffer (zero is
2783 no longer an error code) or a negative error code. Also if the
2784 input buffer is NULL and length 0 finalisation should be performed.
2788 * If a candidate issuer certificate is already part of the constructed
2789 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2793 * Improve forward-security support: add functions
2795 void SSL_CTX_set_not_resumable_session_callback(
2796 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2797 void SSL_set_not_resumable_session_callback(
2798 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2800 for use by SSL/TLS servers; the callback function will be called whenever a
2801 new session is created, and gets to decide whether the session may be
2802 cached to make it resumable (return 0) or not (return 1). (As by the
2803 SSL/TLS protocol specifications, the session_id sent by the server will be
2804 empty to indicate that the session is not resumable; also, the server will
2805 not generate RFC 4507 (RFC 5077) session tickets.)
2807 A simple reasonable callback implementation is to return is_forward_secure.
2808 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2809 by the SSL/TLS server library, indicating whether it can provide forward
2812 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
2814 * New -verify_name option in command line utilities to set verification
2819 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2820 Add CMAC pkey methods.
2824 * Experimental renegotiation in s_server -www mode. If the client
2825 browses /reneg connection is renegotiated. If /renegcert it is
2826 renegotiated requesting a certificate.
2830 * Add an "external" session cache for debugging purposes to s_server. This
2831 should help trace issues which normally are only apparent in deployed
2832 multi-process servers.
2836 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2837 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2838 BIO_set_cipher() and some obscure PEM functions were changed so they
2839 can now return an error. The RAND changes required a change to the
2840 RAND_METHOD structure.
2844 * New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2845 a gcc attribute to warn if the result of a function is ignored. This
2846 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2847 whose return value is often ignored.
2851 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2852 These allow SCTs (signed certificate timestamps) to be requested and
2853 validated when establishing a connection.
2855 *Rob Percival <robpercival@google.com>*
2857 ### Changes between 1.0.2g and 1.0.2h [3 May 2016] ###
2859 * Prevent padding oracle in AES-NI CBC MAC check
2861 A MITM attacker can use a padding oracle attack to decrypt traffic
2862 when the connection uses an AES CBC cipher and the server support
2865 This issue was introduced as part of the fix for Lucky 13 padding
2866 attack (CVE-2013-0169). The padding check was rewritten to be in
2867 constant time by making sure that always the same bytes are read and
2868 compared against either the MAC or padding bytes. But it no longer
2869 checked that there was enough data to have both the MAC and padding
2872 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2877 * Fix EVP_EncodeUpdate overflow
2879 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2880 Base64 encoding of binary data. If an attacker is able to supply very large
2881 amounts of input data then a length check can overflow resulting in a heap
2884 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2885 the PEM_write_bio* family of functions. These are mainly used within the
2886 OpenSSL command line applications, so any application which processes data
2887 from an untrusted source and outputs it as a PEM file should be considered
2888 vulnerable to this issue. User applications that call these APIs directly
2889 with large amounts of untrusted data may also be vulnerable.
2891 This issue was reported by Guido Vranken.
2896 * Fix EVP_EncryptUpdate overflow
2898 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2899 is able to supply very large amounts of input data after a previous call to
2900 EVP_EncryptUpdate() with a partial block then a length check can overflow
2901 resulting in a heap corruption. Following an analysis of all OpenSSL
2902 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2903 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2904 the first called function after an EVP_EncryptInit(), and therefore that
2905 specific call must be safe. The second form is where the length passed to
2906 EVP_EncryptUpdate() can be seen from the code to be some small value and
2907 therefore there is no possibility of an overflow. Since all instances are
2908 one of these two forms, it is believed that there can be no overflows in
2909 internal code due to this problem. It should be noted that
2910 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2911 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2912 of these calls have also been analysed too and it is believed there are no
2913 instances in internal usage where an overflow could occur.
2915 This issue was reported by Guido Vranken.
2920 * Prevent ASN.1 BIO excessive memory allocation
2922 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2923 a short invalid encoding can cause allocation of large amounts of memory
2924 potentially consuming excessive resources or exhausting memory.
2926 Any application parsing untrusted data through d2i BIO functions is
2927 affected. The memory based functions such as d2i_X509() are *not* affected.
2928 Since the memory based functions are used by the TLS library, TLS
2929 applications are not affected.
2931 This issue was reported by Brian Carpenter.
2938 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2939 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2940 in arbitrary stack data being returned in the buffer.
2942 This issue was reported by Guido Vranken.
2947 * Modify behavior of ALPN to invoke callback after SNI/servername
2948 callback, such that updates to the SSL_CTX affect ALPN.
2952 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2957 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
2958 methods are enabled and ssl2 is disabled the methods return NULL.
2962 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016] ###
2964 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2965 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2966 provide any "EXPORT" or "LOW" strength ciphers.
2970 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2971 is by default disabled at build-time. Builds that are not configured with
2972 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2973 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2974 will need to explicitly call either of:
2976 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2978 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2980 as appropriate. Even if either of those is used, or the application
2981 explicitly uses the version-specific SSLv2_method() or its client and
2982 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2983 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2984 ciphers, and SSLv2 56-bit DES are no longer available.
2989 * Fix a double-free in DSA code
2991 A double free bug was discovered when OpenSSL parses malformed DSA private
2992 keys and could lead to a DoS attack or memory corruption for applications
2993 that receive DSA private keys from untrusted sources. This scenario is
2996 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
3002 * Disable SRP fake user seed to address a server memory leak.
3004 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
3006 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3007 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
3008 was changed to ignore the "fake user" SRP seed, even if the seed
3011 Users should use SRP_VBASE_get1_by_user instead. Note that in
3012 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3013 also that even though configuring the SRP seed attempts to hide
3014 invalid usernames by continuing the handshake with fake
3015 credentials, this behaviour is not constant time and no strong
3016 guarantees are made that the handshake is indistinguishable from
3017 that of a valid user.
3022 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
3024 In the BN_hex2bn function the number of hex digits is calculated using an
3025 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
3026 large values of |i| this can result in |bn_expand| not allocating any
3027 memory because |i * 4| is negative. This can leave the internal BIGNUM data
3028 field as NULL leading to a subsequent NULL ptr deref. For very large values
3029 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
3030 In this case memory is allocated to the internal BIGNUM data field, but it
3031 is insufficiently sized leading to heap corruption. A similar issue exists
3032 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
3033 is ever called by user applications with very large untrusted hex/dec data.
3034 This is anticipated to be a rare occurrence.
3036 All OpenSSL internal usage of these functions use data that is not expected
3037 to be untrusted, e.g. config file data or application command line
3038 arguments. If user developed applications generate config file data based
3039 on untrusted data then it is possible that this could also lead to security
3040 consequences. This is also anticipated to be rare.
3042 This issue was reported to OpenSSL by Guido Vranken.
3047 * Fix memory issues in BIO_*printf functions
3049 The internal |fmtstr| function used in processing a "%s" format string in
3050 the BIO_*printf functions could overflow while calculating the length of a
3051 string and cause an OOB read when printing very long strings.
3053 Additionally the internal |doapr_outch| function can attempt to write to an
3054 OOB memory location (at an offset from the NULL pointer) in the event of a
3055 memory allocation failure. In 1.0.2 and below this could be caused where
3056 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
3057 could be in processing a very long "%s" format string. Memory leaks can
3060 The first issue may mask the second issue dependent on compiler behaviour.
3061 These problems could enable attacks where large amounts of untrusted data
3062 is passed to the BIO_*printf functions. If applications use these functions
3063 in this way then they could be vulnerable. OpenSSL itself uses these
3064 functions when printing out human-readable dumps of ASN.1 data. Therefore
3065 applications that print this data could be vulnerable if the data is from
3066 untrusted sources. OpenSSL command line applications could also be
3067 vulnerable where they print out ASN.1 data, or if untrusted data is passed
3068 as command line arguments.
3070 Libssl is not considered directly vulnerable. Additionally certificates etc
3071 received via remote connections via libssl are also unlikely to be able to
3072 trigger these issues because of message size limits enforced within libssl.
3074 This issue was reported to OpenSSL Guido Vranken.
3079 * Side channel attack on modular exponentiation
3081 A side-channel attack was found which makes use of cache-bank conflicts on
3082 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
3083 of RSA keys. The ability to exploit this issue is limited as it relies on
3084 an attacker who has control of code in a thread running on the same
3085 hyper-threaded core as the victim thread which is performing decryptions.
3087 This issue was reported to OpenSSL by Yuval Yarom, The University of
3088 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
3089 Nadia Heninger, University of Pennsylvania with more information at
3090 http://cachebleed.info.
3095 * Change the req app to generate a 2048-bit RSA/DSA key by default,
3096 if no keysize is specified with default_bits. This fixes an
3097 omission in an earlier change that changed all RSA/DSA key generation
3098 apps to use 2048 bits by default.
3102 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016] ###
3103 * DH small subgroups
3105 Historically OpenSSL only ever generated DH parameters based on "safe"
3106 primes. More recently (in version 1.0.2) support was provided for
3107 generating X9.42 style parameter files such as those required for RFC 5114
3108 support. The primes used in such files may not be "safe". Where an
3109 application is using DH configured with parameters based on primes that are
3110 not "safe" then an attacker could use this fact to find a peer's private
3111 DH exponent. This attack requires that the attacker complete multiple
3112 handshakes in which the peer uses the same private DH exponent. For example
3113 this could be used to discover a TLS server's private DH exponent if it's
3114 reusing the private DH exponent or it's using a static DH ciphersuite.
3116 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
3117 TLS. It is not on by default. If the option is not set then the server
3118 reuses the same private DH exponent for the life of the server process and
3119 would be vulnerable to this attack. It is believed that many popular
3120 applications do set this option and would therefore not be at risk.
3122 The fix for this issue adds an additional check where a "q" parameter is
3123 available (as is the case in X9.42 based parameters). This detects the
3124 only known attack, and is the only possible defense for static DH
3125 ciphersuites. This could have some performance impact.
3127 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
3128 default and cannot be disabled. This could have some performance impact.
3130 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
3135 * SSLv2 doesn't block disabled ciphers
3137 A malicious client can negotiate SSLv2 ciphers that have been disabled on
3138 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
3139 been disabled, provided that the SSLv2 protocol was not also disabled via
3142 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
3143 and Sebastian Schinzel.
3148 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015] ###
3150 * BN_mod_exp may produce incorrect results on x86_64
3152 There is a carry propagating bug in the x86_64 Montgomery squaring
3153 procedure. No EC algorithms are affected. Analysis suggests that attacks
3154 against RSA and DSA as a result of this defect would be very difficult to
3155 perform and are not believed likely. Attacks against DH are considered just
3156 feasible (although very difficult) because most of the work necessary to
3157 deduce information about a private key may be performed offline. The amount
3158 of resources required for such an attack would be very significant and
3159 likely only accessible to a limited number of attackers. An attacker would
3160 additionally need online access to an unpatched system using the target
3161 private key in a scenario with persistent DH parameters and a private
3162 key that is shared between multiple clients. For example this can occur by
3163 default in OpenSSL DHE based SSL/TLS ciphersuites.
3165 This issue was reported to OpenSSL by Hanno Böck.
3170 * Certificate verify crash with missing PSS parameter
3172 The signature verification routines will crash with a NULL pointer
3173 dereference if presented with an ASN.1 signature using the RSA PSS
3174 algorithm and absent mask generation function parameter. Since these
3175 routines are used to verify certificate signature algorithms this can be
3176 used to crash any certificate verification operation and exploited in a
3177 DoS attack. Any application which performs certificate verification is
3178 vulnerable including OpenSSL clients and servers which enable client
3181 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
3186 * X509_ATTRIBUTE memory leak
3188 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
3189 memory. This structure is used by the PKCS#7 and CMS routines so any
3190 application which reads PKCS#7 or CMS data from untrusted sources is
3191 affected. SSL/TLS is not affected.
3193 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
3199 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3200 This changes the decoding behaviour for some invalid messages,
3201 though the change is mostly in the more lenient direction, and
3202 legacy behaviour is preserved as much as possible.
3206 * In DSA_generate_parameters_ex, if the provided seed is too short,
3209 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3211 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015] ###
3213 * Alternate chains certificate forgery
3215 During certificate verification, OpenSSL will attempt to find an
3216 alternative certificate chain if the first attempt to build such a chain
3217 fails. An error in the implementation of this logic can mean that an
3218 attacker could cause certain checks on untrusted certificates to be
3219 bypassed, such as the CA flag, enabling them to use a valid leaf
3220 certificate to act as a CA and "issue" an invalid certificate.
3222 This issue was reported to OpenSSL by Adam Langley/David Benjamin
3227 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015] ###
3229 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
3230 incompatibility in the handling of HMAC. The previous ABI has now been
3235 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015] ###
3237 * Malformed ECParameters causes infinite loop
3239 When processing an ECParameters structure OpenSSL enters an infinite loop
3240 if the curve specified is over a specially malformed binary polynomial
3243 This can be used to perform denial of service against any
3244 system which processes public keys, certificate requests or
3245 certificates. This includes TLS clients and TLS servers with
3246 client authentication enabled.
3248 This issue was reported to OpenSSL by Joseph Barr-Pixton.
3253 * Exploitable out-of-bounds read in X509_cmp_time
3255 X509_cmp_time does not properly check the length of the ASN1_TIME
3256 string and can read a few bytes out of bounds. In addition,
3257 X509_cmp_time accepts an arbitrary number of fractional seconds in the
3260 An attacker can use this to craft malformed certificates and CRLs of
3261 various sizes and potentially cause a segmentation fault, resulting in
3262 a DoS on applications that verify certificates or CRLs. TLS clients
3263 that verify CRLs are affected. TLS clients and servers with client
3264 authentication enabled may be affected if they use custom verification
3267 This issue was reported to OpenSSL by Robert Swiecki (Google), and
3268 independently by Hanno Böck.
3273 * PKCS7 crash with missing EnvelopedContent
3275 The PKCS#7 parsing code does not handle missing inner EncryptedContent
3276 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
3277 with missing content and trigger a NULL pointer dereference on parsing.
3279 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
3280 structures from untrusted sources are affected. OpenSSL clients and
3281 servers are not affected.
3283 This issue was reported to OpenSSL by Michal Zalewski (Google).
3288 * CMS verify infinite loop with unknown hash function
3290 When verifying a signedData message the CMS code can enter an infinite loop
3291 if presented with an unknown hash function OID. This can be used to perform
3292 denial of service against any system which verifies signedData messages using
3294 This issue was reported to OpenSSL by Johannes Bauer.
3299 * Race condition handling NewSessionTicket
3301 If a NewSessionTicket is received by a multi-threaded client when attempting to
3302 reuse a previous ticket then a race condition can occur potentially leading to
3303 a double free of the ticket data.
3308 * Only support 256-bit or stronger elliptic curves with the
3309 'ecdh_auto' setting (server) or by default (client). Of supported
3310 curves, prefer P-256 (both).
3314 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015] ###
3316 * ClientHello sigalgs DoS fix
3318 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
3319 invalid signature algorithms extension a NULL pointer dereference will
3320 occur. This can be exploited in a DoS attack against the server.
3322 This issue was was reported to OpenSSL by David Ramos of Stanford
3326 *Stephen Henson and Matt Caswell*
3328 * Multiblock corrupted pointer fix
3330 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
3331 feature only applies on 64 bit x86 architecture platforms that support AES
3332 NI instructions. A defect in the implementation of "multiblock" can cause
3333 OpenSSL's internal write buffer to become incorrectly set to NULL when
3334 using non-blocking IO. Typically, when the user application is using a
3335 socket BIO for writing, this will only result in a failed connection.
3336 However if some other BIO is used then it is likely that a segmentation
3337 fault will be triggered, thus enabling a potential DoS attack.
3339 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
3344 * Segmentation fault in DTLSv1_listen fix
3346 The DTLSv1_listen function is intended to be stateless and processes the
3347 initial ClientHello from many peers. It is common for user code to loop
3348 over the call to DTLSv1_listen until a valid ClientHello is received with
3349 an associated cookie. A defect in the implementation of DTLSv1_listen means
3350 that state is preserved in the SSL object from one invocation to the next
3351 that can lead to a segmentation fault. Errors processing the initial
3352 ClientHello can trigger this scenario. An example of such an error could be
3353 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
3356 This issue was reported to OpenSSL by Per Allansson.
3361 * Segmentation fault in ASN1_TYPE_cmp fix
3363 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
3364 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
3365 certificate signature algorithm consistency this can be used to crash any
3366 certificate verification operation and exploited in a DoS attack. Any
3367 application which performs certificate verification is vulnerable including
3368 OpenSSL clients and servers which enable client authentication.
3373 * Segmentation fault for invalid PSS parameters fix
3375 The signature verification routines will crash with a NULL pointer
3376 dereference if presented with an ASN.1 signature using the RSA PSS
3377 algorithm and invalid parameters. Since these routines are used to verify
3378 certificate signature algorithms this can be used to crash any
3379 certificate verification operation and exploited in a DoS attack. Any
3380 application which performs certificate verification is vulnerable including
3381 OpenSSL clients and servers which enable client authentication.
3383 This issue was was reported to OpenSSL by Brian Carpenter.
3388 * ASN.1 structure reuse memory corruption fix
3390 Reusing a structure in ASN.1 parsing may allow an attacker to cause
3391 memory corruption via an invalid write. Such reuse is and has been
3392 strongly discouraged and is believed to be rare.
3394 Applications that parse structures containing CHOICE or ANY DEFINED BY
3395 components may be affected. Certificate parsing (d2i_X509 and related
3396 functions) are however not affected. OpenSSL clients and servers are
3402 * PKCS7 NULL pointer dereferences fix
3404 The PKCS#7 parsing code does not handle missing outer ContentInfo
3405 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
3406 missing content and trigger a NULL pointer dereference on parsing.
3408 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
3409 otherwise parse PKCS#7 structures from untrusted sources are
3410 affected. OpenSSL clients and servers are not affected.
3412 This issue was reported to OpenSSL by Michal Zalewski (Google).
3417 * DoS via reachable assert in SSLv2 servers fix
3419 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
3420 servers that both support SSLv2 and enable export cipher suites by sending
3421 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
3423 This issue was discovered by Sean Burford (Google) and Emilia Käsper
3424 (OpenSSL development team).
3429 * Empty CKE with client auth and DHE fix
3431 If client auth is used then a server can seg fault in the event of a DHE
3432 ciphersuite being selected and a zero length ClientKeyExchange message
3433 being sent by the client. This could be exploited in a DoS attack.
3438 * Handshake with unseeded PRNG fix
3440 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
3441 with an unseeded PRNG. The conditions are:
3442 - The client is on a platform where the PRNG has not been seeded
3443 automatically, and the user has not seeded manually
3444 - A protocol specific client method version has been used (i.e. not
3445 SSL_client_methodv23)
3446 - A ciphersuite is used that does not require additional random data from
3447 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
3449 If the handshake succeeds then the client random that has been used will
3450 have been generated from a PRNG with insufficient entropy and therefore the
3451 output may be predictable.
3453 For example using the following command with an unseeded openssl will
3454 succeed on an unpatched platform:
3456 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
3461 * Use After Free following d2i_ECPrivatekey error fix
3463 A malformed EC private key file consumed via the d2i_ECPrivateKey function
3464 could cause a use after free condition. This, in turn, could cause a double
3465 free in several private key parsing functions (such as d2i_PrivateKey
3466 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
3467 for applications that receive EC private keys from untrusted
3468 sources. This scenario is considered rare.
3470 This issue was discovered by the BoringSSL project and fixed in their
3476 * X509_to_X509_REQ NULL pointer deref fix
3478 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
3479 the certificate key is invalid. This function is rarely used in practice.
3481 This issue was discovered by Brian Carpenter.
3486 * Removed the export ciphers from the DEFAULT ciphers
3490 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015] ###
3492 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
3493 ARMv5 through ARMv8, as opposite to "locking" it to single one.
3494 So far those who have to target multiple platforms would compromise
3495 and argue that binary targeting say ARMv5 would still execute on
3496 ARMv8. "Universal" build resolves this compromise by providing
3497 near-optimal performance even on newer platforms.
3501 * Accelerated NIST P-256 elliptic curve implementation for x86_64
3502 (other platforms pending).
3504 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
3506 * Add support for the SignedCertificateTimestampList certificate and
3507 OCSP response extensions from RFC6962.
3511 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3512 for corner cases. (Certain input points at infinity could lead to
3513 bogus results, with non-infinity inputs mapped to infinity too.)
3517 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
3518 This covers AES, SHA256/512 and GHASH. "Initial" means that most
3519 common cases are optimized and there still is room for further
3520 improvements. Vector Permutation AES for Altivec is also added.
3524 * Add support for little-endian ppc64 Linux target.
3526 *Marcelo Cerri (IBM)*
3528 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
3529 SHA1, SHA256 and GHASH. "Initial" means that most common cases
3530 are optimized and there still is room for further improvements.
3531 Both 32- and 64-bit modes are supported.
3533 *Andy Polyakov, Ard Biesheuvel (Linaro)*
3535 * Improved ARMv7 NEON support.
3539 * Support for SPARC Architecture 2011 crypto extensions, first
3540 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
3541 SHA256/512, MD5, GHASH and modular exponentiation.
3543 *Andy Polyakov, David Miller*
3545 * Accelerated modular exponentiation for Intel processors, a.k.a.
3548 *Shay Gueron & Vlad Krasnov (Intel Corp)*
3550 * Support for new and upcoming Intel processors, including AVX2,
3551 BMI and SHA ISA extensions. This includes additional "stitched"
3552 implementations, AESNI-SHA256 and GCM, and multi-buffer support
3555 This work was sponsored by Intel Corp.
3559 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
3560 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
3561 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
3565 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3566 this fixes a limitation in previous versions of OpenSSL.
3570 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3571 MGF1 digest and OAEP label.
3575 * Add EVP support for key wrapping algorithms, to avoid problems with
3576 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3577 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3578 algorithms and include tests cases.
3582 * Add functions to allocate and set the fields of an ECDSA_METHOD
3585 *Douglas E. Engert, Steve Henson*
3587 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
3588 difference in days and seconds between two tm or ASN1_TIME structures.
3592 * Add -rev test option to s_server to just reverse order of characters
3593 received by client and send back to server. Also prints an abbreviated
3594 summary of the connection parameters.
3598 * New option -brief for s_client and s_server to print out a brief summary
3599 of connection parameters.
3603 * Add callbacks for arbitrary TLS extensions.
3605 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
3607 * New option -crl_download in several openssl utilities to download CRLs
3608 from CRLDP extension in certificates.
3612 * New options -CRL and -CRLform for s_client and s_server for CRLs.
3616 * New function X509_CRL_diff to generate a delta CRL from the difference
3617 of two full CRLs. Add support to "crl" utility.
3621 * New functions to set lookup_crls function and to retrieve
3622 X509_STORE from X509_STORE_CTX.
3626 * Print out deprecated issuer and subject unique ID fields in
3631 * Extend OCSP I/O functions so they can be used for simple general purpose
3632 HTTP as well as OCSP. New wrapper function which can be used to download
3633 CRLs using the OCSP API.
3637 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3641 * SSL_CONF* functions. These provide a common framework for application
3642 configuration using configuration files or command lines.
3646 * SSL/TLS tracing code. This parses out SSL/TLS records using the
3647 message callback and prints the results. Needs compile time option
3648 "enable-ssl-trace". New options to s_client and s_server to enable
3653 * New ctrl and macro to retrieve supported points extensions.
3654 Print out extension in s_server and s_client.
3658 * New functions to retrieve certificate signature and signature
3663 * Add functions to retrieve and manipulate the raw cipherlist sent by a
3668 * New Suite B modes for TLS code. These use and enforce the requirements
3669 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3670 only use Suite B curves. The Suite B modes can be set by using the
3671 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3675 * New chain verification flags for Suite B levels of security. Check
3676 algorithms are acceptable when flags are set in X509_verify_cert.
3680 * Make tls1_check_chain return a set of flags indicating checks passed
3681 by a certificate chain. Add additional tests to handle client
3682 certificates: checks for matching certificate type and issuer name
3687 * If an attempt is made to use a signature algorithm not in the peer
3688 preference list abort the handshake. If client has no suitable
3689 signature algorithms in response to a certificate request do not
3690 use the certificate.
3694 * If server EC tmp key is not in client preference list abort handshake.
3698 * Add support for certificate stores in CERT structure. This makes it
3699 possible to have different stores per SSL structure or one store in
3700 the parent SSL_CTX. Include distinct stores for certificate chain
3701 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3702 to build and store a certificate chain in CERT structure: returning
3703 an error if the chain cannot be built: this will allow applications
3704 to test if a chain is correctly configured.
3706 Note: if the CERT based stores are not set then the parent SSL_CTX
3707 store is used to retain compatibility with existing behaviour.
3712 * New function ssl_set_client_disabled to set a ciphersuite disabled
3713 mask based on the current session, check mask when sending client
3714 hello and checking the requested ciphersuite.
3718 * New ctrls to retrieve and set certificate types in a certificate
3719 request message. Print out received values in s_client. If certificate
3720 types is not set with custom values set sensible values based on
3721 supported signature algorithms.
3725 * Support for distinct client and server supported signature algorithms.
3729 * Add certificate callback. If set this is called whenever a certificate
3730 is required by client or server. An application can decide which
3731 certificate chain to present based on arbitrary criteria: for example
3732 supported signature algorithms. Add very simple example to s_server.
3733 This fixes many of the problems and restrictions of the existing client
3734 certificate callback: for example you can now clear an existing
3735 certificate and specify the whole chain.
3739 * Add new "valid_flags" field to CERT_PKEY structure which determines what
3740 the certificate can be used for (if anything). Set valid_flags field
3741 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3742 to have similar checks in it.
3744 Add new "cert_flags" field to CERT structure and include a "strict mode".
3745 This enforces some TLS certificate requirements (such as only permitting
3746 certificate signature algorithms contained in the supported algorithms
3747 extension) which some implementations ignore: this option should be used
3748 with caution as it could cause interoperability issues.
3752 * Update and tidy signature algorithm extension processing. Work out
3753 shared signature algorithms based on preferences and peer algorithms
3754 and print them out in s_client and s_server. Abort handshake if no
3755 shared signature algorithms.
3759 * Add new functions to allow customised supported signature algorithms
3760 for SSL and SSL_CTX structures. Add options to s_client and s_server
3765 * New function SSL_certs_clear() to delete all references to certificates
3766 from an SSL structure. Before this once a certificate had been added
3767 it couldn't be removed.
3771 * Integrate hostname, email address and IP address checking with certificate
3772 verification. New verify options supporting checking in openssl utility.
3776 * Fixes and wildcard matching support to hostname and email checking
3777 functions. Add manual page.
3779 *Florian Weimer (Red Hat Product Security Team)*
3781 * New functions to check a hostname email or IP address against a
3782 certificate. Add options x509 utility to print results of checks against
3787 * Fix OCSP checking.
3789 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
3791 * Initial experimental support for explicitly trusted non-root CAs.
3792 OpenSSL still tries to build a complete chain to a root but if an
3793 intermediate CA has a trust setting included that is used. The first
3794 setting is used: whether to trust (e.g., -addtrust option to the x509
3799 * Add -trusted_first option which attempts to find certificates in the
3800 trusted store even if an untrusted chain is also supplied.
3804 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3805 platform support for Linux and Android.
3809 * Support for linux-x32, ILP32 environment in x86_64 framework.
3813 * Experimental multi-implementation support for FIPS capable OpenSSL.
3814 When in FIPS mode the approved implementations are used as normal,
3815 when not in FIPS mode the internal unapproved versions are used instead.
3816 This means that the FIPS capable OpenSSL isn't forced to use the
3817 (often lower performance) FIPS implementations outside FIPS mode.
3821 * Transparently support X9.42 DH parameters when calling
3822 PEM_read_bio_DHparameters. This means existing applications can handle
3823 the new parameter format automatically.
3827 * Initial experimental support for X9.42 DH parameter format: mainly
3828 to support use of 'q' parameter for RFC5114 parameters.
3832 * Add DH parameters from RFC5114 including test data to dhtest.
3836 * Support for automatic EC temporary key parameter selection. If enabled
3837 the most preferred EC parameters are automatically used instead of
3838 hardcoded fixed parameters. Now a server just has to call:
3839 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3840 support ECDH and use the most appropriate parameters.
3844 * Enhance and tidy EC curve and point format TLS extension code. Use
3845 static structures instead of allocation if default values are used.
3846 New ctrls to set curves we wish to support and to retrieve shared curves.
3847 Print out shared curves in s_server. New options to s_server and s_client
3848 to set list of supported curves.
3852 * New ctrls to retrieve supported signature algorithms and
3853 supported curve values as an array of NIDs. Extend openssl utility
3854 to print out received values.
3858 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3859 between NIDs and the more common NIST names such as "P-256". Enhance
3860 ecparam utility and ECC method to recognise the NIST names for curves.
3864 * Enhance SSL/TLS certificate chain handling to support different
3865 chains for each certificate instead of one chain in the parent SSL_CTX.
3869 * Support for fixed DH ciphersuite client authentication: where both
3870 server and client use DH certificates with common parameters.
3874 * Support for fixed DH ciphersuites: those requiring DH server
3879 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
3881 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3882 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3883 X509_CINF_get_signature were reverted post internal team review.
3885 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015] ###
3887 * Build fixes for the Windows and OpenVMS platforms
3889 *Matt Caswell and Richard Levitte*
3891 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015] ###
3893 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3894 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3895 dereference. This could lead to a Denial Of Service attack. Thanks to
3896 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3901 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3902 dtls1_buffer_record function under certain conditions. In particular this
3903 could occur if an attacker sent repeated DTLS records with the same
3904 sequence number but for the next epoch. The memory leak could be exploited
3905 by an attacker in a Denial of Service attack through memory exhaustion.
3906 Thanks to Chris Mueller for reporting this issue.
3911 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3912 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3913 method would be set to NULL which could later result in a NULL pointer
3914 dereference. Thanks to Frank Schmirler for reporting this issue.
3919 * Abort handshake if server key exchange message is omitted for ephemeral
3922 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3923 reporting this issue.
3928 * Remove non-export ephemeral RSA code on client and server. This code
3929 violated the TLS standard by allowing the use of temporary RSA keys in
3930 non-export ciphersuites and could be used by a server to effectively
3931 downgrade the RSA key length used to a value smaller than the server
3932 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3933 INRIA or reporting this issue.
3938 * Fixed issue where DH client certificates are accepted without verification.
3939 An OpenSSL server will accept a DH certificate for client authentication
3940 without the certificate verify message. This effectively allows a client to
3941 authenticate without the use of a private key. This only affects servers
3942 which trust a client certificate authority which issues certificates
3943 containing DH keys: these are extremely rare and hardly ever encountered.
3944 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3950 * Ensure that the session ID context of an SSL is updated when its
3951 SSL_CTX is updated via SSL_set_SSL_CTX.
3953 The session ID context is typically set from the parent SSL_CTX,
3954 and can vary with the CTX.
3958 * Fix various certificate fingerprint issues.
3960 By using non-DER or invalid encodings outside the signed portion of a
3961 certificate the fingerprint can be changed without breaking the signature.
3962 Although no details of the signed portion of the certificate can be changed
3963 this can cause problems with some applications: e.g. those using the
3964 certificate fingerprint for blacklists.
3966 1. Reject signatures with non zero unused bits.
3968 If the BIT STRING containing the signature has non zero unused bits reject
3969 the signature. All current signature algorithms require zero unused bits.
3971 2. Check certificate algorithm consistency.
3973 Check the AlgorithmIdentifier inside TBS matches the one in the
3974 certificate signature. NB: this will result in signature failure
3975 errors for some broken certificates.
3977 Thanks to Konrad Kraszewski from Google for reporting this issue.
3979 3. Check DSA/ECDSA signatures use DER.
3981 Re-encode DSA/ECDSA signatures and compare with the original received
3982 signature. Return an error if there is a mismatch.
3984 This will reject various cases including garbage after signature
3985 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3986 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3987 (negative or with leading zeroes).
3989 Further analysis was conducted and fixes were developed by Stephen Henson
3990 of the OpenSSL core team.
3996 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3997 results on some platforms, including x86_64. This bug occurs at random
3998 with a very low probability, and is not known to be exploitable in any
3999 way, though its exact impact is difficult to determine. Thanks to Pieter
4000 Wuille (Blockstream) who reported this issue and also suggested an initial
4001 fix. Further analysis was conducted by the OpenSSL development team and
4002 Adam Langley of Google. The final fix was developed by Andy Polyakov of
4003 the OpenSSL core team.
4008 *) Do not resume sessions on the server if the negotiated protocol
4009 version does not match the session's version. Resuming with a different
4010 version, while not strictly forbidden by the RFC, is of questionable
4011 sanity and breaks all known clients.
4013 *David Benjamin, Emilia Käsper*
4015 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
4016 early CCS messages during renegotiation. (Note that because
4017 renegotiation is encrypted, this early CCS was not exploitable.)
4021 *) Tighten client-side session ticket handling during renegotiation:
4022 ensure that the client only accepts a session ticket if the server sends
4023 the extension anew in the ServerHello. Previously, a TLS client would
4024 reuse the old extension state and thus accept a session ticket if one was
4025 announced in the initial ServerHello.
4027 Similarly, ensure that the client requires a session ticket if one
4028 was advertised in the ServerHello. Previously, a TLS client would
4029 ignore a missing NewSessionTicket message.
4033 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014] ###
4037 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
4038 sends a carefully crafted handshake message, to cause OpenSSL to fail
4039 to free up to 64k of memory causing a memory leak. This could be
4040 exploited in a Denial Of Service attack. This issue affects OpenSSL
4041 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
4042 whether SRTP is used or configured. Implementations of OpenSSL that
4043 have been compiled with OPENSSL_NO_SRTP defined are not affected.
4045 The fix was developed by the OpenSSL team.
4050 * Session Ticket Memory Leak.
4052 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
4053 integrity of that ticket is first verified. In the event of a session
4054 ticket integrity check failing, OpenSSL will fail to free memory
4055 causing a memory leak. By sending a large number of invalid session
4056 tickets an attacker could exploit this issue in a Denial Of Service
4062 * Build option no-ssl3 is incomplete.
4064 When OpenSSL is configured with "no-ssl3" as a build option, servers
4065 could accept and complete a SSL 3.0 handshake, and clients could be
4066 configured to send them.
4069 *Akamai and the OpenSSL team*
4071 * Add support for TLS_FALLBACK_SCSV.
4072 Client applications doing fallback retries should call
4073 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
4076 *Adam Langley, Bodo Moeller*
4078 * Add additional DigestInfo checks.
4080 Re-encode DigestInto in DER and check against the original when
4081 verifying RSA signature: this will reject any improperly encoded
4082 DigestInfo structures.
4084 Note: this is a precautionary measure and no attacks are currently known.
4089 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014] ###
4091 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
4092 SRP code can be overrun an internal buffer. Add sanity check that
4093 g, A, B < N to SRP code.
4095 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
4096 Group for discovering this issue.
4101 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
4102 TLS 1.0 instead of higher protocol versions when the ClientHello message
4103 is badly fragmented. This allows a man-in-the-middle attacker to force a
4104 downgrade to TLS 1.0 even if both the server and the client support a
4105 higher protocol version, by modifying the client's TLS records.
4107 Thanks to David Benjamin and Adam Langley (Google) for discovering and
4108 researching this issue.
4113 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
4114 to a denial of service attack. A malicious server can crash the client
4115 with a null pointer dereference (read) by specifying an anonymous (EC)DH
4116 ciphersuite and sending carefully crafted handshake messages.
4118 Thanks to Felix Gröbert (Google) for discovering and researching this
4124 * By sending carefully crafted DTLS packets an attacker could cause openssl
4125 to leak memory. This can be exploited through a Denial of Service attack.
4126 Thanks to Adam Langley for discovering and researching this issue.
4131 * An attacker can force openssl to consume large amounts of memory whilst
4132 processing DTLS handshake messages. This can be exploited through a
4133 Denial of Service attack.
4134 Thanks to Adam Langley for discovering and researching this issue.
4139 * An attacker can force an error condition which causes openssl to crash
4140 whilst processing DTLS packets due to memory being freed twice. This
4141 can be exploited through a Denial of Service attack.
4142 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
4148 * If a multithreaded client connects to a malicious server using a resumed
4149 session and the server sends an ec point format extension it could write
4150 up to 255 bytes to freed memory.
4152 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
4158 * A malicious server can crash an OpenSSL client with a null pointer
4159 dereference (read) by specifying an SRP ciphersuite even though it was not
4160 properly negotiated with the client. This can be exploited through a
4161 Denial of Service attack.
4163 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
4164 discovering and researching this issue.
4169 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
4170 X509_name_oneline, X509_name_print_ex et al. to leak some information
4171 from the stack. Applications may be affected if they echo pretty printing
4172 output to the attacker.
4174 Thanks to Ivan Fratric (Google) for discovering this issue.
4177 *Emilia Käsper, and Steve Henson*
4179 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4180 for corner cases. (Certain input points at infinity could lead to
4181 bogus results, with non-infinity inputs mapped to infinity too.)
4185 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014] ###
4187 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
4188 handshake can force the use of weak keying material in OpenSSL
4189 SSL/TLS clients and servers.
4191 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
4192 researching this issue. (CVE-2014-0224)
4194 *KIKUCHI Masashi, Steve Henson*
4196 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
4197 OpenSSL DTLS client the code can be made to recurse eventually crashing
4200 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
4203 *Imre Rad, Steve Henson*
4205 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
4206 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
4207 client or server. This is potentially exploitable to run arbitrary
4208 code on a vulnerable client or server.
4210 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
4212 *Jüri Aedla, Steve Henson*
4214 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
4215 are subject to a denial of service attack.
4217 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
4218 this issue. (CVE-2014-3470)
4220 *Felix Gröbert, Ivan Fratric, Steve Henson*
4222 * Harmonize version and its documentation. -f flag is used to display
4225 *mancha <mancha1@zoho.com>*
4227 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4228 in i2d_ECPrivateKey.
4230 *mancha <mancha1@zoho.com>*
4232 * Fix some double frees. These are not thought to be exploitable.
4234 *mancha <mancha1@zoho.com>*
4236 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014] ###
4238 * A missing bounds check in the handling of the TLS heartbeat extension
4239 can be used to reveal up to 64k of memory to a connected client or
4242 Thanks for Neel Mehta of Google Security for discovering this bug and to
4243 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4244 preparing the fix (CVE-2014-0160)
4246 *Adam Langley, Bodo Moeller*
4248 * Fix for the attack described in the paper "Recovering OpenSSL
4249 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4250 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4251 http://eprint.iacr.org/2014/140
4253 Thanks to Yuval Yarom and Naomi Benger for discovering this
4254 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
4256 *Yuval Yarom and Naomi Benger*
4258 * TLS pad extension: draft-agl-tls-padding-03
4260 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
4261 TLS client Hello record length value would otherwise be > 255 and
4262 less that 512 pad with a dummy extension containing zeroes so it
4263 is at least 512 bytes long.
4266 *Adam Langley, Steve Henson*
4268 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014] ###
4270 * Fix for TLS record tampering bug. A carefully crafted invalid
4271 handshake could crash OpenSSL with a NULL pointer exception.
4272 Thanks to Anton Johansson for reporting this issues.
4275 * Keep original DTLS digest and encryption contexts in retransmission
4276 structures so we can use the previous session parameters if they need
4277 to be resent. (CVE-2013-6450)
4281 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
4282 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
4283 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
4284 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
4285 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
4286 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
4288 *Rob Stradling, Adam Langley*
4290 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013] ###
4292 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
4293 supporting platforms or when small records were transferred.
4295 *Andy Polyakov, Steve Henson*
4297 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013] ###
4299 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
4301 This addresses the flaw in CBC record processing discovered by
4302 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
4303 at: http://www.isg.rhul.ac.uk/tls/
4305 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4306 Security Group at Royal Holloway, University of London
4307 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
4308 Emilia Käsper for the initial patch.
4311 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
4313 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
4314 ciphersuites which can be exploited in a denial of service attack.
4315 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
4316 and detecting this bug and to Wolfgang Ettlinger
4317 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
4322 * Return an error when checking OCSP signatures when key is NULL.
4323 This fixes a DoS attack. (CVE-2013-0166)
4327 * Make openssl verify return errors.
4329 *Chris Palmer <palmer@google.com> and Ben Laurie*
4331 * Call OCSP Stapling callback after ciphersuite has been chosen, so
4332 the right response is stapled. Also change SSL_get_certificate()
4333 so it returns the certificate actually sent.
4334 See http://rt.openssl.org/Ticket/Display.html?id=2836.
4336 *Rob Stradling <rob.stradling@comodo.com>*
4338 * Fix possible deadlock when decoding public keys.
4342 * Don't use TLS 1.0 record version number in initial client hello
4347 ### Changes between 1.0.1b and 1.0.1c [10 May 2012] ###
4349 * Sanity check record length before skipping explicit IV in TLS
4350 1.2, 1.1 and DTLS to fix DoS attack.
4352 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
4353 fuzzing as a service testing platform.
4358 * Initialise tkeylen properly when encrypting CMS messages.
4359 Thanks to Solar Designer of Openwall for reporting this issue.
4363 * In FIPS mode don't try to use composite ciphers as they are not
4368 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012] ###
4370 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
4371 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
4372 mean any application compiled against OpenSSL 1.0.0 headers setting
4373 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
4374 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
4375 0x10000000L Any application which was previously compiled against
4376 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
4377 will need to be recompiled as a result. Letting be results in
4378 inability to disable specifically TLS 1.1 and in client context,
4379 in unlike event, limit maximum offered version to TLS 1.0 [see below].
4383 * In order to ensure interoperability SSL_OP_NO_protocolX does not
4384 disable just protocol X, but all protocols above X *if* there are
4385 protocols *below* X still enabled. In more practical terms it means
4386 that if application wants to disable TLS1.0 in favor of TLS1.1 and
4387 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
4388 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
4393 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012] ###
4395 * Check for potentially exploitable overflows in asn1_d2i_read_bio
4396 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
4397 in CRYPTO_realloc_clean.
4399 Thanks to Tavis Ormandy, Google Security Team, for discovering this
4400 issue and to Adam Langley <agl@chromium.org> for fixing it.
4403 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
4405 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
4409 * Workarounds for some broken servers that "hang" if a client hello
4410 record length exceeds 255 bytes.
4412 1. Do not use record version number > TLS 1.0 in initial client
4413 hello: some (but not all) hanging servers will now work.
4414 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
4415 the number of ciphers sent in the client hello. This should be
4416 set to an even number, such as 50, for example by passing:
4417 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
4418 Most broken servers should now work.
4419 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
4420 TLS 1.2 client support entirely.
4424 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
4428 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012] ###
4430 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
4431 STRING form instead of a DigestInfo.
4435 * The format used for MDC2 RSA signatures is inconsistent between EVP
4436 and the RSA_sign/RSA_verify functions. This was made more apparent when
4437 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
4438 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
4439 the correct format in RSA_verify so both forms transparently work.
4443 * Some servers which support TLS 1.0 can choke if we initially indicate
4444 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
4445 encrypted premaster secret. As a workaround use the maximum permitted
4446 client version in client hello, this should keep such servers happy
4447 and still work with previous versions of OpenSSL.
4451 * Add support for TLS/DTLS heartbeats.
4453 *Robin Seggelmann <seggelmann@fh-muenster.de>*
4455 * Add support for SCTP.
4457 *Robin Seggelmann <seggelmann@fh-muenster.de>*
4459 * Improved PRNG seeding for VOS.
4461 *Paul Green <Paul.Green@stratus.com>*
4463 * Extensive assembler packs updates, most notably:
4465 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
4466 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
4467 - x86_64: bit-sliced AES implementation;
4468 - ARM: NEON support, contemporary platforms optimizations;
4469 - s390x: z196 support;
4470 - *: GHASH and GF(2^m) multiplication implementations;
4474 * Make TLS-SRP code conformant with RFC 5054 API cleanup
4475 (removal of unnecessary code)
4477 *Peter Sylvester <peter.sylvester@edelweb.fr>*
4479 * Add TLS key material exporter from RFC 5705.
4483 * Add DTLS-SRTP negotiation from RFC 5764.
4487 * Add Next Protocol Negotiation,
4488 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
4489 disabled with a no-npn flag to config or Configure. Code donated
4492 *Adam Langley <agl@google.com> and Ben Laurie*
4494 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
4495 NIST-P256, NIST-P521, with constant-time single point multiplication on
4496 typical inputs. Compiler support for the nonstandard type __uint128_t is
4497 required to use this (present in gcc 4.4 and later, for 64-bit builds).
4498 Code made available under Apache License version 2.0.
4500 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
4501 line to include this in your build of OpenSSL, and run "make depend" (or
4502 "make update"). This enables the following EC_METHODs:
4504 EC_GFp_nistp224_method()
4505 EC_GFp_nistp256_method()
4506 EC_GFp_nistp521_method()
4508 EC_GROUP_new_by_curve_name() will automatically use these (while
4509 EC_GROUP_new_curve_GFp() currently prefers the more flexible
4512 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
4514 * Use type ossl_ssize_t instad of ssize_t which isn't available on
4515 all platforms. Move ssize_t definition from e_os.h to the public
4516 header file e_os2.h as it now appears in public header file cms.h
4520 * New -sigopt option to the ca, req and x509 utilities. Additional
4521 signature parameters can be passed using this option and in
4526 * Add RSA PSS signing function. This will generate and set the
4527 appropriate AlgorithmIdentifiers for PSS based on those in the
4528 corresponding EVP_MD_CTX structure. No application support yet.
4532 * Support for companion algorithm specific ASN1 signing routines.
4533 New function ASN1_item_sign_ctx() signs a pre-initialised
4534 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
4535 the appropriate parameters.
4539 * Add new algorithm specific ASN1 verification initialisation function
4540 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
4541 handling will be the same no matter what EVP_PKEY_METHOD is used.
4542 Add a PSS handler to support verification of PSS signatures: checked
4543 against a number of sample certificates.
4547 * Add signature printing for PSS. Add PSS OIDs.
4549 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
4551 * Add algorithm specific signature printing. An individual ASN1 method
4552 can now print out signatures instead of the standard hex dump.
4554 More complex signatures (e.g. PSS) can print out more meaningful
4555 information. Include DSA version that prints out the signature
4560 * Password based recipient info support for CMS library: implementing
4565 * Split password based encryption into PBES2 and PBKDF2 functions. This
4566 neatly separates the code into cipher and PBE sections and is required
4567 for some algorithms that split PBES2 into separate pieces (such as
4568 password based CMS).
4572 * Session-handling fixes:
4573 - Fix handling of connections that are resuming with a session ID,
4574 but also support Session Tickets.
4575 - Fix a bug that suppressed issuing of a new ticket if the client
4576 presented a ticket with an expired session.
4577 - Try to set the ticket lifetime hint to something reasonable.
4578 - Make tickets shorter by excluding irrelevant information.
4579 - On the client side, don't ignore renewed tickets.
4581 *Adam Langley, Bodo Moeller (Google)*
4583 * Fix PSK session representation.
4587 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
4589 This work was sponsored by Intel.
4593 * Add GCM support to TLS library. Some custom code is needed to split
4594 the IV between the fixed (from PRF) and explicit (from TLS record)
4595 portions. This adds all GCM ciphersuites supported by RFC5288 and
4596 RFC5289. Generalise some AES* cipherstrings to include GCM and
4597 add a special AESGCM string for GCM only.
4601 * Expand range of ctrls for AES GCM. Permit setting invocation
4602 field on decrypt and retrieval of invocation field only on encrypt.
4606 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
4607 As required by RFC5289 these ciphersuites cannot be used if for
4608 versions of TLS earlier than 1.2.
4612 * For FIPS capable OpenSSL interpret a NULL default public key method
4613 as unset and return the appropriate default but do *not* set the default.
4614 This means we can return the appropriate method in applications that
4615 switch between FIPS and non-FIPS modes.
4619 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
4620 ENGINE is used then we cannot handle that in the FIPS module so we
4621 keep original code iff non-FIPS operations are allowed.
4625 * Add -attime option to openssl utilities.
4627 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
4629 * Redirect DSA and DH operations to FIPS module in FIPS mode.
4633 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
4634 FIPS EC methods unconditionally for now.
4638 * New build option no-ec2m to disable characteristic 2 code.
4642 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
4643 all cases can be covered as some introduce binary incompatibilities.
4647 * Redirect RSA operations to FIPS module including keygen,
4648 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
4652 * Add similar low level API blocking to ciphers.
4656 * Low level digest APIs are not approved in FIPS mode: any attempt
4657 to use these will cause a fatal error. Applications that *really* want
4658 to use them can use the private_* version instead.
4662 * Redirect cipher operations to FIPS module for FIPS builds.
4666 * Redirect digest operations to FIPS module for FIPS builds.
4670 * Update build system to add "fips" flag which will link in fipscanister.o
4671 for static and shared library builds embedding a signature if needed.
4675 * Output TLS supported curves in preference order instead of numerical
4676 order. This is currently hardcoded for the highest order curves first.
4677 This should be configurable so applications can judge speed vs strength.
4681 * Add TLS v1.2 server support for client authentication.
4685 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
4690 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
4691 FIPS modules versions.
4695 * Add TLS v1.2 client side support for client authentication. Keep cache
4696 of handshake records longer as we don't know the hash algorithm to use
4697 until after the certificate request message is received.
4701 * Initial TLS v1.2 client support. Add a default signature algorithms
4702 extension including all the algorithms we support. Parse new signature
4703 format in client key exchange. Relax some ECC signing restrictions for
4704 TLS v1.2 as indicated in RFC5246.
4708 * Add server support for TLS v1.2 signature algorithms extension. Switch
4709 to new signature format when needed using client digest preference.
4710 All server ciphersuites should now work correctly in TLS v1.2. No client
4711 support yet and no support for client certificates.
4715 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
4716 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
4717 ciphersuites. At present only RSA key exchange ciphersuites work with
4718 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
4719 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
4720 and version checking.
4724 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
4725 with this defined it will not be affected by any changes to ssl internal
4726 structures. Add several utility functions to allow openssl application
4727 to work with OPENSSL_NO_SSL_INTERN defined.
4731 * A long standing patch to add support for SRP from EdelWeb (Peter
4732 Sylvester and Christophe Renou) was integrated.
4733 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
4734 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
4737 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
4741 * Permit abbreviated handshakes when renegotiating using the function
4742 SSL_renegotiate_abbreviated().
4744 *Robin Seggelmann <seggelmann@fh-muenster.de>*
4746 * Add call to ENGINE_register_all_complete() to
4747 ENGINE_load_builtin_engines(), so some implementations get used
4748 automatically instead of needing explicit application support.
4752 * Add support for TLS key exporter as described in RFC5705.
4754 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
4756 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
4757 a few changes are required:
4759 Add SSL_OP_NO_TLSv1_1 flag.
4760 Add TLSv1_1 methods.
4761 Update version checking logic to handle version 1.1.
4762 Add explicit IV handling (ported from DTLS code).
4763 Add command line options to s_client/s_server.
4767 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012] ###
4769 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4770 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4771 content decryption and always return the same error. Note: this attack
4772 needs on average 2^20 messages so it only affects automated senders. The
4773 old behaviour can be re-enabled in the CMS code by setting the
4774 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4775 an MMA defence is not necessary.
4776 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4777 this issue. (CVE-2012-0884)
4781 * Fix CVE-2011-4619: make sure we really are receiving a
4782 client hello before rejecting multiple SGC restarts. Thanks to
4783 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4787 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012] ###
4789 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4790 Thanks to Antonio Martin, Enterprise Secure Access Research and
4791 Development, Cisco Systems, Inc. for discovering this bug and
4792 preparing a fix. (CVE-2012-0050)
4796 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012] ###
4798 * Nadhem Alfardan and Kenny Paterson have discovered an extension
4799 of the Vaudenay padding oracle attack on CBC mode encryption
4800 which enables an efficient plaintext recovery attack against
4801 the OpenSSL implementation of DTLS. Their attack exploits timing
4802 differences arising during decryption processing. A research
4803 paper describing this attack can be found at:
4804 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4805 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4806 Security Group at Royal Holloway, University of London
4807 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4808 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4809 for preparing the fix. (CVE-2011-4108)
4811 *Robin Seggelmann, Michael Tuexen*
4813 * Clear bytes used for block padding of SSL 3.0 records.
4816 *Adam Langley (Google)*
4818 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4819 Kadianakis <desnacked@gmail.com> for discovering this issue and
4820 Adam Langley for preparing the fix. (CVE-2011-4619)
4822 *Adam Langley (Google)*
4824 * Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4826 *Andrey Kulikov <amdeich@gmail.com>*
4828 * Prevent malformed RFC3779 data triggering an assertion failure.
4829 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4830 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4832 *Rob Austein <sra@hactrn.net>*
4834 * Improved PRNG seeding for VOS.
4836 *Paul Green <Paul.Green@stratus.com>*
4838 * Fix ssl_ciph.c set-up race.
4840 *Adam Langley (Google)*
4842 * Fix spurious failures in ecdsatest.c.
4844 *Emilia Käsper (Google)*
4846 * Fix the BIO_f_buffer() implementation (which was mixing different
4847 interpretations of the '..._len' fields).
4849 *Adam Langley (Google)*
4851 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4852 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4853 threads won't reuse the same blinding coefficients.
4855 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4856 lock to call BN_BLINDING_invert_ex, and avoids one use of
4857 BN_BLINDING_update for each BN_BLINDING structure (previously,
4858 the last update always remained unused).
4860 *Emilia Käsper (Google)*
4862 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4864 *Bob Buckholz (Google)*
4866 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011] ###
4868 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4869 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4871 *Kaspar Brand <ossl@velox.ch>*
4873 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
4874 for multi-threaded use of ECDH. (CVE-2011-3210)
4876 *Adam Langley (Google)*
4878 * Fix x509_name_ex_d2i memory leak on bad inputs.
4882 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4883 signature public key algorithm by using OID xref utilities instead.
4884 Before this you could only use some ECC ciphersuites with SHA1 only.
4888 * Add protection against ECDSA timing attacks as mentioned in the paper
4889 by Billy Bob Brumley and Nicola Tuveri, see:
4891 http://eprint.iacr.org/2011/232.pdf
4894 *Billy Bob Brumley and Nicola Tuveri*
4896 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011] ###
4898 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4900 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
4902 * Fix bug in string printing code: if *any* escaping is enabled we must
4903 escape the escape character (backslash) or the resulting string is
4908 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010] ###
4910 * Disable code workaround for ancient and obsolete Netscape browsers
4911 and servers: an attacker can use it in a ciphersuite downgrade attack.
4912 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4916 * Fixed J-PAKE implementation error, originally discovered by
4917 Sebastien Martini, further info and confirmation from Stefan
4918 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4922 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010] ###
4924 * Fix extension code to avoid race conditions which can result in a buffer
4925 overrun vulnerability: resumed sessions must not be modified as they can
4926 be shared by multiple threads. CVE-2010-3864
4930 * Fix WIN32 build system to correctly link an ENGINE directory into
4935 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010] ###
4937 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4940 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
4942 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010] ###
4944 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4945 context. The operation can be customised via the ctrl mechanism in
4946 case ENGINEs want to include additional functionality.
4950 * Tolerate yet another broken PKCS#8 key format: private key value negative.
4954 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4955 output hashes compatible with older versions of OpenSSL.
4957 *Willy Weisz <weisz@vcpc.univie.ac.at>*
4959 * Fix compression algorithm handling: if resuming a session use the
4960 compression algorithm of the resumed session instead of determining
4961 it from client hello again. Don't allow server to change algorithm.
4965 * Add load_crls() function to apps tidying load_certs() too. Add option
4966 to verify utility to allow additional CRLs to be included.
4970 * Update OCSP request code to permit adding custom headers to the request:
4971 some responders need this.
4975 * The function EVP_PKEY_sign() returns <=0 on error: check return code
4978 *Julia Lawall <julia@diku.dk>*
4980 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
4981 needlessly dereferenced structures, used obsolete functions and
4982 didn't handle all updated verify codes correctly.
4986 * Disable MD2 in the default configuration.
4990 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4991 indicate the initial BIO being pushed or popped. This makes it possible
4992 to determine whether the BIO is the one explicitly called or as a result
4993 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4994 it handles reference counts correctly and doesn't zero out the I/O bio
4995 when it is not being explicitly popped. WARNING: applications which
4996 included workarounds for the old buggy behaviour will need to be modified
4997 or they could free up already freed BIOs.
5001 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
5002 renaming to all platforms (within the 0.9.8 branch, this was
5003 done conditionally on Netware platforms to avoid a name clash).
5005 *Guenter <lists@gknw.net>*
5007 * Add ECDHE and PSK support to DTLS.
5009 *Michael Tuexen <tuexen@fh-muenster.de>*
5011 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
5016 * Add "missing" function EVP_MD_flags() (without this the only way to
5017 retrieve a digest flags is by accessing the structure directly. Update
5018 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
5019 or cipher is registered as in the "from" argument. Print out all
5020 registered digests in the dgst usage message instead of manually
5021 attempting to work them out.
5025 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
5026 this allows the use of compression and extensions. Change default cipher
5027 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
5028 by default unless an application cipher string requests it.
5032 * Alter match criteria in PKCS12_parse(). It used to try to use local
5033 key ids to find matching certificates and keys but some PKCS#12 files
5034 don't follow the (somewhat unwritten) rules and this strategy fails.
5035 Now just gather all certificates together and the first private key
5036 then look for the first certificate that matches the key.
5040 * Support use of registered digest and cipher names for dgst and cipher
5041 commands instead of having to add each one as a special case. So now
5048 openssl dgst -sha256 foo
5050 and this works for ENGINE based algorithms too.
5055 * Update Gost ENGINE to support parameter files.
5057 *Victor B. Wagner <vitus@cryptocom.ru>*
5059 * Support GeneralizedTime in ca utility.
5061 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
5063 * Enhance the hash format used for certificate directory links. The new
5064 form uses the canonical encoding (meaning equivalent names will work
5065 even if they aren't identical) and uses SHA1 instead of MD5. This form
5066 is incompatible with the older format and as a result c_rehash should
5067 be used to rebuild symbolic links.
5071 * Make PKCS#8 the default write format for private keys, replacing the
5072 traditional format. This form is standardised, more secure and doesn't
5073 include an implicit MD5 dependency.
5077 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
5078 committed to OpenSSL should pass this lot as a minimum.
5082 * Add session ticket override functionality for use by EAP-FAST.
5084 *Jouni Malinen <j@w1.fi>*
5086 * Modify HMAC functions to return a value. Since these can be implemented
5087 in an ENGINE errors can occur.
5091 * Type-checked OBJ_bsearch_ex.
5095 * Type-checked OBJ_bsearch. Also some constification necessitated
5096 by type-checking. Still to come: TXT_DB, bsearch(?),
5097 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
5102 * New function OPENSSL_gmtime_adj() to add a specific number of days and
5103 seconds to a tm structure directly, instead of going through OS
5104 specific date routines. This avoids any issues with OS routines such
5105 as the year 2038 bug. New *_adj() functions for ASN1 time structures
5106 and X509_time_adj_ex() to cover the extended range. The existing
5107 X509_time_adj() is still usable and will no longer have any date issues.
5111 * Delta CRL support. New use deltas option which will attempt to locate
5112 and search any appropriate delta CRLs available.
5114 This work was sponsored by Google.
5118 * Support for CRLs partitioned by reason code. Reorganise CRL processing
5119 code and add additional score elements. Validate alternate CRL paths
5120 as part of the CRL checking and indicate a new error "CRL path validation
5121 error" in this case. Applications wanting additional details can use
5122 the verify callback and check the new "parent" field. If this is not
5123 NULL CRL path validation is taking place. Existing applications won't
5124 see this because it requires extended CRL support which is off by
5127 This work was sponsored by Google.
5131 * Support for freshest CRL extension.
5133 This work was sponsored by Google.
5137 * Initial indirect CRL support. Currently only supported in the CRLs
5138 passed directly and not via lookup. Process certificate issuer
5139 CRL entry extension and lookup CRL entries by bother issuer name
5140 and serial number. Check and process CRL issuer entry in IDP extension.
5142 This work was sponsored by Google.
5146 * Add support for distinct certificate and CRL paths. The CRL issuer
5147 certificate is validated separately in this case. Only enabled if
5148 an extended CRL support flag is set: this flag will enable additional
5149 CRL functionality in future.
5151 This work was sponsored by Google.
5155 * Add support for policy mappings extension.
5157 This work was sponsored by Google.
5161 * Fixes to pathlength constraint, self issued certificate handling,
5162 policy processing to align with RFC3280 and PKITS tests.
5164 This work was sponsored by Google.
5168 * Support for name constraints certificate extension. DN, email, DNS
5169 and URI types are currently supported.
5171 This work was sponsored by Google.
5175 * To cater for systems that provide a pointer-based thread ID rather
5176 than numeric, deprecate the current numeric thread ID mechanism and
5177 replace it with a structure and associated callback type. This
5178 mechanism allows a numeric "hash" to be extracted from a thread ID in
5179 either case, and on platforms where pointers are larger than 'long',
5180 mixing is done to help ensure the numeric 'hash' is usable even if it
5181 can't be guaranteed unique. The default mechanism is to use "&errno"
5182 as a pointer-based thread ID to distinguish between threads.
5184 Applications that want to provide their own thread IDs should now use
5185 CRYPTO_THREADID_set_callback() to register a callback that will call
5186 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
5188 Note that ERR_remove_state() is now deprecated, because it is tied
5189 to the assumption that thread IDs are numeric. ERR_remove_state(0)
5190 to free the current thread's error state should be replaced by
5191 ERR_remove_thread_state(NULL).
5193 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
5194 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
5195 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
5196 application was previously providing a numeric thread callback that
5197 was inappropriate for distinguishing threads, then uniqueness might
5198 have been obtained with &errno that happened immediately in the
5199 intermediate development versions of OpenSSL; this is no longer the
5200 case, the numeric thread callback will now override the automatic use
5203 *Geoff Thorpe, with help from Bodo Moeller*
5205 * Initial support for different CRL issuing certificates. This covers a
5206 simple case where the self issued certificates in the chain exist and
5207 the real CRL issuer is higher in the existing chain.
5209 This work was sponsored by Google.
5213 * Removed effectively defunct crypto/store from the build.
5217 * Revamp of STACK to provide stronger type-checking. Still to come:
5218 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
5219 ASN1_STRING, CONF_VALUE.
5223 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
5224 RAM on SSL connections. This option can save about 34k per idle SSL.
5228 * Revamp of LHASH to provide stronger type-checking. Still to come:
5229 STACK, TXT_DB, bsearch, qsort.
5233 * Initial support for Cryptographic Message Syntax (aka CMS) based
5234 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
5235 support for data, signedData, compressedData, digestedData and
5236 encryptedData, envelopedData types included. Scripts to check against
5237 RFC4134 examples draft and interop and consistency checks of many
5238 content types and variants.
5242 * Add options to enc utility to support use of zlib compression BIO.
5246 * Extend mk1mf to support importing of options and assembly language
5247 files from Configure script, currently only included in VC-WIN32.
5248 The assembly language rules can now optionally generate the source
5249 files from the associated perl scripts.
5253 * Implement remaining functionality needed to support GOST ciphersuites.
5254 Interop testing has been performed using CryptoPro implementations.
5256 *Victor B. Wagner <vitus@cryptocom.ru>*
5258 * s390x assembler pack.
5262 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
5267 * Implement Opaque PRF Input TLS extension as specified in
5268 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
5269 official specification yet and no extension type assignment by
5270 IANA exists, this extension (for now) will have to be explicitly
5271 enabled when building OpenSSL by providing the extension number
5272 to use. For example, specify an option
5274 -DTLSEXT_TYPE_opaque_prf_input=0x9527
5276 to the "config" or "Configure" script to enable the extension,
5277 assuming extension number 0x9527 (which is a completely arbitrary
5278 and unofficial assignment based on the MD5 hash of the Internet
5279 Draft). Note that by doing so, you potentially lose
5280 interoperability with other TLS implementations since these might
5281 be using the same extension number for other purposes.
5283 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
5284 opaque PRF input value to use in the handshake. This will create
5285 an internal copy of the length-'len' string at 'src', and will
5286 return non-zero for success.
5288 To get more control and flexibility, provide a callback function
5291 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
5292 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
5296 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
5299 Callback function 'cb' will be called in handshakes, and is
5300 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
5301 Argument 'arg' is for application purposes (the value as given to
5302 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
5303 be provided to the callback function). The callback function
5304 has to return non-zero to report success: usually 1 to use opaque
5305 PRF input just if possible, or 2 to enforce use of the opaque PRF
5306 input. In the latter case, the library will abort the handshake
5307 if opaque PRF input is not successfully negotiated.
5309 Arguments 'peerinput' and 'len' given to the callback function
5310 will always be NULL and 0 in the case of a client. A server will
5311 see the client's opaque PRF input through these variables if
5312 available (NULL and 0 otherwise). Note that if the server
5313 provides an opaque PRF input, the length must be the same as the
5314 length of the client's opaque PRF input.
5316 Note that the callback function will only be called when creating
5317 a new session (session resumption can resume whatever was
5318 previously negotiated), and will not be called in SSL 2.0
5319 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
5320 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
5321 for applications that need to enforce opaque PRF input.
5326 * Update ssl code to support digests other than SHA1+MD5 for handshake
5330 *Victor B. Wagner <vitus@cryptocom.ru>*
5332 * Add RFC4507 support to OpenSSL. This includes the corrections in
5333 RFC4507bis. The encrypted ticket format is an encrypted encoded
5334 SSL_SESSION structure, that way new session features are automatically
5337 If a client application caches session in an SSL_SESSION structure
5338 support is transparent because tickets are now stored in the encoded
5341 The SSL_CTX structure automatically generates keys for ticket
5342 protection in servers so again support should be possible
5343 with no application modification.
5345 If a client or server wishes to disable RFC4507 support then the option
5346 SSL_OP_NO_TICKET can be set.
5348 Add a TLS extension debugging callback to allow the contents of any client
5349 or server extensions to be examined.
5351 This work was sponsored by Google.
5355 * Final changes to avoid use of pointer pointer casts in OpenSSL.
5356 OpenSSL should now compile cleanly on gcc 4.2
5358 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
5360 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
5361 support including streaming MAC support: this is required for GOST
5362 ciphersuite support.
5364 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
5366 * Add option -stream to use PKCS#7 streaming in smime utility. New
5367 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
5368 to output in BER and PEM format.
5372 * Experimental support for use of HMAC via EVP_PKEY interface. This
5373 allows HMAC to be handled via the EVP_DigestSign*() interface. The
5374 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
5375 ENGINE support for HMAC keys which are unextractable. New -mac and
5376 -macopt options to dgst utility.
5380 * New option -sigopt to dgst utility. Update dgst to use
5381 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
5382 alternative signing parameters such as X9.31 or PSS in the dgst
5387 * Change ssl_cipher_apply_rule(), the internal function that does
5388 the work each time a ciphersuite string requests enabling
5389 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
5390 removing ("!foo+bar") a class of ciphersuites: Now it maintains
5391 the order of disabled ciphersuites such that those ciphersuites
5392 that most recently went from enabled to disabled not only stay
5393 in order with respect to each other, but also have higher priority
5394 than other disabled ciphersuites the next time ciphersuites are
5397 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
5398 the same ciphersuites as with "HIGH" alone, but in a specific
5399 order where the PSK ciphersuites come first (since they are the
5400 most recently disabled ciphersuites when "HIGH" is parsed).
5402 Also, change ssl_create_cipher_list() (using this new
5403 functionality) such that between otherwise identical
5404 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
5409 * Change ssl_create_cipher_list() so that it automatically
5410 arranges the ciphersuites in reasonable order before starting
5411 to process the rule string. Thus, the definition for "DEFAULT"
5412 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
5413 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
5414 This makes it much easier to arrive at a reasonable default order
5415 in applications for which anonymous ciphers are OK (meaning
5416 that you can't actually use DEFAULT).
5418 *Bodo Moeller; suggested by Victor Duchovni*
5420 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
5421 processing) into multiple integers instead of setting
5422 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
5423 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
5424 (These masks as well as the individual bit definitions are hidden
5425 away into the non-exported interface ssl/ssl_locl.h, so this
5426 change to the definition of the SSL_CIPHER structure shouldn't
5427 affect applications.) This give us more bits for each of these
5428 categories, so there is no longer a need to coagulate AES128 and
5429 AES256 into a single algorithm bit, and to coagulate Camellia128
5430 and Camellia256 into a single algorithm bit, which has led to all
5433 Thus, among other things, the kludge introduced in 0.9.7m and
5434 0.9.8e for masking out AES256 independently of AES128 or masking
5435 out Camellia256 independently of AES256 is not needed here in 0.9.9.
5437 With the change, we also introduce new ciphersuite aliases that
5438 so far were missing: "AES128", "AES256", "CAMELLIA128", and
5443 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
5444 Use the leftmost N bytes of the signature input if the input is
5445 larger than the prime q (with N being the size in bytes of q).
5449 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
5450 it yet and it is largely untested.
5454 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
5458 * Initial incomplete changes to avoid need for function casts in OpenSSL
5459 some compilers (gcc 4.2 and later) reject their use. Safestack is
5460 reimplemented. Update ASN1 to avoid use of legacy functions.
5464 * Win32/64 targets are linked with Winsock2.
5468 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
5469 to external functions. This can be used to increase CRL handling
5470 efficiency especially when CRLs are very large by (for example) storing
5471 the CRL revoked certificates in a database.
5475 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
5476 new CRLs added to a directory can be used. New command line option
5477 -verify_return_error to s_client and s_server. This causes real errors
5478 to be returned by the verify callback instead of carrying on no matter
5479 what. This reflects the way a "real world" verify callback would behave.
5483 * GOST engine, supporting several GOST algorithms and public key formats.
5484 Kindly donated by Cryptocom.
5488 * Partial support for Issuing Distribution Point CRL extension. CRLs
5489 partitioned by DP are handled but no indirect CRL or reason partitioning
5490 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
5491 selected via a scoring technique which handles IDP and AKID in CRLs.
5495 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
5496 will ultimately be used for all verify operations: this will remove the
5497 X509_STORE dependency on certificate verification and allow alternative
5498 lookup methods. X509_STORE based implementations of these two callbacks.
5502 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
5503 Modify get_crl() to find a valid (unexpired) CRL if possible.
5507 * New function X509_CRL_match() to check if two CRLs are identical. Normally
5508 this would be called X509_CRL_cmp() but that name is already used by
5509 a function that just compares CRL issuer names. Cache several CRL
5510 extensions in X509_CRL structure and cache CRLDP in X509.
5514 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
5515 this maps equivalent X509_NAME structures into a consistent structure.
5516 Name comparison can then be performed rapidly using memcmp().
5520 * Non-blocking OCSP request processing. Add -timeout option to ocsp
5525 * Allow digests to supply their own micalg string for S/MIME type using
5526 the ctrl EVP_MD_CTRL_MICALG.
5530 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
5531 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
5532 ctrl. It can then customise the structure before and/or after signing
5537 * New function OBJ_add_sigid() to allow application defined signature OIDs
5538 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
5539 to free up any added signature OIDs.
5543 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
5544 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
5545 digest and cipher tables. New options added to openssl utility:
5546 list-message-digest-algorithms and list-cipher-algorithms.
5550 * Change the array representation of binary polynomials: the list
5551 of degrees of non-zero coefficients is now terminated with -1.
5552 Previously it was terminated with 0, which was also part of the
5553 value; thus, the array representation was not applicable to
5554 polynomials where t^0 has coefficient zero. This change makes
5555 the array representation useful in a more general context.
5559 * Various modifications and fixes to SSL/TLS cipher string
5560 handling. For ECC, the code now distinguishes between fixed ECDH
5561 with RSA certificates on the one hand and with ECDSA certificates
5562 on the other hand, since these are separate ciphersuites. The
5563 unused code for Fortezza ciphersuites has been removed.
5565 For consistency with EDH, ephemeral ECDH is now called "EECDH"
5566 (not "ECDHE"). For consistency with the code for DH
5567 certificates, use of ECDH certificates is now considered ECDH
5568 authentication, not RSA or ECDSA authentication (the latter is
5569 merely the CA's signing algorithm and not actively used in the
5572 The temporary ciphersuite alias "ECCdraft" is no longer
5573 available, and ECC ciphersuites are no longer excluded from "ALL"
5574 and "DEFAULT". The following aliases now exist for RFC 4492
5575 ciphersuites, most of these by analogy with the DH case:
5577 kECDHr - ECDH cert, signed with RSA
5578 kECDHe - ECDH cert, signed with ECDSA
5579 kECDH - ECDH cert (signed with either RSA or ECDSA)
5580 kEECDH - ephemeral ECDH
5581 ECDH - ECDH cert or ephemeral ECDH
5587 AECDH - anonymous ECDH
5588 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
5593 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
5594 Use correct micalg parameters depending on digest(s) in signed message.
5598 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
5599 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
5603 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
5604 an engine to register a method. Add ENGINE lookups for methods and
5605 functional reference processing.
5609 * New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
5610 EVP_{Sign,Verify}* which allow an application to customise the signature
5615 * New -resign option to smime utility. This adds one or more signers
5616 to an existing PKCS#7 signedData structure. Also -md option to use an
5617 alternative message digest algorithm for signing.
5621 * Tidy up PKCS#7 routines and add new functions to make it easier to
5622 create PKCS7 structures containing multiple signers. Update smime
5623 application to support multiple signers.
5627 * New -macalg option to pkcs12 utility to allow setting of an alternative
5632 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
5633 Reorganize PBE internals to lookup from a static table using NIDs,
5634 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
5635 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
5636 PRF which will be automatically used with PBES2.
5640 * Replace the algorithm specific calls to generate keys in "req" with the
5645 * Update PKCS#7 enveloped data routines to use new API. This is now
5646 supported by any public key method supporting the encrypt operation. A
5647 ctrl is added to allow the public key algorithm to examine or modify
5648 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
5653 * Add a ctrl to asn1 method to allow a public key algorithm to express
5654 a default digest type to use. In most cases this will be SHA1 but some
5655 algorithms (such as GOST) need to specify an alternative digest. The
5656 return value indicates how strong the preference is 1 means optional and
5657 2 is mandatory (that is it is the only supported type). Modify
5658 ASN1_item_sign() to accept a NULL digest argument to indicate it should
5659 use the default md. Update openssl utilities to use the default digest
5660 type for signing if it is not explicitly indicated.
5664 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
5665 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
5666 signing method from the key type. This effectively removes the link
5667 between digests and public key types.
5671 * Add an OID cross reference table and utility functions. Its purpose is to
5672 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
5673 rsaEncryption. This will allow some of the algorithm specific hackery
5674 needed to use the correct OID to be removed.
5678 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
5679 structures for PKCS7_sign(). They are now set up by the relevant public
5684 * Add provisional EC pkey method with support for ECDSA and ECDH.
5688 * Add support for key derivation (agreement) in the API, DH method and
5693 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
5694 public and private key formats. As a side effect these add additional
5695 command line functionality not previously available: DSA signatures can be
5696 generated and verified using pkeyutl and DH key support and generation in
5703 *Oliver Tappe <zooey@hirschkaefer.de>*
5705 * New make target "install_html_docs" installs HTML renditions of the
5708 *Oliver Tappe <zooey@hirschkaefer.de>*
5710 * New utility "genpkey" this is analogous to "genrsa" etc except it can
5711 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
5712 support key and parameter generation and add initial key generation
5713 functionality for RSA.
5717 * Add functions for main EVP_PKEY_method operations. The undocumented
5718 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
5719 EVP_PKEY_{encrypt,decrypt}_old.
5723 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
5724 key API, doesn't do much yet.
5728 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
5729 public key algorithms. New option to openssl utility:
5730 "list-public-key-algorithms" to print out info.
5734 * Implement the Supported Elliptic Curves Extension for
5735 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5739 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
5740 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
5744 * New utilities pkey and pkeyparam. These are similar to algorithm specific
5745 utilities such as rsa, dsa, dsaparam etc except they process any key
5750 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
5751 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
5752 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
5757 * Initial support for pluggable public key ASN1.
5758 De-spaghettify the public key ASN1 handling. Move public and private
5759 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
5760 algorithm specific handling to a single module within the relevant
5761 algorithm directory. Add functions to allow (near) opaque processing
5762 of public and private key structures.
5766 * Implement the Supported Point Formats Extension for
5767 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5771 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
5772 for the psk identity [hint] and the psk callback functions to the
5773 SSL_SESSION, SSL and SSL_CTX structure.
5776 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
5780 SSL_CTX_use_psk_identity_hint
5781 SSL_get_psk_identity_hint
5782 SSL_get_psk_identity
5783 SSL_use_psk_identity_hint
5786 *Mika Kousa and Pasi Eronen of Nokia Corporation*
5788 * Add RFC 3161 compliant time stamp request creation, response generation
5789 and response verification functionality.
5791 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
5793 * Add initial support for TLS extensions, specifically for the server_name
5794 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5795 have new members for a host name. The SSL data structure has an
5796 additional member SSL_CTX *initial_ctx so that new sessions can be
5797 stored in that context to allow for session resumption, even after the
5798 SSL has been switched to a new SSL_CTX in reaction to a client's
5799 server_name extension.
5801 New functions (subject to change):
5803 SSL_get_servername()
5804 SSL_get_servername_type()
5807 New CTRL codes and macros (subject to change):
5809 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5810 - SSL_CTX_set_tlsext_servername_callback()
5811 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5812 - SSL_CTX_set_tlsext_servername_arg()
5813 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5815 openssl s_client has a new '-servername ...' option.
5817 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5818 '-key2 ...', '-servername_fatal' (subject to change). This allows
5819 testing the HostName extension for a specific single host name ('-cert'
5820 and '-key' remain fallbacks for handshakes without HostName
5821 negotiation). If the unrecognized_name alert has to be sent, this by
5822 default is a warning; it becomes fatal with the '-servername_fatal'
5826 *Peter Sylvester, Remy Allais, Christophe Renou*
5828 * Whirlpool hash implementation is added.
5832 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
5833 bn(64,32). Because of instruction set limitations it doesn't have
5834 any negative impact on performance. This was done mostly in order
5835 to make it possible to share assembler modules, such as bn_mul_mont
5836 implementations, between 32- and 64-bit builds without hassle.
5840 * Move code previously exiled into file crypto/ec/ec2_smpt.c
5841 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
5846 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
5847 dedicated Montgomery multiplication procedure, is introduced.
5848 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
5849 "64-bit" performance on certain 32-bit targets.
5853 * New option SSL_OP_NO_COMP to disable use of compression selectively
5854 in SSL structures. New SSL ctrl to set maximum send fragment size.
5855 Save memory by setting the I/O buffer sizes dynamically instead of
5856 using the maximum available value.
5860 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
5861 in addition to the text details.
5865 * Very, very preliminary EXPERIMENTAL support for printing of general
5866 ASN1 structures. This currently produces rather ugly output and doesn't
5867 handle several customised structures at all.
5871 * Integrated support for PVK file format and some related formats such
5872 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
5873 these in the 'rsa' and 'dsa' utilities.
5877 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
5881 * Remove the ancient ASN1_METHOD code. This was only ever used in one
5882 place for the (very old) "NETSCAPE" format certificates which are now
5883 handled using new ASN1 code equivalents.
5887 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
5888 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
5889 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
5893 * Modify CRL distribution points extension code to print out previously
5894 unsupported fields. Enhance extension setting code to allow setting of
5899 * Add print and set support for Issuing Distribution Point CRL extension.
5903 * Change 'Configure' script to enable Camellia by default.
5907 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010] ###
5909 * When rejecting SSL/TLS records due to an incorrect version number, never
5910 update s->server with a new major version number. As of
5911 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5912 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5913 the previous behavior could result in a read attempt at NULL when
5914 receiving specific incorrect SSL/TLS records once record payload
5915 protection is active. (CVE-2010-0740)
5917 *Bodo Moeller, Adam Langley <agl@chromium.org>*
5919 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5920 could be crashed if the relevant tables were not present (e.g. chrooted).
5922 *Tomas Hoger <thoger@redhat.com>*
5924 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010] ###
5926 * Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5928 *Martin Olsson, Neel Mehta*
5930 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
5931 accommodate for stack sorting, always a write lock!).
5935 * On some versions of WIN32 Heap32Next is very slow. This can cause
5936 excessive delays in the RAND_poll(): over a minute. As a workaround
5937 include a time check in the inner Heap32Next loop too.
5941 * The code that handled flushing of data in SSL/TLS originally used the
5942 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5943 the problem outlined in PR#1949. The fix suggested there however can
5944 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5945 of Apache). So instead simplify the code to flush unconditionally.
5946 This should be fine since flushing with no data to flush is a no op.
5950 * Handle TLS versions 2.0 and later properly and correctly use the
5951 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5952 off ancient servers have a habit of sticking around for a while...
5956 * Modify compression code so it frees up structures without using the
5957 ex_data callbacks. This works around a problem where some applications
5958 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5959 restarting) then use compression (e.g. SSL with compression) later.
5960 This results in significant per-connection memory leaks and
5961 has caused some security issues including CVE-2008-1678 and
5966 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5967 change when encrypting or decrypting.
5971 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5972 connect and renegotiate with servers which do not support RI.
5973 Until RI is more widely deployed this option is enabled by default.
5977 * Add "missing" ssl ctrls to clear options and mode.
5981 * If client attempts to renegotiate and doesn't support RI respond with
5982 a no_renegotiation alert as required by RFC5746. Some renegotiating
5983 TLS clients will continue a connection gracefully when they receive
5984 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5985 waiting for a server hello which it will never receive. Now we treat a
5986 received no_renegotiation alert as a fatal error. This is because
5987 applications requesting a renegotiation might well expect it to succeed
5988 and would have no code in place to handle the server denying it so the
5989 only safe thing to do is to terminate the connection.
5993 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5994 peer supports secure renegotiation and 0 otherwise. Print out peer
5995 renegotiation support in s_client/s_server.
5999 * Replace the highly broken and deprecated SPKAC certification method with
6000 the updated NID creation version. This should correctly handle UTF8.
6004 * Implement RFC5746. Re-enable renegotiation but require the extension
6005 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
6006 turns out to be a bad idea. It has been replaced by
6007 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
6008 SSL_CTX_set_options(). This is really not recommended unless you
6009 know what you are doing.
6011 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
6013 * Fixes to stateless session resumption handling. Use initial_ctx when
6014 issuing and attempting to decrypt tickets in case it has changed during
6015 servername handling. Use a non-zero length session ID when attempting
6016 stateless session resumption: this makes it possible to determine if
6017 a resumption has occurred immediately after receiving server hello
6018 (several places in OpenSSL subtly assume this) instead of later in
6023 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
6024 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
6025 fixes for a few places where the return code is not checked
6028 *Julia Lawall <julia@diku.dk>*
6030 * Add --strict-warnings option to Configure script to include devteam
6031 warnings in other configurations.
6035 * Add support for --libdir option and LIBDIR variable in makefiles. This
6036 makes it possible to install openssl libraries in locations which
6037 have names other than "lib", for example "/usr/lib64" which some
6040 *Steve Henson, based on patch from Jeremy Utley*
6042 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
6043 X690 8.9.12 and can produce some misleading textual output of OIDs.
6045 *Steve Henson, reported by Dan Kaminsky*
6047 * Delete MD2 from algorithm tables. This follows the recommendation in
6048 several standards that it is not used in new applications due to
6049 several cryptographic weaknesses. For binary compatibility reasons
6050 the MD2 API is still compiled in by default.
6054 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
6059 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
6060 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
6063 *Guenter <lists@gknw.net>*
6065 * Fix the server certificate chain building code to use X509_verify_cert(),
6066 it used to have an ad-hoc builder which was unable to cope with anything
6067 other than a simple chain.
6069 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
6071 * Don't check self signed certificate signatures in X509_verify_cert()
6072 by default (a flag can override this): it just wastes time without
6073 adding any security. As a useful side effect self signed root CAs
6074 with non-FIPS digests are now usable in FIPS mode.
6078 * In dtls1_process_out_of_seq_message() the check if the current message
6079 is already buffered was missing. For every new message was memory
6080 allocated, allowing an attacker to perform an denial of service attack
6081 with sending out of seq handshake messages until there is no memory
6082 left. Additionally every future message was buffered, even if the
6083 sequence number made no sense and would be part of another handshake.
6084 So only messages with sequence numbers less than 10 in advance will be
6085 buffered. (CVE-2009-1378)
6087 *Robin Seggelmann, discovered by Daniel Mentz*
6089 * Records are buffered if they arrive with a future epoch to be
6090 processed after finishing the corresponding handshake. There is
6091 currently no limitation to this buffer allowing an attacker to perform
6092 a DOS attack with sending records with future epochs until there is no
6093 memory left. This patch adds the pqueue_size() function to determine
6094 the size of a buffer and limits the record buffer to 100 entries.
6097 *Robin Seggelmann, discovered by Daniel Mentz*
6099 * Keep a copy of frag->msg_header.frag_len so it can be used after the
6100 parent structure is freed. (CVE-2009-1379)
6104 * Handle non-blocking I/O properly in SSL_shutdown() call.
6106 *Darryl Miles <darryl-mailinglists@netbauds.net>*
6110 *Ilya O. <vrghost@gmail.com>*
6112 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009] ###
6114 * Disable renegotiation completely - this fixes a severe security
6115 problem (CVE-2009-3555) at the cost of breaking all
6116 renegotiation. Renegotiation can be re-enabled by setting
6117 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
6118 run-time. This is really not recommended unless you know what
6123 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009] ###
6125 * Don't set val to NULL when freeing up structures, it is freed up by
6126 underlying code. If sizeof(void *) > sizeof(long) this can result in
6127 zeroing past the valid field. (CVE-2009-0789)
6129 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
6131 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
6132 checked correctly. This would allow some invalid signed attributes to
6133 appear to verify correctly. (CVE-2009-0591)
6135 *Ivan Nestlerode <inestlerode@us.ibm.com>*
6137 * Reject UniversalString and BMPString types with invalid lengths. This
6138 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
6139 a legal length. (CVE-2009-0590)
6143 * Set S/MIME signing as the default purpose rather than setting it
6144 unconditionally. This allows applications to override it at the store
6149 * Permit restricted recursion of ASN1 strings. This is needed in practice
6150 to handle some structures.
6154 * Improve efficiency of mem_gets: don't search whole buffer each time
6157 *Jeremy Shapiro <jnshapir@us.ibm.com>*
6159 * New -hex option for openssl rand.
6163 * Print out UTF8String and NumericString when parsing ASN1.
6167 * Support NumericString type for name components.
6171 * Allow CC in the environment to override the automatically chosen
6172 compiler. Note that nothing is done to ensure flags work with the
6177 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009] ###
6179 * Properly check EVP_VerifyFinal() and similar return values
6182 *Ben Laurie, Bodo Moeller, Google Security Team*
6184 * Enable TLS extensions by default.
6188 * Allow the CHIL engine to be loaded, whether the application is
6189 multithreaded or not. (This does not release the developer from the
6190 obligation to set up the dynamic locking callbacks.)
6192 *Sander Temme <sander@temme.net>*
6194 * Use correct exit code if there is an error in dgst command.
6196 *Steve Henson; problem pointed out by Roland Dirlewanger*
6198 * Tweak Configure so that you need to say "experimental-jpake" to enable
6199 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
6203 * Add experimental JPAKE support, including demo authentication in
6204 s_client and s_server.
6208 * Set the comparison function in v3_addr_canonize().
6210 *Rob Austein <sra@hactrn.net>*
6212 * Add support for XMPP STARTTLS in s_client.
6214 *Philip Paeps <philip@freebsd.org>*
6216 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
6217 to ensure that even with this option, only ciphersuites in the
6218 server's preference list will be accepted. (Note that the option
6219 applies only when resuming a session, so the earlier behavior was
6220 just about the algorithm choice for symmetric cryptography.)
6224 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008] ###
6226 * Fix NULL pointer dereference if a DTLS server received
6227 ChangeCipherSpec as first record (CVE-2009-1386).
6231 * Fix a state transition in s3_srvr.c and d1_srvr.c
6232 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
6236 * The fix in 0.9.8c that supposedly got rid of unsafe
6237 double-checked locking was incomplete for RSA blinding,
6238 addressing just one layer of what turns out to have been
6239 doubly unsafe triple-checked locking.
6241 So now fix this for real by retiring the MONT_HELPER macro
6242 in crypto/rsa/rsa_eay.c.
6245 *Bodo Moeller; problem pointed out by Marius Schilder*
6247 * Various precautionary measures:
6249 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
6251 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
6252 (NB: This would require knowledge of the secret session ticket key
6253 to exploit, in which case you'd be SOL either way.)
6255 - Change bn_nist.c so that it will properly handle input BIGNUMs
6256 outside the expected range.
6258 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
6262 *Neel Mehta, Bodo Moeller*
6264 * Allow engines to be "soft loaded" - i.e. optionally don't die if
6265 the load fails. Useful for distros.
6267 *Ben Laurie and the FreeBSD team*
6269 * Add support for Local Machine Keyset attribute in PKCS#12 files.
6273 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
6277 * Expand ENGINE to support engine supplied SSL client certificate functions.
6279 This work was sponsored by Logica.
6283 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
6284 keystores. Support for SSL/TLS client authentication too.
6285 Not compiled unless enable-capieng specified to Configure.
6287 This work was sponsored by Logica.
6291 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
6292 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
6293 attribute creation routines such as certificate requests and PKCS#12
6298 ### Changes between 0.9.8g and 0.9.8h [28 May 2008] ###
6300 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
6301 handshake which could lead to a client crash as found using the
6302 Codenomicon TLS test suite (CVE-2008-1672)
6304 *Steve Henson, Mark Cox*
6306 * Fix double free in TLS server name extensions which could lead to
6307 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
6311 * Clear error queue in SSL_CTX_use_certificate_chain_file()
6313 Clear the error queue to ensure that error entries left from
6314 older function calls do not interfere with the correct operation.
6316 *Lutz Jaenicke, Erik de Castro Lopo*
6318 * Remove root CA certificates of commercial CAs:
6320 The OpenSSL project does not recommend any specific CA and does not
6321 have any policy with respect to including or excluding any CA.
6322 Therefore it does not make any sense to ship an arbitrary selection
6323 of root CA certificates with the OpenSSL software.
6327 * RSA OAEP patches to fix two separate invalid memory reads.
6328 The first one involves inputs when 'lzero' is greater than
6329 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
6330 before the beginning of from). The second one involves inputs where
6331 the 'db' section contains nothing but zeroes (there is a one-byte
6332 invalid read after the end of 'db').
6334 *Ivan Nestlerode <inestlerode@us.ibm.com>*
6336 * Partial backport from 0.9.9-dev:
6338 Introduce bn_mul_mont (dedicated Montgomery multiplication
6339 procedure) as a candidate for BIGNUM assembler implementation.
6340 While 0.9.9-dev uses assembler for various architectures, only
6341 x86_64 is available by default here in the 0.9.8 branch, and
6342 32-bit x86 is available through a compile-time setting.
6344 To try the 32-bit x86 assembler implementation, use Configure
6345 option "enable-montasm" (which exists only for this backport).
6347 As "enable-montasm" for 32-bit x86 disclaims code stability
6348 anyway, in this constellation we activate additional code
6349 backported from 0.9.9-dev for further performance improvements,
6350 namely BN_from_montgomery_word. (To enable this otherwise,
6351 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
6354 *Andy Polyakov (backport partially by Bodo Moeller)*
6356 * Add TLS session ticket callback. This allows an application to set
6357 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
6358 values. This is useful for key rollover for example where several key
6359 sets may exist with different names.
6363 * Reverse ENGINE-internal logic for caching default ENGINE handles.
6364 This was broken until now in 0.9.8 releases, such that the only way
6365 a registered ENGINE could be used (assuming it initialises
6366 successfully on the host) was to explicitly set it as the default
6367 for the relevant algorithms. This is in contradiction with 0.9.7
6368 behaviour and the documentation. With this fix, when an ENGINE is
6369 registered into a given algorithm's table of implementations, the
6370 'uptodate' flag is reset so that auto-discovery will be used next
6371 time a new context for that algorithm attempts to select an
6374 *Ian Lister (tweaked by Geoff Thorpe)*
6376 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
6377 implementation in the following ways:
6379 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
6382 Lack of BER streaming support means one pass streaming processing is
6383 only supported if data is detached: setting the streaming flag is
6384 ignored for embedded content.
6386 CMS support is disabled by default and must be explicitly enabled
6387 with the enable-cms configuration option.
6391 * Update the GMP engine glue to do direct copies between BIGNUM and
6392 mpz_t when openssl and GMP use the same limb size. Otherwise the
6393 existing "conversion via a text string export" trick is still used.
6395 *Paul Sheer <paulsheer@gmail.com>*
6397 * Zlib compression BIO. This is a filter BIO which compressed and
6398 uncompresses any data passed through it.
6402 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
6403 RFC3394 compatible AES key wrapping.
6407 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
6408 sets string data without copying. X509_ALGOR_set0() and
6409 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
6410 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
6411 from an X509_ATTRIBUTE structure optionally checking it occurs only
6412 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
6417 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
6418 to get the expected BN_FLG_CONSTTIME behavior.
6420 *Bodo Moeller (Google)*
6424 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
6425 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
6426 - added some more tests to do_tests.pl
6427 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
6428 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
6429 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
6430 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
6431 - various changes to netware.pl to enable gcc-cross builds on Win32
6433 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
6434 - various changes to fix missing prototype warnings
6435 - fixed x86nasm.pl to create correct asm files for NASM COFF output
6436 - added AES, WHIRLPOOL and CPUID assembler code to build files
6437 - added missing AES assembler make rules to mk1mf.pl
6438 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
6440 *Guenter Knauf <eflash@gmx.net>*
6442 * Implement certificate status request TLS extension defined in RFC3546.
6443 A client can set the appropriate parameters and receive the encoded
6444 OCSP response via a callback. A server can query the supplied parameters
6445 and set the encoded OCSP response in the callback. Add simplified examples
6446 to s_client and s_server.
6450 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007] ###
6453 + Binary incompatibility of ssl_ctx_st structure
6454 + DTLS interoperation with non-compliant servers
6455 + Don't call get_session_cb() without proposed session
6456 + Fix ia64 assembler code
6458 *Andy Polyakov, Steve Henson*
6460 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007] ###
6462 * DTLS Handshake overhaul. There were longstanding issues with
6463 OpenSSL DTLS implementation, which were making it impossible for
6464 RFC 4347 compliant client to communicate with OpenSSL server.
6465 Unfortunately just fixing these incompatibilities would "cut off"
6466 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
6467 server keeps tolerating non RFC compliant syntax. The opposite is
6468 not true, 0.9.8f client can not communicate with earlier server.
6469 This update even addresses CVE-2007-4995.
6473 * Changes to avoid need for function casts in OpenSSL: some compilers
6474 (gcc 4.2 and later) reject their use.
6475 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
6478 * Add RFC4507 support to OpenSSL. This includes the corrections in
6479 RFC4507bis. The encrypted ticket format is an encrypted encoded
6480 SSL_SESSION structure, that way new session features are automatically
6483 If a client application caches session in an SSL_SESSION structure
6484 support is transparent because tickets are now stored in the encoded
6487 The SSL_CTX structure automatically generates keys for ticket
6488 protection in servers so again support should be possible
6489 with no application modification.
6491 If a client or server wishes to disable RFC4507 support then the option
6492 SSL_OP_NO_TICKET can be set.
6494 Add a TLS extension debugging callback to allow the contents of any client
6495 or server extensions to be examined.
6497 This work was sponsored by Google.
6501 * Add initial support for TLS extensions, specifically for the server_name
6502 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
6503 have new members for a host name. The SSL data structure has an
6504 additional member SSL_CTX *initial_ctx so that new sessions can be
6505 stored in that context to allow for session resumption, even after the
6506 SSL has been switched to a new SSL_CTX in reaction to a client's
6507 server_name extension.
6509 New functions (subject to change):
6511 SSL_get_servername()
6512 SSL_get_servername_type()
6515 New CTRL codes and macros (subject to change):
6517 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
6518 - SSL_CTX_set_tlsext_servername_callback()
6519 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
6520 - SSL_CTX_set_tlsext_servername_arg()
6521 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
6523 openssl s_client has a new '-servername ...' option.
6525 openssl s_server has new options '-servername_host ...', '-cert2 ...',
6526 '-key2 ...', '-servername_fatal' (subject to change). This allows
6527 testing the HostName extension for a specific single host name ('-cert'
6528 and '-key' remain fallbacks for handshakes without HostName
6529 negotiation). If the unrecognized_name alert has to be sent, this by
6530 default is a warning; it becomes fatal with the '-servername_fatal'
6534 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
6536 * Add AES and SSE2 assembly language support to VC++ build.
6540 * Mitigate attack on final subtraction in Montgomery reduction.
6544 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
6545 (which previously caused an internal error).
6549 * Squeeze another 10% out of IGE mode when in != out.
6553 * AES IGE mode speedup.
6555 *Dean Gaudet (Google)*
6557 * Add the Korean symmetric 128-bit cipher SEED (see
6558 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
6559 add SEED ciphersuites from RFC 4162:
6561 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
6562 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
6563 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
6564 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
6566 To minimize changes between patchlevels in the OpenSSL 0.9.8
6567 series, SEED remains excluded from compilation unless OpenSSL
6568 is configured with 'enable-seed'.
6570 *KISA, Bodo Moeller*
6572 * Mitigate branch prediction attacks, which can be practical if a
6573 single processor is shared, allowing a spy process to extract
6574 information. For detailed background information, see
6575 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
6576 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
6577 and Necessary Software Countermeasures"). The core of the change
6578 are new versions BN_div_no_branch() and
6579 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
6580 respectively, which are slower, but avoid the security-relevant
6581 conditional branches. These are automatically called by BN_div()
6582 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
6583 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
6584 remove a conditional branch.
6586 BN_FLG_CONSTTIME is the new name for the previous
6587 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
6588 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
6589 in the exponent causes BN_mod_exp_mont() to use the alternative
6590 implementation in BN_mod_exp_mont_consttime().) The old name
6591 remains as a deprecated alias.
6593 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
6594 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
6595 constant-time implementations for more than just exponentiation.
6596 Here too the old name is kept as a deprecated alias.
6598 BN_BLINDING_new() will now use BN_dup() for the modulus so that
6599 the BN_BLINDING structure gets an independent copy of the
6600 modulus. This means that the previous "BIGNUM *m" argument to
6601 BN_BLINDING_new() and to BN_BLINDING_create_param() now
6602 essentially becomes "const BIGNUM *m", although we can't actually
6603 change this in the header file before 0.9.9. It allows
6604 RSA_setup_blinding() to use BN_with_flags() on the modulus to
6605 enable BN_FLG_CONSTTIME.
6608 *Matthew D Wood (Intel Corp)*
6610 * In the SSL/TLS server implementation, be strict about session ID
6611 context matching (which matters if an application uses a single
6612 external cache for different purposes). Previously,
6613 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
6614 set. This did ensure strict client verification, but meant that,
6615 with applications using a single external cache for quite
6616 different requirements, clients could circumvent ciphersuite
6617 restrictions for a given session ID context by starting a session
6618 in a different context.
6622 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6623 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6624 authentication-only ciphersuites.
6628 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
6629 not complete and could lead to a possible single byte overflow
6630 (CVE-2007-5135) [Ben Laurie]
6632 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007] ###
6634 * Since AES128 and AES256 (and similarly Camellia128 and
6635 Camellia256) share a single mask bit in the logic of
6636 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6637 kludge to work properly if AES128 is available and AES256 isn't
6638 (or if Camellia128 is available and Camellia256 isn't).
6642 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
6643 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
6644 When a point or a seed is encoded in a BIT STRING, we need to
6645 prevent the removal of trailing zero bits to get the proper DER
6646 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
6647 of a NamedBitList, for which trailing 0 bits need to be removed.)
6651 * Have SSL/TLS server implementation tolerate "mismatched" record
6652 protocol version while receiving ClientHello even if the
6653 ClientHello is fragmented. (The server can't insist on the
6654 particular protocol version it has chosen before the ServerHello
6655 message has informed the client about his choice.)
6659 * Add RFC 3779 support.
6661 *Rob Austein for ARIN, Ben Laurie*
6663 * Load error codes if they are not already present instead of using a
6664 static variable. This allows them to be cleanly unloaded and reloaded.
6665 Improve header file function name parsing.
6669 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
6670 or CAPABILITY handshake as required by RFCs.
6674 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006] ###
6676 * Introduce limits to prevent malicious keys being able to
6677 cause a denial of service. (CVE-2006-2940)
6679 *Steve Henson, Bodo Moeller*
6681 * Fix ASN.1 parsing of certain invalid structures that can result
6682 in a denial of service. (CVE-2006-2937) [Steve Henson]
6684 * Fix buffer overflow in SSL_get_shared_ciphers() function.
6685 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6687 * Fix SSL client code which could crash if connecting to a
6688 malicious SSLv2 server. (CVE-2006-4343)
6690 *Tavis Ormandy and Will Drewry, Google Security Team*
6692 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
6693 match only those. Before that, "AES256-SHA" would be interpreted
6694 as a pattern and match "AES128-SHA" too (since AES128-SHA got
6695 the same strength classification in 0.9.7h) as we currently only
6696 have a single AES bit in the ciphersuite description bitmap.
6697 That change, however, also applied to ciphersuite strings such as
6698 "RC4-MD5" that intentionally matched multiple ciphersuites --
6699 namely, SSL 2.0 ciphersuites in addition to the more common ones
6700 from SSL 3.0/TLS 1.0.
6702 So we change the selection algorithm again: Naming an explicit
6703 ciphersuite selects this one ciphersuite, and any other similar
6704 ciphersuite (same bitmap) from *other* protocol versions.
6705 Thus, "RC4-MD5" again will properly select both the SSL 2.0
6706 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
6708 Since SSL 2.0 does not have any ciphersuites for which the
6709 128/256 bit distinction would be relevant, this works for now.
6710 The proper fix will be to use different bits for AES128 and
6711 AES256, which would have avoided the problems from the beginning;
6712 however, bits are scarce, so we can only do this in a new release
6713 (not just a patchlevel) when we can change the SSL_CIPHER
6714 definition to split the single 'unsigned long mask' bitmap into
6715 multiple values to extend the available space.
6720 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006] ###
6722 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6723 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6725 * Add AES IGE and biIGE modes.
6729 * Change the Unix randomness entropy gathering to use poll() when
6730 possible instead of select(), since the latter has some
6731 undesirable limitations.
6733 *Darryl Miles via Richard Levitte and Bodo Moeller*
6735 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
6736 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
6737 cannot be implicitly activated as part of, e.g., the "AES" alias.
6738 However, please upgrade to OpenSSL 0.9.9[-dev] for
6739 non-experimental use of the ECC ciphersuites to get TLS extension
6740 support, which is required for curve and point format negotiation
6741 to avoid potential handshake problems.
6745 * Disable rogue ciphersuites:
6747 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6748 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6749 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6751 The latter two were purportedly from
6752 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6755 Also deactivate the remaining ciphersuites from
6756 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6757 unofficial, and the ID has long expired.
6761 * Fix RSA blinding Heisenbug (problems sometimes occurred on
6762 dual-core machines) and other potential thread-safety issues.
6766 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
6767 versions), which is now available for royalty-free use
6768 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
6769 Also, add Camellia TLS ciphersuites from RFC 4132.
6771 To minimize changes between patchlevels in the OpenSSL 0.9.8
6772 series, Camellia remains excluded from compilation unless OpenSSL
6773 is configured with 'enable-camellia'.
6777 * Disable the padding bug check when compression is in use. The padding
6778 bug check assumes the first packet is of even length, this is not
6779 necessarily true if compression is enabled and can result in false
6780 positives causing handshake failure. The actual bug test is ancient
6781 code so it is hoped that implementations will either have fixed it by
6782 now or any which still have the bug do not support compression.
6786 ### Changes between 0.9.8a and 0.9.8b [04 May 2006] ###
6788 * When applying a cipher rule check to see if string match is an explicit
6789 cipher suite and only match that one cipher suite if it is.
6793 * Link in manifests for VC++ if needed.
6795 *Austin Ziegler <halostatue@gmail.com>*
6797 * Update support for ECC-based TLS ciphersuites according to
6798 draft-ietf-tls-ecc-12.txt with proposed changes (but without
6799 TLS extensions, which are supported starting with the 0.9.9
6800 branch, not in the OpenSSL 0.9.8 branch).
6804 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
6805 opaque EVP_CIPHER_CTX handling.
6809 * Fixes and enhancements to zlib compression code. We now only use
6810 "zlib1.dll" and use the default __cdecl calling convention on Win32
6811 to conform with the standards mentioned here:
6812 http://www.zlib.net/DLL_FAQ.txt
6813 Static zlib linking now works on Windows and the new --with-zlib-include
6814 --with-zlib-lib options to Configure can be used to supply the location
6815 of the headers and library. Gracefully handle case where zlib library
6820 * Several fixes and enhancements to the OID generation code. The old code
6821 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
6822 handle numbers larger than ULONG_MAX, truncated printing and had a
6823 non standard OBJ_obj2txt() behaviour.
6827 * Add support for building of engines under engine/ as shared libraries
6828 under VC++ build system.
6832 * Corrected the numerous bugs in the Win32 path splitter in DSO.
6833 Hopefully, we will not see any false combination of paths any more.
6837 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005] ###
6839 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6840 (part of SSL_OP_ALL). This option used to disable the
6841 countermeasure against man-in-the-middle protocol-version
6842 rollback in the SSL 2.0 server implementation, which is a bad
6843 idea. (CVE-2005-2969)
6845 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6846 for Information Security, National Institute of Advanced Industrial
6847 Science and Technology [AIST*, Japan)]
6849 * Add two function to clear and return the verify parameter flags.
6853 * Keep cipherlists sorted in the source instead of sorting them at
6854 runtime, thus removing the need for a lock.
6858 * Avoid some small subgroup attacks in Diffie-Hellman.
6860 *Nick Mathewson and Ben Laurie*
6862 * Add functions for well-known primes.
6866 * Extended Windows CE support.
6868 *Satoshi Nakamura and Andy Polyakov*
6870 * Initialize SSL_METHOD structures at compile time instead of during
6871 runtime, thus removing the need for a lock.
6875 * Make PKCS7_decrypt() work even if no certificate is supplied by
6876 attempting to decrypt each encrypted key in turn. Add support to
6881 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005] ###
6883 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
6886 * Add libcrypto.pc and libssl.pc for those who feel they need them.
6890 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
6891 key into the same file any more.
6895 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
6899 * Add -utf8 command line and config file option to 'ca'.
6901 *Stefan <stf@udoma.org*
6903 * Removed the macro des_crypt(), as it seems to conflict with some
6904 libraries. Use DES_crypt().
6908 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
6909 involves renaming the source and generated shared-libs for
6910 both. The engines will accept the corrected or legacy ids
6911 ('ncipher' and '4758_cca' respectively) when binding. NB,
6912 this only applies when building 'shared'.
6914 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
6916 * Add attribute functions to EVP_PKEY structure. Modify
6917 PKCS12_create() to recognize a CSP name attribute and
6918 use it. Make -CSP option work again in pkcs12 utility.
6922 * Add new functionality to the bn blinding code:
6923 - automatic re-creation of the BN_BLINDING parameters after
6924 a fixed number of uses (currently 32)
6925 - add new function for parameter creation
6926 - introduce flags to control the update behaviour of the
6927 BN_BLINDING parameters
6928 - hide BN_BLINDING structure
6929 Add a second BN_BLINDING slot to the RSA structure to improve
6930 performance when a single RSA object is shared among several
6935 * Add support for DTLS.
6937 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
6939 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
6940 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
6944 * Remove buggy and incomplete DH cert support from
6945 ssl/ssl_rsa.c and ssl/s3_both.c
6949 * Use SHA-1 instead of MD5 as the default digest algorithm for
6950 the apps/openssl applications.
6954 * Compile clean with "-Wall -Wmissing-prototypes
6955 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
6956 DEBUG_SAFESTACK must also be set.
6960 * Change ./Configure so that certain algorithms can be disabled by default.
6961 The new counterpiece to "no-xxx" is "enable-xxx".
6963 The patented RC5 and MDC2 algorithms will now be disabled unless
6964 "enable-rc5" and "enable-mdc2", respectively, are specified.
6966 (IDEA remains enabled despite being patented. This is because IDEA
6967 is frequently required for interoperability, and there is no license
6968 fee for non-commercial use. As before, "no-idea" can be used to
6969 avoid this algorithm.)
6974 * Add processing of proxy certificates (see RFC 3820). This work was
6975 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
6976 EGEE (Enabling Grids for E-science in Europe).
6980 * RC4 performance overhaul on modern architectures/implementations, such
6981 as Intel P4, IA-64 and AMD64.
6985 * New utility extract-section.pl. This can be used specify an alternative
6986 section number in a pod file instead of having to treat each file as
6987 a separate case in Makefile. This can be done by adding two lines to the
6990 =for comment openssl_section:XXX
6992 The blank line is mandatory.
6997 * New arguments -certform, -keyform and -pass for s_client and s_server
6998 to allow alternative format key and certificate files and passphrase
7003 * New structure X509_VERIFY_PARAM which combines current verify parameters,
7004 update associated structures and add various utility functions.
7006 Add new policy related verify parameters, include policy checking in
7007 standard verify code. Enhance 'smime' application with extra parameters
7008 to support policy checking and print out.
7012 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
7013 Nehemiah processors. These extensions support AES encryption in hardware
7014 as well as RNG (though RNG support is currently disabled).
7016 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
7018 * Deprecate BN_[get|set]_params() functions (they were ignored internally).
7022 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
7024 *Andy Polyakov and a number of other people*
7026 * Improved PowerPC platform support. Most notably BIGNUM assembler
7027 implementation contributed by IBM.
7029 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
7031 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
7032 exponent rather than 'unsigned long'. There is a corresponding change to
7033 the new 'rsa_keygen' element of the RSA_METHOD structure.
7035 *Jelte Jansen, Geoff Thorpe*
7037 * Functionality for creating the initial serial number file is now
7038 moved from CA.pl to the 'ca' utility with a new option -create_serial.
7040 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
7041 number file to 1, which is bound to cause problems. To avoid
7042 the problems while respecting compatibility between different 0.9.7
7043 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
7044 CA.pl for serial number initialization. With the new release 0.9.8,
7045 we can fix the problem directly in the 'ca' utility.)
7049 * Reduced header interdependencies by declaring more opaque objects in
7050 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
7051 give fewer recursive includes, which could break lazy source code - so
7052 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
7053 developers should define this symbol when building and using openssl to
7054 ensure they track the recommended behaviour, interfaces, [etc], but
7055 backwards-compatible behaviour prevails when this isn't defined.
7059 * New function X509_POLICY_NODE_print() which prints out policy nodes.
7063 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
7064 This will generate a random key of the appropriate length based on the
7065 cipher context. The EVP_CIPHER can provide its own random key generation
7066 routine to support keys of a specific form. This is used in the des and
7067 3des routines to generate a key of the correct parity. Update S/MIME
7068 code to use new functions and hence generate correct parity DES keys.
7069 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
7070 valid (weak or incorrect parity).
7074 * Add a local set of CRLs that can be used by X509_verify_cert() as well
7075 as looking them up. This is useful when the verified structure may contain
7076 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
7077 present unless the new PKCS7_NO_CRL flag is asserted.
7081 * Extend ASN1 oid configuration module. It now additionally accepts the
7084 shortName = some long name, 1.2.3.4
7088 * Reimplemented the BN_CTX implementation. There is now no more static
7089 limitation on the number of variables it can handle nor the depth of the
7090 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
7091 information can now expand as required, and rather than having a single
7092 static array of bignums, BN_CTX now uses a linked-list of such arrays
7093 allowing it to expand on demand whilst maintaining the usefulness of
7094 BN_CTX's "bundling".
7098 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
7099 to allow all RSA operations to function using a single BN_CTX.
7103 * Preliminary support for certificate policy evaluation and checking. This
7104 is initially intended to pass the tests outlined in "Conformance Testing
7105 of Relying Party Client Certificate Path Processing Logic" v1.07.
7109 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
7110 remained unused and not that useful. A variety of other little bignum
7111 tweaks and fixes have also been made continuing on from the audit (see
7116 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
7117 associated ASN1, EVP and SSL functions and old ASN1 macros.
7121 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
7122 and this should never fail. So the return value from the use of
7123 BN_set_word() (which can fail due to needless expansion) is now deprecated;
7124 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
7128 * BN_CTX_get() should return zero-valued bignums, providing the same
7129 initialised value as BN_new().
7131 *Geoff Thorpe, suggested by Ulf Möller*
7133 * Support for inhibitAnyPolicy certificate extension.
7137 * An audit of the BIGNUM code is underway, for which debugging code is
7138 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
7139 is considered valid when processing BIGNUMs, and causes execution to
7140 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
7141 further steps are taken to deliberately pollute unused data in BIGNUM
7142 structures to try and expose faulty code further on. For now, openssl will
7143 (in its default mode of operation) continue to tolerate the inconsistent
7144 forms that it has tolerated in the past, but authors and packagers should
7145 consider trying openssl and their own applications when compiled with
7146 these debugging symbols defined. It will help highlight potential bugs in
7147 their own code, and will improve the test coverage for OpenSSL itself. At
7148 some point, these tighter rules will become openssl's default to improve
7149 maintainability, though the assert()s and other overheads will remain only
7150 in debugging configurations. See bn.h for more details.
7152 *Geoff Thorpe, Nils Larsch, Ulf Möller*
7154 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
7155 that can only be obtained through BN_CTX_new() (which implicitly
7156 initialises it). The presence of this function only made it possible
7157 to overwrite an existing structure (and cause memory leaks).
7161 * Because of the callback-based approach for implementing LHASH as a
7162 template type, lh_insert() adds opaque objects to hash-tables and
7163 lh_doall() or lh_doall_arg() are typically used with a destructor callback
7164 to clean up those corresponding objects before destroying the hash table
7165 (and losing the object pointers). So some over-zealous constifications in
7166 LHASH have been relaxed so that lh_insert() does not take (nor store) the
7167 objects as "const" and the lh_doall[_arg] callback wrappers are not
7168 prototyped to have "const" restrictions on the object pointers they are
7169 given (and so aren't required to cast them away any more).
7173 * The tmdiff.h API was so ugly and minimal that our own timing utility
7174 (speed) prefers to use its own implementation. The two implementations
7175 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
7176 its object type properly exposed (MS_TM) instead of casting to/from "char
7177 *". This may still change yet if someone realises MS_TM and "ms_time_***"
7178 aren't necessarily the greatest nomenclatures - but this is what was used
7179 internally to the implementation so I've used that for now.
7183 * Ensure that deprecated functions do not get compiled when
7184 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
7185 the self-tests were still using deprecated key-generation functions so
7186 these have been updated also.
7190 * Reorganise PKCS#7 code to separate the digest location functionality
7191 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
7192 New function PKCS7_set_digest() to set the digest type for PKCS#7
7193 digestedData type. Add additional code to correctly generate the
7194 digestedData type and add support for this type in PKCS7 initialization
7199 * New function PKCS7_set0_type_other() this initializes a PKCS7
7200 structure of type "other".
7204 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
7205 sure the loop does correctly stop and breaking ("division by zero")
7206 modulus operations are not performed. The (pre-generated) prime
7207 table crypto/bn/bn_prime.h was already correct, but it could not be
7208 re-generated on some platforms because of the "division by zero"
7209 situation in the script.
7211 *Ralf S. Engelschall*
7213 * Update support for ECC-based TLS ciphersuites according to
7214 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
7215 SHA-1 now is only used for "small" curves (where the
7216 representation of a field element takes up to 24 bytes); for
7217 larger curves, the field element resulting from ECDH is directly
7218 used as premaster secret.
7220 *Douglas Stebila (Sun Microsystems Laboratories)*
7222 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
7223 curve secp160r1 to the tests.
7225 *Douglas Stebila (Sun Microsystems Laboratories)*
7227 * Add the possibility to load symbols globally with DSO.
7229 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
7231 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
7232 control of the error stack.
7236 * Add support for STORE in ENGINE.
7240 * Add the STORE type. The intention is to provide a common interface
7241 to certificate and key stores, be they simple file-based stores, or
7242 HSM-type store, or LDAP stores, or...
7243 NOTE: The code is currently UNTESTED and isn't really used anywhere.
7247 * Add a generic structure called OPENSSL_ITEM. This can be used to
7248 pass a list of arguments to any function as well as provide a way
7249 for a function to pass data back to the caller.
7253 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
7254 works like BUF_strdup() but can be used to duplicate a portion of
7255 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
7260 * Add the function sk_find_ex() which works like sk_find(), but will
7261 return an index to an element even if an exact match couldn't be
7262 found. The index is guaranteed to point at the element where the
7263 searched-for key would be inserted to preserve sorting order.
7267 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
7268 takes an extra flags argument for optional functionality. Currently,
7269 the following flags are defined:
7271 OBJ_BSEARCH_VALUE_ON_NOMATCH
7272 This one gets OBJ_bsearch_ex() to return a pointer to the first
7273 element where the comparing function returns a negative or zero
7276 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
7277 This one gets OBJ_bsearch_ex() to return a pointer to the first
7278 element where the comparing function returns zero. This is useful
7279 if there are more than one element where the comparing function
7284 * Make it possible to create self-signed certificates with 'openssl ca'
7285 in such a way that the self-signed certificate becomes part of the
7286 CA database and uses the same mechanisms for serial number generation
7287 as all other certificate signing. The new flag '-selfsign' enables
7288 this functionality. Adapt CA.sh and CA.pl.in.
7292 * Add functionality to check the public key of a certificate request
7293 against a given private. This is useful to check that a certificate
7294 request can be signed by that key (self-signing).
7298 * Make it possible to have multiple active certificates with the same
7299 subject in the CA index file. This is done only if the keyword
7300 'unique_subject' is set to 'no' in the main CA section (default
7301 if 'CA_default') of the configuration file. The value is saved
7302 with the database itself in a separate index attribute file,
7303 named like the index file with '.attr' appended to the name.
7307 * Generate multi-valued AVAs using '+' notation in config files for
7312 * Support for nameConstraints certificate extension.
7316 * Support for policyConstraints certificate extension.
7320 * Support for policyMappings certificate extension.
7324 * Make sure the default DSA_METHOD implementation only uses its
7325 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
7326 and change its own handlers to be NULL so as to remove unnecessary
7327 indirection. This lets alternative implementations fallback to the
7328 default implementation more easily.
7332 * Support for directoryName in GeneralName related extensions
7337 * Make it possible to link applications using Makefile.shared.
7338 Make that possible even when linking against static libraries!
7342 * Support for single pass processing for S/MIME signing. This now
7343 means that S/MIME signing can be done from a pipe, in addition
7344 cleartext signing (multipart/signed type) is effectively streaming
7345 and the signed data does not need to be all held in memory.
7347 This is done with a new flag PKCS7_STREAM. When this flag is set
7348 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
7349 is done after the data is output (and digests calculated) in
7350 SMIME_write_PKCS7().
7354 * Add full support for -rpath/-R, both in shared libraries and
7355 applications, at least on the platforms where it's known how
7360 * In crypto/ec/ec_mult.c, implement fast point multiplication with
7361 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
7362 will now compute a table of multiples of the generator that
7363 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
7364 faster (notably in the case of a single point multiplication,
7365 scalar * generator).
7367 *Nils Larsch, Bodo Moeller*
7369 * IPv6 support for certificate extensions. The various extensions
7370 which use the IP:a.b.c.d can now take IPv6 addresses using the
7371 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
7376 * Added an ENGINE that implements RSA by performing private key
7377 exponentiations with the GMP library. The conversions to and from
7378 GMP's mpz_t format aren't optimised nor are any montgomery forms
7379 cached, and on x86 it appears OpenSSL's own performance has caught up.
7380 However there are likely to be other architectures where GMP could
7381 provide a boost. This ENGINE is not built in by default, but it can be
7382 specified at Configure time and should be accompanied by the necessary
7383 linker additions, eg;
7384 ./config -DOPENSSL_USE_GMP -lgmp
7388 * "openssl engine" will not display ENGINE/DSO load failure errors when
7389 testing availability of engines with "-t" - the old behaviour is
7390 produced by increasing the feature's verbosity with "-tt".
7394 * ECDSA routines: under certain error conditions uninitialized BN objects
7395 could be freed. Solution: make sure initialization is performed early
7396 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
7401 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
7402 and DH_METHOD (eg. by ENGINE implementations) to override the normal
7403 software implementations. For DSA and DH, parameter generation can
7404 also be overridden by providing the appropriate method callbacks.
7408 * Change the "progress" mechanism used in key-generation and
7409 primality testing to functions that take a new BN_GENCB pointer in
7410 place of callback/argument pairs. The new API functions have "_ex"
7411 postfixes and the older functions are reimplemented as wrappers for
7412 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
7413 declarations of the old functions to help (graceful) attempts to
7414 migrate to the new functions. Also, the new key-generation API
7415 functions operate on a caller-supplied key-structure and return
7416 success/failure rather than returning a key or NULL - this is to
7417 help make "keygen" another member function of RSA_METHOD etc.
7419 Example for using the new callback interface:
7421 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
7425 BN_GENCB_set(&my_cb, my_callback, my_arg);
7427 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
7428 /* For the meaning of a, b in calls to my_callback(), see the
7429 * documentation of the function that calls the callback.
7430 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
7431 * my_callback should return 1 if it wants BN_is_prime_ex()
7432 * to continue, or 0 to stop.
7437 * Change the ZLIB compression method to be stateful, and make it
7438 available to TLS with the number defined in
7439 draft-ietf-tls-compression-04.txt.
7443 * Add the ASN.1 structures and functions for CertificatePair, which
7444 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
7446 CertificatePair ::= SEQUENCE {
7447 forward [0] Certificate OPTIONAL,
7448 reverse [1] Certificate OPTIONAL,
7449 -- at least one of the pair shall be present -- }
7451 Also implement the PEM functions to read and write certificate
7452 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
7454 This needed to be defined, mostly for the sake of the LDAP
7455 attribute crossCertificatePair, but may prove useful elsewhere as
7460 * Make it possible to inhibit symlinking of shared libraries in
7461 Makefile.shared, for Cygwin's sake.
7465 * Extend the BIGNUM API by creating a function
7466 void BN_set_negative(BIGNUM *a, int neg);
7467 and a macro that behave like
7468 int BN_is_negative(const BIGNUM *a);
7470 to avoid the need to access 'a->neg' directly in applications.
7474 * Implement fast modular reduction for pseudo-Mersenne primes
7475 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
7476 EC_GROUP_new_curve_GFp() will now automatically use this
7479 *Nils Larsch <nla@trustcenter.de>*
7481 * Add new lock type (CRYPTO_LOCK_BN).
7485 * Change the ENGINE framework to automatically load engines
7486 dynamically from specific directories unless they could be
7487 found to already be built in or loaded. Move all the
7488 current engines except for the cryptodev one to a new
7490 The engines in engines/ are built as shared libraries if
7491 the "shared" options was given to ./Configure or ./config.
7492 Otherwise, they are inserted in libcrypto.a.
7493 /usr/local/ssl/engines is the default directory for dynamic
7494 engines, but that can be overridden at configure time through
7495 the usual use of --prefix and/or --openssldir, and at run
7496 time with the environment variable OPENSSL_ENGINES.
7498 *Geoff Thorpe and Richard Levitte*
7500 * Add Makefile.shared, a helper makefile to build shared
7501 libraries. Adapt Makefile.org.
7505 * Add version info to Win32 DLLs.
7507 *Peter 'Luna' Runestig" <peter@runestig.com>*
7509 * Add new 'medium level' PKCS#12 API. Certificates and keys
7510 can be added using this API to created arbitrary PKCS#12
7511 files while avoiding the low level API.
7513 New options to PKCS12_create(), key or cert can be NULL and
7514 will then be omitted from the output file. The encryption
7515 algorithm NIDs can be set to -1 for no encryption, the mac
7516 iteration count can be set to 0 to omit the mac.
7518 Enhance pkcs12 utility by making the -nokeys and -nocerts
7519 options work when creating a PKCS#12 file. New option -nomac
7520 to omit the mac, NONE can be set for an encryption algorithm.
7521 New code is modified to use the enhanced PKCS12_create()
7522 instead of the low level API.
7526 * Extend ASN1 encoder to support indefinite length constructed
7527 encoding. This can output sequences tags and octet strings in
7528 this form. Modify pk7_asn1.c to support indefinite length
7529 encoding. This is experimental and needs additional code to
7530 be useful, such as an ASN1 bio and some enhanced streaming
7533 Extend template encode functionality so that tagging is passed
7534 down to the template encoder.
7538 * Let 'openssl req' fail if an argument to '-newkey' is not
7539 recognized instead of using RSA as a default.
7543 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
7544 As these are not official, they are not included in "ALL";
7545 the "ECCdraft" ciphersuite group alias can be used to select them.
7547 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
7549 * Add ECDH engine support.
7551 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
7553 * Add ECDH in new directory crypto/ecdh/.
7555 *Douglas Stebila (Sun Microsystems Laboratories)*
7557 * Let BN_rand_range() abort with an error after 100 iterations
7558 without success (which indicates a broken PRNG).
7562 * Change BN_mod_sqrt() so that it verifies that the input value
7563 is really the square of the return value. (Previously,
7564 BN_mod_sqrt would show GIGO behaviour.)
7568 * Add named elliptic curves over binary fields from X9.62, SECG,
7569 and WAP/WTLS; add OIDs that were still missing.
7571 *Sheueling Chang Shantz and Douglas Stebila
7572 (Sun Microsystems Laboratories)*
7574 * Extend the EC library for elliptic curves over binary fields
7575 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
7578 EC_GF2m_simple_method
7582 EC_GROUP_new_curve_GF2m
7583 EC_GROUP_set_curve_GF2m
7584 EC_GROUP_get_curve_GF2m
7585 EC_POINT_set_affine_coordinates_GF2m
7586 EC_POINT_get_affine_coordinates_GF2m
7587 EC_POINT_set_compressed_coordinates_GF2m
7589 Point compression for binary fields is disabled by default for
7590 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
7593 As binary polynomials are represented as BIGNUMs, various members
7594 of the EC_GROUP and EC_POINT data structures can be shared
7595 between the implementations for prime fields and binary fields;
7596 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
7597 are essentially identical to their ..._GFp counterparts.
7598 (For simplicity, the '..._GFp' prefix has been dropped from
7599 various internal method names.)
7601 An internal 'field_div' method (similar to 'field_mul' and
7602 'field_sqr') has been added; this is used only for binary fields.
7604 *Sheueling Chang Shantz and Douglas Stebila
7605 (Sun Microsystems Laboratories)*
7607 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
7608 through methods ('mul', 'precompute_mult').
7610 The generic implementations (now internally called 'ec_wNAF_mul'
7611 and 'ec_wNAF_precomputed_mult') remain the default if these
7612 methods are undefined.
7614 *Sheueling Chang Shantz and Douglas Stebila
7615 (Sun Microsystems Laboratories)*
7617 * New function EC_GROUP_get_degree, which is defined through
7618 EC_METHOD. For curves over prime fields, this returns the bit
7619 length of the modulus.
7621 *Sheueling Chang Shantz and Douglas Stebila
7622 (Sun Microsystems Laboratories)*
7624 * New functions EC_GROUP_dup, EC_POINT_dup.
7625 (These simply call ..._new and ..._copy).
7627 *Sheueling Chang Shantz and Douglas Stebila
7628 (Sun Microsystems Laboratories)*
7630 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
7631 Polynomials are represented as BIGNUMs (where the sign bit is not
7632 used) in the following functions [macros]:
7635 BN_GF2m_sub [= BN_GF2m_add]
7636 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
7637 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
7638 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
7640 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
7641 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
7642 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
7643 BN_GF2m_cmp [= BN_ucmp]
7645 (Note that only the 'mod' functions are actually for fields GF(2^m).
7646 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
7648 For some functions, an the irreducible polynomial defining a
7649 field can be given as an 'unsigned int[]' with strictly
7650 decreasing elements giving the indices of those bits that are set;
7651 i.e., p[] represents the polynomial
7652 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
7654 p[0] > p[1] > ... > p[k] = 0.
7655 This applies to the following functions:
7660 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
7661 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
7663 BN_GF2m_mod_sqrt_arr
7664 BN_GF2m_mod_solve_quad_arr
7668 Conversion can be performed by the following functions:
7673 bntest.c has additional tests for binary polynomial arithmetic.
7675 Two implementations for BN_GF2m_mod_div() are available.
7676 The default algorithm simply uses BN_GF2m_mod_inv() and
7677 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
7678 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
7679 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
7681 *Sheueling Chang Shantz and Douglas Stebila
7682 (Sun Microsystems Laboratories)*
7684 * Add new error code 'ERR_R_DISABLED' that can be used when some
7685 functionality is disabled at compile-time.
7687 *Douglas Stebila <douglas.stebila@sun.com>*
7689 * Change default behaviour of 'openssl asn1parse' so that more
7690 information is visible when viewing, e.g., a certificate:
7692 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
7693 mode the content of non-printable OCTET STRINGs is output in a
7694 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
7695 avoid the appearance of a printable string.
7697 *Nils Larsch <nla@trustcenter.de>*
7699 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
7701 EC_GROUP_set_asn1_flag()
7702 EC_GROUP_get_asn1_flag()
7703 EC_GROUP_set_point_conversion_form()
7704 EC_GROUP_get_point_conversion_form()
7705 These control ASN1 encoding details:
7706 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
7707 has been set to OPENSSL_EC_NAMED_CURVE.
7708 - Points are encoded in uncompressed form by default; options for
7709 asn1_for are as for point2oct, namely
7710 POINT_CONVERSION_COMPRESSED
7711 POINT_CONVERSION_UNCOMPRESSED
7712 POINT_CONVERSION_HYBRID
7714 Also add 'seed' and 'seed_len' members to EC_GROUP with access
7717 EC_GROUP_get0_seed()
7718 EC_GROUP_get_seed_len()
7719 This is used only for ASN1 purposes (so far).
7721 *Nils Larsch <nla@trustcenter.de>*
7723 * Add 'field_type' member to EC_METHOD, which holds the NID
7724 of the appropriate field type OID. The new function
7725 EC_METHOD_get_field_type() returns this value.
7727 *Nils Larsch <nla@trustcenter.de>*
7732 EC_POINT_point2hex()
7733 EC_POINT_hex2point()
7734 providing useful interfaces to EC_POINT_point2oct() and
7735 EC_POINT_oct2point().
7737 *Nils Larsch <nla@trustcenter.de>*
7739 * Change internals of the EC library so that the functions
7740 EC_GROUP_set_generator()
7741 EC_GROUP_get_generator()
7742 EC_GROUP_get_order()
7743 EC_GROUP_get_cofactor()
7744 are implemented directly in crypto/ec/ec_lib.c and not dispatched
7745 to methods, which would lead to unnecessary code duplication when
7746 adding different types of curves.
7748 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
7750 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
7751 arithmetic, and such that modified wNAFs are generated
7752 (which avoid length expansion in many cases).
7756 * Add a function EC_GROUP_check_discriminant() (defined via
7757 EC_METHOD) that verifies that the curve discriminant is non-zero.
7759 Add a function EC_GROUP_check() that makes some sanity tests
7760 on a EC_GROUP, its generator and order. This includes
7761 EC_GROUP_check_discriminant().
7763 *Nils Larsch <nla@trustcenter.de>*
7765 * Add ECDSA in new directory crypto/ecdsa/.
7767 Add applications 'openssl ecparam' and 'openssl ecdsa'
7768 (these are based on 'openssl dsaparam' and 'openssl dsa').
7770 ECDSA support is also included in various other files across the
7771 library. Most notably,
7772 - 'openssl req' now has a '-newkey ecdsa:file' option;
7773 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
7774 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
7775 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
7776 them suitable for ECDSA where domain parameters must be
7777 extracted before the specific public key;
7778 - ECDSA engine support has been added.
7780 *Nils Larsch <nla@trustcenter.de>*
7782 * Include some named elliptic curves, and add OIDs from X9.62,
7783 SECG, and WAP/WTLS. Each curve can be obtained from the new
7785 EC_GROUP_new_by_curve_name(),
7786 and the list of available named curves can be obtained with
7787 EC_get_builtin_curves().
7788 Also add a 'curve_name' member to EC_GROUP objects, which can be
7790 EC_GROUP_set_curve_name()
7791 EC_GROUP_get_curve_name()
7793 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
7795 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7796 was actually never needed) and in BN_mul(). The removal in BN_mul()
7797 required a small change in bn_mul_part_recursive() and the addition
7798 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7799 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7800 bn_sub_words() and bn_add_words() except they take arrays with
7805 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007] ###
7807 * Cleanse PEM buffers before freeing them since they may contain
7810 *Benjamin Bennett <ben@psc.edu>*
7812 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
7813 a ciphersuite string such as "DEFAULT:RSA" cannot enable
7814 authentication-only ciphersuites.
7818 * Since AES128 and AES256 share a single mask bit in the logic of
7819 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
7820 kludge to work properly if AES128 is available and AES256 isn't.
7824 * Expand security boundary to match 1.1.1 module.
7828 * Remove redundant features: hash file source, editing of test vectors
7829 modify fipsld to use external fips_premain.c signature.
7833 * New perl script mkfipsscr.pl to create shell scripts or batch files to
7834 run algorithm test programs.
7838 * Make algorithm test programs more tolerant of whitespace.
7842 * Have SSL/TLS server implementation tolerate "mismatched" record
7843 protocol version while receiving ClientHello even if the
7844 ClientHello is fragmented. (The server can't insist on the
7845 particular protocol version it has chosen before the ServerHello
7846 message has informed the client about his choice.)
7850 * Load error codes if they are not already present instead of using a
7851 static variable. This allows them to be cleanly unloaded and reloaded.
7855 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006] ###
7857 * Introduce limits to prevent malicious keys being able to
7858 cause a denial of service. (CVE-2006-2940)
7860 *Steve Henson, Bodo Moeller*
7862 * Fix ASN.1 parsing of certain invalid structures that can result
7863 in a denial of service. (CVE-2006-2937) [Steve Henson]
7865 * Fix buffer overflow in SSL_get_shared_ciphers() function.
7866 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
7868 * Fix SSL client code which could crash if connecting to a
7869 malicious SSLv2 server. (CVE-2006-4343)
7871 *Tavis Ormandy and Will Drewry, Google Security Team*
7873 * Change ciphersuite string processing so that an explicit
7874 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
7875 will no longer include "AES128-SHA"), and any other similar
7876 ciphersuite (same bitmap) from *other* protocol versions (so that
7877 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
7878 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
7879 changes from 0.9.8b and 0.9.8d.
7883 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006] ###
7885 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
7886 (CVE-2006-4339) [Ben Laurie and Google Security Team]
7888 * Change the Unix randomness entropy gathering to use poll() when
7889 possible instead of select(), since the latter has some
7890 undesirable limitations.
7892 *Darryl Miles via Richard Levitte and Bodo Moeller*
7894 * Disable rogue ciphersuites:
7896 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
7897 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
7898 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
7900 The latter two were purportedly from
7901 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
7904 Also deactivate the remaining ciphersuites from
7905 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
7906 unofficial, and the ID has long expired.
7910 * Fix RSA blinding Heisenbug (problems sometimes occurred on
7911 dual-core machines) and other potential thread-safety issues.
7915 ### Changes between 0.9.7i and 0.9.7j [04 May 2006] ###
7917 * Adapt fipsld and the build system to link against the validated FIPS
7918 module in FIPS mode.
7922 * Fixes for VC++ 2005 build under Windows.
7926 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
7927 from a Windows bash shell such as MSYS. It is autodetected from the
7928 "config" script when run from a VC++ environment. Modify standard VC++
7929 build to use fipscanister.o from the GNU make build.
7933 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005] ###
7935 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
7936 The value now differs depending on if you build for FIPS or not.
7937 BEWARE! A program linked with a shared FIPSed libcrypto can't be
7938 safely run with a non-FIPSed libcrypto, as it may crash because of
7939 the difference induced by this change.
7943 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005] ###
7945 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
7946 (part of SSL_OP_ALL). This option used to disable the
7947 countermeasure against man-in-the-middle protocol-version
7948 rollback in the SSL 2.0 server implementation, which is a bad
7949 idea. (CVE-2005-2969)
7951 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
7952 for Information Security, National Institute of Advanced Industrial
7953 Science and Technology [AIST*, Japan)]
7955 * Minimal support for X9.31 signatures and PSS padding modes. This is
7956 mainly for FIPS compliance and not fully integrated at this stage.
7960 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
7961 the exponentiation using a fixed-length exponent. (Otherwise,
7962 the information leaked through timing could expose the secret key
7963 after many signatures; cf. Bleichenbacher's attack on DSA with
7968 * Make a new fixed-window mod_exp implementation the default for
7969 RSA, DSA, and DH private-key operations so that the sequence of
7970 squares and multiplies and the memory access pattern are
7971 independent of the particular secret key. This will mitigate
7972 cache-timing and potential related attacks.
7974 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
7975 and this is automatically used by BN_mod_exp_mont() if the new flag
7976 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
7977 will use this BN flag for private exponents unless the flag
7978 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
7979 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
7982 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
7984 * Change the client implementation for SSLv23_method() and
7985 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
7986 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
7987 (Previously, the SSL 2.0 backwards compatible Client Hello
7988 message format would be used even with SSL_OP_NO_SSLv2.)
7992 * Add support for smime-type MIME parameter in S/MIME messages which some
7997 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
7998 a threadsafe manner. Modify rsa code to use new function and add calls
7999 to dsa and dh code (which had race conditions before).
8003 * Include the fixed error library code in the C error file definitions
8004 instead of fixing them up at runtime. This keeps the error code
8005 structures constant.
8009 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005] ###
8011 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
8014 * Fixes for newer kerberos headers. NB: the casts are needed because
8015 the 'length' field is signed on one version and unsigned on another
8016 with no (?) obvious way to tell the difference, without these VC++
8017 complains. Also the "definition" of FAR (blank) is no longer included
8018 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
8019 some needed definitions.
8023 * Undo Cygwin change.
8027 * Added support for proxy certificates according to RFC 3820.
8028 Because they may be a security thread to unaware applications,
8029 they must be explicitly allowed in run-time. See
8030 docs/HOWTO/proxy_certificates.txt for further information.
8034 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005] ###
8036 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
8037 server and client random values. Previously
8038 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
8039 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
8041 This change has negligible security impact because:
8043 1. Server and client random values still have 24 bytes of pseudo random
8046 2. Server and client random values are sent in the clear in the initial
8049 3. The master secret is derived using the premaster secret (48 bytes in
8050 size for static RSA ciphersuites) as well as client server and random
8053 The OpenSSL team would like to thank the UK NISCC for bringing this issue
8056 *Stephen Henson, reported by UK NISCC*
8058 * Use Windows randomness collection on Cygwin.
8062 * Fix hang in EGD/PRNGD query when communication socket is closed
8063 prematurely by EGD/PRNGD.
8065 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
8067 * Prompt for pass phrases when appropriate for PKCS12 input format.
8071 * Back-port of selected performance improvements from development
8072 branch, as well as improved support for PowerPC platforms.
8076 * Add lots of checks for memory allocation failure, error codes to indicate
8077 failure and freeing up memory if a failure occurs.
8079 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
8081 * Add new -passin argument to dgst.
8085 * Perform some character comparisons of different types in X509_NAME_cmp:
8086 this is needed for some certificates that re-encode DNs into UTF8Strings
8087 (in violation of RFC3280) and can't or won't issue name rollover
8092 * Make an explicit check during certificate validation to see that
8093 the CA setting in each certificate on the chain is correct. As a
8094 side effect always do the following basic checks on extensions,
8095 not just when there's an associated purpose to the check:
8097 - if there is an unhandled critical extension (unless the user
8098 has chosen to ignore this fault)
8099 - if the path length has been exceeded (if one is set at all)
8100 - that certain extensions fit the associated purpose (if one has
8105 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004] ###
8107 * Avoid a race condition when CRLs are checked in a multi threaded
8108 environment. This would happen due to the reordering of the revoked
8109 entries during signature checking and serial number lookup. Now the
8110 encoding is cached and the serial number sort performed under a lock.
8111 Add new STACK function sk_is_sorted().
8115 * Add Delta CRL to the extension code.
8119 * Various fixes to s3_pkt.c so alerts are sent properly.
8121 *David Holmes <d.holmes@f5.com>*
8123 * Reduce the chances of duplicate issuer name and serial numbers (in
8124 violation of RFC3280) using the OpenSSL certificate creation utilities.
8125 This is done by creating a random 64 bit value for the initial serial
8126 number when a serial number file is created or when a self signed
8127 certificate is created using 'openssl req -x509'. The initial serial
8128 number file is created using 'openssl x509 -next_serial' in CA.pl
8129 rather than being initialized to 1.
8133 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004] ###
8135 * Fix null-pointer assignment in do_change_cipher_spec() revealed
8136 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8138 *Joe Orton, Steve Henson*
8140 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
8143 *Joe Orton, Steve Henson*
8145 * Make it possible to have multiple active certificates with the same
8146 subject in the CA index file. This is done only if the keyword
8147 'unique_subject' is set to 'no' in the main CA section (default
8148 if 'CA_default') of the configuration file. The value is saved
8149 with the database itself in a separate index attribute file,
8150 named like the index file with '.attr' appended to the name.
8154 * X509 verify fixes. Disable broken certificate workarounds when
8155 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
8156 keyUsage extension present. Don't accept CRLs with unhandled critical
8157 extensions: since verify currently doesn't process CRL extensions this
8158 rejects a CRL with *any* critical extensions. Add new verify error codes
8163 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
8164 A clarification of RFC2560 will require the use of OCTET STRINGs and
8165 some implementations cannot handle the current raw format. Since OpenSSL
8166 copies and compares OCSP nonces as opaque blobs without any attempt at
8167 parsing them this should not create any compatibility issues.
8171 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
8172 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
8173 this HMAC (and other) operations are several times slower than OpenSSL
8178 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
8180 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
8182 * Use the correct content when signing type "other".
8186 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003] ###
8188 * Fix various bugs revealed by running the NISCC test suite:
8190 Stop out of bounds reads in the ASN1 code when presented with
8191 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8193 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
8195 If verify callback ignores invalid public key errors don't try to check
8196 certificate signature with the NULL public key.
8201 * New -ignore_err option in ocsp application to stop the server
8202 exiting on the first error in a request.
8206 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8207 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8212 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8213 extra data after the compression methods not only for TLS 1.0
8214 but also for SSL 3.0 (as required by the specification).
8216 *Bodo Moeller; problem pointed out by Matthias Loepfe*
8218 * Change X509_certificate_type() to mark the key as exported/exportable
8219 when it's 512 *bits* long, not 512 bytes.
8223 * Change AES_cbc_encrypt() so it outputs exact multiple of
8224 blocks during encryption.
8228 * Various fixes to base64 BIO and non blocking I/O. On write
8229 flushes were not handled properly if the BIO retried. On read
8230 data was not being buffered properly and had various logic bugs.
8231 This also affects blocking I/O when the data being decoded is a
8236 * Various S/MIME bugfixes and compatibility changes:
8237 output correct application/pkcs7 MIME type if
8238 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
8239 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
8240 of files as .eml work). Correctly handle very long lines in MIME
8245 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003] ###
8247 * Countermeasure against the Klima-Pokorny-Rosa extension of
8248 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8249 a protocol version number mismatch like a decryption error
8250 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8254 * Turn on RSA blinding by default in the default implementation
8255 to avoid a timing attack. Applications that don't want it can call
8256 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8257 They would be ill-advised to do so in most cases.
8259 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
8261 * Change RSA blinding code so that it works when the PRNG is not
8262 seeded (in this case, the secret RSA exponent is abused as
8263 an unpredictable seed -- if it is not unpredictable, there
8264 is no point in blinding anyway). Make RSA blinding thread-safe
8265 by remembering the creator's thread ID in rsa->blinding and
8266 having all other threads use local one-time blinding factors
8267 (this requires more computation than sharing rsa->blinding, but
8268 avoids excessive locking; and if an RSA object is not shared
8269 between threads, blinding will still be very fast).
8273 * Fixed a typo bug that would cause ENGINE_set_default() to set an
8274 ENGINE as defaults for all supported algorithms irrespective of
8275 the 'flags' parameter. 'flags' is now honoured, so applications
8276 should make sure they are passing it correctly.
8280 * Target "mingw" now allows native Windows code to be generated in
8281 the Cygwin environment as well as with the MinGW compiler.
8285 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003] ###
8287 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8288 via timing by performing a MAC computation even if incorrect
8289 block cipher padding has been found. This is a countermeasure
8290 against active attacks where the attacker has to distinguish
8291 between bad padding and a MAC verification error. (CVE-2003-0078)
8293 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8294 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8295 Martin Vuagnoux (EPFL, Ilion)*
8297 * Make the no-err option work as intended. The intention with no-err
8298 is not to have the whole error stack handling routines removed from
8299 libcrypto, it's only intended to remove all the function name and
8300 reason texts, thereby removing some of the footprint that may not
8301 be interesting if those errors aren't displayed anyway.
8303 NOTE: it's still possible for any application or module to have its
8304 own set of error texts inserted. The routines are there, just not
8305 used by default when no-err is given.
8309 * Add support for FreeBSD on IA64.
8311 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
8313 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
8314 Kerberos function mit_des_cbc_cksum(). Before this change,
8315 the value returned by DES_cbc_cksum() was like the one from
8316 mit_des_cbc_cksum(), except the bytes were swapped.
8318 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
8320 * Allow an application to disable the automatic SSL chain building.
8321 Before this a rather primitive chain build was always performed in
8322 ssl3_output_cert_chain(): an application had no way to send the
8323 correct chain if the automatic operation produced an incorrect result.
8325 Now the chain builder is disabled if either:
8327 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
8329 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
8331 The reasoning behind this is that an application would not want the
8332 auto chain building to take place if extra chain certificates are
8333 present and it might also want a means of sending no additional
8334 certificates (for example the chain has two certificates and the
8339 * Add the possibility to build without the ENGINE framework.
8341 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
8343 * Under Win32 gmtime() can return NULL: check return value in
8344 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
8348 * DSA routines: under certain error conditions uninitialized BN objects
8349 could be freed. Solution: make sure initialization is performed early
8350 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
8351 Nils Larsch <nla@trustcenter.de> via PR#459)
8355 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
8356 checked on reconnect on the client side, therefore session resumption
8357 could still fail with a "ssl session id is different" error. This
8358 behaviour is masked when SSL_OP_ALL is used due to
8359 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
8360 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
8361 followup to PR #377.
8365 * IA-32 assembler support enhancements: unified ELF targets, support
8366 for SCO/Caldera platforms, fix for Cygwin shared build.
8370 * Add support for FreeBSD on sparc64. As a consequence, support for
8371 FreeBSD on non-x86 processors is separate from x86 processors on
8372 the config script, much like the NetBSD support.
8374 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
8376 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002] ###
8378 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
8381 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
8382 code (06) was taken as the first octet of the session ID and the last
8383 octet was ignored consequently. As a result SSLv2 client side session
8384 caching could not have worked due to the session ID mismatch between
8386 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
8391 * Change the declaration of needed Kerberos libraries to use EX_LIBS
8392 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
8397 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
8398 seems that in spite of existing for more than a year, many application
8399 author have done nothing to provide the necessary callbacks, which
8400 means that this particular engine will not work properly anywhere.
8401 This is a very unfortunate situation which forces us, in the name
8402 of usability, to give the hw_ncipher.c a static lock, which is part
8404 NOTE: This is for the 0.9.7 series ONLY. This hack will never
8405 appear in 0.9.8 or later. We EXPECT application authors to have
8406 dealt properly with this when 0.9.8 is released (unless we actually
8407 make such changes in the libcrypto locking code that changes will
8408 have to be made anyway).
8412 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
8413 octets have been read, EOF or an error occurs. Without this change
8414 some truncated ASN1 structures will not produce an error.
8418 * Disable Heimdal support, since it hasn't been fully implemented.
8419 Still give the possibility to force the use of Heimdal, but with
8420 warnings and a request that patches get sent to openssl-dev.
8424 * Add the VC-CE target, introduce the WINCE sysname, and add
8425 INSTALL.WCE and appropriate conditionals to make it build.
8427 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
8429 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
8430 cygssl-x.y.z.dll, where x, y and z are the major, minor and
8431 edit numbers of the version.
8433 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
8435 * Introduce safe string copy and catenation functions
8436 (BUF_strlcpy() and BUF_strlcat()).
8438 *Ben Laurie (CHATS) and Richard Levitte*
8440 * Avoid using fixed-size buffers for one-line DNs.
8442 *Ben Laurie (CHATS)*
8444 * Add BUF_MEM_grow_clean() to avoid information leakage when
8445 resizing buffers containing secrets, and use where appropriate.
8447 *Ben Laurie (CHATS)*
8449 * Avoid using fixed size buffers for configuration file location.
8451 *Ben Laurie (CHATS)*
8453 * Avoid filename truncation for various CA files.
8455 *Ben Laurie (CHATS)*
8457 * Use sizeof in preference to magic numbers.
8459 *Ben Laurie (CHATS)*
8461 * Avoid filename truncation in cert requests.
8463 *Ben Laurie (CHATS)*
8465 * Add assertions to check for (supposedly impossible) buffer
8468 *Ben Laurie (CHATS)*
8470 * Don't cache truncated DNS entries in the local cache (this could
8471 potentially lead to a spoofing attack).
8473 *Ben Laurie (CHATS)*
8475 * Fix various buffers to be large enough for hex/decimal
8476 representations in a platform independent manner.
8478 *Ben Laurie (CHATS)*
8480 * Add CRYPTO_realloc_clean() to avoid information leakage when
8481 resizing buffers containing secrets, and use where appropriate.
8483 *Ben Laurie (CHATS)*
8485 * Add BIO_indent() to avoid much slightly worrying code to do
8488 *Ben Laurie (CHATS)*
8490 * Convert sprintf()/BIO_puts() to BIO_printf().
8492 *Ben Laurie (CHATS)*
8494 * buffer_gets() could terminate with the buffer only half
8497 *Ben Laurie (CHATS)*
8499 * Add assertions to prevent user-supplied crypto functions from
8500 overflowing internal buffers by having large block sizes, etc.
8502 *Ben Laurie (CHATS)*
8504 * New OPENSSL_assert() macro (similar to assert(), but enabled
8507 *Ben Laurie (CHATS)*
8509 * Eliminate unused copy of key in RC4.
8511 *Ben Laurie (CHATS)*
8513 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
8515 *Ben Laurie (CHATS)*
8517 * Fix off-by-one error in EGD path.
8519 *Ben Laurie (CHATS)*
8521 * If RANDFILE path is too long, ignore instead of truncating.
8523 *Ben Laurie (CHATS)*
8525 * Eliminate unused and incorrectly sized X.509 structure
8528 *Ben Laurie (CHATS)*
8530 * Eliminate unused and dangerous function knumber().
8532 *Ben Laurie (CHATS)*
8534 * Eliminate unused and dangerous structure, KSSL_ERR.
8536 *Ben Laurie (CHATS)*
8538 * Protect against overlong session ID context length in an encoded
8539 session object. Since these are local, this does not appear to be
8542 *Ben Laurie (CHATS)*
8544 * Change from security patch (see 0.9.6e below) that did not affect
8545 the 0.9.6 release series:
8547 Remote buffer overflow in SSL3 protocol - an attacker could
8548 supply an oversized master key in Kerberos-enabled versions.
8551 *Ben Laurie (CHATS)*
8553 * Change the SSL kerb5 codes to match RFC 2712.
8557 * Make -nameopt work fully for req and add -reqopt switch.
8559 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
8561 * The "block size" for block ciphers in CFB and OFB mode should be 1.
8563 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
8565 * Make sure tests can be performed even if the corresponding algorithms
8566 have been removed entirely. This was also the last step to make
8567 OpenSSL compilable with DJGPP under all reasonable conditions.
8569 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
8571 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
8572 to allow version independent disabling of normally unselected ciphers,
8573 which may be activated as a side-effect of selecting a single cipher.
8575 (E.g., cipher list string "RSA" enables ciphersuites that are left
8576 out of "ALL" because they do not provide symmetric encryption.
8577 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
8579 *Lutz Jaenicke, Bodo Moeller*
8581 * Add appropriate support for separate platform-dependent build
8582 directories. The recommended way to make a platform-dependent
8583 build directory is the following (tested on Linux), maybe with
8586 # Place yourself outside of the OpenSSL source tree. In
8587 # this example, the environment variable OPENSSL_SOURCE
8588 # is assumed to contain the absolute OpenSSL source directory.
8589 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
8590 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
8591 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
8592 mkdir -p `dirname $F`
8593 ln -s $OPENSSL_SOURCE/$F $F
8596 To be absolutely sure not to disturb the source tree, a "make clean"
8597 is a good thing. If it isn't successful, don't worry about it,
8598 it probably means the source directory is very clean.
8602 * Make sure any ENGINE control commands make local copies of string
8603 pointers passed to them whenever necessary. Otherwise it is possible
8604 the caller may have overwritten (or deallocated) the original string
8605 data when a later ENGINE operation tries to use the stored values.
8607 *Götz Babin-Ebell <babinebell@trustcenter.de>*
8609 * Improve diagnostics in file reading and command-line digests.
8611 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
8613 * Add AES modes CFB and OFB to the object database. Correct an
8614 error in AES-CFB decryption.
8618 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
8619 allows existing EVP_CIPHER_CTX structures to be reused after
8620 calling EVP_*Final(). This behaviour is used by encryption
8621 BIOs and some applications. This has the side effect that
8622 applications must explicitly clean up cipher contexts with
8623 EVP_CIPHER_CTX_cleanup() or they will leak memory.
8627 * Check the values of dna and dnb in bn_mul_recursive before calling
8628 bn_mul_comba (a non zero value means the a or b arrays do not contain
8629 n2 elements) and fallback to bn_mul_normal if either is not zero.
8633 * Fix escaping of non-ASCII characters when using the -subj option
8634 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
8638 * Make object definitions compliant to LDAP (RFC2256): SN is the short
8639 form for "surname", serialNumber has no short form.
8640 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
8641 therefore remove "mail" short name for "internet 7".
8642 The OID for unique identifiers in X509 certificates is
8643 x500UniqueIdentifier, not uniqueIdentifier.
8644 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
8648 * Add an "init" command to the ENGINE config module and auto initialize
8649 ENGINEs. Without any "init" command the ENGINE will be initialized
8650 after all ctrl commands have been executed on it. If init=1 the
8651 ENGINE is initialized at that point (ctrls before that point are run
8652 on the uninitialized ENGINE and after on the initialized one). If
8653 init=0 then the ENGINE will not be initialized at all.
8657 * Fix the 'app_verify_callback' interface so that the user-defined
8658 argument is actually passed to the callback: In the
8659 SSL_CTX_set_cert_verify_callback() prototype, the callback
8660 declaration has been changed from
8663 int (*cb)(X509_STORE_CTX *,void *);
8664 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
8665 i=s->ctx->app_verify_callback(&ctx)
8666 has been changed into
8667 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
8669 To update applications using SSL_CTX_set_cert_verify_callback(),
8670 a dummy argument can be added to their callback functions.
8672 *D. K. Smetters <smetters@parc.xerox.com>*
8674 * Added the '4758cca' ENGINE to support IBM 4758 cards.
8676 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
8678 * Add and OPENSSL_LOAD_CONF define which will cause
8679 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
8680 This allows older applications to transparently support certain
8681 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
8682 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
8683 load the config file and OPENSSL_add_all_algorithms_conf() which will
8684 always load it have also been added.
8688 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
8689 Adjust NIDs and EVP layer.
8691 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
8693 * Config modules support in openssl utility.
8695 Most commands now load modules from the config file,
8696 though in a few (such as version) this isn't done
8697 because it couldn't be used for anything.
8699 In the case of ca and req the config file used is
8700 the same as the utility itself: that is the -config
8701 command line option can be used to specify an
8706 * Move default behaviour from OPENSSL_config(). If appname is NULL
8707 use "openssl_conf" if filename is NULL use default openssl config file.
8711 * Add an argument to OPENSSL_config() to allow the use of an alternative
8712 config section name. Add a new flag to tolerate a missing config file
8713 and move code to CONF_modules_load_file().
8717 * Support for crypto accelerator cards from Accelerated Encryption
8718 Processing, www.aep.ie. (Use engine 'aep')
8719 The support was copied from 0.9.6c [engine] and adapted/corrected
8720 to work with the new engine framework.
8722 *AEP Inc. and Richard Levitte*
8724 * Support for SureWare crypto accelerator cards from Baltimore
8725 Technologies. (Use engine 'sureware')
8726 The support was copied from 0.9.6c [engine] and adapted
8727 to work with the new engine framework.
8731 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
8732 make the newer ENGINE framework commands for the CHIL engine work.
8734 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
8736 * Make it possible to produce shared libraries on ReliantUNIX.
8738 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
8740 * Add the configuration target debug-linux-ppro.
8741 Make 'openssl rsa' use the general key loading routines
8742 implemented in apps.c, and make those routines able to
8743 handle the key format FORMAT_NETSCAPE and the variant
8746 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
8748 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8750 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
8752 * Add -keyform to rsautl, and document -engine.
8754 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
8756 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
8757 BIO_R_NO_SUCH_FILE error code rather than the generic
8758 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
8764 ERR_peek_last_error_line
8765 ERR_peek_last_error_line_data.
8766 These are similar to
8769 ERR_peek_error_line_data,
8770 but report on the latest error recorded rather than the first one
8771 still in the error queue.
8773 *Ben Laurie, Bodo Moeller*
8775 * default_algorithms option in ENGINE config module. This allows things
8777 default_algorithms = ALL
8778 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
8782 * Preliminary ENGINE config module.
8786 * New experimental application configuration code.
8790 * Change the AES code to follow the same name structure as all other
8791 symmetric ciphers, and behave the same way. Move everything to
8792 the directory crypto/aes, thereby obsoleting crypto/rijndael.
8794 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
8796 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
8798 *Ben Laurie and Theo de Raadt*
8800 * Add option to output public keys in req command.
8802 *Massimiliano Pala madwolf@openca.org*
8804 * Use wNAFs in EC_POINTs_mul() for improved efficiency
8805 (up to about 10% better than before for P-192 and P-224).
8809 * New functions/macros
8811 SSL_CTX_set_msg_callback(ctx, cb)
8812 SSL_CTX_set_msg_callback_arg(ctx, arg)
8813 SSL_set_msg_callback(ssl, cb)
8814 SSL_set_msg_callback_arg(ssl, arg)
8816 to request calling a callback function
8818 void cb(int write_p, int version, int content_type,
8819 const void *buf, size_t len, SSL *ssl, void *arg)
8821 whenever a protocol message has been completely received
8822 (write_p == 0) or sent (write_p == 1). Here 'version' is the
8823 protocol version according to which the SSL library interprets
8824 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
8825 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
8826 the content type as defined in the SSL 3.0/TLS 1.0 protocol
8827 specification (change_cipher_spec(20), alert(21), handshake(22)).
8828 'buf' and 'len' point to the actual message, 'ssl' to the
8829 SSL object, and 'arg' is the application-defined value set by
8830 SSL[_CTX]_set_msg_callback_arg().
8832 'openssl s_client' and 'openssl s_server' have new '-msg' options
8833 to enable a callback that displays all protocol messages.
8837 * Change the shared library support so shared libraries are built as
8838 soon as the corresponding static library is finished, and thereby get
8839 openssl and the test programs linked against the shared library.
8840 This still only happens when the keyword "shard" has been given to
8841 the configuration scripts.
8843 NOTE: shared library support is still an experimental thing, and
8844 backward binary compatibility is still not guaranteed.
8846 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
8848 * Add support for Subject Information Access extension.
8850 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
8852 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
8853 additional bytes when new memory had to be allocated, not just
8854 when reusing an existing buffer.
8858 * New command line and configuration option 'utf8' for the req command.
8859 This allows field values to be specified as UTF8 strings.
8863 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
8864 runs for the former and machine-readable output for the latter.
8868 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
8869 of the e-mail address in the DN (i.e., it will go into a certificate
8870 extension only). The new configuration file option 'email_in_dn = no'
8871 has the same effect.
8873 *Massimiliano Pala madwolf@openca.org*
8875 * Change all functions with names starting with des_ to be starting
8876 with DES_ instead. Add wrappers that are compatible with libdes,
8877 but are named _ossl_old_des_*. Finally, add macros that map the
8878 des_* symbols to the corresponding _ossl_old_des_* if libdes
8879 compatibility is desired. If OpenSSL 0.9.6c compatibility is
8880 desired, the des_* symbols will be mapped to DES_*, with one
8883 Since we provide two compatibility mappings, the user needs to
8884 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
8885 compatibility is desired. The default (i.e., when that macro
8886 isn't defined) is OpenSSL 0.9.6c compatibility.
8888 There are also macros that enable and disable the support of old
8889 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
8890 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
8891 are defined, the default will apply: to support the old des routines.
8893 In either case, one must include openssl/des.h to get the correct
8894 definitions. Do not try to just include openssl/des_old.h, that
8897 NOTE: This is a major break of an old API into a new one. Software
8898 authors are encouraged to switch to the DES_ style functions. Some
8899 time in the future, des_old.h and the libdes compatibility functions
8900 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
8901 default), and then completely removed.
8905 * Test for certificates which contain unsupported critical extensions.
8906 If such a certificate is found during a verify operation it is
8907 rejected by default: this behaviour can be overridden by either
8908 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
8909 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
8910 X509_supported_extension() has also been added which returns 1 if a
8911 particular extension is supported.
8915 * Modify the behaviour of EVP cipher functions in similar way to digests
8916 to retain compatibility with existing code.
8920 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
8921 compatibility with existing code. In particular the 'ctx' parameter does
8922 not have to be to be initialized before the call to EVP_DigestInit() and
8923 it is tidied up after a call to EVP_DigestFinal(). New function
8924 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
8925 EVP_MD_CTX_copy() changed to not require the destination to be
8926 initialized valid and new function EVP_MD_CTX_copy_ex() added which
8927 requires the destination to be valid.
8929 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
8930 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
8934 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
8935 so that complete 'Handshake' protocol structures are kept in memory
8936 instead of overwriting 'msg_type' and 'length' with 'body' data.
8940 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
8942 *Massimo Santin via Richard Levitte*
8944 * Major restructuring to the underlying ENGINE code. This includes
8945 reduction of linker bloat, separation of pure "ENGINE" manipulation
8946 (initialisation, etc) from functionality dealing with implementations
8947 of specific crypto interfaces. This change also introduces integrated
8948 support for symmetric ciphers and digest implementations - so ENGINEs
8949 can now accelerate these by providing EVP_CIPHER and EVP_MD
8950 implementations of their own. This is detailed in crypto/engine/README
8951 as it couldn't be adequately described here. However, there are a few
8952 API changes worth noting - some RSA, DSA, DH, and RAND functions that
8953 were changed in the original introduction of ENGINE code have now
8954 reverted back - the hooking from this code to ENGINE is now a good
8955 deal more passive and at run-time, operations deal directly with
8956 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
8957 dereferencing through an ENGINE pointer any more. Also, the ENGINE
8958 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
8959 they were not being used by the framework as there is no concept of a
8960 BIGNUM_METHOD and they could not be generalised to the new
8961 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
8962 ENGINE_cpy() has been removed as it cannot be consistently defined in
8967 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
8971 * Change mkdef.pl to sort symbols that get the same entry number,
8972 and make sure the automatically generated functions ERR_load_*
8973 become part of libeay.num as well.
8977 * New function SSL_renegotiate_pending(). This returns true once
8978 renegotiation has been requested (either SSL_renegotiate() call
8979 or HelloRequest/ClientHello received from the peer) and becomes
8980 false once a handshake has been completed.
8981 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
8982 sends a HelloRequest, but does not ensure that a handshake takes
8983 place. SSL_renegotiate_pending() is useful for checking if the
8984 client has followed the request.)
8988 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
8989 By default, clients may request session resumption even during
8990 renegotiation (if session ID contexts permit); with this option,
8991 session resumption is possible only in the first handshake.
8993 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
8994 more bits available for options that should not be part of
8995 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
8999 * Add some demos for certificate and certificate request creation.
9003 * Make maximum certificate chain size accepted from the peer application
9004 settable (SSL*_get/set_max_cert_list()), as proposed by
9005 "Douglas E. Engert" <deengert@anl.gov>.
9009 * Add support for shared libraries for Unixware-7
9010 (Boyd Lynn Gerber <gerberb@zenez.com>).
9014 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
9015 be done prior to destruction. Use this to unload error strings from
9016 ENGINEs that load their own error strings. NB: This adds two new API
9017 functions to "get" and "set" this destroy handler in an ENGINE.
9021 * Alter all existing ENGINE implementations (except "openssl" and
9022 "openbsd") to dynamically instantiate their own error strings. This
9023 makes them more flexible to be built both as statically-linked ENGINEs
9024 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
9025 Also, add stub code to each that makes building them as self-contained
9026 shared-libraries easier (see README.ENGINE).
9030 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
9031 implementations into applications that are completely implemented in
9032 self-contained shared-libraries. The "dynamic" ENGINE exposes control
9033 commands that can be used to configure what shared-library to load and
9034 to control aspects of the way it is handled. Also, made an update to
9035 the README.ENGINE file that brings its information up-to-date and
9036 provides some information and instructions on the "dynamic" ENGINE
9037 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
9041 * Make it possible to unload ranges of ERR strings with a new
9042 "ERR_unload_strings" function.
9046 * Add a copy() function to EVP_MD.
9050 * Make EVP_MD routines take a context pointer instead of just the
9051 md_data void pointer.
9055 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
9056 that the digest can only process a single chunk of data
9057 (typically because it is provided by a piece of
9058 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
9059 is only going to provide a single chunk of data, and hence the
9060 framework needn't accumulate the data for oneshot drivers.
9064 * As with "ERR", make it possible to replace the underlying "ex_data"
9065 functions. This change also alters the storage and management of global
9066 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
9067 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
9068 index counters. The API functions that use this state have been changed
9069 to take a "class_index" rather than pointers to the class's local STACK
9070 and counter, and there is now an API function to dynamically create new
9071 classes. This centralisation allows us to (a) plug a lot of the
9072 thread-safety problems that existed, and (b) makes it possible to clean
9073 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
9074 such data would previously have always leaked in application code and
9075 workarounds were in place to make the memory debugging turn a blind eye
9076 to it. Application code that doesn't use this new function will still
9077 leak as before, but their memory debugging output will announce it now
9078 rather than letting it slide.
9080 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
9081 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
9082 has a return value to indicate success or failure.
9086 * Make it possible to replace the underlying "ERR" functions such that the
9087 global state (2 LHASH tables and 2 locks) is only used by the "default"
9088 implementation. This change also adds two functions to "get" and "set"
9089 the implementation prior to it being automatically set the first time
9090 any other ERR function takes place. Ie. an application can call "get",
9091 pass the return value to a module it has just loaded, and that module
9092 can call its own "set" function using that value. This means the
9093 module's "ERR" operations will use (and modify) the error state in the
9094 application and not in its own statically linked copy of OpenSSL code.
9098 * Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
9099 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
9100 the operation, and provides a more encapsulated way for external code
9101 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
9102 to use these functions rather than manually incrementing the counts.
9104 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
9108 * Add EVP test program.
9112 * Add symmetric cipher support to ENGINE. Expect the API to change!
9116 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
9117 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
9118 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
9119 These allow a CRL to be built without having to access X509_CRL fields
9120 directly. Modify 'ca' application to use new functions.
9124 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
9125 bug workarounds. Rollback attack detection is a security feature.
9126 The problem will only arise on OpenSSL servers when TLSv1 is not
9127 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
9128 Software authors not wanting to support TLSv1 will have special reasons
9129 for their choice and can explicitly enable this option.
9131 *Bodo Moeller, Lutz Jaenicke*
9133 * Rationalise EVP so it can be extended: don't include a union of
9134 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
9135 (similar to those existing for EVP_CIPHER_CTX).
9140 EVP_MD_CTX_init(&md); /* new function call */
9141 EVP_DigestInit(&md, EVP_sha1());
9142 EVP_DigestUpdate(&md, in, len);
9143 EVP_DigestFinal(&md, out, NULL);
9144 EVP_MD_CTX_cleanup(&md); /* new function call */
9149 * Make DES key schedule conform to the usual scheme, as well as
9150 correcting its structure. This means that calls to DES functions
9151 now have to pass a pointer to a des_key_schedule instead of a
9152 plain des_key_schedule (which was actually always a pointer
9155 des_key_schedule ks;
9157 des_set_key_checked(..., &ks);
9158 des_ncbc_encrypt(..., &ks, ...);
9160 (Note that a later change renames 'des_...' into 'DES_...'.)
9164 * Initial reduction of linker bloat: the use of some functions, such as
9165 PEM causes large amounts of unused functions to be linked in due to
9166 poor organisation. For example pem_all.c contains every PEM function
9167 which has a knock on effect of linking in large amounts of (unused)
9168 ASN1 code. Grouping together similar functions and splitting unrelated
9169 functions prevents this.
9173 * Cleanup of EVP macros.
9177 * Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
9178 correct _ecb suffix.
9182 * Add initial OCSP responder support to ocsp application. The
9183 revocation information is handled using the text based index
9184 use by the ca application. The responder can either handle
9185 requests generated internally, supplied in files (for example
9186 via a CGI script) or using an internal minimal server.
9190 * Add configuration choices to get zlib compression for TLS.
9194 * Changes to Kerberos SSL for RFC 2712 compliance:
9195 1. Implemented real KerberosWrapper, instead of just using
9196 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
9197 2. Implemented optional authenticator field of KerberosWrapper.
9199 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
9200 and authenticator structs; see crypto/krb5/.
9202 Generalized Kerberos calls to support multiple Kerberos libraries.
9203 *Vern Staats <staatsvr@asc.hpc.mil>,
9204 Jeffrey Altman <jaltman@columbia.edu>
9205 via Richard Levitte*
9207 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
9208 already does with RSA. testdsa.h now has 'priv_key/pub_key'
9209 values for each of the key sizes rather than having just
9210 parameters (and 'speed' generating keys each time).
9214 * Speed up EVP routines.
9217 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
9218 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
9219 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
9220 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
9222 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
9223 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
9224 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
9227 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
9229 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
9233 * Added the OS2-EMX target.
9235 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
9237 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
9238 to support NCONF routines in extension code. New function CONF_set_nconf()
9239 to allow functions which take an NCONF to also handle the old LHASH
9240 structure: this means that the old CONF compatible routines can be
9241 retained (in particular wrt extensions) without having to duplicate the
9242 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
9246 * Enhance the general user interface with mechanisms for inner control
9247 and with possibilities to have yes/no kind of prompts.
9251 * Change all calls to low level digest routines in the library and
9252 applications to use EVP. Add missing calls to HMAC_cleanup() and
9253 don't assume HMAC_CTX can be copied using memcpy().
9255 *Verdon Walker <VWalker@novell.com>, Steve Henson*
9257 * Add the possibility to control engines through control names but with
9258 arbitrary arguments instead of just a string.
9259 Change the key loaders to take a UI_METHOD instead of a callback
9260 function pointer. NOTE: this breaks binary compatibility with earlier
9261 versions of OpenSSL [engine].
9262 Adapt the nCipher code for these new conditions and add a card insertion
9267 * Enhance the general user interface with mechanisms to better support
9268 dialog box interfaces, application-defined prompts, the possibility
9269 to use defaults (for example default passwords from somewhere else)
9270 and interrupts/cancellations.
9274 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
9275 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
9279 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
9280 tidy up some unnecessarily weird code in 'sk_new()').
9282 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
9284 * Change the key loading routines for ENGINEs to use the same kind
9285 callback (pem_password_cb) as all other routines that need this
9290 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
9291 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
9292 than this minimum value is recommended.
9296 * New random seeder for OpenVMS, using the system process statistics
9297 that are easily reachable.
9301 * Windows apparently can't transparently handle global
9302 variables defined in DLLs. Initialisations such as:
9304 const ASN1_ITEM *it = &ASN1_INTEGER_it;
9306 won't compile. This is used by the any applications that need to
9307 declare their own ASN1 modules. This was fixed by adding the option
9308 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
9309 needed for static libraries under Win32.
9313 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
9314 setting of purpose and trust fields. New X509_STORE trust and
9315 purpose functions and tidy up setting in other SSL functions.
9319 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
9320 structure. These are inherited by X509_STORE_CTX when it is
9321 initialised. This allows various defaults to be set in the
9322 X509_STORE structure (such as flags for CRL checking and custom
9323 purpose or trust settings) for functions which only use X509_STORE_CTX
9324 internally such as S/MIME.
9326 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
9327 trust settings if they are not set in X509_STORE. This allows X509_STORE
9328 purposes and trust (in S/MIME for example) to override any set by default.
9330 Add command line options for CRL checking to smime, s_client and s_server
9335 * Initial CRL based revocation checking. If the CRL checking flag(s)
9336 are set then the CRL is looked up in the X509_STORE structure and
9337 its validity and signature checked, then if the certificate is found
9338 in the CRL the verify fails with a revoked error.
9340 Various new CRL related callbacks added to X509_STORE_CTX structure.
9342 Command line options added to 'verify' application to support this.
9344 This needs some additional work, such as being able to handle multiple
9345 CRLs with different times, extension based lookup (rather than just
9346 by subject name) and ultimately more complete V2 CRL extension
9351 * Add a general user interface API (crypto/ui/). This is designed
9352 to replace things like des_read_password and friends (backward
9353 compatibility functions using this new API are provided).
9354 The purpose is to remove prompting functions from the DES code
9355 section as well as provide for prompting through dialog boxes in
9356 a window system and the like.
9360 * Add "ex_data" support to ENGINE so implementations can add state at a
9361 per-structure level rather than having to store it globally.
9365 * Make it possible for ENGINE structures to be copied when retrieved by
9366 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
9367 This causes the "original" ENGINE structure to act like a template,
9368 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
9369 operational state can be localised to each ENGINE structure, despite the
9370 fact they all share the same "methods". New ENGINE structures returned in
9371 this case have no functional references and the return value is the single
9372 structural reference. This matches the single structural reference returned
9373 by ENGINE_by_id() normally, when it is incremented on the pre-existing
9378 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
9379 needs to match any other type at all we need to manually clear the
9384 * Changes to the "openssl engine" utility to include;
9385 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
9386 about an ENGINE's available control commands.
9387 - executing control commands from command line arguments using the
9388 '-pre' and '-post' switches. '-post' is only used if '-t' is
9389 specified and the ENGINE is successfully initialised. The syntax for
9390 the individual commands are colon-separated, for example;
9391 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
9395 * New dynamic control command support for ENGINEs. ENGINEs can now
9396 declare their own commands (numbers), names (strings), descriptions,
9397 and input types for run-time discovery by calling applications. A
9398 subset of these commands are implicitly classed as "executable"
9399 depending on their input type, and only these can be invoked through
9400 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
9401 can be based on user input, config files, etc). The distinction is
9402 that "executable" commands cannot return anything other than a boolean
9403 result and can only support numeric or string input, whereas some
9404 discoverable commands may only be for direct use through
9405 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
9406 pointers, or other custom uses. The "executable" commands are to
9407 support parameterisations of ENGINE behaviour that can be
9408 unambiguously defined by ENGINEs and used consistently across any
9409 OpenSSL-based application. Commands have been added to all the
9410 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
9411 control over shared-library paths without source code alterations.
9415 * Changed all ENGINE implementations to dynamically allocate their
9416 ENGINEs rather than declaring them statically. Apart from this being
9417 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
9418 this also allows the implementations to compile without using the
9419 internal engine_int.h header.
9423 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
9424 'const' value. Any code that should be able to modify a RAND_METHOD
9425 should already have non-const pointers to it (ie. they should only
9426 modify their own ones).
9430 * Made a variety of little tweaks to the ENGINE code.
9431 - "atalla" and "ubsec" string definitions were moved from header files
9432 to C code. "nuron" string definitions were placed in variables
9433 rather than hard-coded - allowing parameterisation of these values
9434 later on via ctrl() commands.
9435 - Removed unused "#if 0"'d code.
9436 - Fixed engine list iteration code so it uses ENGINE_free() to release
9437 structural references.
9438 - Constified the RAND_METHOD element of ENGINE structures.
9439 - Constified various get/set functions as appropriate and added
9440 missing functions (including a catch-all ENGINE_cpy that duplicates
9441 all ENGINE values onto a new ENGINE except reference counts/state).
9442 - Removed NULL parameter checks in get/set functions. Setting a method
9443 or function to NULL is a way of cancelling out a previously set
9444 value. Passing a NULL ENGINE parameter is just plain stupid anyway
9445 and doesn't justify the extra error symbols and code.
9446 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
9447 flags from engine_int.h to engine.h.
9448 - Changed prototypes for ENGINE handler functions (init(), finish(),
9449 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
9453 * Implement binary inversion algorithm for BN_mod_inverse in addition
9454 to the algorithm using long division. The binary algorithm can be
9455 used only if the modulus is odd. On 32-bit systems, it is faster
9456 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
9457 roughly 5-15% for 256-bit moduli), so we use it only for moduli
9458 up to 450 bits. In 64-bit environments, the binary algorithm
9459 appears to be advantageous for much longer moduli; here we use it
9460 for moduli up to 2048 bits.
9464 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
9465 could not support the combine flag in choice fields.
9469 * Add a 'copy_extensions' option to the 'ca' utility. This copies
9470 extensions from a certificate request to the certificate.
9474 * Allow multiple 'certopt' and 'nameopt' options to be separated
9475 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
9476 file: this allows the display of the certificate about to be
9477 signed to be customised, to allow certain fields to be included
9478 or excluded and extension details. The old system didn't display
9479 multicharacter strings properly, omitted fields not in the policy
9480 and couldn't display additional details such as extensions.
9484 * Function EC_POINTs_mul for multiple scalar multiplication
9485 of an arbitrary number of elliptic curve points
9486 \sum scalars[i]*points[i],
9487 optionally including the generator defined for the EC_GROUP:
9488 scalar*generator + \sum scalars[i]*points[i].
9490 EC_POINT_mul is a simple wrapper function for the typical case
9491 that the point list has just one item (besides the optional
9496 * First EC_METHODs for curves over GF(p):
9498 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
9499 operations and provides various method functions that can also
9500 operate with faster implementations of modular arithmetic.
9502 EC_GFp_mont_method() reuses most functions that are part of
9503 EC_GFp_simple_method, but uses Montgomery arithmetic.
9505 *Bodo Moeller; point addition and point doubling
9506 implementation directly derived from source code provided by
9507 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
9509 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
9510 crypto/ec/ec_lib.c):
9512 Curves are EC_GROUP objects (with an optional group generator)
9513 based on EC_METHODs that are built into the library.
9515 Points are EC_POINT objects based on EC_GROUP objects.
9517 Most of the framework would be able to handle curves over arbitrary
9518 finite fields, but as there are no obvious types for fields other
9519 than GF(p), some functions are limited to that for now.
9523 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
9524 that the file contains a complete HTTP response.
9528 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
9529 change the def and num file printf format specifier from "%-40sXXX"
9530 to "%-39s XXX". The latter will always guarantee a space after the
9531 field while the former will cause them to run together if the field
9532 is 40 of more characters long.
9536 * Constify the cipher and digest 'method' functions and structures
9537 and modify related functions to take constant EVP_MD and EVP_CIPHER
9542 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
9543 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
9547 * Modify EVP_Digest*() routines so they now return values. Although the
9548 internal software routines can never fail additional hardware versions
9553 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
9555 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
9556 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
9559 ERR_R_NESTED_ASN1_ERROR
9561 ERR_R_MISSING_ASN1_EOS
9562 were 4 .. 9, conflicting with
9563 ERR_LIB_RSA (= ERR_R_RSA_LIB)
9565 ERR_LIB_PEM (= ERR_R_PEM_LIB).
9566 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
9568 Add new error code 'ERR_R_INTERNAL_ERROR'.
9572 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
9577 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
9578 sets the subject name for a new request or supersedes the
9579 subject name in a given request. Formats that can be parsed are
9580 'CN=Some Name, OU=myOU, C=IT'
9582 'CN=Some Name/OU=myOU/C=IT'.
9584 Add options '-batch' and '-verbose' to 'openssl req'.
9586 *Massimiliano Pala <madwolf@hackmasters.net>*
9588 * Introduce the possibility to access global variables through
9589 functions on platform were that's the best way to handle exporting
9590 global variables in shared libraries. To enable this functionality,
9591 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
9592 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
9593 is normally done by Configure or something similar).
9595 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
9596 in the source file (foo.c) like this:
9598 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
9599 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
9601 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
9602 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
9604 OPENSSL_DECLARE_GLOBAL(int,foo);
9605 #define foo OPENSSL_GLOBAL_REF(foo)
9606 OPENSSL_DECLARE_GLOBAL(double,bar);
9607 #define bar OPENSSL_GLOBAL_REF(bar)
9609 The #defines are very important, and therefore so is including the
9610 header file everywhere where the defined globals are used.
9612 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
9613 of ASN.1 items, but that structure is a bit different.
9615 The largest change is in util/mkdef.pl which has been enhanced with
9616 better and easier to understand logic to choose which symbols should
9617 go into the Windows .def files as well as a number of fixes and code
9618 cleanup (among others, algorithm keywords are now sorted
9619 lexicographically to avoid constant rewrites).
9623 * In BN_div() keep a copy of the sign of 'num' before writing the
9624 result to 'rm' because if rm==num the value will be overwritten
9625 and produce the wrong result if 'num' is negative: this caused
9626 problems with BN_mod() and BN_nnmod().
9630 * Function OCSP_request_verify(). This checks the signature on an
9631 OCSP request and verifies the signer certificate. The signer
9632 certificate is just checked for a generic purpose and OCSP request
9637 * Add OCSP_check_validity() function to check the validity of OCSP
9638 responses. OCSP responses are prepared in real time and may only
9639 be a few seconds old. Simply checking that the current time lies
9640 between thisUpdate and nextUpdate max reject otherwise valid responses
9641 caused by either OCSP responder or client clock inaccuracy. Instead
9642 we allow thisUpdate and nextUpdate to fall within a certain period of
9643 the current time. The age of the response can also optionally be
9644 checked. Two new options -validity_period and -status_age added to
9649 * If signature or public key algorithm is unrecognized print out its
9650 OID rather that just UNKNOWN.
9654 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
9655 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
9656 ID to be generated from the issuer certificate alone which can then be
9657 passed to OCSP_id_issuer_cmp().
9661 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
9662 ASN1 modules to export functions returning ASN1_ITEM pointers
9663 instead of the ASN1_ITEM structures themselves. This adds several
9664 new macros which allow the underlying ASN1 function/structure to
9665 be accessed transparently. As a result code should not use ASN1_ITEM
9666 references directly (such as &X509_it) but instead use the relevant
9667 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
9668 use of the new ASN1 code on platforms where exporting structures
9669 is problematical (for example in shared libraries) but exporting
9670 functions returning pointers to structures is not.
9674 * Add support for overriding the generation of SSL/TLS session IDs.
9675 These callbacks can be registered either in an SSL_CTX or per SSL.
9676 The purpose of this is to allow applications to control, if they wish,
9677 the arbitrary values chosen for use as session IDs, particularly as it
9678 can be useful for session caching in multiple-server environments. A
9679 command-line switch for testing this (and any client code that wishes
9680 to use such a feature) has been added to "s_server".
9682 *Geoff Thorpe, Lutz Jaenicke*
9684 * Modify mkdef.pl to recognise and parse preprocessor conditionals
9685 of the form '#if defined(...) || defined(...) || ...' and
9686 '#if !defined(...) && !defined(...) && ...'. This also avoids
9687 the growing number of special cases it was previously handling.
9691 * Make all configuration macros available for application by making
9692 sure they are available in opensslconf.h, by giving them names starting
9693 with "OPENSSL_" to avoid conflicts with other packages and by making
9694 sure e_os2.h will cover all platform-specific cases together with
9696 Additionally, it is now possible to define configuration/platform-
9697 specific names (called "system identities"). In the C code, these
9698 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
9699 macro with the name beginning with "OPENSSL_SYS_", which is determined
9700 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
9705 * New option -set_serial to 'req' and 'x509' this allows the serial
9706 number to use to be specified on the command line. Previously self
9707 signed certificates were hard coded with serial number 0 and the
9708 CA options of 'x509' had to use a serial number in a file which was
9713 * New options to 'ca' utility to support V2 CRL entry extensions.
9714 Currently CRL reason, invalidity date and hold instruction are
9715 supported. Add new CRL extensions to V3 code and some new objects.
9719 * New function EVP_CIPHER_CTX_set_padding() this is used to
9720 disable standard block padding (aka PKCS#5 padding) in the EVP
9721 API, which was previously mandatory. This means that the data is
9722 not padded in any way and so the total length much be a multiple
9723 of the block size, otherwise an error occurs.
9727 * Initial (incomplete) OCSP SSL support.
9731 * New function OCSP_parse_url(). This splits up a URL into its host,
9732 port and path components: primarily to parse OCSP URLs. New -url
9733 option to ocsp utility.
9737 * New nonce behavior. The return value of OCSP_check_nonce() now
9738 reflects the various checks performed. Applications can decide
9739 whether to tolerate certain situations such as an absent nonce
9740 in a response when one was present in a request: the ocsp application
9741 just prints out a warning. New function OCSP_add1_basic_nonce()
9742 this is to allow responders to include a nonce in a response even if
9743 the request is nonce-less.
9747 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
9748 skipped when using openssl x509 multiple times on a single input file,
9749 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
9753 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
9754 set string type: to handle setting ASN1_TIME structures. Fix ca
9755 utility to correctly initialize revocation date of CRLs.
9759 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
9760 the clients preferred ciphersuites and rather use its own preferences.
9761 Should help to work around M$ SGC (Server Gated Cryptography) bug in
9762 Internet Explorer by ensuring unchanged hash method during stepup.
9763 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
9767 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
9768 to aes and add a new 'exist' option to print out symbols that don't
9773 * Additional options to ocsp utility to allow flags to be set and
9774 additional certificates supplied.
9778 * Add the option -VAfile to 'openssl ocsp', so the user can give the
9779 OCSP client a number of certificate to only verify the response
9784 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
9785 handle the new API. Currently only ECB, CBC modes supported. Add new
9788 Add TLS AES ciphersuites as described in RFC3268, "Advanced
9789 Encryption Standard (AES) Ciphersuites for Transport Layer
9790 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
9791 not enabled by default and were not part of the "ALL" ciphersuite
9792 alias because they were not yet official; they could be
9793 explicitly requested by specifying the "AESdraft" ciphersuite
9794 group alias. In the final release of OpenSSL 0.9.7, the group
9795 alias is called "AES" and is part of "ALL".)
9797 *Ben Laurie, Steve Henson, Bodo Moeller*
9799 * New function OCSP_copy_nonce() to copy nonce value (if present) from
9800 request to response.
9804 * Functions for OCSP responders. OCSP_request_onereq_count(),
9805 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
9806 extract information from a certificate request. OCSP_response_create()
9807 creates a response and optionally adds a basic response structure.
9808 OCSP_basic_add1_status() adds a complete single response to a basic
9809 response and returns the OCSP_SINGLERESP structure just added (to allow
9810 extensions to be included for example). OCSP_basic_add1_cert() adds a
9811 certificate to a basic response and OCSP_basic_sign() signs a basic
9812 response with various flags. New helper functions ASN1_TIME_check()
9813 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
9814 (converts ASN1_TIME to GeneralizedTime).
9818 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
9819 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
9820 structure from a certificate. X509_pubkey_digest() digests the public_key
9821 contents: this is used in various key identifiers.
9825 * Make sk_sort() tolerate a NULL argument.
9827 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
9829 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
9830 passed by the function are trusted implicitly. If any of them signed the
9831 response then it is assumed to be valid and is not verified.
9835 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
9836 to data. This was previously part of the PKCS7 ASN1 code. This
9837 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
9838 *Steve Henson, reported by Kenneth R. Robinette
9839 <support@securenetterm.com>*
9841 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
9842 routines: without these tracing memory leaks is very painful.
9843 Fix leaks in PKCS12 and PKCS7 routines.
9847 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
9848 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
9849 effectively meant GeneralizedTime would never be used. Now it
9850 is initialised to -1 but X509_time_adj() now has to check the value
9851 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
9852 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
9853 *Steve Henson, reported by Kenneth R. Robinette
9854 <support@securenetterm.com>*
9856 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
9857 result in a zero length in the ASN1_INTEGER structure which was
9858 not consistent with the structure when d2i_ASN1_INTEGER() was used
9859 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
9860 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
9861 where it did not print out a minus for negative ASN1_INTEGER.
9865 * Add summary printout to ocsp utility. The various functions which
9866 convert status values to strings have been renamed to:
9867 OCSP_response_status_str(), OCSP_cert_status_str() and
9868 OCSP_crl_reason_str() and are no longer static. New options
9869 to verify nonce values and to disable verification. OCSP response
9870 printout format cleaned up.
9874 * Add additional OCSP certificate checks. These are those specified
9875 in RFC2560. This consists of two separate checks: the CA of the
9876 certificate being checked must either be the OCSP signer certificate
9877 or the issuer of the OCSP signer certificate. In the latter case the
9878 OCSP signer certificate must contain the OCSP signing extended key
9879 usage. This check is performed by attempting to match the OCSP
9880 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
9881 in the OCSP_CERTID structures of the response.
9885 * Initial OCSP certificate verification added to OCSP_basic_verify()
9886 and related routines. This uses the standard OpenSSL certificate
9887 verify routines to perform initial checks (just CA validity) and
9888 to obtain the certificate chain. Then additional checks will be
9889 performed on the chain. Currently the root CA is checked to see
9890 if it is explicitly trusted for OCSP signing. This is used to set
9891 a root CA as a global signing root: that is any certificate that
9892 chains to that CA is an acceptable OCSP signing certificate.
9896 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
9897 extensions from a separate configuration file.
9898 As when reading extensions from the main configuration file,
9899 the '-extensions ...' option may be used for specifying the
9902 *Massimiliano Pala <madwolf@comune.modena.it>*
9904 * New OCSP utility. Allows OCSP requests to be generated or
9905 read. The request can be sent to a responder and the output
9906 parsed, outputed or printed in text form. Not complete yet:
9907 still needs to check the OCSP response validity.
9911 * New subcommands for 'openssl ca':
9912 'openssl ca -status <serial>' prints the status of the cert with
9913 the given serial number (according to the index file).
9914 'openssl ca -updatedb' updates the expiry status of certificates
9917 *Massimiliano Pala <madwolf@comune.modena.it>*
9919 * New '-newreq-nodes' command option to CA.pl. This is like
9920 '-newreq', but calls 'openssl req' with the '-nodes' option
9921 so that the resulting key is not encrypted.
9923 *Damien Miller <djm@mindrot.org>*
9925 * New configuration for the GNU Hurd.
9927 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
9929 * Initial code to implement OCSP basic response verify. This
9930 is currently incomplete. Currently just finds the signer's
9931 certificate and verifies the signature on the response.
9935 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
9936 value of OPENSSLDIR. This is available via the new '-d' option
9937 to 'openssl version', and is also included in 'openssl version -a'.
9941 * Allowing defining memory allocation callbacks that will be given
9942 file name and line number information in additional arguments
9943 (a const char* and an int). The basic functionality remains, as
9944 well as the original possibility to just replace malloc(),
9945 realloc() and free() by functions that do not know about these
9946 additional arguments. To register and find out the current
9947 settings for extended allocation functions, the following
9948 functions are provided:
9950 CRYPTO_set_mem_ex_functions
9951 CRYPTO_set_locked_mem_ex_functions
9952 CRYPTO_get_mem_ex_functions
9953 CRYPTO_get_locked_mem_ex_functions
9955 These work the same way as CRYPTO_set_mem_functions and friends.
9956 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
9957 extended allocation function is enabled.
9958 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
9959 a conventional allocation function is enabled.
9961 *Richard Levitte, Bodo Moeller*
9963 * Finish off removing the remaining LHASH function pointer casts.
9964 There should no longer be any prototype-casting required when using
9965 the LHASH abstraction, and any casts that remain are "bugs". See
9966 the callback types and macros at the head of lhash.h for details
9967 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
9971 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
9972 If /dev/[u]random devices are not available or do not return enough
9973 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
9975 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
9976 /etc/entropy will be queried once each in this sequence, querying stops
9977 when enough entropy was collected without querying more sockets.
9981 * Change the Unix RAND_poll() variant to be able to poll several
9982 random devices, as specified by DEVRANDOM, until a sufficient amount
9983 of data has been collected. We spend at most 10 ms on each file
9984 (select timeout) and read in non-blocking mode. DEVRANDOM now
9985 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
9986 (previously it was just the string "/dev/urandom"), so on typical
9987 platforms the 10 ms delay will never occur.
9988 Also separate out the Unix variant to its own file, rand_unix.c.
9989 For VMS, there's a currently-empty rand_vms.c.
9993 * Move OCSP client related routines to ocsp_cl.c. These
9994 provide utility functions which an application needing
9995 to issue a request to an OCSP responder and analyse the
9996 response will typically need: as opposed to those which an
9997 OCSP responder itself would need which will be added later.
9999 OCSP_request_sign() signs an OCSP request with an API similar
10000 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
10001 response. OCSP_response_get1_basic() extracts basic response
10002 from response. OCSP_resp_find_status(): finds and extracts status
10003 information from an OCSP_CERTID structure (which will be created
10004 when the request structure is built). These are built from lower
10005 level functions which work on OCSP_SINGLERESP structures but
10006 won't normally be used unless the application wishes to examine
10007 extensions in the OCSP response for example.
10009 Replace nonce routines with a pair of functions.
10010 OCSP_request_add1_nonce() adds a nonce value and optionally
10011 generates a random value. OCSP_check_nonce() checks the
10012 validity of the nonce in an OCSP response.
10016 * Change function OCSP_request_add() to OCSP_request_add0_id().
10017 This doesn't copy the supplied OCSP_CERTID and avoids the
10018 need to free up the newly created id. Change return type
10019 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
10020 This can then be used to add extensions to the request.
10021 Deleted OCSP_request_new(), since most of its functionality
10022 is now in OCSP_REQUEST_new() (and the case insensitive name
10023 clash) apart from the ability to set the request name which
10024 will be added elsewhere.
10028 * Update OCSP API. Remove obsolete extensions argument from
10029 various functions. Extensions are now handled using the new
10030 OCSP extension code. New simple OCSP HTTP function which
10031 can be used to send requests and parse the response.
10035 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
10036 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
10037 uses the special reorder version of SET OF to sort the attributes
10038 and reorder them to match the encoded order. This resolves a long
10039 standing problem: a verify on a PKCS7 structure just after signing
10040 it used to fail because the attribute order did not match the
10041 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
10042 it uses the received order. This is necessary to tolerate some broken
10043 software that does not order SET OF. This is handled by encoding
10044 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
10045 to produce the required SET OF.
10049 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
10050 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
10051 files to get correct declarations of the ASN.1 item variables.
10055 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
10056 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
10057 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
10058 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
10059 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
10060 ASN1_ITEM and no wrapper functions.
10064 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
10065 replace the old function pointer based I/O routines. Change most of
10066 the *_d2i_bio() and *_d2i_fp() functions to use these.
10070 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
10071 lines, recognize more "algorithms" that can be deselected, and make
10072 it complain about algorithm deselection that isn't recognised.
10076 * New ASN1 functions to handle dup, sign, verify, digest, pack and
10077 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
10078 to use new functions. Add NO_ASN1_OLD which can be set to remove
10079 some old style ASN1 functions: this can be used to determine if old
10080 code will still work when these eventually go away.
10084 * New extension functions for OCSP structures, these follow the
10085 same conventions as certificates and CRLs.
10089 * New function X509V3_add1_i2d(). This automatically encodes and
10090 adds an extension. Its behaviour can be customised with various
10091 flags to append, replace or delete. Various wrappers added for
10092 certificates and CRLs.
10096 * Fix to avoid calling the underlying ASN1 print routine when
10097 an extension cannot be parsed. Correct a typo in the
10098 OCSP_SERVICELOC extension. Tidy up print OCSP format.
10102 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
10103 entries for variables.
10107 * Add functionality to apps/openssl.c for detecting locking
10108 problems: As the program is single-threaded, all we have
10109 to do is register a locking callback using an array for
10110 storing which locks are currently held by the program.
10114 * Use a lock around the call to CRYPTO_get_ex_new_index() in
10115 SSL_get_ex_data_X509_STORE_idx(), which is used in
10116 ssl_verify_cert_chain() and thus can be called at any time
10117 during TLS/SSL handshakes so that thread-safety is essential.
10118 Unfortunately, the ex_data design is not at all suited
10119 for multi-threaded use, so it probably should be abolished.
10123 * Added Broadcom "ubsec" ENGINE to OpenSSL.
10125 *Broadcom, tweaked and integrated by Geoff Thorpe*
10127 * Move common extension printing code to new function
10128 X509V3_print_extensions(). Reorganise OCSP print routines and
10129 implement some needed OCSP ASN1 functions. Add OCSP extensions.
10133 * New function X509_signature_print() to remove duplication in some
10138 * Add a special meaning when SET OF and SEQUENCE OF flags are both
10139 set (this was treated exactly the same as SET OF previously). This
10140 is used to reorder the STACK representing the structure to match the
10141 encoding. This will be used to get round a problem where a PKCS7
10142 structure which was signed could not be verified because the STACK
10143 order did not reflect the encoded order.
10147 * Reimplement the OCSP ASN1 module using the new code.
10151 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
10152 for its ASN1 operations. The old style function pointers still exist
10153 for now but they will eventually go away.
10157 * Merge in replacement ASN1 code from the ASN1 branch. This almost
10158 completely replaces the old ASN1 functionality with a table driven
10159 encoder and decoder which interprets an ASN1_ITEM structure describing
10160 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
10161 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
10162 has also been converted to the new form.
10166 * Change BN_mod_exp_recp so that negative moduli are tolerated
10167 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
10168 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
10169 for negative moduli.
10173 * Fix BN_uadd and BN_usub: Always return non-negative results instead
10174 of not touching the result's sign bit.
10178 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
10183 * Changed the LHASH code to use prototypes for callbacks, and created
10184 macros to declare and implement thin (optionally static) functions
10185 that provide type-safety and avoid function pointer casting for the
10186 type-specific callbacks.
10190 * Added Kerberos Cipher Suites to be used with TLS, as written in
10192 *Veers Staats <staatsvr@asc.hpc.mil>,
10193 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
10195 * Reformat the FAQ so the different questions and answers can be divided
10196 in sections depending on the subject.
10200 * Have the zlib compression code load ZLIB.DLL dynamically under
10205 * New function BN_mod_sqrt for computing square roots modulo a prime
10206 (using the probabilistic Tonelli-Shanks algorithm unless
10207 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
10208 be handled deterministically).
10210 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
10212 * Make BN_mod_inverse faster by explicitly handling small quotients
10213 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
10214 512 bits], about 30% for larger ones [1024 or 2048 bits].)
10218 * New function BN_kronecker.
10222 * Fix BN_gcd so that it works on negative inputs; the result is
10223 positive unless both parameters are zero.
10224 Previously something reasonably close to an infinite loop was
10225 possible because numbers could be growing instead of shrinking
10226 in the implementation of Euclid's algorithm.
10230 * Fix BN_is_word() and BN_is_one() macros to take into account the
10231 sign of the number in question.
10233 Fix BN_is_word(a,w) to work correctly for w == 0.
10235 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
10236 because its test if the absolute value of 'a' equals 'w'.
10237 Note that BN_abs_is_word does *not* handle w == 0 reliably;
10238 it exists mostly for use in the implementations of BN_is_zero(),
10239 BN_is_one(), and BN_is_word().
10243 * New function BN_swap.
10247 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
10248 the exponentiation functions are more likely to produce reasonable
10249 results on negative inputs.
10253 * Change BN_mod_mul so that the result is always non-negative.
10254 Previously, it could be negative if one of the factors was negative;
10255 I don't think anyone really wanted that behaviour.
10259 * Move BN_mod_... functions into new file crypto/bn/bn_mod.c
10260 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
10261 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
10262 and add new functions:
10271 BN_mod_lshift1_quick
10273 BN_mod_lshift_quick
10275 These functions always generate non-negative results.
10277 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
10278 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
10280 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
10281 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
10282 be reduced modulo m.
10284 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
10287 The following entry accidentally appeared in the CHANGES file
10288 distributed with OpenSSL 0.9.7. The modifications described in
10289 it do *not* apply to OpenSSL 0.9.7.
10291 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10292 was actually never needed) and in BN_mul(). The removal in BN_mul()
10293 required a small change in bn_mul_part_recursive() and the addition
10294 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10295 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10296 bn_sub_words() and bn_add_words() except they take arrays with
10302 * In 'openssl passwd', verify passwords read from the terminal
10303 unless the '-salt' option is used (which usually means that
10304 verification would just waste user's time since the resulting
10305 hash is going to be compared with some given password hash)
10306 or the new '-noverify' option is used.
10308 This is an incompatible change, but it does not affect
10309 non-interactive use of 'openssl passwd' (passwords on the command
10310 line, '-stdin' option, '-in ...' option) and thus should not
10311 cause any problems.
10315 * Remove all references to RSAref, since there's no more need for it.
10319 * Make DSO load along a path given through an environment variable
10320 (SHLIB_PATH) with shl_load().
10324 * Constify the ENGINE code as a result of BIGNUM constification.
10325 Also constify the RSA code and most things related to it. In a
10326 few places, most notable in the depth of the ASN.1 code, ugly
10327 casts back to non-const were required (to be solved at a later
10332 * Make it so the openssl application has all engines loaded by default.
10336 * Constify the BIGNUM routines a little more.
10340 * Add the following functions:
10342 ENGINE_load_cswift()
10344 ENGINE_load_atalla()
10345 ENGINE_load_nuron()
10346 ENGINE_load_builtin_engines()
10348 That way, an application can itself choose if external engines that
10349 are built-in in OpenSSL shall ever be used or not. The benefit is
10350 that applications won't have to be linked with libdl or other dso
10351 libraries unless it's really needed.
10353 Changed 'openssl engine' to load all engines on demand.
10354 Changed the engine header files to avoid the duplication of some
10355 declarations (they differed!).
10359 * 'openssl engine' can now list capabilities.
10363 * Better error reporting in 'openssl engine'.
10367 * Never call load_dh_param(NULL) in s_server.
10371 * Add engine application. It can currently list engines by name and
10372 identity, and test if they are actually available.
10376 * Improve RPM specification file by forcing symbolic linking and making
10377 sure the installed documentation is also owned by root.root.
10379 *Damien Miller <djm@mindrot.org>*
10381 * Give the OpenSSL applications more possibilities to make use of
10382 keys (public as well as private) handled by engines.
10386 * Add OCSP code that comes from CertCo.
10390 * Add VMS support for the Rijndael code.
10394 * Added untested support for Nuron crypto accelerator.
10398 * Add support for external cryptographic devices. This code was
10399 previously distributed separately as the "engine" branch.
10401 *Geoff Thorpe, Richard Levitte*
10403 * Rework the filename-translation in the DSO code. It is now possible to
10404 have far greater control over how a "name" is turned into a filename
10405 depending on the operating environment and any oddities about the
10406 different shared library filenames on each system.
10410 * Support threads on FreeBSD-elf in Configure.
10414 * Fix for SHA1 assembly problem with MASM: it produces
10415 warnings about corrupt line number information when assembling
10416 with debugging information. This is caused by the overlapping
10419 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
10422 NCONF_get_number() has no error checking at all. As a replacement,
10423 NCONF_get_number_e() is defined (_e for "error checking") and is
10424 promoted strongly. The old NCONF_get_number is kept around for
10425 binary backward compatibility.
10426 Make it possible for methods to load from something other than a BIO,
10427 by providing a function pointer that is given a name instead of a BIO.
10428 For example, this could be used to load configuration data from an
10433 * Fix for non blocking accept BIOs. Added new I/O special reason
10434 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
10435 with non blocking I/O was not possible because no retry code was
10436 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
10441 * Added the beginnings of Rijndael support.
10445 * Fix for bug in DirectoryString mask setting. Add support for
10446 X509_NAME_print_ex() in 'req' and X509_print_ex() function
10447 to allow certificate printing to more controllable, additional
10448 'certopt' option to 'x509' to allow new printing options to be
10453 * Clean old EAY MD5 hack from e_os.h.
10457 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004] ###
10459 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10460 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
10462 *Joe Orton, Steve Henson*
10464 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003] ###
10466 * Fix additional bug revealed by the NISCC test suite:
10468 Stop bug triggering large recursion when presented with
10469 certain ASN.1 tags (CVE-2003-0851)
10473 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003] ###
10475 * Fix various bugs revealed by running the NISCC test suite:
10477 Stop out of bounds reads in the ASN1 code when presented with
10478 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10480 If verify callback ignores invalid public key errors don't try to check
10481 certificate signature with the NULL public key.
10486 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10487 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10492 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10493 extra data after the compression methods not only for TLS 1.0
10494 but also for SSL 3.0 (as required by the specification).
10496 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10498 * Change X509_certificate_type() to mark the key as exported/exportable
10499 when it's 512 *bits* long, not 512 bytes.
10503 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003] ###
10505 * Countermeasure against the Klima-Pokorny-Rosa extension of
10506 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10507 a protocol version number mismatch like a decryption error
10508 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10512 * Turn on RSA blinding by default in the default implementation
10513 to avoid a timing attack. Applications that don't want it can call
10514 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10515 They would be ill-advised to do so in most cases.
10517 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10519 * Change RSA blinding code so that it works when the PRNG is not
10520 seeded (in this case, the secret RSA exponent is abused as
10521 an unpredictable seed -- if it is not unpredictable, there
10522 is no point in blinding anyway). Make RSA blinding thread-safe
10523 by remembering the creator's thread ID in rsa->blinding and
10524 having all other threads use local one-time blinding factors
10525 (this requires more computation than sharing rsa->blinding, but
10526 avoids excessive locking; and if an RSA object is not shared
10527 between threads, blinding will still be very fast).
10531 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003] ###
10533 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10534 via timing by performing a MAC computation even if incorrect
10535 block cipher padding has been found. This is a countermeasure
10536 against active attacks where the attacker has to distinguish
10537 between bad padding and a MAC verification error. (CVE-2003-0078)
10539 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10540 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10541 Martin Vuagnoux (EPFL, Ilion)*
10543 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002] ###
10545 * New function OPENSSL_cleanse(), which is used to cleanse a section of
10546 memory from its contents. This is done with a counter that will
10547 place alternating values in each byte. This can be used to solve
10548 two issues: 1) the removal of calls to memset() by highly optimizing
10549 compilers, and 2) cleansing with other values than 0, since those can
10550 be read through on certain media, for example a swap space on disk.
10554 * Bugfix: client side session caching did not work with external caching,
10555 because the session->cipher setting was not restored when reloading
10556 from the external cache. This problem was masked, when
10557 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
10558 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
10562 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
10563 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
10565 *Zeev Lieber <zeev-l@yahoo.com>*
10567 * Undo an undocumented change introduced in 0.9.6e which caused
10568 repeated calls to OpenSSL_add_all_ciphers() and
10569 OpenSSL_add_all_digests() to be ignored, even after calling
10574 * Change the default configuration reader to deal with last line not
10575 being properly terminated.
10579 * Change X509_NAME_cmp() so it applies the special rules on handling
10580 DN values that are of type PrintableString, as well as RDNs of type
10581 emailAddress where the value has the type ia5String.
10583 *stefank@valicert.com via Richard Levitte*
10585 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
10586 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
10587 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
10588 the bitwise-OR of the two for use by the majority of applications
10589 wanting this behaviour, and update the docs. The documented
10590 behaviour and actual behaviour were inconsistent and had been
10591 changing anyway, so this is more a bug-fix than a behavioural
10594 *Geoff Thorpe, diagnosed by Nadav Har'El*
10596 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
10597 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
10601 * Fix initialization code race conditions in
10602 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
10603 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
10604 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
10605 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
10606 ssl2_get_cipher_by_char(),
10607 ssl3_get_cipher_by_char().
10609 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
10611 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
10612 the cached sessions are flushed, as the remove_cb() might use ex_data
10613 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
10614 (see [openssl.org #212]).
10616 *Geoff Thorpe, Lutz Jaenicke*
10618 * Fix typo in OBJ_txt2obj which incorrectly passed the content
10619 length, instead of the encoding length to d2i_ASN1_OBJECT.
10623 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002] ###
10625 * [In 0.9.6g-engine release:]
10626 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
10628 *Lynn Gazis <lgazis@rainbow.com>*
10630 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002] ###
10632 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
10633 and get fix the header length calculation.
10634 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
10635 Alon Kantor <alonk@checkpoint.com> (and others),
10638 * Use proper error handling instead of 'assertions' in buffer
10639 overflow checks added in 0.9.6e. This prevents DoS (the
10640 assertions could call abort()).
10642 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
10644 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002] ###
10646 * Add various sanity checks to asn1_get_length() to reject
10647 the ASN1 length bytes if they exceed sizeof(long), will appear
10648 negative or the content length exceeds the length of the
10651 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
10653 * Fix cipher selection routines: ciphers without encryption had no flags
10654 for the cipher strength set and where therefore not handled correctly
10655 by the selection routines (PR #130).
10659 * Fix EVP_dsa_sha macro.
10664 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
10665 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
10666 that was added in OpenSSL 0.9.6d.
10668 As the countermeasure turned out to be incompatible with some
10669 broken SSL implementations, the new option is part of SSL_OP_ALL.
10670 SSL_OP_ALL is usually employed when compatibility with weird SSL
10671 implementations is desired (e.g. '-bugs' option to 's_client' and
10672 's_server'), so the new option is automatically set in many
10677 * Changes in security patch:
10679 Changes marked "(CHATS)" were sponsored by the Defense Advanced
10680 Research Projects Agency (DARPA) and Air Force Research Laboratory,
10681 Air Force Materiel Command, USAF, under agreement number
10684 * Add various sanity checks to asn1_get_length() to reject
10685 the ASN1 length bytes if they exceed sizeof(long), will appear
10686 negative or the content length exceeds the length of the
10687 supplied buffer. (CVE-2002-0659)
10689 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
10691 * Assertions for various potential buffer overflows, not known to
10692 happen in practice.
10694 *Ben Laurie (CHATS)*
10696 * Various temporary buffers to hold ASCII versions of integers were
10697 too small for 64 bit platforms. (CVE-2002-0655)
10698 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
10700 * Remote buffer overflow in SSL3 protocol - an attacker could
10701 supply an oversized session ID to a client. (CVE-2002-0656)
10703 [Ben Laurie (CHATS)*
10705 * Remote buffer overflow in SSL2 protocol - an attacker could
10706 supply an oversized client master key. (CVE-2002-0656)
10708 *Ben Laurie (CHATS)*
10710 ### Changes between 0.9.6c and 0.9.6d [9 May 2002] ###
10712 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
10713 encoded as NULL) with id-dsa-with-sha1.
10715 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
10717 * Check various X509_...() return values in apps/req.c.
10719 *Nils Larsch <nla@trustcenter.de>*
10721 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
10722 an end-of-file condition would erroneously be flagged, when the CRLF
10723 was just at the end of a processed block. The bug was discovered when
10724 processing data through a buffering memory BIO handing the data to a
10725 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
10726 <ptsekov@syntrex.com> and Nedelcho Stanev.
10730 * Implement a countermeasure against a vulnerability recently found
10731 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
10732 before application data chunks to avoid the use of known IVs
10733 with data potentially chosen by the attacker.
10737 * Fix length checks in ssl3_get_client_hello().
10741 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
10742 to prevent ssl3_read_internal() from incorrectly assuming that
10743 ssl3_read_bytes() found application data while handshake
10744 processing was enabled when in fact s->s3->in_read_app_data was
10745 merely automatically cleared during the initial handshake.
10747 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
10749 * Fix object definitions for Private and Enterprise: they were not
10750 recognized in their shortname (=lowercase) representation. Extend
10751 obj_dat.pl to issue an error when using undefined keywords instead
10752 of silently ignoring the problem (Svenning Sorensen
10753 <sss@sss.dnsalias.net>).
10757 * Fix DH_generate_parameters() so that it works for 'non-standard'
10758 generators, i.e. generators other than 2 and 5. (Previously, the
10759 code did not properly initialise the 'add' and 'rem' values to
10760 BN_generate_prime().)
10762 In the new general case, we do not insist that 'generator' is
10763 actually a primitive root: This requirement is rather pointless;
10764 a generator of the order-q subgroup is just as good, if not
10769 * Map new X509 verification errors to alerts. Discovered and submitted by
10770 Tom Wu <tom@arcot.com>.
10774 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
10775 returning non-zero before the data has been completely received
10776 when using non-blocking I/O.
10778 *Bodo Moeller; problem pointed out by John Hughes*
10780 * Some of the ciphers missed the strength entry (SSL_LOW etc).
10782 *Ben Laurie, Lutz Jaenicke*
10784 * Fix bug in SSL_clear(): bad sessions were not removed (found by
10785 Yoram Zahavi <YoramZ@gilian.com>).
10789 * Add information about CygWin 1.3 and on, and preserve proper
10790 configuration for the versions before that.
10792 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
10794 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
10795 check whether we deal with a copy of a session and do not delete from
10796 the cache in this case. Problem reported by "Izhar Shoshani Levi"
10797 <izhar@checkpoint.com>.
10801 * Do not store session data into the internal session cache, if it
10802 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
10803 flag is set). Proposed by Aslam <aslam@funk.com>.
10807 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
10812 * [In 0.9.6d-engine release:]
10813 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
10815 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
10817 * Add the configuration target linux-s390x.
10819 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
10821 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
10822 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
10823 variable as an indication that a ClientHello message has been
10824 received. As the flag value will be lost between multiple
10825 invocations of ssl3_accept when using non-blocking I/O, the
10826 function may not be aware that a handshake has actually taken
10827 place, thus preventing a new session from being added to the
10830 To avoid this problem, we now set s->new_session to 2 instead of
10831 using a local variable.
10833 *Lutz Jaenicke, Bodo Moeller*
10835 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
10836 if the SSL_R_LENGTH_MISMATCH error is detected.
10838 *Geoff Thorpe, Bodo Moeller*
10840 * New 'shared_ldflag' column in Configure platform table.
10844 * Fix EVP_CIPHER_mode macro.
10846 *"Dan S. Camper" <dan@bti.net>*
10848 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
10849 type, we must throw them away by setting rr->length to 0.
10851 *D P Chang <dpc@qualys.com>*
10853 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001] ###
10855 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
10856 <Dominikus.Scherkl@biodata.com>. (The previous implementation
10857 worked incorrectly for those cases where range = 10..._2 and
10858 3*range is two bits longer than range.)
10862 * Only add signing time to PKCS7 structures if it is not already
10867 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
10868 OBJ_ld_ce should be OBJ_id_ce.
10869 Also some ip-pda OIDs in crypto/objects/objects.txt were
10870 incorrect (cf. RFC 3039).
10872 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
10874 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
10875 returns early because it has nothing to do.
10877 *Andy Schneider <andy.schneider@bjss.co.uk>*
10879 * [In 0.9.6c-engine release:]
10880 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
10882 *Andy Schneider <andy.schneider@bjss.co.uk>*
10884 * [In 0.9.6c-engine release:]
10885 Add support for Cryptographic Appliance's keyserver technology.
10886 (Use engine 'keyclient')
10888 *Cryptographic Appliances and Geoff Thorpe*
10890 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
10891 is called via tools/c89.sh because arguments have to be
10892 rearranged (all '-L' options must appear before the first object
10895 *Richard Shapiro <rshapiro@abinitio.com>*
10897 * [In 0.9.6c-engine release:]
10898 Add support for Broadcom crypto accelerator cards, backported
10901 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
10903 * [In 0.9.6c-engine release:]
10904 Add support for SureWare crypto accelerator cards from
10905 Baltimore Technologies. (Use engine 'sureware')
10907 *Baltimore Technologies and Mark Cox*
10909 * [In 0.9.6c-engine release:]
10910 Add support for crypto accelerator cards from Accelerated
10911 Encryption Processing, www.aep.ie. (Use engine 'aep')
10913 *AEP Inc. and Mark Cox*
10915 * Add a configuration entry for gcc on UnixWare.
10917 *Gary Benson <gbenson@redhat.com>*
10919 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
10920 messages are stored in a single piece (fixed-length part and
10921 variable-length part combined) and fix various bugs found on the way.
10925 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
10926 instead. BIO_gethostbyname() does not know what timeouts are
10927 appropriate, so entries would stay in cache even when they have
10929 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
10931 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
10932 faced with a pathologically small ClientHello fragment that does
10933 not contain client_version: Instead of aborting with an error,
10934 simply choose the highest available protocol version (i.e.,
10935 TLS 1.0 unless it is disabled). In practice, ClientHello
10936 messages are never sent like this, but this change gives us
10937 strictly correct behaviour at least for TLS.
10941 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
10942 never resets s->method to s->ctx->method when called from within
10943 one of the SSL handshake functions.
10945 *Bodo Moeller; problem pointed out by Niko Baric*
10947 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
10948 (sent using the client's version number) if client_version is
10949 smaller than the protocol version in use. Also change
10950 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
10951 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
10952 the client will at least see that alert.
10956 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
10961 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
10962 client receives HelloRequest while in a handshake.
10964 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
10966 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
10967 should end in 'break', not 'goto end' which circumvents various
10968 cleanups done in state SSL_ST_OK. But session related stuff
10969 must be disabled for SSL_ST_OK in the case that we just sent a
10972 Also avoid some overhead by not calling ssl_init_wbio_buffer()
10973 before just sending a HelloRequest.
10975 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
10977 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
10978 reveal whether illegal block cipher padding was found or a MAC
10979 verification error occurred. (Neither SSLerr() codes nor alerts
10980 are directly visible to potential attackers, but the information
10981 may leak via logfiles.)
10983 Similar changes are not required for the SSL 2.0 implementation
10984 because the number of padding bytes is sent in clear for SSL 2.0,
10985 and the extra bytes are just ignored. However ssl/s2_pkt.c
10986 failed to verify that the purported number of padding bytes is in
10991 * Add OpenUNIX-8 support including shared libraries
10992 (Boyd Lynn Gerber <gerberb@zenez.com>).
10996 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
10997 'wristwatch attack' using huge encoding parameters (cf.
10998 James H. Manger's CRYPTO 2001 paper). Note that the
10999 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
11000 encoding parameters and hence was not vulnerable.
11004 * BN_sqr() bug fix.
11006 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
11008 * Rabin-Miller test analyses assume uniformly distributed witnesses,
11009 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
11010 followed by modular reduction.
11012 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
11014 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
11015 equivalent based on BN_pseudo_rand() instead of BN_rand().
11019 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
11020 This function was broken, as the check for a new client hello message
11021 to handle SGC did not allow these large messages.
11022 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
11026 * Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
11030 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
11031 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
11035 * Rework the configuration and shared library support for Tru64 Unix.
11036 The configuration part makes use of modern compiler features and
11037 still retains old compiler behavior for those that run older versions
11038 of the OS. The shared library support part includes a variant that
11039 uses the RPATH feature, and is available through the special
11040 configuration target "alpha-cc-rpath", which will never be selected
11043 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
11045 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
11046 with the same message size as in ssl3_get_certificate_request().
11047 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
11048 messages might inadvertently be reject as too long.
11050 *Petr Lampa <lampa@fee.vutbr.cz>*
11052 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
11056 * Modified SSL library such that the verify_callback that has been set
11057 specificly for an SSL object with SSL_set_verify() is actually being
11058 used. Before the change, a verify_callback set with this function was
11059 ignored and the verify_callback() set in the SSL_CTX at the time of
11060 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
11061 to allow the necessary settings.
11065 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
11066 explicitly to NULL, as at least on Solaris 8 this seems not always to be
11067 done automatically (in contradiction to the requirements of the C
11068 standard). This made problems when used from OpenSSH.
11072 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
11073 dh->length and always used
11075 BN_rand_range(priv_key, dh->p).
11077 BN_rand_range() is not necessary for Diffie-Hellman, and this
11078 specific range makes Diffie-Hellman unnecessarily inefficient if
11079 dh->length (recommended exponent length) is much smaller than the
11080 length of dh->p. We could use BN_rand_range() if the order of
11081 the subgroup was stored in the DH structure, but we only have
11086 BN_rand(priv_key, l, ...)
11088 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
11095 RSA_eay_public_encrypt
11096 RSA_eay_private_decrypt
11097 RSA_eay_private_encrypt (signing)
11098 RSA_eay_public_decrypt (signature verification)
11100 (default implementations for RSA_public_encrypt,
11101 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
11102 always reject numbers >= n.
11106 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
11107 to synchronize access to 'locking_thread'. This is necessary on
11108 systems where access to 'locking_thread' (an 'unsigned long'
11109 variable) is not atomic.
11113 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
11114 *before* setting the 'crypto_lock_rand' flag. The previous code had
11115 a race condition if 0 is a valid thread ID.
11117 *Travis Vitek <vitek@roguewave.com>*
11119 * Add support for shared libraries under Irix.
11121 *Albert Chin-A-Young <china@thewrittenword.com>*
11123 * Add configuration option to build on Linux on both big-endian and
11124 little-endian MIPS.
11126 *Ralf Baechle <ralf@uni-koblenz.de>*
11128 * Add the possibility to create shared libraries on HP-UX.
11132 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001] ###
11134 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
11135 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
11136 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
11137 PRNG state recovery was possible based on the output of
11138 one PRNG request appropriately sized to gain knowledge on
11139 'md' followed by enough consecutive 1-byte PRNG requests
11140 to traverse all of 'state'.
11142 1. When updating 'md_local' (the current thread's copy of 'md')
11143 during PRNG output generation, hash all of the previous
11144 'md_local' value, not just the half used for PRNG output.
11146 2. Make the number of bytes from 'state' included into the hash
11147 independent from the number of PRNG bytes requested.
11149 The first measure alone would be sufficient to avoid
11150 Markku-Juhani's attack. (Actually it had never occurred
11151 to me that the half of 'md_local' used for chaining was the
11152 half from which PRNG output bytes were taken -- I had always
11153 assumed that the secret half would be used.) The second
11154 measure makes sure that additional data from 'state' is never
11155 mixed into 'md_local' in small portions; this heuristically
11156 further strengthens the PRNG.
11160 * Fix crypto/bn/asm/mips3.s.
11164 * When only the key is given to "enc", the IV is undefined. Print out
11165 an error message in this case.
11169 * Handle special case when X509_NAME is empty in X509 printing routines.
11173 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
11174 positive and less than q.
11178 * Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
11179 used: it isn't thread safe and the add_lock_callback should handle
11182 *Paul Rose <Paul.Rose@bridge.com>*
11184 * Verify that incoming data obeys the block size in
11185 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
11191 *Ulf Möller, Bodo Möller*
11193 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
11194 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
11195 when fixing the server behaviour for backwards-compatible 'client
11196 hello' messages. (Note that the attack is impractical against
11197 SSL 3.0 and TLS 1.0 anyway because length and version checking
11198 means that the probability of guessing a valid ciphertext is
11199 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
11202 Before 0.9.5, the countermeasure (hide the error by generating a
11203 random 'decryption result') did not work properly because
11204 ERR_clear_error() was missing, meaning that SSL_get_error() would
11205 detect the supposedly ignored error.
11207 Both problems are now fixed.
11211 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
11212 (previously it was 1024).
11216 * Fix for compatibility mode trust settings: ignore trust settings
11217 unless some valid trust or reject settings are present.
11221 * Fix for blowfish EVP: its a variable length cipher.
11225 * Fix various bugs related to DSA S/MIME verification. Handle missing
11226 parameters in DSA public key structures and return an error in the
11227 DSA routines if parameters are absent.
11231 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
11232 in the current directory if neither $RANDFILE nor $HOME was set.
11233 RAND_file_name() in 0.9.6a returned NULL in this case. This has
11234 caused some confusion to Windows users who haven't defined $HOME.
11235 Thus RAND_file_name() is changed again: e_os.h can define a
11236 DEFAULT_HOME, which will be used if $HOME is not set.
11237 For Windows, we use "C:"; on other platforms, we still require
11238 environment variables.
11240 * Move 'if (!initialized) RAND_poll()' into regions protected by
11241 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
11242 having multiple threads call RAND_poll() concurrently.
11246 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
11247 combination of a flag and a thread ID variable.
11248 Otherwise while one thread is in ssleay_rand_bytes (which sets the
11249 flag), *other* threads can enter ssleay_add_bytes without obeying
11250 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
11251 that they do not hold after the first thread unsets add_do_not_lock).
11255 * Change bctest again: '-x' expressions are not available in all
11256 versions of 'test'.
11260 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001] ###
11262 * Fix a couple of memory leaks in PKCS7_dataDecode()
11264 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
11266 * Change Configure and Makefiles to provide EXE_EXT, which will contain
11267 the default extension for executables, if any. Also, make the perl
11268 scripts that use symlink() to test if it really exists and use "cp"
11269 if it doesn't. All this made OpenSSL compilable and installable in
11274 * Fix for asn1_GetSequence() for indefinite length constructed data.
11275 If SEQUENCE is length is indefinite just set c->slen to the total
11276 amount of data available.
11278 *Steve Henson, reported by shige@FreeBSD.org*
11280 *This change does not apply to 0.9.7.*
11282 * Change bctest to avoid here-documents inside command substitution
11283 (workaround for FreeBSD /bin/sh bug).
11284 For compatibility with Ultrix, avoid shell functions (introduced
11285 in the bctest version that searches along $PATH).
11289 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
11290 with des_encrypt() defined on some operating systems, like Solaris
11295 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
11296 On the Importance of Eliminating Errors in Cryptographic
11297 Computations, J. Cryptology 14 (2001) 2, 101-119,
11298 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
11302 * MIPS assembler BIGNUM division bug fix.
11306 * Disabled incorrect Alpha assembler code.
11310 * Fix PKCS#7 decode routines so they correctly update the length
11311 after reading an EOC for the EXPLICIT tag.
11315 *This change does not apply to 0.9.7.*
11317 * Fix bug in PKCS#12 key generation routines. This was triggered
11318 if a 3DES key was generated with a 0 initial byte. Include
11319 PKCS12_BROKEN_KEYGEN compilation option to retain the old
11320 (but broken) behaviour.
11324 * Enhance bctest to search for a working bc along $PATH and print
11327 *Tim Rice <tim@multitalents.net> via Richard Levitte*
11329 * Fix memory leaks in err.c: free err_data string if necessary;
11330 don't write to the wrong index in ERR_set_error_data.
11334 * Implement ssl23_peek (analogous to ssl23_read), which previously
11339 * Replace rdtsc with _emit statements for VC++ version 5.
11341 *Jeremy Cooper <jeremy@baymoo.org>*
11343 * Make it possible to reuse SSLv2 sessions.
11347 * In copy_email() check for >= 0 as a return value for
11348 X509_NAME_get_index_by_NID() since 0 is a valid index.
11350 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
11352 * Avoid coredump with unsupported or invalid public keys by checking if
11353 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
11354 PKCS7_verify() fails with non detached data.
11358 * Don't use getenv in library functions when run as setuid/setgid.
11359 New function OPENSSL_issetugid().
11363 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
11364 due to incorrect handling of multi-threading:
11366 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
11368 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
11370 3. Count how many times MemCheck_off() has been called so that
11371 nested use can be treated correctly. This also avoids
11372 inband-signalling in the previous code (which relied on the
11373 assumption that thread ID 0 is impossible).
11377 * Add "-rand" option also to s_client and s_server.
11381 * Fix CPU detection on Irix 6.x.
11382 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
11383 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
11385 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
11390 *This change does not apply to 0.9.7.*
11392 * Use the cached encoding of an X509_NAME structure rather than
11393 copying it. This is apparently the reason for the libsafe "errors"
11394 but the code is actually correct.
11398 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
11399 Bleichenbacher's DSA attack.
11400 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
11401 to be set and top=0 forces the highest bit to be set; top=-1 is new
11402 and leaves the highest bit random.
11404 *Ulf Moeller, Bodo Moeller*
11406 * In the NCONF_...-based implementations for CONF_... queries
11407 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
11408 a temporary CONF structure with the data component set to NULL
11409 (which gives segmentation faults in lh_retrieve).
11410 Instead, use NULL for the CONF pointer in CONF_get_string and
11411 CONF_get_number (which may use environment variables) and directly
11412 return NULL from CONF_get_section.
11416 * Fix potential buffer overrun for EBCDIC.
11420 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
11421 keyUsage if basicConstraints absent for a CA.
11425 * Make SMIME_write_PKCS7() write mail header values with a format that
11426 is more generally accepted (no spaces before the semicolon), since
11427 some programs can't parse those values properly otherwise. Also make
11428 sure BIO's that break lines after each write do not create invalid
11433 * Make the CRL encoding routines work with empty SEQUENCE OF. The
11434 macros previously used would not encode an empty SEQUENCE OF
11435 and break the signature.
11439 *This change does not apply to 0.9.7.*
11441 * Zero the premaster secret after deriving the master secret in
11446 * Add some EVP_add_digest_alias registrations (as found in
11447 OpenSSL_add_all_digests()) to SSL_library_init()
11448 aka OpenSSL_add_ssl_algorithms(). This provides improved
11449 compatibility with peers using X.509 certificates
11450 with unconventional AlgorithmIdentifier OIDs.
11454 * Fix for Irix with NO_ASM.
11456 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
11458 * ./config script fixes.
11460 *Ulf Moeller, Richard Levitte*
11462 * Fix 'openssl passwd -1'.
11466 * Change PKCS12_key_gen_asc() so it can cope with non null
11467 terminated strings whose length is passed in the passlen
11468 parameter, for example from PEM callbacks. This was done
11469 by adding an extra length parameter to asc2uni().
11471 *Steve Henson, reported by <oddissey@samsung.co.kr>*
11473 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
11474 call failed, free the DSA structure.
11478 * Fix to uni2asc() to cope with zero length Unicode strings.
11479 These are present in some PKCS#12 files.
11483 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
11484 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
11485 when writing a 32767 byte record.
11487 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
11489 * In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
11490 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
11492 (RSA objects have a reference count access to which is protected
11493 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
11494 so they are meant to be shared between threads.)
11495 *Bodo Moeller, Geoff Thorpe; original patch submitted by
11496 "Reddie, Steven" <Steven.Reddie@ca.com>*
11498 * Fix a deadlock in CRYPTO_mem_leaks().
11502 * Use better test patterns in bntest.
11506 * rand_win.c fix for Borland C.
11510 * BN_rshift bugfix for n == 0.
11514 * Add a 'bctest' script that checks for some known 'bc' bugs
11515 so that 'make test' does not abort just because 'bc' is broken.
11519 * Store verify_result within SSL_SESSION also for client side to
11520 avoid potential security hole. (Re-used sessions on the client side
11521 always resulted in verify_result==X509_V_OK, not using the original
11522 result of the server certificate verification.)
11526 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
11527 SSL3_RT_APPLICATION_DATA, return 0.
11528 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
11533 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
11534 releases, have been re-implemented by renaming the previous
11535 implementations of ssl2_read and ssl3_read to ssl2_read_internal
11536 and ssl3_read_internal, respectively, and adding 'peek' parameters
11537 to them. The new ssl[23]_{read,peek} functions are calls to
11538 ssl[23]_read_internal with the 'peek' flag set appropriately.
11539 A 'peek' parameter has also been added to ssl3_read_bytes, which
11540 does the actual work for ssl3_read_internal.
11544 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
11545 the method-specific "init()" handler. Also clean up ex_data after
11546 calling the method-specific "finish()" handler. Previously, this was
11547 happening the other way round.
11551 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
11552 The previous value, 12, was not always sufficient for BN_mod_exp().
11556 * Make sure that shared libraries get the internal name engine with
11557 the full version number and not just 0. This should mark the
11558 shared libraries as not backward compatible. Of course, this should
11559 be changed again when we can guarantee backward binary compatibility.
11563 * Fix typo in get_cert_by_subject() in by_dir.c
11565 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
11567 * Rework the system to generate shared libraries:
11569 - Make note of the expected extension for the shared libraries and
11570 if there is a need for symbolic links from for example libcrypto.so.0
11571 to libcrypto.so.0.9.7. There is extended info in Configure for
11574 - Make as few rebuilds of the shared libraries as possible.
11576 - Still avoid linking the OpenSSL programs with the shared libraries.
11578 - When installing, install the shared libraries separately from the
11583 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
11585 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
11586 and not in SSL_clear because the latter is also used by the
11587 accept/connect functions; previously, the settings made by
11588 SSL_set_read_ahead would be lost during the handshake.
11590 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
11592 * Correct util/mkdef.pl to be selective about disabled algorithms.
11593 Previously, it would create entries for disabled algorithms no
11598 * Added several new manual pages for SSL_* function.
11602 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000] ###
11604 * In ssl23_get_client_hello, generate an error message when faced
11605 with an initial SSL 3.0/TLS record that is too small to contain the
11606 first two bytes of the ClientHello message, i.e. client_version.
11607 (Note that this is a pathologic case that probably has never happened
11608 in real life.) The previous approach was to use the version number
11609 from the record header as a substitute; but our protocol choice
11610 should not depend on that one because it is not authenticated
11611 by the Finished messages.
11615 * More robust randomness gathering functions for Windows.
11617 *Jeffrey Altman <jaltman@columbia.edu>*
11619 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
11620 not set then we don't setup the error code for issuer check errors
11621 to avoid possibly overwriting other errors which the callback does
11622 handle. If an application does set the flag then we assume it knows
11623 what it is doing and can handle the new informational codes
11628 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
11629 a general "ANY" type, as such it should be able to decode anything
11630 including tagged types. However it didn't check the class so it would
11631 wrongly interpret tagged types in the same way as their universal
11632 counterpart and unknown types were just rejected. Changed so that the
11633 tagged and unknown types are handled in the same way as a SEQUENCE:
11634 that is the encoding is stored intact. There is also a new type
11635 "V_ASN1_OTHER" which is used when the class is not universal, in this
11636 case we have no idea what the actual type is so we just lump them all
11641 * On VMS, stdout may very well lead to a file that is written to
11642 in a record-oriented fashion. That means that every write() will
11643 write a separate record, which will be read separately by the
11644 programs trying to read from it. This can be very confusing.
11646 The solution is to put a BIO filter in the way that will buffer
11647 text until a linefeed is reached, and then write everything a
11648 line at a time, so every record written will be an actual line,
11649 not chunks of lines and not (usually doesn't happen, but I've
11650 seen it once) several lines in one record. BIO_f_linebuffer() is
11653 Currently, it's a VMS-only method, because that's where it has
11654 been tested well enough.
11658 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
11659 it can return incorrect results.
11660 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
11661 but it was in 0.9.6-beta[12].)
11665 * Disable the check for content being present when verifying detached
11666 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
11667 include zero length content when signing messages.
11671 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
11672 BIO_ctrl (for BIO pairs).
11676 * Add DSO method for VMS.
11680 * Bug fix: Montgomery multiplication could produce results with the
11685 * Add RPM specification openssl.spec and modify it to build three
11686 packages. The default package contains applications, application
11687 documentation and run-time libraries. The devel package contains
11688 include files, static libraries and function documentation. The
11689 doc package contains the contents of the doc directory. The original
11690 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
11694 * Add a large number of documentation files for many SSL routines.
11696 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
11698 * Add a configuration entry for Sony News 4.
11700 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
11702 * Don't set the two most significant bits to one when generating a
11703 random number < q in the DSA library.
11707 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
11708 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
11709 the underlying transport is blocking) if a handshake took place.
11710 (The default behaviour is needed by applications such as s_client
11711 and s_server that use select() to determine when to use SSL_read;
11712 but for applications that know in advance when to expect data, it
11713 just makes things more complicated.)
11717 * Add RAND_egd_bytes(), which gives control over the number of bytes read
11722 * Add a few more EBCDIC conditionals that make `req' and `x509'
11723 work better on such systems.
11725 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
11727 * Add two demo programs for PKCS12_parse() and PKCS12_create().
11728 Update PKCS12_parse() so it copies the friendlyName and the
11729 keyid to the certificates aux info.
11733 * Fix bug in PKCS7_verify() which caused an infinite loop
11734 if there was more than one signature.
11736 *Sven Uszpelkat <su@celocom.de>*
11738 * Major change in util/mkdef.pl to include extra information
11739 about each symbol, as well as presenting variables as well
11740 as functions. This change means that there's n more need
11741 to rebuild the .num files when some algorithms are excluded.
11745 * Allow the verify time to be set by an application,
11746 rather than always using the current time.
11750 * Phase 2 verify code reorganisation. The certificate
11751 verify code now looks up an issuer certificate by a
11752 number of criteria: subject name, authority key id
11753 and key usage. It also verifies self signed certificates
11754 by the same criteria. The main comparison function is
11755 X509_check_issued() which performs these checks.
11757 Lot of changes were necessary in order to support this
11758 without completely rewriting the lookup code.
11760 Authority and subject key identifier are now cached.
11762 The LHASH 'certs' is X509_STORE has now been replaced
11763 by a STACK_OF(X509_OBJECT). This is mainly because an
11764 LHASH can't store or retrieve multiple objects with
11765 the same hash value.
11767 As a result various functions (which were all internal
11768 use only) have changed to handle the new X509_STORE
11769 structure. This will break anything that messed round
11770 with X509_STORE internally.
11772 The functions X509_STORE_add_cert() now checks for an
11773 exact match, rather than just subject name.
11775 The X509_STORE API doesn't directly support the retrieval
11776 of multiple certificates matching a given criteria, however
11777 this can be worked round by performing a lookup first
11778 (which will fill the cache with candidate certificates)
11779 and then examining the cache for matches. This is probably
11780 the best we can do without throwing out X509_LOOKUP
11781 entirely (maybe later...).
11783 The X509_VERIFY_CTX structure has been enhanced considerably.
11785 All certificate lookup operations now go via a get_issuer()
11786 callback. Although this currently uses an X509_STORE it
11787 can be replaced by custom lookups. This is a simple way
11788 to bypass the X509_STORE hackery necessary to make this
11789 work and makes it possible to use more efficient techniques
11790 in future. A very simple version which uses a simple
11791 STACK for its trusted certificate store is also provided
11792 using X509_STORE_CTX_trusted_stack().
11794 The verify_cb() and verify() callbacks now have equivalents
11795 in the X509_STORE_CTX structure.
11797 X509_STORE_CTX also has a 'flags' field which can be used
11798 to customise the verify behaviour.
11802 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
11803 excludes S/MIME capabilities.
11807 * When a certificate request is read in keep a copy of the
11808 original encoding of the signed data and use it when outputting
11809 again. Signatures then use the original encoding rather than
11810 a decoded, encoded version which may cause problems if the
11811 request is improperly encoded.
11815 * For consistency with other BIO_puts implementations, call
11816 buffer_write(b, ...) directly in buffer_puts instead of calling
11819 In BIO_puts, increment b->num_write as in BIO_write.
11821 *Peter.Sylvester@EdelWeb.fr*
11823 * Fix BN_mul_word for the case where the word is 0. (We have to use
11824 BN_zero, we may not return a BIGNUM with an array consisting of
11825 words set to zero.)
11829 * Avoid calling abort() from within the library when problems are
11830 detected, except if preprocessor symbols have been defined
11831 (such as REF_CHECK, BN_DEBUG etc.).
11835 * New openssl application 'rsautl'. This utility can be
11836 used for low level RSA operations. DER public key
11837 BIO/fp routines also added.
11841 * New Configure entry and patches for compiling on QNX 4.
11843 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
11845 * A demo state-machine implementation was sponsored by
11846 Nuron (http://www.nuron.com/) and is now available in
11847 demos/state_machine.
11851 * New options added to the 'dgst' utility for signature
11852 generation and verification.
11856 * Unrecognized PKCS#7 content types are now handled via a
11857 catch all ASN1_TYPE structure. This allows unsupported
11858 types to be stored as a "blob" and an application can
11859 encode and decode it manually.
11863 * Fix various signed/unsigned issues to make a_strex.c
11864 compile under VC++.
11866 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
11868 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
11869 length if passed a buffer. ASN1_INTEGER_to_BN failed
11870 if passed a NULL BN and its argument was negative.
11872 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
11874 * Modification to PKCS#7 encoding routines to output definite
11875 length encoding. Since currently the whole structures are in
11876 memory there's not real point in using indefinite length
11877 constructed encoding. However if OpenSSL is compiled with
11878 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
11882 * Added BIO_vprintf() and BIO_vsnprintf().
11886 * Added more prefixes to parse for in the strings written
11887 through a logging bio, to cover all the levels that are available
11888 through syslog. The prefixes are now:
11890 PANIC, EMERG, EMR => LOG_EMERG
11891 ALERT, ALR => LOG_ALERT
11892 CRIT, CRI => LOG_CRIT
11893 ERROR, ERR => LOG_ERR
11894 WARNING, WARN, WAR => LOG_WARNING
11895 NOTICE, NOTE, NOT => LOG_NOTICE
11896 INFO, INF => LOG_INFO
11897 DEBUG, DBG => LOG_DEBUG
11899 and as before, if none of those prefixes are present at the
11900 beginning of the string, LOG_ERR is chosen.
11902 On Win32, the LOG_* levels are mapped according to this:
11904 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
11905 LOG_WARNING => EVENTLOG_WARNING_TYPE
11906 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
11911 * Made it possible to reconfigure with just the configuration
11912 argument "reconf" or "reconfigure". The command line arguments
11913 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
11914 and are retrieved from there when reconfiguring.
11920 *Assar Westerlund <assar@sics.se>, Richard Levitte*
11922 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
11926 * The obj_dat.pl script was messing up the sorting of object
11927 names. The reason was that it compared the quoted version
11928 of strings as a result "OCSP" > "OCSP Signing" because
11929 " > SPACE. Changed script to store unquoted versions of
11930 names and add quotes on output. It was also omitting some
11931 names from the lookup table if they were given a default
11932 value (that is if SN is missing it is given the same
11933 value as LN and vice versa), these are now added on the
11934 grounds that if an object has a name we should be able to
11935 look it up. Finally added warning output when duplicate
11936 short or long names are found.
11940 * Changes needed for Tandem NSK.
11942 *Scott Uroff <scott@xypro.com>*
11944 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
11945 RSA_padding_check_SSLv23(), special padding was never detected
11946 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
11947 version rollback attacks was not effective.
11949 In s23_clnt.c, don't use special rollback-attack detection padding
11950 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
11951 client; similarly, in s23_srvr.c, don't do the rollback check if
11952 SSL 2.0 is the only protocol enabled in the server.
11956 * Make it possible to get hexdumps of unprintable data with 'openssl
11957 asn1parse'. By implication, the functions ASN1_parse_dump() and
11958 BIO_dump_indent() are added.
11962 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
11963 these print out strings and name structures based on various
11964 flags including RFC2253 support and proper handling of
11965 multibyte characters. Added options to the 'x509' utility
11966 to allow the various flags to be set.
11970 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
11971 Also change the functions X509_cmp_current_time() and
11972 X509_gmtime_adj() work with an ASN1_TIME structure,
11973 this will enable certificates using GeneralizedTime in validity
11974 dates to be checked.
11978 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
11979 negative public key encodings) on by default,
11980 NO_NEG_PUBKEY_BUG can be set to disable it.
11984 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
11985 content octets. An i2c_ASN1_OBJECT is unnecessary because
11986 the encoding can be trivially obtained from the structure.
11990 * crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
11991 not read locks (CRYPTO_r_[un]lock).
11995 * A first attempt at creating official support for shared
11996 libraries through configuration. I've kept it so the
11997 default is static libraries only, and the OpenSSL programs
11998 are always statically linked for now, but there are
11999 preparations for dynamic linking in place.
12000 This has been tested on Linux and Tru64.
12004 * Randomness polling function for Win9x, as described in:
12005 Peter Gutmann, Software Generation of Practically Strong
12010 * Fix so PRNG is seeded in req if using an already existing
12015 * New options to smime application. -inform and -outform
12016 allow alternative formats for the S/MIME message including
12017 PEM and DER. The -content option allows the content to be
12018 specified separately. This should allow things like Netscape
12019 form signing output easier to verify.
12023 * Fix the ASN1 encoding of tags using the 'long form'.
12027 * New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
12028 STRING types. These convert content octets to and from the
12029 underlying type. The actual tag and length octets are
12030 already assumed to have been read in and checked. These
12031 are needed because all other string types have virtually
12032 identical handling apart from the tag. By having versions
12033 of the ASN1 functions that just operate on content octets
12034 IMPLICIT tagging can be handled properly. It also allows
12035 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
12036 and ASN1_INTEGER are identical apart from the tag.
12040 * Change the handling of OID objects as follows:
12042 - New object identifiers are inserted in objects.txt, following
12043 the syntax given in objects.README.
12044 - objects.pl is used to process obj_mac.num and create a new
12046 - obj_dat.pl is used to create a new obj_dat.h, using the data in
12049 This is currently kind of a hack, and the perl code in objects.pl
12050 isn't very elegant, but it works as I intended. The simplest way
12051 to check that it worked correctly is to look in obj_dat.h and
12052 check the array nid_objs and make sure the objects haven't moved
12053 around (this is important!). Additions are OK, as well as
12054 consistent name changes.
12058 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
12062 * Addition of the command line parameter '-rand file' to 'openssl req'.
12063 The given file adds to whatever has already been seeded into the
12064 random pool through the RANDFILE configuration file option or
12065 environment variable, or the default random state file.
12069 * mkstack.pl now sorts each macro group into lexical order.
12070 Previously the output order depended on the order the files
12071 appeared in the directory, resulting in needless rewriting
12076 * Patches to make OpenSSL compile under Win32 again. Mostly
12077 work arounds for the VC++ problem that it treats func() as
12078 func(void). Also stripped out the parts of mkdef.pl that
12079 added extra typesafe functions: these no longer exist.
12083 * Reorganisation of the stack code. The macros are now all
12084 collected in safestack.h . Each macro is defined in terms of
12085 a "stack macro" of the form SKM_<name>(type, a, b). The
12086 DEBUG_SAFESTACK is now handled in terms of function casts,
12087 this has the advantage of retaining type safety without the
12088 use of additional functions. If DEBUG_SAFESTACK is not defined
12089 then the non typesafe macros are used instead. Also modified the
12090 mkstack.pl script to handle the new form. Needs testing to see
12091 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
12092 the default if no major problems. Similar behaviour for ASN1_SET_OF
12093 and PKCS12_STACK_OF.
12097 * When some versions of IIS use the 'NET' form of private key the
12098 key derivation algorithm is different. Normally MD5(password) is
12099 used as a 128 bit RC4 key. In the modified case
12100 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
12101 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
12102 as the old Netscape_RSA functions except they have an additional
12103 'sgckey' parameter which uses the modified algorithm. Also added
12104 an -sgckey command line option to the rsa utility. Thanks to
12105 Adrian Peck <bertie@ncipher.com> for posting details of the modified
12106 algorithm to openssl-dev.
12110 * The evp_local.h macros were using 'c.##kname' which resulted in
12111 invalid expansion on some systems (SCO 5.0.5 for example).
12112 Corrected to 'c.kname'.
12114 *Phillip Porch <root@theporch.com>*
12116 * New X509_get1_email() and X509_REQ_get1_email() functions that return
12117 a STACK of email addresses from a certificate or request, these look
12118 in the subject name and the subject alternative name extensions and
12119 omit any duplicate addresses.
12123 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
12124 This makes DSA verification about 2 % faster.
12128 * Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
12129 (meaning that now 2^5 values will be precomputed, which is only 4 KB
12130 plus overhead for 1024 bit moduli).
12131 This makes exponentiations about 0.5 % faster for 1024 bit
12132 exponents (as measured by "openssl speed rsa2048").
12136 * Rename memory handling macros to avoid conflicts with other
12138 Malloc => OPENSSL_malloc
12139 Malloc_locked => OPENSSL_malloc_locked
12140 Realloc => OPENSSL_realloc
12141 Free => OPENSSL_free
12145 * New function BN_mod_exp_mont_word for small bases (roughly 15%
12146 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
12150 * CygWin32 support.
12152 *John Jarvie <jjarvie@newsguy.com>*
12154 * The type-safe stack code has been rejigged. It is now only compiled
12155 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
12156 by default all type-specific stack functions are "#define"d back to
12157 standard stack functions. This results in more streamlined output
12158 but retains the type-safety checking possibilities of the original
12163 * The STACK code has been cleaned up, and certain type declarations
12164 that didn't make a lot of sense have been brought in line. This has
12165 also involved a cleanup of sorts in safestack.h to more correctly
12166 map type-safe stack functions onto their plain stack counterparts.
12167 This work has also resulted in a variety of "const"ifications of
12168 lots of the code, especially "_cmp" operations which should normally
12169 be prototyped with "const" parameters anyway.
12173 * When generating bytes for the first time in md_rand.c, 'stir the pool'
12174 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
12175 (The PRNG state consists of two parts, the large pool 'state' and 'md',
12176 where all of 'md' is used each time the PRNG is used, but 'state'
12177 is used only indexed by a cyclic counter. As entropy may not be
12178 well distributed from the beginning, 'md' is important as a
12179 chaining variable. However, the output function chains only half
12180 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
12181 all of 'md', and seeding with STATE_SIZE dummy bytes will result
12182 in all of 'state' being rewritten, with the new values depending
12183 on virtually all of 'md'. This overcomes the 80 bit limitation.)
12187 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
12188 the handshake is continued after ssl_verify_cert_chain();
12189 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
12190 can lead to 'unexplainable' connection aborts later.
12192 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
12194 * Major EVP API cipher revision.
12195 Add hooks for extra EVP features. This allows various cipher
12196 parameters to be set in the EVP interface. Support added for variable
12197 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
12198 setting of RC2 and RC5 parameters.
12200 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
12203 Remove lots of duplicated code from the EVP library. For example *every*
12204 cipher init() function handles the 'iv' in the same way according to the
12205 cipher mode. They also all do nothing if the 'key' parameter is NULL and
12206 for CFB and OFB modes they zero ctx->num.
12208 New functionality allows removal of S/MIME code RC2 hack.
12210 Most of the routines have the same form and so can be declared in terms
12213 By shifting this to the top level EVP_CipherInit() it can be removed from
12214 all individual ciphers. If the cipher wants to handle IVs or keys
12215 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
12218 Change lots of functions like EVP_EncryptUpdate() to now return a
12219 value: although software versions of the algorithms cannot fail
12220 any installed hardware versions can.
12224 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
12225 this option is set, tolerate broken clients that send the negotiated
12226 protocol version number instead of the requested protocol version
12231 * Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
12232 i.e. non-zero for export ciphersuites, zero otherwise.
12233 Previous versions had this flag inverted, inconsistent with
12234 rsa_tmp_cb (..._TMP_RSA_CB).
12236 *Bodo Moeller; problem reported by Amit Chopra*
12238 * Add missing DSA library text string. Work around for some IIS
12239 key files with invalid SEQUENCE encoding.
12243 * Add a document (doc/standards.txt) that list all kinds of standards
12244 and so on that are implemented in OpenSSL.
12248 * Enhance c_rehash script. Old version would mishandle certificates
12249 with the same subject name hash and wouldn't handle CRLs at all.
12250 Added -fingerprint option to crl utility, to support new c_rehash
12255 * Eliminate non-ANSI declarations in crypto.h and stack.h.
12259 * Fix for SSL server purpose checking. Server checking was
12260 rejecting certificates which had extended key usage present
12261 but no ssl client purpose.
12263 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
12265 * Make PKCS#12 code work with no password. The PKCS#12 spec
12266 is a little unclear about how a blank password is handled.
12267 Since the password in encoded as a BMPString with terminating
12268 double NULL a zero length password would end up as just the
12269 double NULL. However no password at all is different and is
12270 handled differently in the PKCS#12 key generation code. NS
12271 treats a blank password as zero length. MSIE treats it as no
12272 password on export: but it will try both on import. We now do
12273 the same: PKCS12_parse() tries zero length and no password if
12274 the password is set to "" or NULL (NULL is now a valid password:
12275 it wasn't before) as does the pkcs12 application.
12279 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
12280 perror when PEM_read_bio_X509_REQ fails, the error message must
12281 be obtained from the error queue.
12285 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
12286 it in ERR_remove_state if appropriate, and change ERR_get_state
12287 accordingly to avoid race conditions (this is necessary because
12288 thread_hash is no longer constant once set).
12292 * Bugfix for linux-elf makefile.one.
12296 * RSA_get_default_method() will now cause a default
12297 RSA_METHOD to be chosen if one doesn't exist already.
12298 Previously this was only set during a call to RSA_new()
12299 or RSA_new_method(NULL) meaning it was possible for
12300 RSA_get_default_method() to return NULL.
12304 * Added native name translation to the existing DSO code
12305 that will convert (if the flag to do so is set) filenames
12306 that are sufficiently small and have no path information
12307 into a canonical native form. Eg. "blah" converted to
12308 "libblah.so" or "blah.dll" etc.
12312 * New function ERR_error_string_n(e, buf, len) which is like
12313 ERR_error_string(e, buf), but writes at most 'len' bytes
12314 including the 0 terminator. For ERR_error_string_n, 'buf'
12317 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
12319 * CONF library reworked to become more general. A new CONF
12320 configuration file reader "class" is implemented as well as a
12321 new functions (NCONF_*, for "New CONF") to handle it. The now
12322 old CONF_* functions are still there, but are reimplemented to
12323 work in terms of the new functions. Also, a set of functions
12324 to handle the internal storage of the configuration data is
12325 provided to make it easier to write new configuration file
12326 reader "classes" (I can definitely see something reading a
12327 configuration file in XML format, for example), called _CONF_*,
12328 or "the configuration storage API"...
12330 The new configuration file reading functions are:
12332 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
12333 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
12335 NCONF_default, NCONF_WIN32
12337 NCONF_dump_fp, NCONF_dump_bio
12339 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
12340 NCONF_new creates a new CONF object. This works in the same way
12341 as other interfaces in OpenSSL, like the BIO interface.
12342 NCONF_dump_* dump the internal storage of the configuration file,
12343 which is useful for debugging. All other functions take the same
12344 arguments as the old CONF_* functions wth the exception of the
12345 first that must be a `CONF *' instead of a `LHASH *'.
12347 To make it easer to use the new classes with the old CONF_* functions,
12348 the function CONF_set_default_method is provided.
12352 * Add '-tls1' option to 'openssl ciphers', which was already
12353 mentioned in the documentation but had not been implemented.
12354 (This option is not yet really useful because even the additional
12355 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
12359 * Initial DSO code added into libcrypto for letting OpenSSL (and
12360 OpenSSL-based applications) load shared libraries and bind to
12361 them in a portable way.
12363 *Geoff Thorpe, with contributions from Richard Levitte*
12365 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000] ###
12367 * Make sure _lrotl and _lrotr are only used with MSVC.
12369 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
12370 (the default implementation of RAND_status).
12372 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
12373 to '-clrext' (= clear extensions), as intended and documented.
12374 *Bodo Moeller; inconsistency pointed out by Michael Attili
12375 <attili@amaxo.com>*
12377 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
12378 was larger than the MD block size.
12380 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
12382 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
12383 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
12384 using the passed key: if the passed key was a private key the result
12385 of X509_print(), for example, would be to print out all the private key
12390 * des_quad_cksum() byte order bug fix.
12391 *Ulf Möller, using the problem description in krb4-0.9.7, where
12392 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
12394 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
12397 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
12399 * For easily testing in shell scripts whether some command
12400 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
12401 returns with exit code 0 iff no command of the given name is available.
12402 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
12403 the output goes to stdout and nothing is printed to stderr.
12404 Additional arguments are always ignored.
12406 Since for each cipher there is a command of the same name,
12407 the 'no-cipher' compilation switches can be tested this way.
12409 ('openssl no-XXX' is not able to detect pseudo-commands such
12410 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
12414 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
12418 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
12419 is set; it will be thrown away anyway because each handshake creates
12421 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
12422 to parameters -- in previous versions (since OpenSSL 0.9.3) the
12423 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
12424 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
12428 * New s_client option -ign_eof: EOF at stdin is ignored, and
12429 'Q' and 'R' lose their special meanings (quit/renegotiate).
12430 This is part of what -quiet does; unlike -quiet, -ign_eof
12431 does not suppress any output.
12435 * Add compatibility options to the purpose and trust code. The
12436 purpose X509_PURPOSE_ANY is "any purpose" which automatically
12437 accepts a certificate or CA, this was the previous behaviour,
12438 with all the associated security issues.
12440 X509_TRUST_COMPAT is the old trust behaviour: only and
12441 automatically trust self signed roots in certificate store. A
12442 new trust setting X509_TRUST_DEFAULT is used to specify that
12443 a purpose has no associated trust setting and it should instead
12444 use the value in the default purpose.
12448 * Fix the PKCS#8 DSA private key code so it decodes keys again
12449 and fix a memory leak.
12453 * In util/mkerr.pl (which implements 'make errors'), preserve
12454 reason strings from the previous version of the .c file, as
12455 the default to have only downcase letters (and digits) in
12456 automatically generated reasons codes is not always appropriate.
12460 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
12461 using strerror. Previously, ERR_reason_error_string() returned
12462 library names as reason strings for SYSerr; but SYSerr is a special
12463 case where small numbers are errno values, not library numbers.
12467 * Add '-dsaparam' option to 'openssl dhparam' application. This
12468 converts DSA parameters into DH parameters. (When creating parameters,
12469 DSA_generate_parameters is used.)
12473 * Include 'length' (recommended exponent length) in C code generated
12474 by 'openssl dhparam -C'.
12478 * The second argument to set_label in perlasm was already being used
12479 so couldn't be used as a "file scope" flag. Moved to third argument
12484 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
12485 instead of RAND_bytes for encryption IVs and salts.
12489 * Include RAND_status() into RAND_METHOD instead of implementing
12490 it only for md_rand.c Otherwise replacing the PRNG by calling
12491 RAND_set_rand_method would be impossible.
12495 * Don't let DSA_generate_key() enter an infinite loop if the random
12496 number generation fails.
12500 * New 'rand' application for creating pseudo-random output.
12504 * Added configuration support for Linux/IA64
12506 *Rolf Haberrecker <rolf@suse.de>*
12508 * Assembler module support for Mingw32.
12512 * Shared library support for HPUX (in shlib/).
12514 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
12516 * Shared library support for Solaris gcc.
12518 *Lutz Behnke <behnke@trustcenter.de>*
12520 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000] ###
12522 * PKCS7_encrypt() was adding text MIME headers twice because they
12523 were added manually and by SMIME_crlf_copy().
12527 * In bntest.c don't call BN_rand with zero bits argument.
12529 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
12531 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
12532 case was implemented. This caused BN_div_recp() to fail occasionally.
12536 * Add an optional second argument to the set_label() in the perl
12537 assembly language builder. If this argument exists and is set
12538 to 1 it signals that the assembler should use a symbol whose
12539 scope is the entire file, not just the current function. This
12540 is needed with MASM which uses the format label:: for this scope.
12542 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
12544 * Change the ASN1 types so they are typedefs by default. Before
12545 almost all types were #define'd to ASN1_STRING which was causing
12546 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
12551 * Change names of new functions to the new get1/get0 naming
12552 convention: After 'get1', the caller owns a reference count
12553 and has to call ..._free; 'get0' returns a pointer to some
12554 data structure without incrementing reference counters.
12555 (Some of the existing 'get' functions increment a reference
12556 counter, some don't.)
12557 Similarly, 'set1' and 'add1' functions increase reference
12558 counters or duplicate objects.
12562 * Allow for the possibility of temp RSA key generation failure:
12563 the code used to assume it always worked and crashed on failure.
12567 * Fix potential buffer overrun problem in BIO_printf().
12568 *Ulf Möller, using public domain code by Patrick Powell; problem
12569 pointed out by David Sacerdote <das33@cornell.edu>*
12571 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
12572 RAND_egd() and RAND_status(). In the command line application,
12573 the EGD socket can be specified like a seed file using RANDFILE
12578 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
12579 Some CAs (e.g. Verisign) distribute certificates in this form.
12583 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
12584 list to exclude them. This means that no special compilation option
12585 is needed to use anonymous DH: it just needs to be included in the
12590 * Change the EVP_MD_CTX_type macro so its meaning consistent with
12591 EVP_MD_type. The old functionality is available in a new macro called
12592 EVP_MD_md(). Change code that uses it and update docs.
12596 * ..._ctrl functions now have corresponding ..._callback_ctrl functions
12597 where the 'void *' argument is replaced by a function pointer argument.
12598 Previously 'void *' was abused to point to functions, which works on
12599 many platforms, but is not correct. As these functions are usually
12600 called by macros defined in OpenSSL header files, most source code
12601 should work without changes.
12605 * <openssl/opensslconf.h> (which is created by Configure) now contains
12606 sections with information on -D... compiler switches used for
12607 compiling the library so that applications can see them. To enable
12608 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
12609 must be defined. E.g.,
12610 #define OPENSSL_ALGORITHM_DEFINES
12611 #include <openssl/opensslconf.h>
12612 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
12614 *Richard Levitte, Ulf and Bodo Möller*
12616 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
12621 * Change the 'other' type in certificate aux info to a STACK_OF
12622 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
12623 the required ASN1 format: arbitrary types determined by an OID.
12627 * Add some PEM_write_X509_REQ_NEW() functions and a command line
12628 argument to 'req'. This is not because the function is newer or
12629 better than others it just uses the work 'NEW' in the certificate
12630 request header lines. Some software needs this.
12634 * Reorganise password command line arguments: now passwords can be
12635 obtained from various sources. Delete the PEM_cb function and make
12636 it the default behaviour: i.e. if the callback is NULL and the
12637 usrdata argument is not NULL interpret it as a null terminated pass
12638 phrase. If usrdata and the callback are NULL then the pass phrase
12639 is prompted for as usual.
12643 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
12644 the support is automatically enabled. The resulting binaries will
12645 autodetect the card and use it if present.
12647 *Ben Laurie and Compaq Inc.*
12649 * Work around for Netscape hang bug. This sends certificate request
12650 and server done in one record. Since this is perfectly legal in the
12651 SSL/TLS protocol it isn't a "bug" option and is on by default. See
12652 the bugs/SSLv3 entry for more info.
12656 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
12660 * Add -rand argument to smime and pkcs12 applications and read/write
12665 * New 'passwd' tool for crypt(3) and apr1 password hashes.
12669 * Add command line password options to the remaining applications.
12673 * Bug fix for BN_div_recp() for numerators with an even number of
12678 * More tests in bntest.c, and changed test_bn output.
12682 * ./config recognizes MacOS X now.
12686 * Bug fix for BN_div() when the first words of num and divisor are
12687 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
12691 * Add support for various broken PKCS#8 formats, and command line
12692 options to produce them.
12696 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
12697 get temporary BIGNUMs from a BN_CTX.
12701 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
12706 * Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
12707 include a #define from the old name to the new. The original intent
12708 was that statically linked binaries could for example just call
12709 SSLeay_add_all_ciphers() to just add ciphers to the table and not
12710 link with digests. This never worked because SSLeay_add_all_digests()
12711 and SSLeay_add_all_ciphers() were in the same source file so calling
12712 one would link with the other. They are now in separate source files.
12716 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
12720 * Use a less unusual form of the Miller-Rabin primality test (it used
12721 a binary algorithm for exponentiation integrated into the Miller-Rabin
12722 loop, our standard modexp algorithms are faster).
12726 * Support for the EBCDIC character set completed.
12728 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
12730 * Source code cleanups: use const where appropriate, eliminate casts,
12731 use void * instead of char * in lhash.
12735 * Bugfix: ssl3_send_server_key_exchange was not restartable
12736 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
12737 this the server could overwrite ephemeral keys that the client
12742 * Turn DSA_is_prime into a macro that calls BN_is_prime,
12743 using 50 iterations of the Rabin-Miller test.
12745 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
12746 iterations of the Rabin-Miller test as required by the appendix
12747 to FIPS PUB 186[-1]) instead of DSA_is_prime.
12748 As BN_is_prime_fasttest includes trial division, DSA parameter
12749 generation becomes much faster.
12751 This implies a change for the callback functions in DSA_is_prime
12752 and DSA_generate_parameters: The callback function is called once
12753 for each positive witness in the Rabin-Miller test, not just
12754 occasionally in the inner loop; and the parameters to the
12755 callback function now provide an iteration count for the outer
12756 loop rather than for the current invocation of the inner loop.
12757 DSA_generate_parameters additionally can call the callback
12758 function with an 'iteration count' of -1, meaning that a
12759 candidate has passed the trial division test (when q is generated
12760 from an application-provided seed, trial division is skipped).
12764 * New function BN_is_prime_fasttest that optionally does trial
12765 division before starting the Rabin-Miller test and has
12766 an additional BN_CTX * argument (whereas BN_is_prime always
12767 has to allocate at least one BN_CTX).
12768 'callback(1, -1, cb_arg)' is called when a number has passed the
12769 trial division stage.
12773 * Fix for bug in CRL encoding. The validity dates weren't being handled
12778 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
12782 * New function BN_pseudo_rand().
12786 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
12787 bignum version of BN_from_montgomery() with the working code from
12788 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
12793 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
12794 made it impossible to use the same SSL_SESSION data structure in
12795 SSL2 clients in multiple threads.
12799 * The return value of RAND_load_file() no longer counts bytes obtained
12800 by stat(). RAND_load_file(..., -1) is new and uses the complete file
12801 to seed the PRNG (previously an explicit byte count was required).
12803 *Ulf Möller, Bodo Möller*
12805 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
12806 used (char *) instead of (void *) and had casts all over the place.
12810 * Make BN_generate_prime() return NULL on error if ret!=NULL.
12814 * Retain source code compatibility for BN_prime_checks macro:
12815 BN_is_prime(..., BN_prime_checks, ...) now uses
12816 BN_prime_checks_for_size to determine the appropriate number of
12817 Rabin-Miller iterations.
12821 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
12822 DH_CHECK_P_NOT_SAFE_PRIME.
12823 (Check if this is true? OpenPGP calls them "strong".)
12827 * Merge the functionality of "dh" and "gendh" programs into a new program
12828 "dhparam". The old programs are retained for now but will handle DH keys
12829 (instead of parameters) in future.
12833 * Make the ciphers, s_server and s_client programs check the return values
12834 when a new cipher list is set.
12838 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
12839 ciphers. Before when the 56bit ciphers were enabled the sorting was
12842 The syntax for the cipher sorting has been extended to support sorting by
12843 cipher-strength (using the strength_bits hard coded in the tables).
12844 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
12846 Fix a bug in the cipher-command parser: when supplying a cipher command
12847 string with an "undefined" symbol (neither command nor alphanumeric
12848 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
12849 an error is flagged.
12851 Due to the strength-sorting extension, the code of the
12852 ssl_create_cipher_list() function was completely rearranged. I hope that
12853 the readability was also increased :-)
12855 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
12857 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
12858 for the first serial number and places 2 in the serial number file. This
12859 avoids problems when the root CA is created with serial number zero and
12860 the first user certificate has the same issuer name and serial number
12865 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
12866 the new code. Add documentation for this stuff.
12870 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
12871 X509_*() to X509at_*() on the grounds that they don't handle X509
12872 structures and behave in an analogous way to the X509v3 functions:
12873 they shouldn't be called directly but wrapper functions should be used
12876 So we also now have some wrapper functions that call the X509at functions
12877 when passed certificate requests. (TO DO: similar things can be done with
12878 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
12879 things. Some of these need some d2i or i2d and print functionality
12880 because they handle more complex structures.)
12884 * Add missing #ifndefs that caused missing symbols when building libssl
12885 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
12886 NO_RSA in ssl/s2*.c.
12888 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
12890 * Precautions against using the PRNG uninitialized: RAND_bytes() now
12891 has a return value which indicates the quality of the random data
12892 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
12893 error queue. New function RAND_pseudo_bytes() generates output that is
12894 guaranteed to be unique but not unpredictable. RAND_add is like
12895 RAND_seed, but takes an extra argument for an entropy estimate
12896 (RAND_seed always assumes full entropy).
12900 * Do more iterations of Rabin-Miller probable prime test (specifically,
12901 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
12902 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
12903 in crypto/bn/bn_prime.c for the complete table). This guarantees a
12904 false-positive rate of at most 2^-80 for random input.
12908 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
12912 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
12913 in the 0.9.5 release), this returns the chain
12914 from an X509_CTX structure with a dup of the stack and all
12915 the X509 reference counts upped: so the stack will exist
12916 after X509_CTX_cleanup() has been called. Modify pkcs12.c
12919 Also make SSL_SESSION_print() print out the verify return
12924 * Add manpage for the pkcs12 command. Also change the default
12925 behaviour so MAC iteration counts are used unless the new
12926 -nomaciter option is used. This improves file security and
12927 only older versions of MSIE (4.0 for example) need it.
12931 * Honor the no-xxx Configure options when creating .DEF files.
12935 * Add PKCS#10 attributes to field table: challengePassword,
12936 unstructuredName and unstructuredAddress. These are taken from
12937 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
12938 international characters are used.
12940 More changes to X509_ATTRIBUTE code: allow the setting of types
12941 based on strings. Remove the 'loc' parameter when adding
12942 attributes because these will be a SET OF encoding which is sorted
12947 * Initial changes to the 'req' utility to allow request generation
12948 automation. This will allow an application to just generate a template
12949 file containing all the field values and have req construct the
12952 Initial support for X509_ATTRIBUTE handling. Stacks of these are
12953 used all over the place including certificate requests and PKCS#7
12954 structures. They are currently handled manually where necessary with
12955 some primitive wrappers for PKCS#7. The new functions behave in a
12956 manner analogous to the X509 extension functions: they allow
12957 attributes to be looked up by NID and added.
12959 Later something similar to the X509V3 code would be desirable to
12960 automatically handle the encoding, decoding and printing of the
12961 more complex types. The string types like challengePassword can
12962 be handled by the string table functions.
12964 Also modified the multi byte string table handling. Now there is
12965 a 'global mask' which masks out certain types. The table itself
12966 can use the flag STABLE_NO_MASK to ignore the mask setting: this
12967 is useful when for example there is only one permissible type
12968 (as in countryName) and using the mask might result in no valid
12973 * Clean up 'Finished' handling, and add functions SSL_get_finished and
12974 SSL_get_peer_finished to allow applications to obtain the latest
12975 Finished messages sent to the peer or expected from the peer,
12976 respectively. (SSL_get_peer_finished is usually the Finished message
12977 actually received from the peer, otherwise the protocol will be aborted.)
12979 As the Finished message are message digests of the complete handshake
12980 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
12981 be used for external authentication procedures when the authentication
12982 provided by SSL/TLS is not desired or is not enough.
12986 * Enhanced support for Alpha Linux is added. Now ./config checks if
12987 the host supports BWX extension and if Compaq C is present on the
12988 $PATH. Just exploiting of the BWX extension results in 20-30%
12989 performance kick for some algorithms, e.g. DES and RC4 to mention
12990 a couple. Compaq C in turn generates ~20% faster code for MD5 and
12995 * Add support for MS "fast SGC". This is arguably a violation of the
12996 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
12997 weak crypto and after checking the certificate is SGC a second one
12998 with strong crypto. MS SGC stops the first handshake after receiving
12999 the server certificate message and sends a second client hello. Since
13000 a server will typically do all the time consuming operations before
13001 expecting any further messages from the client (server key exchange
13002 is the most expensive) there is little difference between the two.
13004 To get OpenSSL to support MS SGC we have to permit a second client
13005 hello message after we have sent server done. In addition we have to
13006 reset the MAC if we do get this second client hello.
13010 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
13011 if a DER encoded private key is RSA or DSA traditional format. Changed
13012 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
13013 format DER encoded private key. Newer code should use PKCS#8 format which
13014 has the key type encoded in the ASN1 structure. Added DER private key
13015 support to pkcs8 application.
13019 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
13020 ciphersuites has been selected (as required by the SSL 3/TLS 1
13021 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
13022 is set, we interpret this as a request to violate the specification
13023 (the worst that can happen is a handshake failure, and 'correct'
13024 behaviour would result in a handshake failure anyway).
13028 * In SSL_CTX_add_session, take into account that there might be multiple
13029 SSL_SESSION structures with the same session ID (e.g. when two threads
13030 concurrently obtain them from an external cache).
13031 The internal cache can handle only one SSL_SESSION with a given ID,
13032 so if there's a conflict, we now throw out the old one to achieve
13037 * Add OIDs for idea and blowfish in CBC mode. This will allow both
13038 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
13039 some routines that use cipher OIDs: some ciphers do not have OIDs
13040 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
13045 * Simplify the trust setting structure and code. Now we just have
13046 two sequences of OIDs for trusted and rejected settings. These will
13047 typically have values the same as the extended key usage extension
13048 and any application specific purposes.
13050 The trust checking code now has a default behaviour: it will just
13051 check for an object with the same NID as the passed id. Functions can
13052 be provided to override either the default behaviour or the behaviour
13053 for a given id. SSL client, server and email already have functions
13054 in place for compatibility: they check the NID and also return "trusted"
13055 if the certificate is self signed.
13059 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
13060 traditional format into an EVP_PKEY structure.
13064 * Add a password callback function PEM_cb() which either prompts for
13065 a password if usr_data is NULL or otherwise assumes it is a null
13066 terminated password. Allow passwords to be passed on command line
13067 environment or config files in a few more utilities.
13071 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
13072 keys. Add some short names for PKCS#8 PBE algorithms and allow them
13073 to be specified on the command line for the pkcs8 and pkcs12 utilities.
13074 Update documentation.
13078 * Support for ASN1 "NULL" type. This could be handled before by using
13079 ASN1_TYPE but there wasn't any function that would try to read a NULL
13080 and produce an error if it couldn't. For compatibility we also have
13081 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
13082 don't allocate anything because they don't need to.
13086 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
13089 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
13091 * Rebuild of the memory allocation routines used by OpenSSL code and
13092 possibly others as well. The purpose is to make an interface that
13093 provide hooks so anyone can build a separate set of allocation and
13094 deallocation routines to be used by OpenSSL, for example memory
13095 pool implementations, or something else, which was previously hard
13096 since Malloc(), Realloc() and Free() were defined as macros having
13097 the values malloc, realloc and free, respectively (except for Win32
13098 compilations). The same is provided for memory debugging code.
13099 OpenSSL already comes with functionality to find memory leaks, but
13100 this gives people a chance to debug other memory problems.
13102 With these changes, a new set of functions and macros have appeared:
13104 CRYPTO_set_mem_debug_functions() [F]
13105 CRYPTO_get_mem_debug_functions() [F]
13106 CRYPTO_dbg_set_options() [F]
13107 CRYPTO_dbg_get_options() [F]
13108 CRYPTO_malloc_debug_init() [M]
13110 The memory debug functions are NULL by default, unless the library
13111 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
13112 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
13113 gives the standard debugging functions that come with OpenSSL) or
13114 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
13115 provided by the library user) must be used. When the standard
13116 debugging functions are used, CRYPTO_dbg_set_options can be used to
13117 request additional information:
13118 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
13119 the CRYPTO_MDEBUG_xxx macro when compiling the library.
13121 Also, things like CRYPTO_set_mem_functions will always give the
13122 expected result (the new set of functions is used for allocation
13123 and deallocation) at all times, regardless of platform and compiler
13126 To finish it up, some functions that were never use in any other
13127 way than through macros have a new API and new semantic:
13129 CRYPTO_dbg_malloc()
13130 CRYPTO_dbg_realloc()
13133 All macros of value have retained their old syntax.
13135 *Richard Levitte and Bodo Moeller*
13137 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
13138 ordering of SMIMECapabilities wasn't in "strength order" and there
13139 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
13144 * Some ASN1 types with illegal zero length encoding (INTEGER,
13145 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
13147 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
13149 * Merge in my S/MIME library for OpenSSL. This provides a simple
13150 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
13151 functionality to handle multipart/signed properly) and a utility
13152 called 'smime' to call all this stuff. This is based on code I
13153 originally wrote for Celo who have kindly allowed it to be
13154 included in OpenSSL.
13158 * Add variants des_set_key_checked and des_set_key_unchecked of
13159 des_set_key (aka des_key_sched). Global variable des_check_key
13160 decides which of these is called by des_set_key; this way
13161 des_check_key behaves as it always did, but applications and
13162 the library itself, which was buggy for des_check_key == 1,
13163 have a cleaner way to pick the version they need.
13167 * New function PKCS12_newpass() which changes the password of a
13172 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
13173 dynamic mix. In both cases the ids can be used as an index into the
13174 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
13175 functions so they accept a list of the field values and the
13176 application doesn't need to directly manipulate the X509_TRUST
13181 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
13186 * Modify the way the V3 extension code looks up extensions. This now
13187 works in a similar way to the object code: we have some "standard"
13188 extensions in a static table which is searched with OBJ_bsearch()
13189 and the application can add dynamic ones if needed. The file
13190 crypto/x509v3/ext_dat.h now has the info: this file needs to be
13191 updated whenever a new extension is added to the core code and kept
13192 in ext_nid order. There is a simple program 'tabtest.c' which checks
13193 this. New extensions are not added too often so this file can readily
13194 be maintained manually.
13196 There are two big advantages in doing things this way. The extensions
13197 can be looked up immediately and no longer need to be "added" using
13198 X509V3_add_standard_extensions(): this function now does nothing.
13199 *Side note: I get *lots* of email saying the extension code doesn't
13200 work because people forget to call this function*
13201 Also no dynamic allocation is done unless new extensions are added:
13202 so if we don't add custom extensions there is no need to call
13203 X509V3_EXT_cleanup().
13207 * Modify enc utility's salting as follows: make salting the default. Add a
13208 magic header, so unsalted files fail gracefully instead of just decrypting
13209 to garbage. This is because not salting is a big security hole, so people
13210 should be discouraged from doing it.
13214 * Fixes and enhancements to the 'x509' utility. It allowed a message
13215 digest to be passed on the command line but it only used this
13216 parameter when signing a certificate. Modified so all relevant
13217 operations are affected by the digest parameter including the
13218 -fingerprint and -x509toreq options. Also -x509toreq choked if a
13219 DSA key was used because it didn't fix the digest.
13223 * Initial certificate chain verify code. Currently tests the untrusted
13224 certificates for consistency with the verify purpose (which is set
13225 when the X509_STORE_CTX structure is set up) and checks the pathlength.
13227 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
13228 this is because it will reject chains with invalid extensions whereas
13229 every previous version of OpenSSL and SSLeay made no checks at all.
13231 Trust code: checks the root CA for the relevant trust settings. Trust
13232 settings have an initial value consistent with the verify purpose: e.g.
13233 if the verify purpose is for SSL client use it expects the CA to be
13234 trusted for SSL client use. However the default value can be changed to
13235 permit custom trust settings: one example of this would be to only trust
13236 certificates from a specific "secure" set of CAs.
13238 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
13239 which should be used for version portability: especially since the
13240 verify structure is likely to change more often now.
13242 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
13243 to set them. If not set then assume SSL clients will verify SSL servers
13246 Two new options to the verify program: -untrusted allows a set of
13247 untrusted certificates to be passed in and -purpose which sets the
13248 intended purpose of the certificate. If a purpose is set then the
13249 new chain verify code is used to check extension consistency.
13253 * Support for the authority information access extension.
13257 * Modify RSA and DSA PEM read routines to transparently handle
13258 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
13259 public keys in a format compatible with certificate
13260 SubjectPublicKeyInfo structures. Unfortunately there were already
13261 functions called *_PublicKey_* which used various odd formats so
13262 these are retained for compatibility: however the DSA variants were
13263 never in a public release so they have been deleted. Changed dsa/rsa
13264 utilities to handle the new format: note no releases ever handled public
13265 keys so we should be OK.
13267 The primary motivation for this change is to avoid the same fiasco
13268 that dogs private keys: there are several incompatible private key
13269 formats some of which are standard and some OpenSSL specific and
13270 require various evil hacks to allow partial transparent handling and
13271 even then it doesn't work with DER formats. Given the option anything
13272 other than PKCS#8 should be dumped: but the other formats have to
13273 stay in the name of compatibility.
13275 With public keys and the benefit of hindsight one standard format
13276 is used which works with EVP_PKEY, RSA or DSA structures: though
13277 it clearly returns an error if you try to read the wrong kind of key.
13279 Added a -pubkey option to the 'x509' utility to output the public key.
13280 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
13281 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
13282 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
13283 that do the same as the EVP_PKEY_assign_*() except they up the
13284 reference count of the added key (they don't "swallow" the
13289 * Fixes to crypto/x509/by_file.c the code to read in certificates and
13290 CRLs would fail if the file contained no certificates or no CRLs:
13291 added a new function to read in both types and return the number
13292 read: this means that if none are read it will be an error. The
13293 DER versions of the certificate and CRL reader would always fail
13294 because it isn't possible to mix certificates and CRLs in DER format
13295 without choking one or the other routine. Changed this to just read
13296 a certificate: this is the best we can do. Also modified the code
13297 in apps/verify.c to take notice of return codes: it was previously
13298 attempting to read in certificates from NULL pointers and ignoring
13299 any errors: this is one reason why the cert and CRL reader seemed
13300 to work. It doesn't check return codes from the default certificate
13301 routines: these may well fail if the certificates aren't installed.
13305 * Code to support otherName option in GeneralName.
13309 * First update to verify code. Change the verify utility
13310 so it warns if it is passed a self signed certificate:
13311 for consistency with the normal behaviour. X509_verify
13312 has been modified to it will now verify a self signed
13313 certificate if *exactly* the same certificate appears
13314 in the store: it was previously impossible to trust a
13315 single self signed certificate. This means that:
13316 openssl verify ss.pem
13317 now gives a warning about a self signed certificate but
13318 openssl verify -CAfile ss.pem ss.pem
13323 * For servers, store verify_result in SSL_SESSION data structure
13324 (and add it to external session representation).
13325 This is needed when client certificate verifications fails,
13326 but an application-provided verification callback (set by
13327 SSL_CTX_set_cert_verify_callback) allows accepting the session
13328 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
13329 but returns 1): When the session is reused, we have to set
13330 ssl->verify_result to the appropriate error code to avoid
13333 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
13335 * Fix a bug in the new PKCS#7 code: it didn't consider the
13336 case in PKCS7_dataInit() where the signed PKCS7 structure
13337 didn't contain any existing data because it was being created.
13339 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
13341 * Add a salt to the key derivation routines in enc.c. This
13342 forms the first 8 bytes of the encrypted file. Also add a
13343 -S option to allow a salt to be input on the command line.
13347 * New function X509_cmp(). Oddly enough there wasn't a function
13348 to compare two certificates. We do this by working out the SHA1
13349 hash and comparing that. X509_cmp() will be needed by the trust
13354 * SSL_get1_session() is like SSL_get_session(), but increments
13355 the reference count in the SSL_SESSION returned.
13357 *Geoff Thorpe <geoff@eu.c2.net>*
13359 * Fix for 'req': it was adding a null to request attributes.
13360 Also change the X509_LOOKUP and X509_INFO code to handle
13361 certificate auxiliary information.
13365 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
13370 * Add the possibility to add extra information to the memory leak
13371 detecting output, to form tracebacks, showing from where each
13372 allocation was originated: CRYPTO_push_info("constant string") adds
13373 the string plus current file name and line number to a per-thread
13374 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
13375 is like calling CYRPTO_pop_info() until the stack is empty.
13376 Also updated memory leak detection code to be multi-thread-safe.
13380 * Add options -text and -noout to pkcs7 utility and delete the
13381 encryption options which never did anything. Update docs.
13385 * Add options to some of the utilities to allow the pass phrase
13386 to be included on either the command line (not recommended on
13387 OSes like Unix) or read from the environment. Update the
13388 manpages and fix a few bugs.
13392 * Add a few manpages for some of the openssl commands.
13396 * Fix the -revoke option in ca. It was freeing up memory twice,
13397 leaking and not finding already revoked certificates.
13401 * Extensive changes to support certificate auxiliary information.
13402 This involves the use of X509_CERT_AUX structure and X509_AUX
13403 functions. An X509_AUX function such as PEM_read_X509_AUX()
13404 can still read in a certificate file in the usual way but it
13405 will also read in any additional "auxiliary information". By
13406 doing things this way a fair degree of compatibility can be
13407 retained: existing certificates can have this information added
13408 using the new 'x509' options.
13410 Current auxiliary information includes an "alias" and some trust
13411 settings. The trust settings will ultimately be used in enhanced
13412 certificate chain verification routines: currently a certificate
13413 can only be trusted if it is self signed and then it is trusted
13418 * Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
13419 The problem was that one of the replacement routines had not been working
13420 since SSLeay releases. For now the offending routine has been replaced
13421 with non-optimised assembler. Even so, this now gives around 95%
13422 performance improvement for 1024 bit RSA signs.
13426 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
13427 handling. Most clients have the effective key size in bits equal to
13428 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
13429 A few however don't do this and instead use the size of the decrypted key
13430 to determine the RC2 key length and the AlgorithmIdentifier to determine
13431 the effective key length. In this case the effective key length can still
13432 be 40 bits but the key length can be 168 bits for example. This is fixed
13433 by manually forcing an RC2 key into the EVP_PKEY structure because the
13434 EVP code can't currently handle unusual RC2 key sizes: it always assumes
13435 the key length and effective key length are equal.
13439 * Add a bunch of functions that should simplify the creation of
13440 X509_NAME structures. Now you should be able to do:
13441 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
13442 and have it automatically work out the correct field type and fill in
13443 the structures. The more adventurous can try:
13444 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
13445 and it will (hopefully) work out the correct multibyte encoding.
13449 * Change the 'req' utility to use the new field handling and multibyte
13450 copy routines. Before the DN field creation was handled in an ad hoc
13451 way in req, ca, and x509 which was rather broken and didn't support
13452 BMPStrings or UTF8Strings. Since some software doesn't implement
13453 BMPStrings or UTF8Strings yet, they can be enabled using the config file
13454 using the dirstring_type option. See the new comment in the default
13455 openssl.cnf for more info.
13459 * Make crypto/rand/md_rand.c more robust:
13460 - Assure unique random numbers after fork().
13461 - Make sure that concurrent threads access the global counter and
13462 md serializably so that we never lose entropy in them
13463 or use exactly the same state in multiple threads.
13464 Access to the large state is not always serializable because
13465 the additional locking could be a performance killer, and
13466 md should be large enough anyway.
13470 * New file apps/app_rand.c with commonly needed functionality
13471 for handling the random seed file.
13473 Use the random seed file in some applications that previously did not:
13475 dsaparam -genkey (which also ignored its '-rand' option),
13478 x509 (when signing).
13479 Except on systems with /dev/urandom, it is crucial to have a random
13480 seed file at least for key creation, DSA signing, and for DH exchanges;
13481 for RSA signatures we could do without one.
13483 gendh and gendsa (unlike genrsa) used to read only the first byte
13484 of each file listed in the '-rand' option. The function as previously
13485 found in genrsa is now in app_rand.c and is used by all programs
13486 that support '-rand'.
13490 * In RAND_write_file, use mode 0600 for creating files;
13491 don't just chmod when it may be too late.
13495 * Report an error from X509_STORE_load_locations
13496 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
13500 * New function ASN1_mbstring_copy() this copies a string in either
13501 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
13502 into an ASN1_STRING type. A mask of permissible types is passed
13503 and it chooses the "minimal" type to use or an error if not type
13508 * Add function equivalents to the various macros in asn1.h. The old
13509 macros are retained with an M_ prefix. Code inside the library can
13510 use the M_ macros. External code (including the openssl utility)
13511 should *NOT* in order to be "shared library friendly".
13515 * Add various functions that can check a certificate's extensions
13516 to see if it usable for various purposes such as SSL client,
13517 server or S/MIME and CAs of these types. This is currently
13518 VERY EXPERIMENTAL but will ultimately be used for certificate chain
13519 verification. Also added a -purpose flag to x509 utility to
13520 print out all the purposes.
13524 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
13529 * New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
13530 for, obtain and decode and extension and obtain its critical flag.
13531 This allows all the necessary extension code to be handled in a
13532 single function call.
13536 * RC4 tune-up featuring 30-40% performance improvement on most RISC
13537 platforms. See crypto/rc4/rc4_enc.c for further details.
13541 * New -noout option to asn1parse. This causes no output to be produced
13542 its main use is when combined with -strparse and -out to extract data
13543 from a file (which may not be in ASN.1 format).
13547 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
13548 when producing the local key id.
13550 *Richard Levitte <levitte@stacken.kth.se>*
13552 * New option -dhparam in s_server. This allows a DH parameter file to be
13553 stated explicitly. If it is not stated then it tries the first server
13554 certificate file. The previous behaviour hard coded the filename
13559 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
13560 a public key to be input or output. For example:
13561 openssl rsa -in key.pem -pubout -out pubkey.pem
13562 Also added necessary DSA public key functions to handle this.
13566 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
13567 in the message. This was handled by allowing
13568 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
13570 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
13572 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
13573 to the end of the strings whereas this didn't. This would cause problems
13574 if strings read with d2i_ASN1_bytes() were later modified.
13576 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
13578 * Fix for base64 decode bug. When a base64 bio reads only one line of
13579 data and it contains EOF it will end up returning an error. This is
13580 caused by input 46 bytes long. The cause is due to the way base64
13581 BIOs find the start of base64 encoded data. They do this by trying a
13582 trial decode on each line until they find one that works. When they
13583 do a flag is set and it starts again knowing it can pass all the
13584 data directly through the decoder. Unfortunately it doesn't reset
13585 the context it uses. This means that if EOF is reached an attempt
13586 is made to pass two EOFs through the context and this causes the
13587 resulting error. This can also cause other problems as well. As is
13588 usual with these problems it takes *ages* to find and the fix is
13589 trivial: move one line.
13591 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) *
13593 * Ugly workaround to get s_client and s_server working under Windows. The
13594 old code wouldn't work because it needed to select() on sockets and the
13595 tty (for keypresses and to see if data could be written). Win32 only
13596 supports select() on sockets so we select() with a 1s timeout on the
13597 sockets and then see if any characters are waiting to be read, if none
13598 are present then we retry, we also assume we can always write data to
13599 the tty. This isn't nice because the code then blocks until we've
13600 received a complete line of data and it is effectively polling the
13601 keyboard at 1s intervals: however it's quite a bit better than not
13602 working at all :-) A dedicated Windows application might handle this
13603 with an event loop for example.
13607 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
13608 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
13609 will be called when RSA_sign() and RSA_verify() are used. This is useful
13610 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
13611 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
13612 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
13613 This necessitated the support of an extra signature type NID_md5_sha1
13614 for SSL signatures and modifications to the SSL library to use it instead
13615 of calling RSA_public_decrypt() and RSA_private_encrypt().
13619 * Add new -verify -CAfile and -CApath options to the crl program, these
13620 will lookup a CRL issuers certificate and verify the signature in a
13621 similar way to the verify program. Tidy up the crl program so it
13622 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
13623 less strict. It will now permit CRL extensions even if it is not
13624 a V2 CRL: this will allow it to tolerate some broken CRLs.
13628 * Initialize all non-automatic variables each time one of the openssl
13629 sub-programs is started (this is necessary as they may be started
13630 multiple times from the "OpenSSL>" prompt).
13632 *Lennart Bang, Bodo Moeller*
13634 * Preliminary compilation option RSA_NULL which disables RSA crypto without
13635 removing all other RSA functionality (this is what NO_RSA does). This
13636 is so (for example) those in the US can disable those operations covered
13637 by the RSA patent while allowing storage and parsing of RSA keys and RSA
13642 * Non-copying interface to BIO pairs.
13643 (still largely untested)
13647 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
13648 ASCII string. This was handled independently in various places before.
13652 * New functions UTF8_getc() and UTF8_putc() that parse and generate
13653 UTF8 strings a character at a time.
13657 * Use client_version from client hello to select the protocol
13658 (s23_srvr.c) and for RSA client key exchange verification
13659 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
13663 * Add various utility functions to handle SPKACs, these were previously
13664 handled by poking round in the structure internals. Added new function
13665 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
13666 print, verify and generate SPKACs. Based on an original idea from
13667 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
13671 * RIPEMD160 is operational on all platforms and is back in 'make test'.
13675 * Allow the config file extension section to be overwritten on the
13676 command line. Based on an original idea from Massimiliano Pala
13677 <madwolf@comune.modena.it>. The new option is called -extensions
13678 and can be applied to ca, req and x509. Also -reqexts to override
13679 the request extensions in req and -crlexts to override the crl extensions
13684 * Add new feature to the SPKAC handling in ca. Now you can include
13685 the same field multiple times by preceding it by "XXXX." for example:
13688 this is the same syntax as used in the req config file.
13692 * Allow certificate extensions to be added to certificate requests. These
13693 are specified in a 'req_extensions' option of the req section of the
13694 config file. They can be printed out with the -text option to req but
13695 are otherwise ignored at present.
13699 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
13700 data read consists of only the final block it would not decrypted because
13701 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
13702 A misplaced 'break' also meant the decrypted final block might not be
13703 copied until the next read.
13707 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
13708 a few extra parameters to the DH structure: these will be useful if
13709 for example we want the value of 'q' or implement X9.42 DH.
13713 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
13714 provides hooks that allow the default DSA functions or functions on a
13715 "per key" basis to be replaced. This allows hardware acceleration and
13716 hardware key storage to be handled without major modification to the
13717 library. Also added low level modexp hooks and CRYPTO_EX structure and
13718 associated functions.
13722 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
13723 as "read only": it can't be written to and the buffer it points to will
13724 not be freed. Reading from a read only BIO is much more efficient than
13725 a normal memory BIO. This was added because there are several times when
13726 an area of memory needs to be read from a BIO. The previous method was
13727 to create a memory BIO and write the data to it, this results in two
13728 copies of the data and an O(n^2) reading algorithm. There is a new
13729 function BIO_new_mem_buf() which creates a read only memory BIO from
13730 an area of memory. Also modified the PKCS#7 routines to use read only
13735 * Bugfix: ssl23_get_client_hello did not work properly when called in
13736 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
13737 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
13738 but a retry condition occurred while trying to read the rest.
13742 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
13743 NID_pkcs7_encrypted by default: this was wrong since this should almost
13744 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
13745 the encrypted data type: this is a more sensible place to put it and it
13746 allows the PKCS#12 code to be tidied up that duplicated this
13751 * Changed obj_dat.pl script so it takes its input and output files on
13752 the command line. This should avoid shell escape redirection problems
13757 * Initial support for certificate extension requests, these are included
13758 in things like Xenroll certificate requests. Included functions to allow
13759 extensions to be obtained and added.
13763 * -crlf option to s_client and s_server for sending newlines as
13764 CRLF (as required by many protocols).
13768 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999] ###
13770 * Install libRSAglue.a when OpenSSL is built with RSAref.
13772 *Ralf S. Engelschall*
13774 * A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
13776 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
13778 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
13783 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
13784 DH parameters/keys (q is lost during that conversion, but the resulting
13785 DH parameters contain its length).
13787 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
13788 much faster than DH_generate_parameters (which creates parameters
13789 where p = 2*q + 1), and also the smaller q makes DH computations
13790 much more efficient (160-bit exponentiation instead of 1024-bit
13791 exponentiation); so this provides a convenient way to support DHE
13792 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
13793 utter importance to use
13794 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
13796 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
13797 when such DH parameters are used, because otherwise small subgroup
13798 attacks may become possible!
13802 * Avoid memory leak in i2d_DHparams.
13806 * Allow the -k option to be used more than once in the enc program:
13807 this allows the same encrypted message to be read by multiple recipients.
13811 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
13812 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
13813 it will always use the numerical form of the OID, even if it has a short
13818 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
13819 method only got called if p,q,dmp1,dmq1,iqmp components were present,
13820 otherwise bn_mod_exp was called. In the case of hardware keys for example
13821 no private key components need be present and it might store extra data
13822 in the RSA structure, which cannot be accessed from bn_mod_exp.
13823 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
13824 private key operations.
13828 * Added support for SPARC Linux.
13832 * pem_password_cb function type incompatibly changed from
13833 typedef int pem_password_cb(char *buf, int size, int rwflag);
13835 ....(char *buf, int size, int rwflag, void *userdata);
13836 so that applications can pass data to their callbacks:
13837 The PEM[_ASN1]_{read,write}... functions and macros now take an
13838 additional void * argument, which is just handed through whenever
13839 the password callback is called.
13841 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
13843 New function SSL_CTX_set_default_passwd_cb_userdata.
13845 Compatibility note: As many C implementations push function arguments
13846 onto the stack in reverse order, the new library version is likely to
13847 interoperate with programs that have been compiled with the old
13848 pem_password_cb definition (PEM_whatever takes some data that
13849 happens to be on the stack as its last argument, and the callback
13850 just ignores this garbage); but there is no guarantee whatsoever that
13853 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
13854 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
13855 problems not only on Windows, but also on some Unix platforms.
13856 To avoid problematic command lines, these definitions are now in an
13857 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
13858 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
13862 * MIPS III/IV assembler module is reimplemented.
13866 * More DES library cleanups: remove references to srand/rand and
13867 delete an unused file.
13871 * Add support for the free Netwide assembler (NASM) under Win32,
13872 since not many people have MASM (ml) and it can be hard to obtain.
13873 This is currently experimental but it seems to work OK and pass all
13874 the tests. Check out INSTALL.W32 for info.
13878 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
13879 without temporary keys kept an extra copy of the server key,
13880 and connections with temporary keys did not free everything in case
13885 * New function RSA_check_key and new openssl rsa option -check
13886 for verifying the consistency of RSA keys.
13888 *Ulf Moeller, Bodo Moeller*
13890 * Various changes to make Win32 compile work:
13891 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
13892 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
13893 comparison" warnings.
13894 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
13898 * Add a debugging option to PKCS#5 v2 key generation function: when
13899 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
13900 derived keys are printed to stderr.
13904 * Copy the flags in ASN1_STRING_dup().
13906 *Roman E. Pavlov <pre@mo.msk.ru>*
13908 * The x509 application mishandled signing requests containing DSA
13909 keys when the signing key was also DSA and the parameters didn't match.
13911 It was supposed to omit the parameters when they matched the signing key:
13912 the verifying software was then supposed to automatically use the CA's
13913 parameters if they were absent from the end user certificate.
13915 Omitting parameters is no longer recommended. The test was also
13916 the wrong way round! This was probably due to unusual behaviour in
13917 EVP_cmp_parameters() which returns 1 if the parameters match.
13918 This meant that parameters were omitted when they *didn't* match and
13919 the certificate was useless. Certificates signed with 'ca' didn't have
13922 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
13924 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
13925 The interface is as follows:
13926 Applications can use
13927 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
13928 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
13929 "off" is now the default.
13930 The library internally uses
13931 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
13932 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
13933 to disable memory-checking temporarily.
13935 Some inconsistent states that previously were possible (and were
13936 even the default) are now avoided.
13938 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
13939 with each memory chunk allocated; this is occasionally more helpful
13940 than just having a counter.
13942 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
13944 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
13949 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
13950 which largely parallels "options", but is for changing API behaviour,
13951 whereas "options" are about protocol behaviour.
13952 Initial "mode" flags are:
13954 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
13955 a single record has been written.
13956 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
13957 retries use the same buffer location.
13958 (But all of the contents must be
13963 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
13966 * Fix problems with no-hmac etc.
13968 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
13970 * New functions RSA_get_default_method(), RSA_set_method() and
13971 RSA_get_method(). These allows replacement of RSA_METHODs without having
13972 to mess around with the internals of an RSA structure.
13976 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
13977 Also really enable memory leak checks in openssl.c and in some
13980 *Chad C. Mulligan, Bodo Moeller*
13982 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
13983 up the length of negative integers. This has now been simplified to just
13984 store the length when it is first determined and use it later, rather
13985 than trying to keep track of where data is copied and updating it to
13987 *Steve Henson, reported by Brien Wheeler
13988 <bwheeler@authentica-security.com>*
13990 * Add a new function PKCS7_signatureVerify. This allows the verification
13991 of a PKCS#7 signature but with the signing certificate passed to the
13992 function itself. This contrasts with PKCS7_dataVerify which assumes the
13993 certificate is present in the PKCS#7 structure. This isn't always the
13994 case: certificates can be omitted from a PKCS#7 structure and be
13995 distributed by "out of band" means (such as a certificate database).
13999 * Complete the PEM_* macros with DECLARE_PEM versions to replace the
14000 function prototypes in pem.h, also change util/mkdef.pl to add the
14001 necessary function names.
14005 * mk1mf.pl (used by Windows builds) did not properly read the
14006 options set by Configure in the top level Makefile, and Configure
14007 was not even able to write more than one option correctly.
14008 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
14012 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
14013 file to be loaded from a BIO or FILE pointer. The BIO version will
14014 for example allow memory BIOs to contain config info.
14018 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
14019 Whoever hopes to achieve shared-library compatibility across versions
14020 must use this, not the compile-time macro.
14021 (Exercise 0.9.4: Which is the minimum library version required by
14023 Note: All this applies only to multi-threaded programs, others don't
14028 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
14029 through a BIO pair triggered the default case, i.e.
14030 SSLerr(...,SSL_R_UNKNOWN_STATE).
14034 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
14035 can use the SSL library even if none of the specific BIOs is
14040 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
14041 for the encoded length.
14043 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
14045 * Add initial documentation of the X509V3 functions.
14049 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
14050 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
14051 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
14052 secure PKCS#8 private key format with a high iteration count.
14056 * Fix determination of Perl interpreter: A perl or perl5
14057 _directory_ in $PATH was also accepted as the interpreter.
14059 *Ralf S. Engelschall*
14061 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
14062 wrong with it but it was very old and did things like calling
14063 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
14064 unusual formatting.
14068 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
14069 to use the new extension code.
14073 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
14074 with macros. This should make it easier to change their form, add extra
14075 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
14080 * Add to configuration table a new entry that can specify an alternative
14081 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
14082 according to Mark Crispin <MRC@Panda.COM>.
14087 * DES CBC did not update the IV. Weird.
14091 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
14092 Changing the behaviour of the former might break existing programs --
14093 where IV updating is needed, des_ncbc_encrypt can be used.
14096 * When bntest is run from "make test" it drives bc to check its
14097 calculations, as well as internally checking them. If an internal check
14098 fails, it needs to cause bc to give a non-zero result or make test carries
14099 on without noticing the failure. Fixed.
14103 * DES library cleanups.
14107 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
14108 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
14109 ciphers. NOTE: although the key derivation function has been verified
14110 against some published test vectors it has not been extensively tested
14111 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
14116 * Instead of "mkdir -p", which is not fully portable, use new
14117 Perl script "util/mkdir-p.pl".
14121 * Rewrite the way password based encryption (PBE) is handled. It used to
14122 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
14123 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
14124 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
14125 the 'parameter' field of the AlgorithmIdentifier is passed to the
14126 underlying key generation function so it must do its own ASN1 parsing.
14127 This has also changed the EVP_PBE_CipherInit() function which now has a
14128 'parameter' argument instead of literal salt and iteration count values
14129 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
14133 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
14134 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
14135 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
14136 KEY" because this clashed with PKCS#8 unencrypted string. Since this
14137 value was just used as a "magic string" and not used directly its
14138 value doesn't matter.
14142 * Introduce some semblance of const correctness to BN. Shame C doesn't
14147 * "linux-sparc64" configuration (ultrapenguin).
14149 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
14150 "linux-sparc" configuration.
14152 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
14154 * config now generates no-xxx options for missing ciphers.
14158 * Support the EBCDIC character set (work in progress).
14159 File ebcdic.c not yet included because it has a different license.
14161 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14163 * Support BS2000/OSD-POSIX.
14165 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14167 * Make callbacks for key generation use void * instead of char *.
14171 * Make S/MIME samples compile (not yet tested).
14175 * Additional typesafe stacks.
14179 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
14184 ### Changes between 0.9.3 and 0.9.3a [29 May 1999] ###
14186 * New configuration variant "sco5-gcc".
14188 * Updated some demos.
14190 *Sean O Riordain, Wade Scholine*
14192 * Add missing BIO_free at exit of pkcs12 application.
14196 * Fix memory leak in conf.c.
14200 * Updates for Win32 to assembler version of MD5.
14204 * Set #! path to perl in apps/der_chop to where we found it
14205 instead of using a fixed path.
14209 * SHA library changes for irix64-mips4-cc.
14213 * Improvements for VMS support.
14218 ### Changes between 0.9.2b and 0.9.3 [24 May 1999] ###
14220 * Bignum library bug fix. IRIX 6 passes "make test" now!
14221 This also avoids the problems with SC4.2 and unpatched SC5.
14223 *Andy Polyakov <appro@fy.chalmers.se>*
14225 * New functions sk_num, sk_value and sk_set to replace the previous macros.
14226 These are required because of the typesafe stack would otherwise break
14227 existing code. If old code used a structure member which used to be STACK
14228 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
14229 sk_num or sk_value it would produce an error because the num, data members
14230 are not present in STACK_OF. Now it just produces a warning. sk_set
14231 replaces the old method of assigning a value to sk_value
14232 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
14233 that does this will no longer work (and should use sk_set instead) but
14234 this could be regarded as a "questionable" behaviour anyway.
14238 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
14239 correctly handle encrypted S/MIME data.
14243 * Change type of various DES function arguments from des_cblock
14244 (which means, in function argument declarations, pointer to char)
14245 to des_cblock * (meaning pointer to array with 8 char elements),
14246 which allows the compiler to do more typechecking; it was like
14247 that back in SSLeay, but with lots of ugly casts.
14249 Introduce new type const_des_cblock.
14253 * Reorganise the PKCS#7 library and get rid of some of the more obvious
14254 problems: find RecipientInfo structure that matches recipient certificate
14255 and initialise the ASN1 structures properly based on passed cipher.
14259 * Belatedly make the BN tests actually check the results.
14263 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
14264 to and from BNs: it was completely broken. New compilation option
14265 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
14266 key elements as negative integers.
14270 * Reorganize and speed up MD5.
14272 *Andy Polyakov <appro@fy.chalmers.se>*
14276 *Richard Levitte <richard@levitte.org>*
14278 * New option -out to asn1parse to allow the parsed structure to be
14279 output to a file. This is most useful when combined with the -strparse
14280 option to examine the output of things like OCTET STRINGS.
14284 * Make SSL library a little more fool-proof by not requiring any longer
14285 that SSL_set_{accept,connect}_state be called before
14286 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
14287 in many applications because usually everything *appeared* to work as
14288 intended anyway -- now it really works as intended).
14292 * Move openssl.cnf out of lib/.
14296 * Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
14297 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
14298 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
14300 *Ralf S. Engelschall*
14302 * Various fixes to the EVP and PKCS#7 code. It may now be able to
14303 handle PKCS#7 enveloped data properly.
14305 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
14307 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
14308 copying pointers. The cert_st handling is changed by this in
14309 various ways (and thus what used to be known as ctx->default_cert
14310 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
14311 any longer when s->cert does not give us what we need).
14312 ssl_cert_instantiate becomes obsolete by this change.
14313 As soon as we've got the new code right (possibly it already is?),
14314 we have solved a couple of bugs of the earlier code where s->cert
14315 was used as if it could not have been shared with other SSL structures.
14317 Note that using the SSL API in certain dirty ways now will result
14318 in different behaviour than observed with earlier library versions:
14319 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
14320 does not influence s as it used to.
14322 In order to clean up things more thoroughly, inside SSL_SESSION
14323 we don't use CERT any longer, but a new structure SESS_CERT
14324 that holds per-session data (if available); currently, this is
14325 the peer's certificate chain and, for clients, the server's certificate
14326 and temporary key. CERT holds only those values that can have
14327 meaningful defaults in an SSL_CTX.
14331 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
14332 from the internal representation. Various PKCS#7 fixes: remove some
14333 evil casts and set the enc_dig_alg field properly based on the signing
14338 * Allow PKCS#12 password to be set from the command line or the
14339 environment. Let 'ca' get its config file name from the environment
14340 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
14345 * Allow certificate policies extension to use an IA5STRING for the
14346 organization field. This is contrary to the PKIX definition but
14347 VeriSign uses it and IE5 only recognises this form. Document 'x509'
14352 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
14353 without disallowing inline assembler and the like for non-pedantic builds.
14357 * Support Borland C++ builder.
14359 *Janez Jere <jj@void.si>, modified by Ulf Möller*
14365 * SHA-1 cleanups and performance enhancements.
14367 *Andy Polyakov <appro@fy.chalmers.se>*
14369 * Sparc v8plus assembler for the bignum library.
14371 *Andy Polyakov <appro@fy.chalmers.se>*
14373 * Accept any -xxx and +xxx compiler options in Configure.
14377 * Update HPUX configuration.
14381 * Add missing sk_<type>_unshift() function to safestack.h
14383 *Ralf S. Engelschall*
14385 * New function SSL_CTX_use_certificate_chain_file that sets the
14386 "extra_cert"s in addition to the certificate. (This makes sense
14387 only for "PEM" format files, as chains as a whole are not
14392 * Support verify_depth from the SSL API.
14393 x509_vfy.c had what can be considered an off-by-one-error:
14394 Its depth (which was not part of the external interface)
14395 was actually counting the number of certificates in a chain;
14396 now it really counts the depth.
14400 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
14401 instead of X509err, which often resulted in confusing error
14402 messages since the error codes are not globally unique
14403 (e.g. an alleged error in ssl3_accept when a certificate
14404 didn't match the private key).
14406 * New function SSL_CTX_set_session_id_context that allows to set a default
14407 value (so that you don't need SSL_set_session_id_context for each
14408 connection using the SSL_CTX).
14412 * OAEP decoding bug fix.
14416 * Support INSTALL_PREFIX for package builders, as proposed by
14421 * New Configure options "threads" and "no-threads". For systems
14422 where the proper compiler options are known (currently Solaris
14423 and Linux), "threads" is the default.
14427 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
14431 * Install various scripts to $(OPENSSLDIR)/misc, not to
14432 $(INSTALLTOP)/bin -- they shouldn't clutter directories
14433 such as /usr/local/bin.
14437 * "make linux-shared" to build shared libraries.
14439 *Niels Poppe <niels@netbox.org>*
14441 * New Configure option no-<cipher> (rsa, idea, rc5, ...).
14445 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
14446 extension adding in x509 utility.
14450 * Remove NOPROTO sections and error code comments.
14454 * Partial rewrite of the DEF file generator to now parse the ANSI
14459 * New Configure options --prefix=DIR and --openssldir=DIR.
14463 * Complete rewrite of the error code script(s). It is all now handled
14464 by one script at the top level which handles error code gathering,
14465 header rewriting and C source file generation. It should be much better
14466 than the old method: it now uses a modified version of Ulf's parser to
14467 read the ANSI prototypes in all header files (thus the old K&R definitions
14468 aren't needed for error creation any more) and do a better job of
14469 translating function codes into names. The old 'ASN1 error code imbedded
14470 in a comment' is no longer necessary and it doesn't use .err files which
14471 have now been deleted. Also the error code call doesn't have to appear all
14472 on one line (which resulted in some large lines...).
14476 * Change #include filenames from <foo.h> to <openssl/foo.h>.
14480 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
14481 0 (which usually indicates a closed connection), but continue reading.
14485 * Fix some race conditions.
14489 * Add support for CRL distribution points extension. Add Certificate
14490 Policies and CRL distribution points documentation.
14494 * Move the autogenerated header file parts to crypto/opensslconf.h.
14498 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
14499 8 of keying material. Merlin has also confirmed interop with this fix
14500 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
14502 *Merlin Hughes <merlin@baltimore.ie>*
14504 * Fix lots of warnings.
14506 *Richard Levitte <levitte@stacken.kth.se>*
14508 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
14509 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
14511 *Richard Levitte <levitte@stacken.kth.se>*
14513 * Fix problems with sizeof(long) == 8.
14515 *Andy Polyakov <appro@fy.chalmers.se>*
14517 * Change functions to ANSI C.
14521 * Fix typos in error codes.
14523 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
14525 * Remove defunct assembler files from Configure.
14529 * SPARC v8 assembler BIGNUM implementation.
14531 *Andy Polyakov <appro@fy.chalmers.se>*
14533 * Support for Certificate Policies extension: both print and set.
14534 Various additions to support the r2i method this uses.
14538 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
14539 return a const string when you are expecting an allocated buffer.
14543 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
14544 types DirectoryString and DisplayText.
14548 * Add code to allow r2i extensions to access the configuration database,
14549 add an LHASH database driver and add several ctx helper functions.
14553 * Fix an evil bug in bn_expand2() which caused various BN functions to
14554 fail when they extended the size of a BIGNUM.
14558 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
14559 support typesafe stack.
14563 * Fix typo in SSL_[gs]et_options().
14565 *Nils Frostberg <nils@medcom.se>*
14567 * Delete various functions and files that belonged to the (now obsolete)
14568 old X509V3 handling code.
14572 * New Configure option "rsaref".
14576 * Don't auto-generate pem.h.
14580 * Introduce type-safe ASN.1 SETs.
14584 * Convert various additional casted stacks to type-safe STACK_OF() variants.
14586 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
14588 * Introduce type-safe STACKs. This will almost certainly break lots of code
14589 that links with OpenSSL (well at least cause lots of warnings), but fear
14590 not: the conversion is trivial, and it eliminates loads of evil casts. A
14591 few STACKed things have been converted already. Feel free to convert more.
14592 In the fullness of time, I'll do away with the STACK type altogether.
14596 * Add `openssl ca -revoke <certfile>' facility which revokes a certificate
14597 specified in <certfile> by updating the entry in the index.txt file.
14598 This way one no longer has to edit the index.txt file manually for
14599 revoking a certificate. The -revoke option does the gory details now.
14601 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
14603 * Fix `openssl crl -noout -text' combination where `-noout' killed the
14604 `-text' option at all and this way the `-noout -text' combination was
14605 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
14607 *Ralf S. Engelschall*
14609 * Make sure a corresponding plain text error message exists for the
14610 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
14611 verify callback function determined that a certificate was revoked.
14613 *Ralf S. Engelschall*
14615 * Bugfix: In test/testenc, don't test "openssl <cipher>" for
14616 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
14617 all available ciphers including rc5, which was forgotten until now.
14618 In order to let the testing shell script know which algorithms
14619 are available, a new (up to now undocumented) command
14620 "openssl list-cipher-commands" is used.
14624 * Bugfix: s_client occasionally would sleep in select() when
14625 it should have checked SSL_pending() first.
14629 * New functions DSA_do_sign and DSA_do_verify to provide access to
14630 the raw DSA values prior to ASN.1 encoding.
14634 * Tweaks to Configure
14636 *Niels Poppe <niels@netbox.org>*
14638 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
14643 * New variables $(RANLIB) and $(PERL) in the Makefiles.
14647 * New config option to avoid instructions that are illegal on the 80386.
14648 The default code is faster, but requires at least a 486.
14652 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
14653 SSL2_SERVER_VERSION (not used at all) macros, which are now the
14654 same as SSL2_VERSION anyway.
14658 * New "-showcerts" option for s_client.
14662 * Still more PKCS#12 integration. Add pkcs12 application to openssl
14663 application. Various cleanups and fixes.
14667 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
14668 modify error routines to work internally. Add error codes and PBE init
14669 to library startup routines.
14673 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
14674 packing functions to asn1 and evp. Changed function names and error
14675 codes along the way.
14679 * PKCS12 integration: and so it begins... First of several patches to
14680 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
14681 objects to objects.h
14685 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
14686 and display support for Thawte strong extranet extension.
14690 * Add LinuxPPC support.
14692 *Jeff Dubrule <igor@pobox.org>*
14694 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
14695 bn_div_words in alpha.s.
14697 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
14699 * Make sure the RSA OAEP test is skipped under -DRSAref because
14700 OAEP isn't supported when OpenSSL is built with RSAref.
14702 *Ulf Moeller <ulf@fitug.de>*
14704 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
14705 so they no longer are missing under -DNOPROTO.
14707 *Soren S. Jorvang <soren@t.dk>*
14710 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999] ###
14712 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
14713 doesn't work when the session is reused. Coming soon!
14717 * Fix a security hole, that allows sessions to be reused in the wrong
14718 context thus bypassing client cert protection! All software that uses
14719 client certs and session caches in multiple contexts NEEDS PATCHING to
14720 allow session reuse! A fuller solution is in the works.
14722 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
14724 * Some more source tree cleanups (removed obsolete files
14725 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
14726 permission on "config" script to be executable) and a fix for the INSTALL
14729 *Ulf Moeller <ulf@fitug.de>*
14731 * Remove some legacy and erroneous uses of malloc, free instead of
14734 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
14736 * Make rsa_oaep_test return non-zero on error.
14738 *Ulf Moeller <ulf@fitug.de>*
14740 * Add support for native Solaris shared libraries. Configure
14741 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
14742 if someone would make that last step automatic.
14744 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
14746 * ctx_size was not built with the right compiler during "make links". Fixed.
14750 * Change the meaning of 'ALL' in the cipher list. It now means "everything
14751 except NULL ciphers". This means the default cipher list will no longer
14752 enable NULL ciphers. They need to be specifically enabled e.g. with
14753 the string "DEFAULT:eNULL".
14757 * Fix to RSA private encryption routines: if p < q then it would
14758 occasionally produce an invalid result. This will only happen with
14759 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
14763 * Be less restrictive and allow also `perl util/perlpath.pl
14764 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
14765 because this way one can also use an interpreter named `perl5' (which is
14766 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
14767 installed as `perl').
14769 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
14771 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
14773 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
14775 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
14776 advapi32.lib to Win32 build and change the pem test comparison
14777 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
14778 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
14779 and crypto/des/ede_cbcm_enc.c.
14783 * DES quad checksum was broken on big-endian architectures. Fixed.
14787 * Comment out two functions in bio.h that aren't implemented. Fix up the
14788 Win32 test batch file so it (might) work again. The Win32 test batch file
14789 is horrible: I feel ill....
14793 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
14794 in e_os.h. Audit of header files to check ANSI and non ANSI
14795 sections: 10 functions were absent from non ANSI section and not exported
14796 from Windows DLLs. Fixed up libeay.num for new functions.
14800 * Make `openssl version' output lines consistent.
14802 *Ralf S. Engelschall*
14804 * Fix Win32 symbol export lists for BIO functions: Added
14805 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
14806 to ms/libeay{16,32}.def.
14808 *Ralf S. Engelschall*
14810 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
14811 fine under Unix and passes some trivial tests I've now added. But the
14812 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
14813 added to make sure no one expects that this stuff really works in the
14814 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
14815 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
14818 *Ralf S. Engelschall*
14820 * Fix the generation of two part addresses in perl.
14822 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
14824 * Add config entry for Linux on MIPS.
14826 *John Tobey <jtobey@channel1.com>*
14828 * Make links whenever Configure is run, unless we are on Windoze.
14832 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
14833 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
14838 * Add a useful kludge to allow package maintainers to specify compiler and
14839 other platforms details on the command line without having to patch the
14840 Configure script everytime: One now can use ``perl Configure
14841 <id>:<details>'', i.e. platform ids are allowed to have details appended
14842 to them (separated by colons). This is treated as there would be a static
14843 pre-configured entry in Configure's %table under key <id> with value
14844 <details> and ``perl Configure <id>'' is called. So, when you want to
14845 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
14846 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
14847 now, which overrides the FreeBSD-elf entry on-the-fly.
14849 *Ralf S. Engelschall*
14851 * Disable new TLS1 ciphersuites by default: they aren't official yet.
14855 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
14856 on the `perl Configure ...' command line. This way one can compile
14857 OpenSSL libraries with Position Independent Code (PIC) which is needed
14858 for linking it into DSOs.
14860 *Ralf S. Engelschall*
14862 * Remarkably, export ciphers were totally broken and no-one had noticed!
14867 * Cleaned up the LICENSE document: The official contact for any license
14868 questions now is the OpenSSL core team under openssl-core@openssl.org.
14869 And add a paragraph about the dual-license situation to make sure people
14870 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
14871 to the OpenSSL toolkit.
14873 *Ralf S. Engelschall*
14875 * General source tree makefile cleanups: Made `making xxx in yyy...'
14876 display consistent in the source tree and replaced `/bin/rm' by `rm'.
14877 Additionally cleaned up the `make links' target: Remove unnecessary
14878 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
14879 to speed processing and no longer clutter the display with confusing
14880 stuff. Instead only the actually done links are displayed.
14882 *Ralf S. Engelschall*
14884 * Permit null encryption ciphersuites, used for authentication only. It used
14885 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
14886 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
14891 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
14892 signed attributes when verifying signatures (this would break them),
14893 the detached data encoding was wrong and public keys obtained using
14894 X509_get_pubkey() weren't freed.
14898 * Add text documentation for the BUFFER functions. Also added a work around
14899 to a Win95 console bug. This was triggered by the password read stuff: the
14900 last character typed gets carried over to the next fread(). If you were
14901 generating a new cert request using 'req' for example then the last
14902 character of the passphrase would be CR which would then enter the first
14907 * Added the new `Includes OpenSSL Cryptography Software' button as
14908 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
14909 button and can be used by applications based on OpenSSL to show the
14910 relationship to the OpenSSL project.
14912 *Ralf S. Engelschall*
14914 * Remove confusing variables in function signatures in files
14915 ssl/ssl_lib.c and ssl/ssl.h.
14917 *Lennart Bong <lob@kulthea.stacken.kth.se>*
14919 * Don't install bss_file.c under PREFIX/include/
14921 *Lennart Bong <lob@kulthea.stacken.kth.se>*
14923 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
14924 functions that return function pointers and has support for NT specific
14925 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
14926 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
14927 unsigned to signed types: this was killing the Win32 compile.
14931 * Add new certificate file to stack functions,
14932 SSL_add_dir_cert_subjects_to_stack() and
14933 SSL_add_file_cert_subjects_to_stack(). These largely supplant
14934 SSL_load_client_CA_file(), and can be used to add multiple certs easily
14935 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
14936 This means that Apache-SSL and similar packages don't have to mess around
14937 to add as many CAs as they want to the preferred list.
14941 * Experiment with doxygen documentation. Currently only partially applied to
14943 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
14944 openssl.doxy as the configuration file.
14948 * Get rid of remaining C++-style comments which strict C compilers hate.
14950 *Ralf S. Engelschall, pointed out by Carlos Amengual*
14952 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
14953 compiled in by default: it has problems with large keys.
14957 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
14958 DH private keys and/or callback functions which directly correspond to
14959 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
14960 is needed for applications which have to configure certificates on a
14961 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
14963 For the RSA certificate situation is makes no difference, but
14964 for the DSA certificate situation this fixes the "no shared cipher"
14965 problem where the OpenSSL cipher selection procedure failed because the
14966 temporary keys were not overtaken from the context and the API provided
14967 no way to reconfigure them.
14968 The new functions now let applications reconfigure the stuff and they
14969 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
14970 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
14971 non-public-API function ssl_cert_instantiate() is used as a helper
14972 function and also to reduce code redundancy inside ssl_rsa.c.
14974 *Ralf S. Engelschall*
14976 * Move s_server -dcert and -dkey options out of the undocumented feature
14977 area because they are useful for the DSA situation and should be
14978 recognized by the users.
14980 *Ralf S. Engelschall*
14982 * Fix the cipher decision scheme for export ciphers: the export bits are
14983 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
14984 SSL_EXP_MASK. So, the original variable has to be used instead of the
14985 already masked variable.
14987 *Richard Levitte <levitte@stacken.kth.se>*
14989 * Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
14991 *Richard Levitte <levitte@stacken.kth.se>*
14993 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
14994 from `int' to `unsigned int' because it's a length and initialized by
14995 EVP_DigestFinal() which expects an `unsigned int *'.
14997 *Richard Levitte <levitte@stacken.kth.se>*
14999 * Don't hard-code path to Perl interpreter on shebang line of Configure
15000 script. Instead use the usual Shell->Perl transition trick.
15002 *Ralf S. Engelschall*
15004 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
15005 (in addition to RSA certificates) to match the behaviour of `openssl dsa
15006 -noout -modulus' as it's already the case for `openssl rsa -noout
15007 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
15008 currently the public key is printed (a decision which was already done by
15009 `openssl dsa -modulus' in the past) which serves a similar purpose.
15010 Additionally the NO_RSA no longer completely removes the whole -modulus
15011 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
15014 *Ralf S. Engelschall*
15016 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
15017 BIO. See the source (crypto/evp/bio_ok.c) for more info.
15019 *Arne Ansper <arne@ats.cyber.ee>*
15021 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
15022 to be added. Now both 'req' and 'ca' can use new objects defined in the
15027 * Add cool BIO that does syslog (or event log on NT).
15029 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
15031 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
15032 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
15033 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
15034 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
15038 * Add preliminary config info for new extension code.
15042 * Make RSA_NO_PADDING really use no padding.
15044 *Ulf Moeller <ulf@fitug.de>*
15046 * Generate errors when private/public key check is done.
15050 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
15051 for some CRL extensions and new objects added.
15055 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
15056 key usage extension and fuller support for authority key id.
15060 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
15061 padding method for RSA, which is recommended for new applications in PKCS
15062 #1 v2.0 (RFC 2437, October 1998).
15063 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
15064 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
15065 against Bleichbacher's attack on RSA.
15066 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
15069 * Updates to the new SSL compression code
15071 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
15073 * Fix so that the version number in the master secret, when passed
15074 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
15075 (because the server will not accept higher), that the version number
15076 is 0x03,0x01, not 0x03,0x00
15078 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
15080 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
15081 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
15082 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
15086 * Support for RAW extensions where an arbitrary extension can be
15087 created by including its DER encoding. See apps/openssl.cnf for
15092 * Make sure latest Perl versions don't interpret some generated C array
15093 code as Perl array code in the crypto/err/err_genc.pl script.
15095 *Lars Weber <3weber@informatik.uni-hamburg.de>*
15097 * Modify ms/do_ms.bat to not generate assembly language makefiles since
15098 not many people have the assembler. Various Win32 compilation fixes and
15099 update to the INSTALL.W32 file with (hopefully) more accurate Win32
15100 build instructions.
15104 * Modify configure script 'Configure' to automatically create crypto/date.h
15105 file under Win32 and also build pem.h from pem.org. New script
15106 util/mkfiles.pl to create the MINFO file on environments that can't do a
15107 'make files': perl util/mkfiles.pl >MINFO should work.
15111 * Major rework of DES function declarations, in the pursuit of correctness
15112 and purity. As a result, many evil casts evaporated, and some weirdness,
15113 too. You may find this causes warnings in your code. Zapping your evil
15114 casts will probably fix them. Mostly.
15118 * Fix for a typo in asn1.h. Bug fix to object creation script
15119 obj_dat.pl. It considered a zero in an object definition to mean
15120 "end of object": none of the objects in objects.h have any zeros
15121 so it wasn't spotted.
15123 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
15125 * Add support for Triple DES Cipher Block Chaining with Output Feedback
15126 Masking (CBCM). In the absence of test vectors, the best I have been able
15127 to do is check that the decrypt undoes the encrypt, so far. Send me test
15128 vectors if you have them.
15132 * Correct calculation of key length for export ciphers (too much space was
15133 allocated for null ciphers). This has not been tested!
15137 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
15138 message is now correct (it understands "crypto" and "ssl" on its
15139 command line). There is also now an "update" option. This will update
15140 the util/ssleay.num and util/libeay.num files with any new functions.
15142 perl util/mkdef.pl crypto ssl update
15143 it will update them.
15147 * Overhauled the Perl interface (perl/*):
15148 - ported BN stuff to OpenSSL's different BN library
15149 - made the perl/ source tree CVS-aware
15150 - renamed the package from SSLeay to OpenSSL (the files still contain
15151 their history because I've copied them in the repository)
15152 - removed obsolete files (the test scripts will be replaced
15153 by better Test::Harness variants in the future)
15155 *Ralf S. Engelschall*
15157 * First cut for a very conservative source tree cleanup:
15158 1. merge various obsolete readme texts into doc/ssleay.txt
15159 where we collect the old documents and readme texts.
15160 2. remove the first part of files where I'm already sure that we no
15161 longer need them because of three reasons: either they are just temporary
15162 files which were left by Eric or they are preserved original files where
15163 I've verified that the diff is also available in the CVS via "cvs diff
15164 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
15165 the crypto/md/ stuff).
15167 *Ralf S. Engelschall*
15169 * More extension code. Incomplete support for subject and issuer alt
15170 name, issuer and authority key id. Change the i2v function parameters
15171 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
15172 what that's for :-) Fix to ASN1 macro which messed up
15173 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
15177 * Preliminary support for ENUMERATED type. This is largely copied from the
15182 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
15184 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
15186 * Make sure `make rehash' target really finds the `openssl' program.
15188 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
15190 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
15191 like to hear about it if this slows down other processors.
15195 * Add CygWin32 platform information to Configure script.
15197 *Alan Batie <batie@aahz.jf.intel.com>*
15199 * Fixed ms/32all.bat script: `no_asm' -> `no-asm'
15201 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
15203 * New program nseq to manipulate netscape certificate sequences
15207 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
15212 * Fixes to BN code. Previously the default was to define BN_RECURSION
15213 but the BN code had some problems that would cause failures when
15214 doing certificate verification and some other functions.
15216 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
15218 * Add ASN1 and PEM code to support netscape certificate sequences.
15222 * Add ASN1 and PEM code to support netscape certificate sequences.
15226 * Add several PKIX and private extended key usage OIDs.
15230 * Modify the 'ca' program to handle the new extension code. Modify
15231 openssl.cnf for new extension format, add comments.
15235 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
15236 and add a sample to openssl.cnf so req -x509 now adds appropriate
15241 * Continued X509 V3 changes. Add to other makefiles, integrate with the
15242 error code, add initial support to X509_print() and x509 application.
15246 * Takes a deep breath and start adding X509 V3 extension support code. Add
15247 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
15248 stuff is currently isolated and isn't even compiled yet.
15252 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
15253 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
15254 Removed the versions check from X509 routines when loading extensions:
15255 this allows certain broken certificates that don't set the version
15256 properly to be processed.
15260 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
15261 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
15262 can still be regenerated with "make depend".
15266 * Spelling mistake in C version of CAST-128.
15268 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
15270 * Changes to the error generation code. The perl script err-code.pl
15271 now reads in the old error codes and retains the old numbers, only
15272 adding new ones if necessary. It also only changes the .err files if new
15273 codes are added. The makefiles have been modified to only insert errors
15274 when needed (to avoid needlessly modifying header files). This is done
15275 by only inserting errors if the .err file is newer than the auto generated
15276 C file. To rebuild all the error codes from scratch (the old behaviour)
15277 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
15278 or delete all the .err files.
15282 * CAST-128 was incorrectly implemented for short keys. The C version has
15283 been fixed, but is untested. The assembler versions are also fixed, but
15284 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
15285 to regenerate it if needed.
15286 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
15287 Hagino <itojun@kame.net>*
15289 * File was opened incorrectly in randfile.c.
15291 *Ulf Möller <ulf@fitug.de>*
15293 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
15294 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
15295 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
15296 al: it's just almost always a UTCTime. Note this patch adds new error
15297 codes so do a "make errors" if there are problems.
15301 * Correct Linux 1 recognition in config.
15303 *Ulf Möller <ulf@fitug.de>*
15305 * Remove pointless MD5 hash when using DSA keys in ca.
15307 *Anonymous <nobody@replay.com>*
15309 * Generate an error if given an empty string as a cert directory. Also
15310 generate an error if handed NULL (previously returned 0 to indicate an
15311 error, but didn't set one).
15313 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
15315 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
15319 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
15320 parameters. This was causing a warning which killed off the Win32 compile.
15324 * Remove C++ style comments from crypto/bn/bn_local.h.
15326 *Neil Costigan <neil.costigan@celocom.com>*
15328 * The function OBJ_txt2nid was broken. It was supposed to return a nid
15329 based on a text string, looking up short and long names and finally
15330 "dot" format. The "dot" format stuff didn't work. Added new function
15331 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
15332 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
15333 OID is not part of the table.
15337 * Add prototypes to X509 lookup/verify methods, fixing a bug in
15338 X509_LOOKUP_by_alias().
15342 * Sort openssl functions by name.
15346 * Get the gendsa program working (hopefully) and add it to app list. Remove
15347 encryption from sample DSA keys (in case anyone is interested the password
15352 * Make _all_ *_free functions accept a NULL pointer.
15354 *Frans Heymans <fheymans@isaserver.be>*
15356 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
15359 *Anonymous <nobody@replay.com>*
15361 * s_server should send the CAfile as acceptable CAs, not its own cert.
15363 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
15365 * Don't blow it for numeric -newkey arguments to apps/req.
15367 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
15369 * Temp key "for export" tests were wrong in s3_srvr.c.
15371 *Anonymous <nobody@replay.com>*
15373 * Add prototype for temp key callback functions
15374 SSL_CTX_set_tmp_{rsa,dh}_callback().
15378 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
15379 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
15383 * X509_name_add_entry() freed the wrong thing after an error.
15385 *Arne Ansper <arne@ats.cyber.ee>*
15387 * rsa_eay.c would attempt to free a NULL context.
15389 *Arne Ansper <arne@ats.cyber.ee>*
15391 * BIO_s_socket() had a broken should_retry() on Windoze.
15393 *Arne Ansper <arne@ats.cyber.ee>*
15395 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
15397 *Arne Ansper <arne@ats.cyber.ee>*
15399 * Make sure the already existing X509_STORE->depth variable is initialized
15400 in X509_STORE_new(), but document the fact that this variable is still
15401 unused in the certificate verification process.
15403 *Ralf S. Engelschall*
15405 * Fix the various library and apps files to free up pkeys obtained from
15406 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
15410 * Fix reference counting in X509_PUBKEY_get(). This makes
15411 demos/maurice/example2.c work, amongst others, probably.
15413 *Steve Henson and Ben Laurie*
15415 * First cut of a cleanup for apps/. First the `ssleay' program is now named
15416 `openssl' and second, the shortcut symlinks for the `openssl <command>'
15417 are no longer created. This way we have a single and consistent command
15418 line interface `openssl <command>', similar to `cvs <command>'.
15420 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
15422 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
15423 BIT STRING wrapper always have zero unused bits.
15427 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
15431 * Make the top-level INSTALL documentation easier to understand.
15435 * Makefiles updated to exit if an error occurs in a sub-directory
15436 make (including if user presses ^C) [Paul Sutton]
15438 * Make Montgomery context stuff explicit in RSA data structure.
15442 * Fix build order of pem and err to allow for generated pem.h.
15446 * Fix renumbering bug in X509_NAME_delete_entry().
15450 * Enhanced the err-ins.pl script so it makes the error library number
15451 global and can add a library name. This is needed for external ASN1 and
15452 other error libraries.
15456 * Fixed sk_insert which never worked properly.
15460 * Fix ASN1 macros so they can handle indefinite length constructed
15461 EXPLICIT tags. Some non standard certificates use these: they can now
15466 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
15467 into a single doc/ssleay.txt bundle. This way the information is still
15468 preserved but no longer messes up this directory. Now it's new room for
15469 the new set of documentation files.
15471 *Ralf S. Engelschall*
15473 * SETs were incorrectly DER encoded. This was a major pain, because they
15474 shared code with SEQUENCEs, which aren't coded the same. This means that
15475 almost everything to do with SETs or SEQUENCEs has either changed name or
15476 number of arguments.
15478 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
15480 * Fix test data to work with the above.
15484 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
15485 was already fixed by Eric for 0.9.1 it seems.
15487 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
15489 * Autodetect FreeBSD3.
15493 * Fix various bugs in Configure. This affects the following platforms:
15497 unixware-2.0-pentium
15502 * Eliminate generated files from CVS. Reorder tests to regenerate files
15503 before they are needed.
15507 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
15512 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998] ###
15514 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
15515 changed SSLeay to OpenSSL in version strings.
15517 *Ralf S. Engelschall*
15519 * Some fixups to the top-level documents.
15523 * Fixed the nasty bug where rsaref.h was not found under compile-time
15524 because the symlink to include/ was missing.
15526 *Ralf S. Engelschall*
15528 * Incorporated the popular no-RSA/DSA-only patches
15529 which allow to compile a RSA-free SSLeay.
15531 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
15533 * Fixed nasty rehash problem under `make -f Makefile.ssl links'
15534 when "ssleay" is still not found.
15536 *Ralf S. Engelschall*
15538 * Added more platforms to Configure: Cray T3E, HPUX 11,
15540 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
15542 * Updated the README file.
15544 *Ralf S. Engelschall*
15546 * Added various .cvsignore files in the CVS repository subdirs
15547 to make a "cvs update" really silent.
15549 *Ralf S. Engelschall*
15551 * Recompiled the error-definition header files and added
15552 missing symbols to the Win32 linker tables.
15554 *Ralf S. Engelschall*
15556 * Cleaned up the top-level documents;
15557 o new files: CHANGES and LICENSE
15558 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
15559 o merged COPYRIGHT into LICENSE
15560 o removed obsolete TODO file
15561 o renamed MICROSOFT to INSTALL.W32
15563 *Ralf S. Engelschall*
15565 * Removed dummy files from the 0.9.1b source tree:
15566 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
15567 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
15568 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
15569 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
15570 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
15572 *Ralf S. Engelschall*
15574 * Added various platform portability fixes.
15578 * The Genesis of the OpenSSL rpject:
15579 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
15580 Young and Tim J. Hudson created while they were working for C2Net until
15583 *The OpenSSL Project*
15586 ### Changes between 0.9.0b and 0.9.1b [not released] ###
15588 * Updated a few CA certificates under certs/
15592 * Changed some BIGNUM api stuff.
15596 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
15597 DGUX x86, Linux Alpha, etc.
15601 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
15602 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
15607 * Add -strparse option to asn1pars program which parses nested
15610 *Dr Stephen Henson <shenson@bigfoot.com>*
15612 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
15616 * DSA fix for "ca" program.
15620 * Added "-genkey" option to "dsaparam" program.
15624 * Added RIPE MD160 (rmd160) message digest.
15628 * Added -a (all) option to "ssleay version" command.
15632 * Added PLATFORM define which is the id given to Configure.
15636 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
15640 * Extended the ASN.1 parser routines.
15644 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
15648 * Added a BN_CTX to the BN library.
15652 * Fixed the weak key values in DES library
15656 * Changed API in EVP library for cipher aliases.
15660 * Added support for RC2/64bit cipher.
15664 * Converted the lhash library to the crypto/mem.c functions.
15668 * Added more recognized ASN.1 object ids.
15672 * Added more RSA padding checks for SSL/TLS.
15676 * Added BIO proxy/filter functionality.
15680 * Added extra_certs to SSL_CTX which can be used
15681 send extra CA certificates to the client in the CA cert chain sending
15682 process. It can be configured with SSL_CTX_add_extra_chain_cert().
15686 * Now Fortezza is denied in the authentication phase because
15687 this is key exchange mechanism is not supported by SSLeay at all.
15691 * Additional PKCS1 checks.
15695 * Support the string "TLSv1" for all TLS v1 ciphers.
15699 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
15700 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
15704 * Fixed a few memory leaks.
15708 * Fixed various code and comment typos.
15712 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
15713 bytes sent in the client random.
15715 *Edward Bishop <ebishop@spyglass.com>*