5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the git commit log; for example,
7 https://github.com/openssl/openssl/commits/ and pick the appropriate
10 Changes between 1.1.1 and 3.0.0 [xx XXX xxxx]
12 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
13 used even when parsing explicit parameters, when loading a serialized key
14 or calling `EC_GROUP_new_from_ecpkparameters()`/
15 `EC_GROUP_new_from_ecparameters()`.
16 This prevents bypass of security hardening and performance gains,
17 especially for curves with specialized EC_METHODs.
18 By default, if a key encoded with explicit parameters is loaded and later
19 serialized, the output is still encoded with explicit parameters, even if
20 internally a "named" EC_GROUP is used for computation.
23 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
24 this change, EC_GROUP_set_generator would accept order and/or cofactor as
25 NULL. After this change, only the cofactor parameter can be NULL. It also
26 does some minimal sanity checks on the passed order.
30 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
31 An attack is simple, if the first CMS_recipientInfo is valid but the
32 second CMS_recipientInfo is chosen ciphertext. If the second
33 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
34 encryption key will be replaced by garbage, and the message cannot be
35 decoded, but if the RSA decryption fails, the correct encryption key is
36 used and the recipient will not notice the attack.
37 As a work around for this potential attack the length of the decrypted
38 key must be equal to the cipher default key length, in case the
39 certifiate is not given and all recipientInfo are tried out.
40 The old behaviour can be re-enabled in the CMS code by setting the
41 CMS_DEBUG_DECRYPT flag.
44 *) Early start up entropy quality from the DEVRANDOM seed source has been
45 improved for older Linux systems. The RAND subsystem will wait for
46 /dev/random to be producing output before seeding from /dev/urandom.
47 The seeded state is stored for future library initialisations using
48 a system global shared memory segment. The shared memory identifier
49 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
50 the desired value. The default identifier is 114.
53 *) Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
54 when primes for RSA keys are computed.
55 Since we previously always generated primes == 2 (mod 3) for RSA keys,
56 the 2-prime and 3-prime RSA modules were easy to distinguish, since
57 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
58 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
59 This avoids possible fingerprinting of newly generated RSA modules.
62 *) Correct the extended master secret constant on EBCDIC systems. Without this
63 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
64 negotiate EMS will fail. Unfortunately this also means that TLS connections
65 between EBCDIC systems with this fix, and EBCDIC systems without this
66 fix will fail if they negotiate EMS.
69 *) Changed the library initialisation so that the config file is now loaded
70 by default. This was already the case for libssl. It now occurs for both
71 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
72 OPENSSL_init_crypto() to suppress automatic loading of a config file.
75 *) Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
76 where the former acts as a replacement for ERR_put_error(), and the
77 latter replaces the combination ERR_put_error()+ERR_add_error_data().
78 ERR_raise_data() adds more flexibility by taking a format string and
79 an arbitrary number of arguments following it, to be processed with
83 *) Introduced a new function, OSSL_PROVIDER_available(), which can be used
84 to check if a named provider is loaded and available. When called, it
85 will also activate all fallback providers if such are still present.
88 *) Enforce a minimum DH modulus size of 512 bits.
91 *) Changed DH parameters to generate the order q subgroup instead of 2q.
92 Previously generated DH parameters are still accepted by DH_check
93 but DH_generate_key works around that by clearing bit 0 of the
94 private key for those. This avoids leaking bit 0 of the private key.
97 *) Significantly reduce secure memory usage by the randomness pools.
100 *) {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
104 *) A new type, EVP_KEYEXCH, has been introduced to represent key exchange
105 algorithms. An implementation of a key exchange algorithm can be obtained
106 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
107 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
108 the older EVP_PKEY_derive_init() function. See the man pages for the new
109 functions for further details.
112 *) The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
115 *) Removed the function names from error messages and deprecated the
118 *) Removed NextStep support and the macro OPENSSL_UNISTD
121 *) Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
122 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
123 Also removed "export var as function" capability; we do not export
124 variables, only functions.
127 *) RC5_32_set_key has been changed to return an int type, with 0 indicating
128 an error and 1 indicating success. In previous versions of OpenSSL this
129 was a void type. If a key was set longer than the maximum possible this
133 *) Support SM2 signing and verification schemes with X509 certificate.
136 *) Use SHA256 as the default digest for TS query in the ts app.
139 *) Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
140 This checks that the salt length is at least 128 bits, the derived key
141 length is at least 112 bits, and that the iteration count is at least 1000.
142 For backwards compatibility these checks are disabled by default in the
143 default provider, but are enabled by default in the fips provider.
144 To enable or disable these checks use the control
145 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
148 *) Default cipher lists/suites are now available via a function, the
149 #defines are deprecated.
152 *) Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
153 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
154 for Windows Store apps easier. Also, the "no-uplink" option has been added.
157 *) Join the directories crypto/x509 and crypto/x509v3
160 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
161 This changes the size when using the genpkey app when no size is given. It
162 fixes an omission in earlier changes that changed all RSA, DSA and DH
163 generation apps to use 2048 bits by default.
166 *) Added command 'openssl kdf' that uses the EVP_KDF API.
169 *) Added command 'openssl mac' that uses the EVP_MAC API.
172 *) Added OPENSSL_info() to get diverse built-in OpenSSL data, such
173 as default directories. Also added the command 'openssl info'
174 for scripting purposes.
177 *) The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
178 deprecated. These undocumented functions were never integrated into the EVP
179 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
180 Bi-directional IGE mode. These modes were never formally standardised and
181 usage of these functions is believed to be very small. In particular
182 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
183 is ever used. The security implications are believed to be minimal, but
184 this issue was never fixed for backwards compatibility reasons. New code
185 should not use these modes.
188 *) Add prediction resistance to the DRBG reseeding process.
191 *) Limit the number of blocks in a data unit for AES-XTS to 2^20 as
192 mandated by IEEE Std 1619-2018.
195 *) Added newline escaping functionality to a filename when using openssl dgst.
196 This output format is to replicate the output format found in the '*sum'
197 checksum programs. This aims to preserve backward compatibility.
198 [Matt Eaton, Richard Levitte, and Paul Dale]
200 *) Removed the heartbeat message in DTLS feature, as it has very
201 little usage and doesn't seem to fulfill a valuable purpose.
202 The configuration option is now deprecated.
205 *) Changed the output of 'openssl {digestname} < file' to display the
206 digest name in its output.
209 *) Added a new generic trace API which provides support for enabling
210 instrumentation through trace output. This feature is mainly intended
211 as an aid for developers and is disabled by default. To utilize it,
212 OpenSSL needs to be configured with the `enable-trace` option.
214 If the tracing API is enabled, the application can activate trace output
215 by registering BIOs as trace channels for a number of tracing and debugging
218 The 'openssl' application has been expanded to enable any of the types
219 available via environment variables defined by the user, and serves as
220 one possible example on how to use this functionality.
221 [Richard Levitte & Matthias St. Pierre]
223 *) Added build tests for C++. These are generated files that only do one
224 thing, to include one public OpenSSL head file each. This tests that
225 the public header files can be usefully included in a C++ application.
227 This test isn't enabled by default. It can be enabled with the option
228 'enable-buildtest-c++'.
231 *) Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
234 *) Add KMAC to EVP_MAC.
237 *) Added property based algorithm implementation selection framework to
241 *) Added SCA hardening for modular field inversion in EC_GROUP through
242 a new dedicated field_inv() pointer in EC_METHOD.
243 This also addresses a leakage affecting conversions from projective
244 to affine coordinates.
245 [Billy Bob Brumley, Nicola Tuveri]
247 *) Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
248 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
249 those algorithms that were already supported through the EVP_PKEY API
250 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
251 and scrypt are now wrappers that call EVP_KDF.
254 *) Build devcrypto engine as a dynamic engine.
257 *) Add keyed BLAKE2 to EVP_MAC.
260 *) Fix a bug in the computation of the endpoint-pair shared secret used
261 by DTLS over SCTP. This breaks interoperability with older versions
262 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
263 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
264 interoperability with such broken implementations. However, enabling
265 this switch breaks interoperability with correct implementations.
267 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
268 re-used X509_PUBKEY object if the second PUBKEY is malformed.
271 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
274 *) Change the license to the Apache License v2.0.
277 *) Change the possible version information given with OPENSSL_API_COMPAT.
278 It may be a pre-3.0.0 style numerical version number as it was defined
279 in 1.1.0, and it may also simply take the major version number.
281 Because of the version numbering of pre-3.0.0 releases, the values 0,
282 1 and 2 are equivalent to 0x00908000L (0.9.8), 0x10000000L (1.0.0) and
283 0x10100000L (1.1.0), respectively.
286 *) Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
288 o Major releases (indicated by incrementing the MAJOR release number)
289 may introduce incompatible API/ABI changes.
290 o Minor releases (indicated by incrementing the MINOR release number)
291 may introduce new features but retain API/ABI compatibility.
292 o Patch releases (indicated by incrementing the PATCH number)
293 are intended for bug fixes and other improvements of existing
294 features only (like improving performance or adding documentation)
295 and retain API/ABI compatibility.
298 *) Add support for RFC5297 SIV mode (siv128), including AES-SIV.
301 *) Remove the 'dist' target and add a tarball building script. The
302 'dist' target has fallen out of use, and it shouldn't be
303 necessary to configure just to create a source distribution.
306 *) Recreate the OS390-Unix config target. It no longer relies on a
307 special script like it did for OpenSSL pre-1.1.0.
310 *) Instead of having the source directories listed in Configure, add
311 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
315 *) Add GMAC to EVP_MAC.
318 *) Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
321 *) Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
322 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
323 to facilitate the continued use of MACs through raw private keys in
324 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
327 *) Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
328 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
331 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
332 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
333 are retained for backwards compatibility.
336 *) AES-XTS mode now enforces that its two keys are different to mitigate
337 the attacked described in "Efficient Instantiations of Tweakable
338 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
339 Details of this attack can be obtained from:
340 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
343 *) Rename the object files, i.e. give them other names than in previous
344 versions. Their names now include the name of the final product, as
345 well as its type mnemonic (bin, lib, shlib).
348 *) Added new option for 'openssl list', '-objects', which will display the
349 list of built in objects, i.e. OIDs with names.
352 *) Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
353 improves application performance by removing data copies and providing
354 applications with zero-copy system calls such as sendfile and splice.
357 Changes between 1.1.1a and 1.1.1b [xx XXX xxxx]
359 *) Change the info callback signals for the start and end of a post-handshake
360 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
361 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
362 confused by this and assume that a TLSv1.2 renegotiation has started. This
363 can break KeyUpdate handling. Instead we no longer signal the start and end
364 of a post handshake message exchange (although the messages themselves are
365 still signalled). This could break some applications that were expecting
366 the old signals. However without this KeyUpdate is not usable for many
370 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
372 *) Timing vulnerability in DSA signature generation
374 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
375 timing side channel attack. An attacker could use variations in the signing
376 algorithm to recover the private key.
378 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
382 *) Timing vulnerability in ECDSA signature generation
384 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
385 timing side channel attack. An attacker could use variations in the signing
386 algorithm to recover the private key.
388 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
392 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
393 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
394 of two gigabytes and the error handling improved.
396 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
397 categorized as a normal bug, not a security issue, because the DRBG reseeds
398 automatically and is fully functional even without additional randomness
399 provided by the application.
401 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
403 *) Add a new ClientHello callback. Provides a callback interface that gives
404 the application the ability to adjust the nascent SSL object at the
405 earliest stage of ClientHello processing, immediately after extensions have
406 been collected but before they have been processed. In particular, this
407 callback can adjust the supported TLS versions in response to the contents
411 *) Add SM2 base algorithm support.
414 *) s390x assembly pack: add (improved) hardware-support for the following
415 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
416 aes-cfb/cfb8, aes-ecb.
419 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
420 parameter is no longer accepted, as it leads to a corrupt table. NULL
421 pem_str is reserved for alias entries only.
424 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
425 step for prime curves. The new implementation is based on formulae from
426 differential addition-and-doubling in homogeneous projective coordinates
427 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
428 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
429 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
430 to work in projective coordinates.
431 [Billy Bob Brumley, Nicola Tuveri]
433 *) Change generating and checking of primes so that the error rate of not
434 being prime depends on the intended use based on the size of the input.
435 For larger primes this will result in more rounds of Miller-Rabin.
436 The maximal error rate for primes with more than 1080 bits is lowered
438 [Kurt Roeckx, Annie Yousar]
440 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
443 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
444 moving between systems, and to avoid confusion when a Windows build is
445 done with mingw vs with MSVC. For POSIX installs, there's still a
446 symlink or copy named 'tsget' to avoid that confusion as well.
449 *) Revert blinding in ECDSA sign and instead make problematic addition
450 length-invariant. Switch even to fixed-length Montgomery multiplication.
453 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
454 step for binary curves. The new implementation is based on formulae from
455 differential addition-and-doubling in mixed Lopez-Dahab projective
456 coordinates, modified to independently blind the operands.
457 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
459 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
460 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
461 EC_METHODs to implement their own specialized "ladder step", to take
462 advantage of more favorable coordinate systems or more efficient
463 differential addition-and-doubling algorithms.
464 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
466 *) Modified the random device based seed sources to keep the relevant
467 file descriptors open rather than reopening them on each access.
468 This allows such sources to operate in a chroot() jail without
469 the associated device nodes being available. This behaviour can be
470 controlled using RAND_keep_random_devices_open().
473 *) Numerous side-channel attack mitigations have been applied. This may have
474 performance impacts for some algorithms for the benefit of improved
475 security. Specific changes are noted in this change log by their respective
479 *) AIX shared library support overhaul. Switch to AIX "natural" way of
480 handling shared libraries, which means collecting shared objects of
481 different versions and bitnesses in one common archive. This allows to
482 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
483 doesn't affect the way 3rd party applications are linked, only how
484 multi-version installation is managed.
487 *) Make ec_group_do_inverse_ord() more robust and available to other
488 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
489 mitigations are applied to the fallback BN_mod_inverse().
490 When using this function rather than BN_mod_inverse() directly, new
491 EC cryptosystem implementations are then safer-by-default.
494 *) Add coordinate blinding for EC_POINT and implement projective
495 coordinate blinding for generic prime curves as a countermeasure to
496 chosen point SCA attacks.
497 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
499 *) Add blinding to ECDSA and DSA signatures to protect against side channel
500 attacks discovered by Keegan Ryan (NCC Group).
503 *) Enforce checking in the pkeyutl command line app to ensure that the input
504 length does not exceed the maximum supported digest length when performing
505 a sign, verify or verifyrecover operation.
508 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
509 I/O in combination with something like select() or poll() will hang. This
510 can be turned off again using SSL_CTX_clear_mode().
511 Many applications do not properly handle non-application data records, and
512 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
513 around the problems in those applications, but can also break some.
514 It's recommended to read the manpages about SSL_read(), SSL_write(),
515 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
516 SSL_CTX_set_read_ahead() again.
519 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
520 now allow empty (zero character) pass phrases.
523 *) Apply blinding to binary field modular inversion and remove patent
524 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
527 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
528 binary and prime elliptic curves.
531 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
532 constant time fixed point multiplication.
535 *) Revise elliptic curve scalar multiplication with timing attack
536 defenses: ec_wNAF_mul redirects to a constant time implementation
537 when computing fixed point and variable point multiplication (which
538 in OpenSSL are mostly used with secret scalars in keygen, sign,
539 ECDH derive operations).
540 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
543 *) Updated CONTRIBUTING
546 *) Updated DRBG / RAND to request nonce and additional low entropy
547 randomness from the system.
548 [Matthias St. Pierre]
550 *) Updated 'openssl rehash' to use OpenSSL consistent default.
553 *) Moved the load of the ssl_conf module to libcrypto, which helps
554 loading engines that libssl uses before libssl is initialised.
557 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
560 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
561 [Ingo Schwarze, Rich Salz]
563 *) Added output of accepting IP address and port for 'openssl s_server'
566 *) Added a new API for TLSv1.3 ciphersuites:
567 SSL_CTX_set_ciphersuites()
568 SSL_set_ciphersuites()
571 *) Memory allocation failures consistently add an error to the error
575 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
576 in libcrypto when run as setuid/setgid.
579 *) Load any config file by default when libssl is used.
582 *) Added new public header file <openssl/rand_drbg.h> and documentation
583 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
584 [Matthias St. Pierre]
586 *) QNX support removed (cannot find contributors to get their approval
587 for the license change).
590 *) TLSv1.3 replay protection for early data has been implemented. See the
591 SSL_read_early_data() man page for further details.
594 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
595 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
596 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
597 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
598 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
599 configuration has been separated out. See the ciphers man page or the
600 SSL_CTX_set_ciphersuites() man page for more information.
603 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
604 in responder mode now supports the new "-multi" option, which
605 spawns the specified number of child processes to handle OCSP
606 requests. The "-timeout" option now also limits the OCSP
607 responder's patience to wait to receive the full client request
608 on a newly accepted connection. Child processes are respawned
609 as needed, and the CA index file is automatically reloaded
610 when changed. This makes it possible to run the "ocsp" responder
611 as a long-running service, making the OpenSSL CA somewhat more
612 feature-complete. In this mode, most diagnostic messages logged
613 after entering the event loop are logged via syslog(3) rather than
617 *) Added support for X448 and Ed448. Heavily based on original work by
621 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
622 objects loaded. This adds the functions OSSL_STORE_expect() and
623 OSSL_STORE_find() as well as needed tools to construct searches and
624 get the search data out of them.
627 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
628 version of OpenSSL should review their configuration settings to ensure
629 that they are still appropriate for TLSv1.3. For further information see:
630 https://wiki.openssl.org/index.php/TLS1.3
633 *) Grand redesign of the OpenSSL random generator
635 The default RAND method now utilizes an AES-CTR DRBG according to
636 NIST standard SP 800-90Ar1. The new random generator is essentially
637 a port of the default random generator from the OpenSSL FIPS 2.0
638 object module. It is a hybrid deterministic random bit generator
639 using an AES-CTR bit stream and which seeds and reseeds itself
640 automatically using trusted system entropy sources.
642 Some of its new features are:
643 o Support for multiple DRBG instances with seed chaining.
644 o The default RAND method makes use of a DRBG.
645 o There is a public and private DRBG instance.
646 o The DRBG instances are fork-safe.
647 o Keep all global DRBG instances on the secure heap if it is enabled.
648 o The public and private DRBG instance are per thread for lock free
650 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
652 *) Changed Configure so it only says what it does and doesn't dump
653 so much data. Instead, ./configdata.pm should be used as a script
654 to display all sorts of configuration data.
657 *) Added processing of "make variables" to Configure.
660 *) Added SHA512/224 and SHA512/256 algorithm support.
663 *) The last traces of Netware support, first removed in 1.1.0, have
667 *) Get rid of Makefile.shared, and in the process, make the processing
668 of certain files (rc.obj, or the .def/.map/.opt files produced from
669 the ordinal files) more visible and hopefully easier to trace and
670 debug (or make silent).
673 *) Make it possible to have environment variable assignments as
674 arguments to config / Configure.
677 *) Add multi-prime RSA (RFC 8017) support.
680 *) Add SM3 implemented according to GB/T 32905-2016
681 [ Jack Lloyd <jack.lloyd@ribose.com>,
682 Ronald Tse <ronald.tse@ribose.com>,
683 Erick Borsboom <erick.borsboom@ribose.com> ]
685 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
686 as documented in RFC6066.
687 Based on a patch from Tomasz Moń
688 [Filipe Raimundo da Silva]
690 *) Add SM4 implemented according to GB/T 32907-2016.
691 [ Jack Lloyd <jack.lloyd@ribose.com>,
692 Ronald Tse <ronald.tse@ribose.com>,
693 Erick Borsboom <erick.borsboom@ribose.com> ]
695 *) Reimplement -newreq-nodes and ERR_error_string_n; the
696 original author does not agree with the license change.
699 *) Add ARIA AEAD TLS support.
702 *) Some macro definitions to support VS6 have been removed. Visual
703 Studio 6 has not worked since 1.1.0
706 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
707 without clearing the errors.
710 *) Add "atfork" functions. If building on a system that without
711 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
712 requirements. The RAND facility now uses/requires this.
718 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
719 not possible to disable entirely. However, it's still possible to
720 disable the console reading UI method, UI_OpenSSL() (use UI_null()
723 To disable, configure with 'no-ui-console'. 'no-ui' is still
724 possible to use as an alias. Check at compile time with the
725 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
726 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
729 *) Add a STORE module, which implements a uniform and URI based reader of
730 stores that can contain keys, certificates, CRLs and numerous other
731 objects. The main API is loosely based on a few stdio functions,
732 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
733 OSSL_STORE_error and OSSL_STORE_close.
734 The implementation uses backends called "loaders" to implement arbitrary
735 URI schemes. There is one built in "loader" for the 'file' scheme.
738 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
739 then adjusted to work on FreeBSD 8.4 as well.
740 Enable by configuring with 'enable-devcryptoeng'. This is done by default
741 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
744 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
745 util/mkerr.pl, which is adapted to allow those prefixes, leading to
746 error code calls like this:
748 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
750 With this change, we claim the namespaces OSSL and OPENSSL in a manner
751 that can be encoded in C. For the foreseeable future, this will only
753 [Richard Levitte and Tim Hudson]
755 *) Removed BSD cryptodev engine.
758 *) Add a build target 'build_all_generated', to build all generated files
759 and only that. This can be used to prepare everything that requires
760 things like perl for a system that lacks perl and then move everything
761 to that system and do the rest of the build there.
764 *) In the UI interface, make it possible to duplicate the user data. This
765 can be used by engines that need to retain the data for a longer time
766 than just the call where this user data is passed.
769 *) Ignore the '-named_curve auto' value for compatibility of applications
771 [Tomas Mraz <tmraz@fedoraproject.org>]
773 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
774 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
775 alerts across multiple records (some of which could be empty). In practice
776 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
777 prohibts this altogether and other libraries (BoringSSL, NSS) do not
778 support this at all. Supporting it adds significant complexity to the
779 record layer, and its removal is unlikely to cause inter-operability
783 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
784 with Z. These are meant to replace LONG and ZLONG and to be size safe.
785 The use of LONG and ZLONG is discouraged and scheduled for deprecation
789 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
790 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
791 [Richard Levitte, Andy Polyakov]
793 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
797 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
798 platform rather than 'mingw'.
801 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
802 success if they are asked to add an object which already exists
803 in the store. This change cascades to other functions which load
804 certificates and CRLs.
807 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
808 facilitate stack unwinding even from assembly subroutines.
811 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
812 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
815 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
816 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
817 which is the minimum version we support.
820 *) Certificate time validation (X509_cmp_time) enforces stricter
821 compliance with RFC 5280. Fractional seconds and timezone offsets
822 are no longer allowed.
825 *) Add support for ARIA
828 *) s_client will now send the Server Name Indication (SNI) extension by
829 default unless the new "-noservername" option is used. The server name is
830 based on the host provided to the "-connect" option unless overridden by
834 *) Add support for SipHash
837 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
838 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
839 prevent issues where no progress is being made and the peer continually
840 sends unrecognised record types, using up resources processing them.
843 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
844 using the algorithm defined in
845 https://www.akkadia.org/drepper/SHA-crypt.txt
848 *) Heartbeat support has been removed; the ABI is changed for now.
849 [Richard Levitte, Rich Salz]
851 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
854 *) The RSA "null" method, which was partially supported to avoid patent
855 issues, has been replaced to always returns NULL.
859 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
861 *) Client DoS due to large DH parameter
863 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
864 malicious server can send a very large prime value to the client. This will
865 cause the client to spend an unreasonably long period of time generating a
866 key for this prime resulting in a hang until the client has finished. This
867 could be exploited in a Denial Of Service attack.
869 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
873 *) Cache timing vulnerability in RSA Key Generation
875 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
876 a cache timing side channel attack. An attacker with sufficient access to
877 mount cache timing attacks during the RSA key generation process could
878 recover the private key.
880 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
881 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
885 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
886 parameter is no longer accepted, as it leads to a corrupt table. NULL
887 pem_str is reserved for alias entries only.
890 *) Revert blinding in ECDSA sign and instead make problematic addition
891 length-invariant. Switch even to fixed-length Montgomery multiplication.
894 *) Change generating and checking of primes so that the error rate of not
895 being prime depends on the intended use based on the size of the input.
896 For larger primes this will result in more rounds of Miller-Rabin.
897 The maximal error rate for primes with more than 1080 bits is lowered
899 [Kurt Roeckx, Annie Yousar]
901 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
904 *) Add blinding to ECDSA and DSA signatures to protect against side channel
905 attacks discovered by Keegan Ryan (NCC Group).
908 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
909 now allow empty (zero character) pass phrases.
912 *) Certificate time validation (X509_cmp_time) enforces stricter
913 compliance with RFC 5280. Fractional seconds and timezone offsets
914 are no longer allowed.
917 *) Fixed a text canonicalisation bug in CMS
919 Where a CMS detached signature is used with text content the text goes
920 through a canonicalisation process first prior to signing or verifying a
921 signature. This process strips trailing space at the end of lines, converts
922 line terminators to CRLF and removes additional trailing line terminators
923 at the end of a file. A bug in the canonicalisation process meant that
924 some characters, such as form-feed, were incorrectly treated as whitespace
925 and removed. This is contrary to the specification (RFC5485). This fix
926 could mean that detached text data signed with an earlier version of
927 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
928 signed with a fixed OpenSSL may fail to verify with an earlier version of
929 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
930 and use the "-binary" flag (for the "cms" command line application) or set
931 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
934 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
936 *) Constructed ASN.1 types with a recursive definition could exceed the stack
938 Constructed ASN.1 types with a recursive definition (such as can be found
939 in PKCS7) could eventually exceed the stack given malicious input with
940 excessive recursion. This could result in a Denial Of Service attack. There
941 are no such structures used within SSL/TLS that come from untrusted sources
942 so this is considered safe.
944 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
949 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
951 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
952 effectively reduced to only comparing the least significant bit of each
953 byte. This allows an attacker to forge messages that would be considered as
954 authenticated in an amount of tries lower than that guaranteed by the
955 security claims of the scheme. The module can only be compiled by the
956 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
958 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
963 *) Add a build target 'build_all_generated', to build all generated files
964 and only that. This can be used to prepare everything that requires
965 things like perl for a system that lacks perl and then move everything
966 to that system and do the rest of the build there.
969 *) Backport SSL_OP_NO_RENGOTIATION
971 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
972 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
973 changes this is no longer possible in 1.1.0. Therefore the new
974 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
975 1.1.0 to provide equivalent functionality.
977 Note that if an application built against 1.1.0h headers (or above) is run
978 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
979 accepted but nothing will happen, i.e. renegotiation will not be prevented.
982 *) Removed the OS390-Unix config target. It relied on a script that doesn't
986 *) rsaz_1024_mul_avx2 overflow bug on x86_64
988 There is an overflow bug in the AVX2 Montgomery multiplication procedure
989 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
990 Analysis suggests that attacks against RSA and DSA as a result of this
991 defect would be very difficult to perform and are not believed likely.
992 Attacks against DH1024 are considered just feasible, because most of the
993 work necessary to deduce information about a private key may be performed
994 offline. The amount of resources required for such an attack would be
995 significant. However, for an attack on TLS to be meaningful, the server
996 would have to share the DH1024 private key among multiple clients, which is
997 no longer an option since CVE-2016-0701.
999 This only affects processors that support the AVX2 but not ADX extensions
1000 like Intel Haswell (4th generation).
1002 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1003 was originally found via the OSS-Fuzz project.
1007 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
1009 *) bn_sqrx8x_internal carry bug on x86_64
1011 There is a carry propagating bug in the x86_64 Montgomery squaring
1012 procedure. No EC algorithms are affected. Analysis suggests that attacks
1013 against RSA and DSA as a result of this defect would be very difficult to
1014 perform and are not believed likely. Attacks against DH are considered just
1015 feasible (although very difficult) because most of the work necessary to
1016 deduce information about a private key may be performed offline. The amount
1017 of resources required for such an attack would be very significant and
1018 likely only accessible to a limited number of attackers. An attacker would
1019 additionally need online access to an unpatched system using the target
1020 private key in a scenario with persistent DH parameters and a private
1021 key that is shared between multiple clients.
1023 This only affects processors that support the BMI1, BMI2 and ADX extensions
1024 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1026 This issue was reported to OpenSSL by the OSS-Fuzz project.
1030 *) Malformed X.509 IPAddressFamily could cause OOB read
1032 If an X.509 certificate has a malformed IPAddressFamily extension,
1033 OpenSSL could do a one-byte buffer overread. The most likely result
1034 would be an erroneous display of the certificate in text format.
1036 This issue was reported to OpenSSL by the OSS-Fuzz project.
1040 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1042 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1043 platform rather than 'mingw'.
1046 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1047 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1048 which is the minimum version we support.
1051 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1053 *) Encrypt-Then-Mac renegotiation crash
1055 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1056 negotiated where it was not in the original handshake (or vice-versa) then
1057 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1058 and servers are affected.
1060 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1064 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1066 *) Truncated packet could crash via OOB read
1068 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1069 cipher is being used, then a truncated packet can cause that host to
1070 perform an out-of-bounds read, usually resulting in a crash.
1072 This issue was reported to OpenSSL by Robert Święcki of Google.
1076 *) Bad (EC)DHE parameters cause a client crash
1078 If a malicious server supplies bad parameters for a DHE or ECDHE key
1079 exchange then this can result in the client attempting to dereference a
1080 NULL pointer leading to a client crash. This could be exploited in a Denial
1083 This issue was reported to OpenSSL by Guido Vranken.
1087 *) BN_mod_exp may produce incorrect results on x86_64
1089 There is a carry propagating bug in the x86_64 Montgomery squaring
1090 procedure. No EC algorithms are affected. Analysis suggests that attacks
1091 against RSA and DSA as a result of this defect would be very difficult to
1092 perform and are not believed likely. Attacks against DH are considered just
1093 feasible (although very difficult) because most of the work necessary to
1094 deduce information about a private key may be performed offline. The amount
1095 of resources required for such an attack would be very significant and
1096 likely only accessible to a limited number of attackers. An attacker would
1097 additionally need online access to an unpatched system using the target
1098 private key in a scenario with persistent DH parameters and a private
1099 key that is shared between multiple clients. For example this can occur by
1100 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1101 similar to CVE-2015-3193 but must be treated as a separate problem.
1103 This issue was reported to OpenSSL by the OSS-Fuzz project.
1107 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1109 *) ChaCha20/Poly1305 heap-buffer-overflow
1111 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1112 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1113 crash. This issue is not considered to be exploitable beyond a DoS.
1115 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1119 *) CMS Null dereference
1121 Applications parsing invalid CMS structures can crash with a NULL pointer
1122 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1123 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1124 structure callback if an attempt is made to free certain invalid encodings.
1125 Only CHOICE structures using a callback which do not handle NULL value are
1128 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1132 *) Montgomery multiplication may produce incorrect results
1134 There is a carry propagating bug in the Broadwell-specific Montgomery
1135 multiplication procedure that handles input lengths divisible by, but
1136 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1137 and DH private keys are impossible. This is because the subroutine in
1138 question is not used in operations with the private key itself and an input
1139 of the attacker's direct choice. Otherwise the bug can manifest itself as
1140 transient authentication and key negotiation failures or reproducible
1141 erroneous outcome of public-key operations with specially crafted input.
1142 Among EC algorithms only Brainpool P-512 curves are affected and one
1143 presumably can attack ECDH key negotiation. Impact was not analyzed in
1144 detail, because pre-requisites for attack are considered unlikely. Namely
1145 multiple clients have to choose the curve in question and the server has to
1146 share the private key among them, neither of which is default behaviour.
1147 Even then only clients that chose the curve will be affected.
1149 This issue was publicly reported as transient failures and was not
1150 initially recognized as a security issue. Thanks to Richard Morgan for
1151 providing reproducible case.
1155 *) Removed automatic addition of RPATH in shared libraries and executables,
1156 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1159 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1161 *) Fix Use After Free for large message sizes
1163 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1164 message larger than approx 16k is received then the underlying buffer to
1165 store the incoming message is reallocated and moved. Unfortunately a
1166 dangling pointer to the old location is left which results in an attempt to
1167 write to the previously freed location. This is likely to result in a
1168 crash, however it could potentially lead to execution of arbitrary code.
1170 This issue only affects OpenSSL 1.1.0a.
1172 This issue was reported to OpenSSL by Robert Święcki.
1176 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1178 *) OCSP Status Request extension unbounded memory growth
1180 A malicious client can send an excessively large OCSP Status Request
1181 extension. If that client continually requests renegotiation, sending a
1182 large OCSP Status Request extension each time, then there will be unbounded
1183 memory growth on the server. This will eventually lead to a Denial Of
1184 Service attack through memory exhaustion. Servers with a default
1185 configuration are vulnerable even if they do not support OCSP. Builds using
1186 the "no-ocsp" build time option are not affected.
1188 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1192 *) SSL_peek() hang on empty record
1194 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1195 sends an empty record. This could be exploited by a malicious peer in a
1196 Denial Of Service attack.
1198 This issue was reported to OpenSSL by Alex Gaynor.
1202 *) Excessive allocation of memory in tls_get_message_header() and
1203 dtls1_preprocess_fragment()
1205 A (D)TLS message includes 3 bytes for its length in the header for the
1206 message. This would allow for messages up to 16Mb in length. Messages of
1207 this length are excessive and OpenSSL includes a check to ensure that a
1208 peer is sending reasonably sized messages in order to avoid too much memory
1209 being consumed to service a connection. A flaw in the logic of version
1210 1.1.0 means that memory for the message is allocated too early, prior to
1211 the excessive message length check. Due to way memory is allocated in
1212 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1213 to service a connection. This could lead to a Denial of Service through
1214 memory exhaustion. However, the excessive message length check still takes
1215 place, and this would cause the connection to immediately fail. Assuming
1216 that the application calls SSL_free() on the failed connection in a timely
1217 manner then the 21Mb of allocated memory will then be immediately freed
1218 again. Therefore the excessive memory allocation will be transitory in
1219 nature. This then means that there is only a security impact if:
1221 1) The application does not call SSL_free() in a timely manner in the event
1222 that the connection fails
1224 2) The application is working in a constrained environment where there is
1225 very little free memory
1227 3) The attacker initiates multiple connection attempts such that there are
1228 multiple connections in a state where memory has been allocated for the
1229 connection; SSL_free() has not yet been called; and there is insufficient
1230 memory to service the multiple requests.
1232 Except in the instance of (1) above any Denial Of Service is likely to be
1233 transitory because as soon as the connection fails the memory is
1234 subsequently freed again in the SSL_free() call. However there is an
1235 increased risk during this period of application crashes due to the lack of
1236 memory - which would then mean a more serious Denial of Service.
1238 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1239 (CVE-2016-6307 and CVE-2016-6308)
1242 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1243 had to be removed. Primary reason is that vendor assembler can't
1244 assemble our modules with -KPIC flag. As result it, assembly
1245 support, was not even available as option. But its lack means
1246 lack of side-channel resistant code, which is incompatible with
1247 security by todays standards. Fortunately gcc is readily available
1248 prepackaged option, which we firmly point at...
1251 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1253 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1254 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1255 (to any value) allows Windows user to access PKCS#12 file generated
1256 with Windows CryptoAPI and protected with non-ASCII password, as well
1257 as files generated under UTF-8 locale on Linux also protected with
1261 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1262 have been disabled by default and removed from DEFAULT, just like RC4.
1263 See the RC4 item below to re-enable both.
1266 *) The method for finding the storage location for the Windows RAND seed file
1267 has changed. First we check %RANDFILE%. If that is not set then we check
1268 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1269 all else fails we fall back to C:\.
1272 *) The EVP_EncryptUpdate() function has had its return type changed from void
1273 to int. A return of 0 indicates and error while a return of 1 indicates
1277 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1278 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1279 off the constant time implementation for RSA, DSA and DH have been made
1280 no-ops and deprecated.
1283 *) Windows RAND implementation was simplified to only get entropy by
1284 calling CryptGenRandom(). Various other RAND-related tickets
1286 [Joseph Wylie Yandle, Rich Salz]
1288 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1289 and OPENSSL_LH_, respectively. The old names are available
1290 with API compatibility. They new names are now completely documented.
1293 *) Unify TYPE_up_ref(obj) methods signature.
1294 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1295 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1296 int (instead of void) like all others TYPE_up_ref() methods.
1297 So now these methods also check the return value of CRYPTO_atomic_add(),
1298 and the validity of object reference counter.
1299 [fdasilvayy@gmail.com]
1301 *) With Windows Visual Studio builds, the .pdb files are installed
1302 alongside the installed libraries and executables. For a static
1303 library installation, ossl_static.pdb is the associate compiler
1304 generated .pdb file to be used when linking programs.
1307 *) Remove openssl.spec. Packaging files belong with the packagers.
1310 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1311 recognise x86_64 architectures automatically. You can still decide
1312 to build for a different bitness with the environment variable
1313 KERNEL_BITS (can be 32 or 64), for example:
1315 KERNEL_BITS=32 ./config
1319 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1320 256 bit AES and HMAC with SHA256.
1323 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1326 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1329 *) To enable users to have their own config files and build file templates,
1330 Configure looks in the directory indicated by the environment variable
1331 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1332 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1333 name and is used as is.
1336 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1337 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1338 X509_CERT_FILE_CTX was removed.
1341 *) "shared" builds are now the default. To create only static libraries use
1342 the "no-shared" Configure option.
1345 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1346 All of these option have not worked for some while and are fundamental
1350 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1351 global cleanup functions are no longer required because they are handled
1352 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1353 Explicitly de-initing can cause problems (e.g. where a library that uses
1354 OpenSSL de-inits, but an application is still using it). The affected
1355 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1356 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1357 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1358 COMP_zlib_cleanup().
1361 *) --strict-warnings no longer enables runtime debugging options
1362 such as REF_DEBUG. Instead, debug options are automatically
1363 enabled with '--debug' builds.
1364 [Andy Polyakov, Emilia Käsper]
1366 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1367 have been moved out of the public header files. New functions for managing
1368 these have been added.
1371 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1372 objects have been moved out of the public header files. New
1373 functions for managing these have been added.
1376 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1377 have been moved out of the public header files. New functions for managing
1378 these have been added.
1381 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1382 moved out of the public header files. New functions for managing these
1386 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1389 *) Removed the mk1mf build scripts.
1392 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1393 it is always safe to #include a header now.
1396 *) Removed the aged BC-32 config and all its supporting scripts
1399 *) Removed support for Ultrix, Netware, and OS/2.
1402 *) Add support for HKDF.
1403 [Alessandro Ghedini]
1405 *) Add support for blake2b and blake2s
1408 *) Added support for "pipelining". Ciphers that have the
1409 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1410 encryptions/decryptions simultaneously. There are currently no built-in
1411 ciphers with this property but the expectation is that engines will be able
1412 to offer it to significantly improve throughput. Support has been extended
1413 into libssl so that multiple records for a single connection can be
1414 processed in one go (for >=TLS 1.1).
1417 *) Added the AFALG engine. This is an async capable engine which is able to
1418 offload work to the Linux kernel. In this initial version it only supports
1419 AES128-CBC. The kernel must be version 4.1.0 or greater.
1422 *) OpenSSL now uses a new threading API. It is no longer necessary to
1423 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1424 are two supported threading models: pthreads and windows threads. It is
1425 also possible to configure OpenSSL at compile time for "no-threads". The
1426 old threading API should no longer be used. The functions have been
1427 replaced with "no-op" compatibility macros.
1428 [Alessandro Ghedini, Matt Caswell]
1430 *) Modify behavior of ALPN to invoke callback after SNI/servername
1431 callback, such that updates to the SSL_CTX affect ALPN.
1434 *) Add SSL_CIPHER queries for authentication and key-exchange.
1437 *) Changes to the DEFAULT cipherlist:
1438 - Prefer (EC)DHE handshakes over plain RSA.
1439 - Prefer AEAD ciphers over legacy ciphers.
1440 - Prefer ECDSA over RSA when both certificates are available.
1441 - Prefer TLSv1.2 ciphers/PRF.
1442 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1446 *) Change the ECC default curve list to be this, in order: x25519,
1447 secp256r1, secp521r1, secp384r1.
1450 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1451 disabled by default. They can be re-enabled using the
1452 enable-weak-ssl-ciphers option to Configure.
1455 *) If the server has ALPN configured, but supports no protocols that the
1456 client advertises, send a fatal "no_application_protocol" alert.
1457 This behaviour is SHALL in RFC 7301, though it isn't universally
1458 implemented by other servers.
1461 *) Add X25519 support.
1462 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1463 for public and private key encoding using the format documented in
1464 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1465 key generation and key derivation.
1467 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1471 *) Deprecate SRP_VBASE_get_by_user.
1472 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1473 In order to fix an unavoidable memory leak (CVE-2016-0798),
1474 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1475 seed, even if the seed is configured.
1477 Users should use SRP_VBASE_get1_by_user instead. Note that in
1478 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1479 also that even though configuring the SRP seed attempts to hide
1480 invalid usernames by continuing the handshake with fake
1481 credentials, this behaviour is not constant time and no strong
1482 guarantees are made that the handshake is indistinguishable from
1483 that of a valid user.
1486 *) Configuration change; it's now possible to build dynamic engines
1487 without having to build shared libraries and vice versa. This
1488 only applies to the engines in engines/, those in crypto/engine/
1489 will always be built into libcrypto (i.e. "static").
1491 Building dynamic engines is enabled by default; to disable, use
1492 the configuration option "disable-dynamic-engine".
1494 The only requirements for building dynamic engines are the
1495 presence of the DSO module and building with position independent
1496 code, so they will also automatically be disabled if configuring
1497 with "disable-dso" or "disable-pic".
1499 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1500 are also taken away from openssl/opensslconf.h, as they are
1504 *) Configuration change; if there is a known flag to compile
1505 position independent code, it will always be applied on the
1506 libcrypto and libssl object files, and never on the application
1507 object files. This means other libraries that use routines from
1508 libcrypto / libssl can be made into shared libraries regardless
1509 of how OpenSSL was configured.
1511 If this isn't desirable, the configuration options "disable-pic"
1512 or "no-pic" can be used to disable the use of PIC. This will
1513 also disable building shared libraries and dynamic engines.
1516 *) Removed JPAKE code. It was experimental and has no wide use.
1519 *) The INSTALL_PREFIX Makefile variable has been renamed to
1520 DESTDIR. That makes for less confusion on what this variable
1521 is for. Also, the configuration option --install_prefix is
1525 *) Heartbeat for TLS has been removed and is disabled by default
1526 for DTLS; configure with enable-heartbeats. Code that uses the
1527 old #define's might need to be updated.
1528 [Emilia Käsper, Rich Salz]
1530 *) Rename REF_CHECK to REF_DEBUG.
1533 *) New "unified" build system
1535 The "unified" build system is aimed to be a common system for all
1536 platforms we support. With it comes new support for VMS.
1538 This system builds supports building in a different directory tree
1539 than the source tree. It produces one Makefile (for unix family
1540 or lookalikes), or one descrip.mms (for VMS).
1542 The source of information to make the Makefile / descrip.mms is
1543 small files called 'build.info', holding the necessary
1544 information for each directory with source to compile, and a
1545 template in Configurations, like unix-Makefile.tmpl or
1548 With this change, the library names were also renamed on Windows
1549 and on VMS. They now have names that are closer to the standard
1550 on Unix, and include the major version number, and in certain
1551 cases, the architecture they are built for. See "Notes on shared
1552 libraries" in INSTALL.
1554 We rely heavily on the perl module Text::Template.
1557 *) Added support for auto-initialisation and de-initialisation of the library.
1558 OpenSSL no longer requires explicit init or deinit routines to be called,
1559 except in certain circumstances. See the OPENSSL_init_crypto() and
1560 OPENSSL_init_ssl() man pages for further information.
1563 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1564 "peer" argument is now expected to be a BIO_ADDR object.
1566 *) Rewrite of BIO networking library. The BIO library lacked consistent
1567 support of IPv6, and adding it required some more extensive
1568 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1569 which hold all types of addresses and chains of address information.
1570 It also introduces a new API, with functions like BIO_socket,
1571 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1572 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1573 have been adapted accordingly.
1576 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1580 *) CRIME protection: disable compression by default, even if OpenSSL is
1581 compiled with zlib enabled. Applications can still enable compression
1582 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1583 using the SSL_CONF library to configure compression.
1586 *) The signature of the session callback configured with
1587 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1588 was explicitly marked as 'const unsigned char*' instead of
1592 *) Always DPURIFY. Remove the use of uninitialized memory in the
1593 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1596 *) Removed many obsolete configuration items, including
1597 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1598 MD2_CHAR, MD2_INT, MD2_LONG
1600 IDEA_SHORT, IDEA_LONG
1601 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1602 [Rich Salz, with advice from Andy Polyakov]
1604 *) Many BN internals have been moved to an internal header file.
1605 [Rich Salz with help from Andy Polyakov]
1607 *) Configuration and writing out the results from it has changed.
1608 Files such as Makefile include/openssl/opensslconf.h and are now
1609 produced through general templates, such as Makefile.in and
1610 crypto/opensslconf.h.in and some help from the perl module
1613 Also, the center of configuration information is no longer
1614 Makefile. Instead, Configure produces a perl module in
1615 configdata.pm which holds most of the config data (in the hash
1616 table %config), the target data that comes from the target
1617 configuration in one of the Configurations/*.conf files (in
1621 *) To clarify their intended purposes, the Configure options
1622 --prefix and --openssldir change their semantics, and become more
1623 straightforward and less interdependent.
1625 --prefix shall be used exclusively to give the location INSTALLTOP
1626 where programs, scripts, libraries, include files and manuals are
1627 going to be installed. The default is now /usr/local.
1629 --openssldir shall be used exclusively to give the default
1630 location OPENSSLDIR where certificates, private keys, CRLs are
1631 managed. This is also where the default openssl.cnf gets
1633 If the directory given with this option is a relative path, the
1634 values of both the --prefix value and the --openssldir value will
1635 be combined to become OPENSSLDIR.
1636 The default for --openssldir is INSTALLTOP/ssl.
1638 Anyone who uses --openssldir to specify where OpenSSL is to be
1639 installed MUST change to use --prefix instead.
1642 *) The GOST engine was out of date and therefore it has been removed. An up
1643 to date GOST engine is now being maintained in an external repository.
1644 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1645 support for GOST ciphersuites (these are only activated if a GOST engine
1649 *) EGD is no longer supported by default; use enable-egd when
1651 [Ben Kaduk and Rich Salz]
1653 *) The distribution now has Makefile.in files, which are used to
1654 create Makefile's when Configure is run. *Configure must be run
1655 before trying to build now.*
1658 *) The return value for SSL_CIPHER_description() for error conditions
1662 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1664 Obtaining and performing DNSSEC validation of TLSA records is
1665 the application's responsibility. The application provides
1666 the TLSA records of its choice to OpenSSL, and these are then
1667 used to authenticate the peer.
1669 The TLSA records need not even come from DNS. They can, for
1670 example, be used to implement local end-entity certificate or
1671 trust-anchor "pinning", where the "pin" data takes the form
1672 of TLSA records, which can augment or replace verification
1673 based on the usual WebPKI public certification authorities.
1676 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1677 continues to support deprecated interfaces in default builds.
1678 However, applications are strongly advised to compile their
1679 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1680 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1681 or the 1.1.0 releases.
1683 In environments in which all applications have been ported to
1684 not use any deprecated interfaces OpenSSL's Configure script
1685 should be used with the --api=1.1.0 option to entirely remove
1686 support for the deprecated features from the library and
1687 unconditionally disable them in the installed headers.
1688 Essentially the same effect can be achieved with the "no-deprecated"
1689 argument to Configure, except that this will always restrict
1690 the build to just the latest API, rather than a fixed API
1693 As applications are ported to future revisions of the API,
1694 they should update their compile-time OPENSSL_API_COMPAT define
1695 accordingly, but in most cases should be able to continue to
1696 compile with later releases.
1698 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1699 0x10000000L and 0x00908000L, respectively. However those
1700 versions did not support the OPENSSL_API_COMPAT feature, and
1701 so applications are not typically tested for explicit support
1702 of just the undeprecated features of either release.
1705 *) Add support for setting the minimum and maximum supported protocol.
1706 It can bet set via the SSL_set_min_proto_version() and
1707 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1708 MaxProtocol. It's recommended to use the new APIs to disable
1709 protocols instead of disabling individual protocols using
1710 SSL_set_options() or SSL_CONF's Protocol. This change also
1711 removes support for disabling TLS 1.2 in the OpenSSL TLS
1712 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1715 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1718 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1719 and integrates ECDSA and ECDH functionality into EC. Implementations can
1720 now redirect key generation and no longer need to convert to or from
1723 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1724 include the ec.h header file instead.
1727 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1728 ciphers who are no longer supported and drops support the ephemeral RSA key
1729 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1732 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1733 opaque. For HMAC_CTX, the following constructors and destructors
1736 HMAC_CTX *HMAC_CTX_new(void);
1737 void HMAC_CTX_free(HMAC_CTX *ctx);
1739 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1740 destroy such methods has been added. See EVP_MD_meth_new(3) and
1741 EVP_CIPHER_meth_new(3) for documentation.
1744 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1745 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1746 EVP_MD_CTX_reset() should be called instead to reinitialise
1747 an already created structure.
1748 2) For consistency with the majority of our object creators and
1749 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1750 EVP_MD_CTX_(new|free). The old names are retained as macros
1751 for deprecated builds.
1754 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1755 cryptographic operations to be performed asynchronously as long as an
1756 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1757 further details. Libssl has also had this capability integrated with the
1758 introduction of the new mode SSL_MODE_ASYNC and associated error
1759 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1760 pages. This work was developed in partnership with Intel Corp.
1763 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1764 always enabled now. If you want to disable the support you should
1765 exclude it using the list of supported ciphers. This also means that the
1766 "-no_ecdhe" option has been removed from s_server.
1769 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1770 SSL_{CTX_}set1_curves() which can set a list.
1773 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1774 curve you want to support using SSL_{CTX_}set1_curves().
1777 *) State machine rewrite. The state machine code has been significantly
1778 refactored in order to remove much duplication of code and solve issues
1779 with the old code (see ssl/statem/README for further details). This change
1780 does have some associated API changes. Notably the SSL_state() function
1781 has been removed and replaced by SSL_get_state which now returns an
1782 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1783 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1787 *) All instances of the string "ssleay" in the public API were replaced
1788 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1789 Some error codes related to internal RSA_eay API's were renamed.
1792 *) The demo files in crypto/threads were moved to demo/threads.
1795 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1797 [Matt Caswell, Rich Salz]
1799 *) New ASN.1 embed macro.
1801 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1802 structure is not allocated: it is part of the parent. That is instead of
1810 This reduces memory fragmentation and make it impossible to accidentally
1811 set a mandatory field to NULL.
1813 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1814 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1815 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1819 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1822 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1823 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1824 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1825 DES and RC4 ciphersuites.
1828 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1829 This changes the decoding behaviour for some invalid messages,
1830 though the change is mostly in the more lenient direction, and
1831 legacy behaviour is preserved as much as possible.
1834 *) Fix no-stdio build.
1835 [ David Woodhouse <David.Woodhouse@intel.com> and also
1836 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1838 *) New testing framework
1839 The testing framework has been largely rewritten and is now using
1840 perl and the perl modules Test::Harness and an extended variant of
1841 Test::More called OpenSSL::Test to do its work. All test scripts in
1842 test/ have been rewritten into test recipes, and all direct calls to
1843 executables in test/Makefile have become individual recipes using the
1844 simplified testing OpenSSL::Test::Simple.
1846 For documentation on our testing modules, do:
1848 perldoc test/testlib/OpenSSL/Test/Simple.pm
1849 perldoc test/testlib/OpenSSL/Test.pm
1853 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1854 are used; the latter aborts on memory leaks (usually checked on exit).
1855 Some undocumented "set malloc, etc., hooks" functions were removed
1856 and others were changed. All are now documented.
1859 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1861 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1863 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1864 from RFC4279, RFC4785, RFC5487, RFC5489.
1866 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1867 original RSA_PSK patch.
1870 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1871 era flag was never set throughout the codebase (only read). Also removed
1872 SSL3_FLAGS_POP_BUFFER which was only used if
1873 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1876 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1877 to be "oneline" instead of "compat".
1880 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1881 not aware of clients that still exhibit this bug, and the workaround
1882 hasn't been working properly for a while.
1885 *) The return type of BIO_number_read() and BIO_number_written() as well as
1886 the corresponding num_read and num_write members in the BIO structure has
1887 changed from unsigned long to uint64_t. On platforms where an unsigned
1888 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1892 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1893 OpenSSL without support for them. It also means that maintaining
1894 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1895 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1898 *) Removed support for the two export grade static DH ciphersuites
1899 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1900 were newly added (along with a number of other static DH ciphersuites) to
1901 1.0.2. However the two export ones have *never* worked since they were
1902 introduced. It seems strange in any case to be adding new export
1903 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1906 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1907 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1908 and turned into macros which simply call the new preferred function names
1909 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1910 should use the new names instead. Also as part of this change the ssl23.h
1911 header file has been removed.
1914 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1915 code and the associated standard is no longer considered fit-for-purpose.
1918 *) RT2547 was closed. When generating a private key, try to make the
1919 output file readable only by the owner. This behavior change might
1920 be noticeable when interacting with other software.
1922 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1926 *) Added HTTP GET support to the ocsp command.
1929 *) Changed default digest for the dgst and enc commands from MD5 to
1933 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1936 *) Added support for TLS extended master secret from
1937 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1938 initial patch which was a great help during development.
1941 *) All libssl internal structures have been removed from the public header
1942 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1943 now redundant). Users should not attempt to access internal structures
1944 directly. Instead they should use the provided API functions.
1947 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1948 Access to deprecated functions can be re-enabled by running config with
1949 "enable-deprecated". In addition applications wishing to use deprecated
1950 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1951 will, by default, disable some transitive includes that previously existed
1952 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1955 *) Added support for OCB mode. OpenSSL has been granted a patent license
1956 compatible with the OpenSSL license for use of OCB. Details are available
1957 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1958 for OCB can be removed by calling config with no-ocb.
1961 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1962 compatible client hello.
1965 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1966 done while fixing the error code for the key-too-small case.
1967 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1969 *) CA.sh has been removed; use CA.pl instead.
1972 *) Removed old DES API.
1975 *) Remove various unsupported platforms:
1981 Sinix/ReliantUNIX RM400
1986 16-bit platforms such as WIN16
1989 *) Clean up OPENSSL_NO_xxx #define's
1990 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1991 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1992 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1993 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1994 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1995 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1996 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1997 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1998 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1999 Remove MS_STATIC; it's a relic from platforms <32 bits.
2002 *) Cleaned up dead code
2003 Remove all but one '#ifdef undef' which is to be looked at.
2006 *) Clean up calling of xxx_free routines.
2007 Just like free(), fix most of the xxx_free routines to accept
2008 NULL. Remove the non-null checks from callers. Save much code.
2011 *) Add secure heap for storage of private keys (when possible).
2012 Add BIO_s_secmem(), CBIGNUM, etc.
2013 Contributed by Akamai Technologies under our Corporate CLA.
2016 *) Experimental support for a new, fast, unbiased prime candidate generator,
2017 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2018 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
2020 *) New output format NSS in the sess_id command line tool. This allows
2021 exporting the session id and the master key in NSS keylog format.
2022 [Martin Kaiser <martin@kaiser.cx>]
2024 *) Harmonize version and its documentation. -f flag is used to display
2026 [mancha <mancha1@zoho.com>]
2028 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2029 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2030 [mancha <mancha1@zoho.com>]
2032 *) Fix some double frees. These are not thought to be exploitable.
2033 [mancha <mancha1@zoho.com>]
2035 *) A missing bounds check in the handling of the TLS heartbeat extension
2036 can be used to reveal up to 64k of memory to a connected client or
2039 Thanks for Neel Mehta of Google Security for discovering this bug and to
2040 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2041 preparing the fix (CVE-2014-0160)
2042 [Adam Langley, Bodo Moeller]
2044 *) Fix for the attack described in the paper "Recovering OpenSSL
2045 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2046 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2047 http://eprint.iacr.org/2014/140
2049 Thanks to Yuval Yarom and Naomi Benger for discovering this
2050 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2051 [Yuval Yarom and Naomi Benger]
2053 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2054 this fixes a limitation in previous versions of OpenSSL.
2057 *) Experimental encrypt-then-mac support.
2059 Experimental support for encrypt then mac from
2060 draft-gutmann-tls-encrypt-then-mac-02.txt
2062 To enable it set the appropriate extension number (0x42 for the test
2063 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2065 For non-compliant peers (i.e. just about everything) this should have no
2068 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2072 *) Add EVP support for key wrapping algorithms, to avoid problems with
2073 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2074 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2075 algorithms and include tests cases.
2078 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2082 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2083 MGF1 digest and OAEP label.
2086 *) Make openssl verify return errors.
2087 [Chris Palmer <palmer@google.com> and Ben Laurie]
2089 *) New function ASN1_TIME_diff to calculate the difference between two
2090 ASN1_TIME structures or one structure and the current time.
2093 *) Update fips_test_suite to support multiple command line options. New
2094 test to induce all self test errors in sequence and check expected
2098 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2099 sign or verify all in one operation.
2102 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2103 test programs and fips_test_suite. Includes functionality to parse
2104 the minimal script output of fipsalgest.pl directly.
2107 *) Add authorisation parameter to FIPS_module_mode_set().
2110 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2113 *) Use separate DRBG fields for internal and external flags. New function
2114 FIPS_drbg_health_check() to perform on demand health checking. Add
2115 generation tests to fips_test_suite with reduced health check interval to
2116 demonstrate periodic health checking. Add "nodh" option to
2117 fips_test_suite to skip very slow DH test.
2120 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2124 *) More extensive health check for DRBG checking many more failure modes.
2125 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2126 combination: call this in fips_test_suite.
2129 *) Add support for canonical generation of DSA parameter 'g'. See
2132 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2133 POST to handle HMAC cases.
2136 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2137 to return numerical and string versions of the FIPS module number.
2140 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2141 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2142 outside the validated module in the FIPS capable OpenSSL.
2145 *) Minor change to DRBG entropy callback semantics. In some cases
2146 there is no multiple of the block length between min_len and
2147 max_len. Allow the callback to return more than max_len bytes
2148 of entropy but discard any extra: it is the callback's responsibility
2149 to ensure that the extra data discarded does not impact the
2150 requested amount of entropy.
2153 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2154 information in FIPS186-3, SP800-57 and SP800-131A.
2157 *) CCM support via EVP. Interface is very similar to GCM case except we
2158 must supply all data in one chunk (i.e. no update, final) and the
2159 message length must be supplied if AAD is used. Add algorithm test
2163 *) Initial version of POST overhaul. Add POST callback to allow the status
2164 of POST to be monitored and/or failures induced. Modify fips_test_suite
2165 to use callback. Always run all selftests even if one fails.
2168 *) XTS support including algorithm test driver in the fips_gcmtest program.
2169 Note: this does increase the maximum key length from 32 to 64 bytes but
2170 there should be no binary compatibility issues as existing applications
2171 will never use XTS mode.
2174 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2175 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2176 performs algorithm blocking for unapproved PRNG types. Also do not
2177 set PRNG type in FIPS_mode_set(): leave this to the application.
2178 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2179 the standard OpenSSL PRNG: set additional data to a date time vector.
2182 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2183 This shouldn't present any incompatibility problems because applications
2184 shouldn't be using these directly and any that are will need to rethink
2185 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2188 *) Extensive self tests and health checking required by SP800-90 DRBG.
2189 Remove strength parameter from FIPS_drbg_instantiate and always
2190 instantiate at maximum supported strength.
2193 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2196 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2199 *) New function DH_compute_key_padded() to compute a DH key and pad with
2200 leading zeroes if needed: this complies with SP800-56A et al.
2203 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2204 anything, incomplete, subject to change and largely untested at present.
2207 *) Modify fipscanisteronly build option to only build the necessary object
2208 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2211 *) Add experimental option FIPSSYMS to give all symbols in
2212 fipscanister.o and FIPS or fips prefix. This will avoid
2213 conflicts with future versions of OpenSSL. Add perl script
2214 util/fipsas.pl to preprocess assembly language source files
2215 and rename any affected symbols.
2218 *) Add selftest checks and algorithm block of non-fips algorithms in
2219 FIPS mode. Remove DES2 from selftests.
2222 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2223 return internal method without any ENGINE dependencies. Add new
2224 tiny fips sign and verify functions.
2227 *) New build option no-ec2m to disable characteristic 2 code.
2230 *) New build option "fipscanisteronly". This only builds fipscanister.o
2231 and (currently) associated fips utilities. Uses the file Makefile.fips
2232 instead of Makefile.org as the prototype.
2235 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2236 Update fips_gcmtest to use IV generator.
2239 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2240 setting output buffer to NULL. The *Final function must be
2241 called although it will not retrieve any additional data. The tag
2242 can be set or retrieved with a ctrl. The IV length is by default 12
2243 bytes (96 bits) but can be set to an alternative value. If the IV
2244 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2248 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2249 underlying do_cipher function handles all cipher semantics itself
2250 including padding and finalisation. This is useful if (for example)
2251 an ENGINE cipher handles block padding itself. The behaviour of
2252 do_cipher is subtly changed if this flag is set: the return value
2253 is the number of characters written to the output buffer (zero is
2254 no longer an error code) or a negative error code. Also if the
2255 input buffer is NULL and length 0 finalisation should be performed.
2258 *) If a candidate issuer certificate is already part of the constructed
2259 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2262 *) Improve forward-security support: add functions
2264 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2265 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2267 for use by SSL/TLS servers; the callback function will be called whenever a
2268 new session is created, and gets to decide whether the session may be
2269 cached to make it resumable (return 0) or not (return 1). (As by the
2270 SSL/TLS protocol specifications, the session_id sent by the server will be
2271 empty to indicate that the session is not resumable; also, the server will
2272 not generate RFC 4507 (RFC 5077) session tickets.)
2274 A simple reasonable callback implementation is to return is_forward_secure.
2275 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2276 by the SSL/TLS server library, indicating whether it can provide forward
2278 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2280 *) New -verify_name option in command line utilities to set verification
2284 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2285 Add CMAC pkey methods.
2288 *) Experimental renegotiation in s_server -www mode. If the client
2289 browses /reneg connection is renegotiated. If /renegcert it is
2290 renegotiated requesting a certificate.
2293 *) Add an "external" session cache for debugging purposes to s_server. This
2294 should help trace issues which normally are only apparent in deployed
2295 multi-process servers.
2298 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2299 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2300 BIO_set_cipher() and some obscure PEM functions were changed so they
2301 can now return an error. The RAND changes required a change to the
2302 RAND_METHOD structure.
2305 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2306 a gcc attribute to warn if the result of a function is ignored. This
2307 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2308 whose return value is often ignored.
2311 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2312 These allow SCTs (signed certificate timestamps) to be requested and
2313 validated when establishing a connection.
2314 [Rob Percival <robpercival@google.com>]
2316 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2318 *) Prevent padding oracle in AES-NI CBC MAC check
2320 A MITM attacker can use a padding oracle attack to decrypt traffic
2321 when the connection uses an AES CBC cipher and the server support
2324 This issue was introduced as part of the fix for Lucky 13 padding
2325 attack (CVE-2013-0169). The padding check was rewritten to be in
2326 constant time by making sure that always the same bytes are read and
2327 compared against either the MAC or padding bytes. But it no longer
2328 checked that there was enough data to have both the MAC and padding
2331 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2335 *) Fix EVP_EncodeUpdate overflow
2337 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2338 Base64 encoding of binary data. If an attacker is able to supply very large
2339 amounts of input data then a length check can overflow resulting in a heap
2342 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2343 the PEM_write_bio* family of functions. These are mainly used within the
2344 OpenSSL command line applications, so any application which processes data
2345 from an untrusted source and outputs it as a PEM file should be considered
2346 vulnerable to this issue. User applications that call these APIs directly
2347 with large amounts of untrusted data may also be vulnerable.
2349 This issue was reported by Guido Vranken.
2353 *) Fix EVP_EncryptUpdate overflow
2355 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2356 is able to supply very large amounts of input data after a previous call to
2357 EVP_EncryptUpdate() with a partial block then a length check can overflow
2358 resulting in a heap corruption. Following an analysis of all OpenSSL
2359 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2360 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2361 the first called function after an EVP_EncryptInit(), and therefore that
2362 specific call must be safe. The second form is where the length passed to
2363 EVP_EncryptUpdate() can be seen from the code to be some small value and
2364 therefore there is no possibility of an overflow. Since all instances are
2365 one of these two forms, it is believed that there can be no overflows in
2366 internal code due to this problem. It should be noted that
2367 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2368 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2369 of these calls have also been analysed too and it is believed there are no
2370 instances in internal usage where an overflow could occur.
2372 This issue was reported by Guido Vranken.
2376 *) Prevent ASN.1 BIO excessive memory allocation
2378 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2379 a short invalid encoding can cause allocation of large amounts of memory
2380 potentially consuming excessive resources or exhausting memory.
2382 Any application parsing untrusted data through d2i BIO functions is
2383 affected. The memory based functions such as d2i_X509() are *not* affected.
2384 Since the memory based functions are used by the TLS library, TLS
2385 applications are not affected.
2387 This issue was reported by Brian Carpenter.
2393 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2394 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2395 in arbitrary stack data being returned in the buffer.
2397 This issue was reported by Guido Vranken.
2401 *) Modify behavior of ALPN to invoke callback after SNI/servername
2402 callback, such that updates to the SSL_CTX affect ALPN.
2405 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2409 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2410 methods are enabled and ssl2 is disabled the methods return NULL.
2413 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2415 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2416 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2417 provide any "EXPORT" or "LOW" strength ciphers.
2420 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2421 is by default disabled at build-time. Builds that are not configured with
2422 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2423 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2424 will need to explicitly call either of:
2426 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2428 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2430 as appropriate. Even if either of those is used, or the application
2431 explicitly uses the version-specific SSLv2_method() or its client and
2432 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2433 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2434 ciphers, and SSLv2 56-bit DES are no longer available.
2438 *) Fix a double-free in DSA code
2440 A double free bug was discovered when OpenSSL parses malformed DSA private
2441 keys and could lead to a DoS attack or memory corruption for applications
2442 that receive DSA private keys from untrusted sources. This scenario is
2445 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2450 *) Disable SRP fake user seed to address a server memory leak.
2452 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2454 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2455 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2456 was changed to ignore the "fake user" SRP seed, even if the seed
2459 Users should use SRP_VBASE_get1_by_user instead. Note that in
2460 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2461 also that even though configuring the SRP seed attempts to hide
2462 invalid usernames by continuing the handshake with fake
2463 credentials, this behaviour is not constant time and no strong
2464 guarantees are made that the handshake is indistinguishable from
2465 that of a valid user.
2469 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2471 In the BN_hex2bn function the number of hex digits is calculated using an
2472 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2473 large values of |i| this can result in |bn_expand| not allocating any
2474 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2475 field as NULL leading to a subsequent NULL ptr deref. For very large values
2476 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2477 In this case memory is allocated to the internal BIGNUM data field, but it
2478 is insufficiently sized leading to heap corruption. A similar issue exists
2479 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2480 is ever called by user applications with very large untrusted hex/dec data.
2481 This is anticipated to be a rare occurrence.
2483 All OpenSSL internal usage of these functions use data that is not expected
2484 to be untrusted, e.g. config file data or application command line
2485 arguments. If user developed applications generate config file data based
2486 on untrusted data then it is possible that this could also lead to security
2487 consequences. This is also anticipated to be rare.
2489 This issue was reported to OpenSSL by Guido Vranken.
2493 *) Fix memory issues in BIO_*printf functions
2495 The internal |fmtstr| function used in processing a "%s" format string in
2496 the BIO_*printf functions could overflow while calculating the length of a
2497 string and cause an OOB read when printing very long strings.
2499 Additionally the internal |doapr_outch| function can attempt to write to an
2500 OOB memory location (at an offset from the NULL pointer) in the event of a
2501 memory allocation failure. In 1.0.2 and below this could be caused where
2502 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2503 could be in processing a very long "%s" format string. Memory leaks can
2506 The first issue may mask the second issue dependent on compiler behaviour.
2507 These problems could enable attacks where large amounts of untrusted data
2508 is passed to the BIO_*printf functions. If applications use these functions
2509 in this way then they could be vulnerable. OpenSSL itself uses these
2510 functions when printing out human-readable dumps of ASN.1 data. Therefore
2511 applications that print this data could be vulnerable if the data is from
2512 untrusted sources. OpenSSL command line applications could also be
2513 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2514 as command line arguments.
2516 Libssl is not considered directly vulnerable. Additionally certificates etc
2517 received via remote connections via libssl are also unlikely to be able to
2518 trigger these issues because of message size limits enforced within libssl.
2520 This issue was reported to OpenSSL Guido Vranken.
2524 *) Side channel attack on modular exponentiation
2526 A side-channel attack was found which makes use of cache-bank conflicts on
2527 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2528 of RSA keys. The ability to exploit this issue is limited as it relies on
2529 an attacker who has control of code in a thread running on the same
2530 hyper-threaded core as the victim thread which is performing decryptions.
2532 This issue was reported to OpenSSL by Yuval Yarom, The University of
2533 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2534 Nadia Heninger, University of Pennsylvania with more information at
2535 http://cachebleed.info.
2539 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2540 if no keysize is specified with default_bits. This fixes an
2541 omission in an earlier change that changed all RSA/DSA key generation
2542 apps to use 2048 bits by default.
2545 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2546 *) DH small subgroups
2548 Historically OpenSSL only ever generated DH parameters based on "safe"
2549 primes. More recently (in version 1.0.2) support was provided for
2550 generating X9.42 style parameter files such as those required for RFC 5114
2551 support. The primes used in such files may not be "safe". Where an
2552 application is using DH configured with parameters based on primes that are
2553 not "safe" then an attacker could use this fact to find a peer's private
2554 DH exponent. This attack requires that the attacker complete multiple
2555 handshakes in which the peer uses the same private DH exponent. For example
2556 this could be used to discover a TLS server's private DH exponent if it's
2557 reusing the private DH exponent or it's using a static DH ciphersuite.
2559 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2560 TLS. It is not on by default. If the option is not set then the server
2561 reuses the same private DH exponent for the life of the server process and
2562 would be vulnerable to this attack. It is believed that many popular
2563 applications do set this option and would therefore not be at risk.
2565 The fix for this issue adds an additional check where a "q" parameter is
2566 available (as is the case in X9.42 based parameters). This detects the
2567 only known attack, and is the only possible defense for static DH
2568 ciphersuites. This could have some performance impact.
2570 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2571 default and cannot be disabled. This could have some performance impact.
2573 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2577 *) SSLv2 doesn't block disabled ciphers
2579 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2580 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2581 been disabled, provided that the SSLv2 protocol was not also disabled via
2584 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2585 and Sebastian Schinzel.
2589 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2591 *) BN_mod_exp may produce incorrect results on x86_64
2593 There is a carry propagating bug in the x86_64 Montgomery squaring
2594 procedure. No EC algorithms are affected. Analysis suggests that attacks
2595 against RSA and DSA as a result of this defect would be very difficult to
2596 perform and are not believed likely. Attacks against DH are considered just
2597 feasible (although very difficult) because most of the work necessary to
2598 deduce information about a private key may be performed offline. The amount
2599 of resources required for such an attack would be very significant and
2600 likely only accessible to a limited number of attackers. An attacker would
2601 additionally need online access to an unpatched system using the target
2602 private key in a scenario with persistent DH parameters and a private
2603 key that is shared between multiple clients. For example this can occur by
2604 default in OpenSSL DHE based SSL/TLS ciphersuites.
2606 This issue was reported to OpenSSL by Hanno Böck.
2610 *) Certificate verify crash with missing PSS parameter
2612 The signature verification routines will crash with a NULL pointer
2613 dereference if presented with an ASN.1 signature using the RSA PSS
2614 algorithm and absent mask generation function parameter. Since these
2615 routines are used to verify certificate signature algorithms this can be
2616 used to crash any certificate verification operation and exploited in a
2617 DoS attack. Any application which performs certificate verification is
2618 vulnerable including OpenSSL clients and servers which enable client
2621 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2625 *) X509_ATTRIBUTE memory leak
2627 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2628 memory. This structure is used by the PKCS#7 and CMS routines so any
2629 application which reads PKCS#7 or CMS data from untrusted sources is
2630 affected. SSL/TLS is not affected.
2632 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2637 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2638 This changes the decoding behaviour for some invalid messages,
2639 though the change is mostly in the more lenient direction, and
2640 legacy behaviour is preserved as much as possible.
2643 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2645 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2647 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2649 *) Alternate chains certificate forgery
2651 During certificate verification, OpenSSL will attempt to find an
2652 alternative certificate chain if the first attempt to build such a chain
2653 fails. An error in the implementation of this logic can mean that an
2654 attacker could cause certain checks on untrusted certificates to be
2655 bypassed, such as the CA flag, enabling them to use a valid leaf
2656 certificate to act as a CA and "issue" an invalid certificate.
2658 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2662 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2664 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2665 incompatibility in the handling of HMAC. The previous ABI has now been
2669 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2671 *) Malformed ECParameters causes infinite loop
2673 When processing an ECParameters structure OpenSSL enters an infinite loop
2674 if the curve specified is over a specially malformed binary polynomial
2677 This can be used to perform denial of service against any
2678 system which processes public keys, certificate requests or
2679 certificates. This includes TLS clients and TLS servers with
2680 client authentication enabled.
2682 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2686 *) Exploitable out-of-bounds read in X509_cmp_time
2688 X509_cmp_time does not properly check the length of the ASN1_TIME
2689 string and can read a few bytes out of bounds. In addition,
2690 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2693 An attacker can use this to craft malformed certificates and CRLs of
2694 various sizes and potentially cause a segmentation fault, resulting in
2695 a DoS on applications that verify certificates or CRLs. TLS clients
2696 that verify CRLs are affected. TLS clients and servers with client
2697 authentication enabled may be affected if they use custom verification
2700 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2701 independently by Hanno Böck.
2705 *) PKCS7 crash with missing EnvelopedContent
2707 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2708 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2709 with missing content and trigger a NULL pointer dereference on parsing.
2711 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2712 structures from untrusted sources are affected. OpenSSL clients and
2713 servers are not affected.
2715 This issue was reported to OpenSSL by Michal Zalewski (Google).
2719 *) CMS verify infinite loop with unknown hash function
2721 When verifying a signedData message the CMS code can enter an infinite loop
2722 if presented with an unknown hash function OID. This can be used to perform
2723 denial of service against any system which verifies signedData messages using
2725 This issue was reported to OpenSSL by Johannes Bauer.
2729 *) Race condition handling NewSessionTicket
2731 If a NewSessionTicket is received by a multi-threaded client when attempting to
2732 reuse a previous ticket then a race condition can occur potentially leading to
2733 a double free of the ticket data.
2737 *) Only support 256-bit or stronger elliptic curves with the
2738 'ecdh_auto' setting (server) or by default (client). Of supported
2739 curves, prefer P-256 (both).
2742 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2744 *) ClientHello sigalgs DoS fix
2746 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2747 invalid signature algorithms extension a NULL pointer dereference will
2748 occur. This can be exploited in a DoS attack against the server.
2750 This issue was was reported to OpenSSL by David Ramos of Stanford
2753 [Stephen Henson and Matt Caswell]
2755 *) Multiblock corrupted pointer fix
2757 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2758 feature only applies on 64 bit x86 architecture platforms that support AES
2759 NI instructions. A defect in the implementation of "multiblock" can cause
2760 OpenSSL's internal write buffer to become incorrectly set to NULL when
2761 using non-blocking IO. Typically, when the user application is using a
2762 socket BIO for writing, this will only result in a failed connection.
2763 However if some other BIO is used then it is likely that a segmentation
2764 fault will be triggered, thus enabling a potential DoS attack.
2766 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2770 *) Segmentation fault in DTLSv1_listen fix
2772 The DTLSv1_listen function is intended to be stateless and processes the
2773 initial ClientHello from many peers. It is common for user code to loop
2774 over the call to DTLSv1_listen until a valid ClientHello is received with
2775 an associated cookie. A defect in the implementation of DTLSv1_listen means
2776 that state is preserved in the SSL object from one invocation to the next
2777 that can lead to a segmentation fault. Errors processing the initial
2778 ClientHello can trigger this scenario. An example of such an error could be
2779 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2782 This issue was reported to OpenSSL by Per Allansson.
2786 *) Segmentation fault in ASN1_TYPE_cmp fix
2788 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2789 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2790 certificate signature algorithm consistency this can be used to crash any
2791 certificate verification operation and exploited in a DoS attack. Any
2792 application which performs certificate verification is vulnerable including
2793 OpenSSL clients and servers which enable client authentication.
2797 *) Segmentation fault for invalid PSS parameters fix
2799 The signature verification routines will crash with a NULL pointer
2800 dereference if presented with an ASN.1 signature using the RSA PSS
2801 algorithm and invalid parameters. Since these routines are used to verify
2802 certificate signature algorithms this can be used to crash any
2803 certificate verification operation and exploited in a DoS attack. Any
2804 application which performs certificate verification is vulnerable including
2805 OpenSSL clients and servers which enable client authentication.
2807 This issue was was reported to OpenSSL by Brian Carpenter.
2811 *) ASN.1 structure reuse memory corruption fix
2813 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2814 memory corruption via an invalid write. Such reuse is and has been
2815 strongly discouraged and is believed to be rare.
2817 Applications that parse structures containing CHOICE or ANY DEFINED BY
2818 components may be affected. Certificate parsing (d2i_X509 and related
2819 functions) are however not affected. OpenSSL clients and servers are
2824 *) PKCS7 NULL pointer dereferences fix
2826 The PKCS#7 parsing code does not handle missing outer ContentInfo
2827 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2828 missing content and trigger a NULL pointer dereference on parsing.
2830 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2831 otherwise parse PKCS#7 structures from untrusted sources are
2832 affected. OpenSSL clients and servers are not affected.
2834 This issue was reported to OpenSSL by Michal Zalewski (Google).
2838 *) DoS via reachable assert in SSLv2 servers fix
2840 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2841 servers that both support SSLv2 and enable export cipher suites by sending
2842 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2844 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2845 (OpenSSL development team).
2849 *) Empty CKE with client auth and DHE fix
2851 If client auth is used then a server can seg fault in the event of a DHE
2852 ciphersuite being selected and a zero length ClientKeyExchange message
2853 being sent by the client. This could be exploited in a DoS attack.
2857 *) Handshake with unseeded PRNG fix
2859 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2860 with an unseeded PRNG. The conditions are:
2861 - The client is on a platform where the PRNG has not been seeded
2862 automatically, and the user has not seeded manually
2863 - A protocol specific client method version has been used (i.e. not
2864 SSL_client_methodv23)
2865 - A ciphersuite is used that does not require additional random data from
2866 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2868 If the handshake succeeds then the client random that has been used will
2869 have been generated from a PRNG with insufficient entropy and therefore the
2870 output may be predictable.
2872 For example using the following command with an unseeded openssl will
2873 succeed on an unpatched platform:
2875 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2879 *) Use After Free following d2i_ECPrivatekey error fix
2881 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2882 could cause a use after free condition. This, in turn, could cause a double
2883 free in several private key parsing functions (such as d2i_PrivateKey
2884 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2885 for applications that receive EC private keys from untrusted
2886 sources. This scenario is considered rare.
2888 This issue was discovered by the BoringSSL project and fixed in their
2893 *) X509_to_X509_REQ NULL pointer deref fix
2895 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2896 the certificate key is invalid. This function is rarely used in practice.
2898 This issue was discovered by Brian Carpenter.
2902 *) Removed the export ciphers from the DEFAULT ciphers
2905 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2907 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2908 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2909 So far those who have to target multiple platforms would compromise
2910 and argue that binary targeting say ARMv5 would still execute on
2911 ARMv8. "Universal" build resolves this compromise by providing
2912 near-optimal performance even on newer platforms.
2915 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2916 (other platforms pending).
2917 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2919 *) Add support for the SignedCertificateTimestampList certificate and
2920 OCSP response extensions from RFC6962.
2923 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2924 for corner cases. (Certain input points at infinity could lead to
2925 bogus results, with non-infinity inputs mapped to infinity too.)
2928 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2929 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2930 common cases are optimized and there still is room for further
2931 improvements. Vector Permutation AES for Altivec is also added.
2934 *) Add support for little-endian ppc64 Linux target.
2935 [Marcelo Cerri (IBM)]
2937 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2938 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2939 are optimized and there still is room for further improvements.
2940 Both 32- and 64-bit modes are supported.
2941 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2943 *) Improved ARMv7 NEON support.
2946 *) Support for SPARC Architecture 2011 crypto extensions, first
2947 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2948 SHA256/512, MD5, GHASH and modular exponentiation.
2949 [Andy Polyakov, David Miller]
2951 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2953 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2955 *) Support for new and upcoming Intel processors, including AVX2,
2956 BMI and SHA ISA extensions. This includes additional "stitched"
2957 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2960 This work was sponsored by Intel Corp.
2963 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2964 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2965 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2968 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2969 this fixes a limitation in previous versions of OpenSSL.
2972 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2973 MGF1 digest and OAEP label.
2976 *) Add EVP support for key wrapping algorithms, to avoid problems with
2977 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2978 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2979 algorithms and include tests cases.
2982 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2984 [Douglas E. Engert, Steve Henson]
2986 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2987 difference in days and seconds between two tm or ASN1_TIME structures.
2990 *) Add -rev test option to s_server to just reverse order of characters
2991 received by client and send back to server. Also prints an abbreviated
2992 summary of the connection parameters.
2995 *) New option -brief for s_client and s_server to print out a brief summary
2996 of connection parameters.
2999 *) Add callbacks for arbitrary TLS extensions.
3000 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
3002 *) New option -crl_download in several openssl utilities to download CRLs
3003 from CRLDP extension in certificates.
3006 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
3009 *) New function X509_CRL_diff to generate a delta CRL from the difference
3010 of two full CRLs. Add support to "crl" utility.
3013 *) New functions to set lookup_crls function and to retrieve
3014 X509_STORE from X509_STORE_CTX.
3017 *) Print out deprecated issuer and subject unique ID fields in
3021 *) Extend OCSP I/O functions so they can be used for simple general purpose
3022 HTTP as well as OCSP. New wrapper function which can be used to download
3023 CRLs using the OCSP API.
3026 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3029 *) SSL_CONF* functions. These provide a common framework for application
3030 configuration using configuration files or command lines.
3033 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
3034 message callback and prints the results. Needs compile time option
3035 "enable-ssl-trace". New options to s_client and s_server to enable
3039 *) New ctrl and macro to retrieve supported points extensions.
3040 Print out extension in s_server and s_client.
3043 *) New functions to retrieve certificate signature and signature
3047 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3051 *) New Suite B modes for TLS code. These use and enforce the requirements
3052 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3053 only use Suite B curves. The Suite B modes can be set by using the
3054 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3057 *) New chain verification flags for Suite B levels of security. Check
3058 algorithms are acceptable when flags are set in X509_verify_cert.
3061 *) Make tls1_check_chain return a set of flags indicating checks passed
3062 by a certificate chain. Add additional tests to handle client
3063 certificates: checks for matching certificate type and issuer name
3067 *) If an attempt is made to use a signature algorithm not in the peer
3068 preference list abort the handshake. If client has no suitable
3069 signature algorithms in response to a certificate request do not
3070 use the certificate.
3073 *) If server EC tmp key is not in client preference list abort handshake.
3076 *) Add support for certificate stores in CERT structure. This makes it
3077 possible to have different stores per SSL structure or one store in
3078 the parent SSL_CTX. Include distinct stores for certificate chain
3079 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3080 to build and store a certificate chain in CERT structure: returning
3081 an error if the chain cannot be built: this will allow applications
3082 to test if a chain is correctly configured.
3084 Note: if the CERT based stores are not set then the parent SSL_CTX
3085 store is used to retain compatibility with existing behaviour.
3089 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3090 mask based on the current session, check mask when sending client
3091 hello and checking the requested ciphersuite.
3094 *) New ctrls to retrieve and set certificate types in a certificate
3095 request message. Print out received values in s_client. If certificate
3096 types is not set with custom values set sensible values based on
3097 supported signature algorithms.
3100 *) Support for distinct client and server supported signature algorithms.
3103 *) Add certificate callback. If set this is called whenever a certificate
3104 is required by client or server. An application can decide which
3105 certificate chain to present based on arbitrary criteria: for example
3106 supported signature algorithms. Add very simple example to s_server.
3107 This fixes many of the problems and restrictions of the existing client
3108 certificate callback: for example you can now clear an existing
3109 certificate and specify the whole chain.
3112 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3113 the certificate can be used for (if anything). Set valid_flags field
3114 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3115 to have similar checks in it.
3117 Add new "cert_flags" field to CERT structure and include a "strict mode".
3118 This enforces some TLS certificate requirements (such as only permitting
3119 certificate signature algorithms contained in the supported algorithms
3120 extension) which some implementations ignore: this option should be used
3121 with caution as it could cause interoperability issues.
3124 *) Update and tidy signature algorithm extension processing. Work out
3125 shared signature algorithms based on preferences and peer algorithms
3126 and print them out in s_client and s_server. Abort handshake if no
3127 shared signature algorithms.
3130 *) Add new functions to allow customised supported signature algorithms
3131 for SSL and SSL_CTX structures. Add options to s_client and s_server
3135 *) New function SSL_certs_clear() to delete all references to certificates
3136 from an SSL structure. Before this once a certificate had been added
3137 it couldn't be removed.
3140 *) Integrate hostname, email address and IP address checking with certificate
3141 verification. New verify options supporting checking in openssl utility.
3144 *) Fixes and wildcard matching support to hostname and email checking
3145 functions. Add manual page.
3146 [Florian Weimer (Red Hat Product Security Team)]
3148 *) New functions to check a hostname email or IP address against a
3149 certificate. Add options x509 utility to print results of checks against
3153 *) Fix OCSP checking.
3154 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3156 *) Initial experimental support for explicitly trusted non-root CAs.
3157 OpenSSL still tries to build a complete chain to a root but if an
3158 intermediate CA has a trust setting included that is used. The first
3159 setting is used: whether to trust (e.g., -addtrust option to the x509
3163 *) Add -trusted_first option which attempts to find certificates in the
3164 trusted store even if an untrusted chain is also supplied.
3167 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3168 platform support for Linux and Android.
3171 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3174 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3175 When in FIPS mode the approved implementations are used as normal,
3176 when not in FIPS mode the internal unapproved versions are used instead.
3177 This means that the FIPS capable OpenSSL isn't forced to use the
3178 (often lower performance) FIPS implementations outside FIPS mode.
3181 *) Transparently support X9.42 DH parameters when calling
3182 PEM_read_bio_DHparameters. This means existing applications can handle
3183 the new parameter format automatically.
3186 *) Initial experimental support for X9.42 DH parameter format: mainly
3187 to support use of 'q' parameter for RFC5114 parameters.
3190 *) Add DH parameters from RFC5114 including test data to dhtest.
3193 *) Support for automatic EC temporary key parameter selection. If enabled
3194 the most preferred EC parameters are automatically used instead of
3195 hardcoded fixed parameters. Now a server just has to call:
3196 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3197 support ECDH and use the most appropriate parameters.
3200 *) Enhance and tidy EC curve and point format TLS extension code. Use
3201 static structures instead of allocation if default values are used.
3202 New ctrls to set curves we wish to support and to retrieve shared curves.
3203 Print out shared curves in s_server. New options to s_server and s_client
3204 to set list of supported curves.
3207 *) New ctrls to retrieve supported signature algorithms and
3208 supported curve values as an array of NIDs. Extend openssl utility
3209 to print out received values.
3212 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3213 between NIDs and the more common NIST names such as "P-256". Enhance
3214 ecparam utility and ECC method to recognise the NIST names for curves.
3217 *) Enhance SSL/TLS certificate chain handling to support different
3218 chains for each certificate instead of one chain in the parent SSL_CTX.
3221 *) Support for fixed DH ciphersuite client authentication: where both
3222 server and client use DH certificates with common parameters.
3225 *) Support for fixed DH ciphersuites: those requiring DH server
3229 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3231 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3232 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3233 X509_CINF_get_signature were reverted post internal team review.
3235 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3237 *) Build fixes for the Windows and OpenVMS platforms
3238 [Matt Caswell and Richard Levitte]
3240 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3242 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3243 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3244 dereference. This could lead to a Denial Of Service attack. Thanks to
3245 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3249 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3250 dtls1_buffer_record function under certain conditions. In particular this
3251 could occur if an attacker sent repeated DTLS records with the same
3252 sequence number but for the next epoch. The memory leak could be exploited
3253 by an attacker in a Denial of Service attack through memory exhaustion.
3254 Thanks to Chris Mueller for reporting this issue.
3258 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3259 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3260 method would be set to NULL which could later result in a NULL pointer
3261 dereference. Thanks to Frank Schmirler for reporting this issue.
3265 *) Abort handshake if server key exchange message is omitted for ephemeral
3268 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3269 reporting this issue.
3273 *) Remove non-export ephemeral RSA code on client and server. This code
3274 violated the TLS standard by allowing the use of temporary RSA keys in
3275 non-export ciphersuites and could be used by a server to effectively
3276 downgrade the RSA key length used to a value smaller than the server
3277 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3278 INRIA or reporting this issue.
3282 *) Fixed issue where DH client certificates are accepted without verification.
3283 An OpenSSL server will accept a DH certificate for client authentication
3284 without the certificate verify message. This effectively allows a client to
3285 authenticate without the use of a private key. This only affects servers
3286 which trust a client certificate authority which issues certificates
3287 containing DH keys: these are extremely rare and hardly ever encountered.
3288 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3293 *) Ensure that the session ID context of an SSL is updated when its
3294 SSL_CTX is updated via SSL_set_SSL_CTX.
3296 The session ID context is typically set from the parent SSL_CTX,
3297 and can vary with the CTX.
3300 *) Fix various certificate fingerprint issues.
3302 By using non-DER or invalid encodings outside the signed portion of a
3303 certificate the fingerprint can be changed without breaking the signature.
3304 Although no details of the signed portion of the certificate can be changed
3305 this can cause problems with some applications: e.g. those using the
3306 certificate fingerprint for blacklists.
3308 1. Reject signatures with non zero unused bits.
3310 If the BIT STRING containing the signature has non zero unused bits reject
3311 the signature. All current signature algorithms require zero unused bits.
3313 2. Check certificate algorithm consistency.
3315 Check the AlgorithmIdentifier inside TBS matches the one in the
3316 certificate signature. NB: this will result in signature failure
3317 errors for some broken certificates.
3319 Thanks to Konrad Kraszewski from Google for reporting this issue.
3321 3. Check DSA/ECDSA signatures use DER.
3323 Re-encode DSA/ECDSA signatures and compare with the original received
3324 signature. Return an error if there is a mismatch.
3326 This will reject various cases including garbage after signature
3327 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3328 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3329 (negative or with leading zeroes).
3331 Further analysis was conducted and fixes were developed by Stephen Henson
3332 of the OpenSSL core team.
3337 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3338 results on some platforms, including x86_64. This bug occurs at random
3339 with a very low probability, and is not known to be exploitable in any
3340 way, though its exact impact is difficult to determine. Thanks to Pieter
3341 Wuille (Blockstream) who reported this issue and also suggested an initial
3342 fix. Further analysis was conducted by the OpenSSL development team and
3343 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3344 the OpenSSL core team.
3348 *) Do not resume sessions on the server if the negotiated protocol
3349 version does not match the session's version. Resuming with a different
3350 version, while not strictly forbidden by the RFC, is of questionable
3351 sanity and breaks all known clients.
3352 [David Benjamin, Emilia Käsper]
3354 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3355 early CCS messages during renegotiation. (Note that because
3356 renegotiation is encrypted, this early CCS was not exploitable.)
3359 *) Tighten client-side session ticket handling during renegotiation:
3360 ensure that the client only accepts a session ticket if the server sends
3361 the extension anew in the ServerHello. Previously, a TLS client would
3362 reuse the old extension state and thus accept a session ticket if one was
3363 announced in the initial ServerHello.
3365 Similarly, ensure that the client requires a session ticket if one
3366 was advertised in the ServerHello. Previously, a TLS client would
3367 ignore a missing NewSessionTicket message.
3370 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3372 *) SRTP Memory Leak.
3374 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3375 sends a carefully crafted handshake message, to cause OpenSSL to fail
3376 to free up to 64k of memory causing a memory leak. This could be
3377 exploited in a Denial Of Service attack. This issue affects OpenSSL
3378 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3379 whether SRTP is used or configured. Implementations of OpenSSL that
3380 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3382 The fix was developed by the OpenSSL team.
3386 *) Session Ticket Memory Leak.
3388 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3389 integrity of that ticket is first verified. In the event of a session
3390 ticket integrity check failing, OpenSSL will fail to free memory
3391 causing a memory leak. By sending a large number of invalid session
3392 tickets an attacker could exploit this issue in a Denial Of Service
3397 *) Build option no-ssl3 is incomplete.
3399 When OpenSSL is configured with "no-ssl3" as a build option, servers
3400 could accept and complete a SSL 3.0 handshake, and clients could be
3401 configured to send them.
3403 [Akamai and the OpenSSL team]
3405 *) Add support for TLS_FALLBACK_SCSV.
3406 Client applications doing fallback retries should call
3407 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3409 [Adam Langley, Bodo Moeller]
3411 *) Add additional DigestInfo checks.
3413 Re-encode DigestInto in DER and check against the original when
3414 verifying RSA signature: this will reject any improperly encoded
3415 DigestInfo structures.
3417 Note: this is a precautionary measure and no attacks are currently known.
3421 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3423 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3424 SRP code can be overrun an internal buffer. Add sanity check that
3425 g, A, B < N to SRP code.
3427 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3428 Group for discovering this issue.
3432 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3433 TLS 1.0 instead of higher protocol versions when the ClientHello message
3434 is badly fragmented. This allows a man-in-the-middle attacker to force a
3435 downgrade to TLS 1.0 even if both the server and the client support a
3436 higher protocol version, by modifying the client's TLS records.
3438 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3439 researching this issue.
3443 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3444 to a denial of service attack. A malicious server can crash the client
3445 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3446 ciphersuite and sending carefully crafted handshake messages.
3448 Thanks to Felix Gröbert (Google) for discovering and researching this
3453 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3454 to leak memory. This can be exploited through a Denial of Service attack.
3455 Thanks to Adam Langley for discovering and researching this issue.
3459 *) An attacker can force openssl to consume large amounts of memory whilst
3460 processing DTLS handshake messages. This can be exploited through a
3461 Denial of Service attack.
3462 Thanks to Adam Langley for discovering and researching this issue.
3466 *) An attacker can force an error condition which causes openssl to crash
3467 whilst processing DTLS packets due to memory being freed twice. This
3468 can be exploited through a Denial of Service attack.
3469 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3474 *) If a multithreaded client connects to a malicious server using a resumed
3475 session and the server sends an ec point format extension it could write
3476 up to 255 bytes to freed memory.
3478 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3483 *) A malicious server can crash an OpenSSL client with a null pointer
3484 dereference (read) by specifying an SRP ciphersuite even though it was not
3485 properly negotiated with the client. This can be exploited through a
3486 Denial of Service attack.
3488 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3489 discovering and researching this issue.
3493 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3494 X509_name_oneline, X509_name_print_ex et al. to leak some information
3495 from the stack. Applications may be affected if they echo pretty printing
3496 output to the attacker.
3498 Thanks to Ivan Fratric (Google) for discovering this issue.
3500 [Emilia Käsper, and Steve Henson]
3502 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3503 for corner cases. (Certain input points at infinity could lead to
3504 bogus results, with non-infinity inputs mapped to infinity too.)
3507 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3509 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3510 handshake can force the use of weak keying material in OpenSSL
3511 SSL/TLS clients and servers.
3513 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3514 researching this issue. (CVE-2014-0224)
3515 [KIKUCHI Masashi, Steve Henson]
3517 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3518 OpenSSL DTLS client the code can be made to recurse eventually crashing
3521 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3523 [Imre Rad, Steve Henson]
3525 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3526 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3527 client or server. This is potentially exploitable to run arbitrary
3528 code on a vulnerable client or server.
3530 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3531 [Jüri Aedla, Steve Henson]
3533 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3534 are subject to a denial of service attack.
3536 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3537 this issue. (CVE-2014-3470)
3538 [Felix Gröbert, Ivan Fratric, Steve Henson]
3540 *) Harmonize version and its documentation. -f flag is used to display
3542 [mancha <mancha1@zoho.com>]
3544 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3545 in i2d_ECPrivateKey.
3546 [mancha <mancha1@zoho.com>]
3548 *) Fix some double frees. These are not thought to be exploitable.
3549 [mancha <mancha1@zoho.com>]
3551 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3553 *) A missing bounds check in the handling of the TLS heartbeat extension
3554 can be used to reveal up to 64k of memory to a connected client or
3557 Thanks for Neel Mehta of Google Security for discovering this bug and to
3558 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3559 preparing the fix (CVE-2014-0160)
3560 [Adam Langley, Bodo Moeller]
3562 *) Fix for the attack described in the paper "Recovering OpenSSL
3563 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3564 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3565 http://eprint.iacr.org/2014/140
3567 Thanks to Yuval Yarom and Naomi Benger for discovering this
3568 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3569 [Yuval Yarom and Naomi Benger]
3571 *) TLS pad extension: draft-agl-tls-padding-03
3573 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3574 TLS client Hello record length value would otherwise be > 255 and
3575 less that 512 pad with a dummy extension containing zeroes so it
3576 is at least 512 bytes long.
3578 [Adam Langley, Steve Henson]
3580 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3582 *) Fix for TLS record tampering bug. A carefully crafted invalid
3583 handshake could crash OpenSSL with a NULL pointer exception.
3584 Thanks to Anton Johansson for reporting this issues.
3587 *) Keep original DTLS digest and encryption contexts in retransmission
3588 structures so we can use the previous session parameters if they need
3589 to be resent. (CVE-2013-6450)
3592 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3593 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3594 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3595 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3596 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3597 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3598 [Rob Stradling, Adam Langley]
3600 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3602 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3603 supporting platforms or when small records were transferred.
3604 [Andy Polyakov, Steve Henson]
3606 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3608 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3610 This addresses the flaw in CBC record processing discovered by
3611 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3612 at: http://www.isg.rhul.ac.uk/tls/
3614 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3615 Security Group at Royal Holloway, University of London
3616 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3617 Emilia Käsper for the initial patch.
3619 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3621 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3622 ciphersuites which can be exploited in a denial of service attack.
3623 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3624 and detecting this bug and to Wolfgang Ettlinger
3625 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3629 *) Return an error when checking OCSP signatures when key is NULL.
3630 This fixes a DoS attack. (CVE-2013-0166)
3633 *) Make openssl verify return errors.
3634 [Chris Palmer <palmer@google.com> and Ben Laurie]
3636 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3637 the right response is stapled. Also change SSL_get_certificate()
3638 so it returns the certificate actually sent.
3639 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3640 [Rob Stradling <rob.stradling@comodo.com>]
3642 *) Fix possible deadlock when decoding public keys.
3645 *) Don't use TLS 1.0 record version number in initial client hello
3649 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3651 *) Sanity check record length before skipping explicit IV in TLS
3652 1.2, 1.1 and DTLS to fix DoS attack.
3654 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3655 fuzzing as a service testing platform.
3659 *) Initialise tkeylen properly when encrypting CMS messages.
3660 Thanks to Solar Designer of Openwall for reporting this issue.
3663 *) In FIPS mode don't try to use composite ciphers as they are not
3667 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3669 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3670 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3671 mean any application compiled against OpenSSL 1.0.0 headers setting
3672 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3673 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3674 0x10000000L Any application which was previously compiled against
3675 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3676 will need to be recompiled as a result. Letting be results in
3677 inability to disable specifically TLS 1.1 and in client context,
3678 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3681 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3682 disable just protocol X, but all protocols above X *if* there are
3683 protocols *below* X still enabled. In more practical terms it means
3684 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3685 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3686 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3690 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3692 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3693 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3694 in CRYPTO_realloc_clean.
3696 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3697 issue and to Adam Langley <agl@chromium.org> for fixing it.
3699 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3701 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3704 *) Workarounds for some broken servers that "hang" if a client hello
3705 record length exceeds 255 bytes.
3707 1. Do not use record version number > TLS 1.0 in initial client
3708 hello: some (but not all) hanging servers will now work.
3709 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3710 the number of ciphers sent in the client hello. This should be
3711 set to an even number, such as 50, for example by passing:
3712 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3713 Most broken servers should now work.
3714 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3715 TLS 1.2 client support entirely.
3718 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3721 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3723 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3724 STRING form instead of a DigestInfo.
3727 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3728 and the RSA_sign/RSA_verify functions. This was made more apparent when
3729 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3730 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3731 the correct format in RSA_verify so both forms transparently work.
3734 *) Some servers which support TLS 1.0 can choke if we initially indicate
3735 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3736 encrypted premaster secret. As a workaround use the maximum permitted
3737 client version in client hello, this should keep such servers happy
3738 and still work with previous versions of OpenSSL.
3741 *) Add support for TLS/DTLS heartbeats.
3742 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3744 *) Add support for SCTP.
3745 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3747 *) Improved PRNG seeding for VOS.
3748 [Paul Green <Paul.Green@stratus.com>]
3750 *) Extensive assembler packs updates, most notably:
3752 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3753 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3754 - x86_64: bit-sliced AES implementation;
3755 - ARM: NEON support, contemporary platforms optimizations;
3756 - s390x: z196 support;
3757 - *: GHASH and GF(2^m) multiplication implementations;
3761 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3762 (removal of unnecessary code)
3763 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3765 *) Add TLS key material exporter from RFC 5705.
3768 *) Add DTLS-SRTP negotiation from RFC 5764.
3771 *) Add Next Protocol Negotiation,
3772 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3773 disabled with a no-npn flag to config or Configure. Code donated
3775 [Adam Langley <agl@google.com> and Ben Laurie]
3777 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3778 NIST-P256, NIST-P521, with constant-time single point multiplication on
3779 typical inputs. Compiler support for the nonstandard type __uint128_t is
3780 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3781 Code made available under Apache License version 2.0.
3783 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3784 line to include this in your build of OpenSSL, and run "make depend" (or
3785 "make update"). This enables the following EC_METHODs:
3787 EC_GFp_nistp224_method()
3788 EC_GFp_nistp256_method()
3789 EC_GFp_nistp521_method()
3791 EC_GROUP_new_by_curve_name() will automatically use these (while
3792 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3794 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3796 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3797 all platforms. Move ssize_t definition from e_os.h to the public
3798 header file e_os2.h as it now appears in public header file cms.h
3801 *) New -sigopt option to the ca, req and x509 utilities. Additional
3802 signature parameters can be passed using this option and in
3806 *) Add RSA PSS signing function. This will generate and set the
3807 appropriate AlgorithmIdentifiers for PSS based on those in the
3808 corresponding EVP_MD_CTX structure. No application support yet.
3811 *) Support for companion algorithm specific ASN1 signing routines.
3812 New function ASN1_item_sign_ctx() signs a pre-initialised
3813 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3814 the appropriate parameters.
3817 *) Add new algorithm specific ASN1 verification initialisation function
3818 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3819 handling will be the same no matter what EVP_PKEY_METHOD is used.
3820 Add a PSS handler to support verification of PSS signatures: checked
3821 against a number of sample certificates.
3824 *) Add signature printing for PSS. Add PSS OIDs.
3825 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3827 *) Add algorithm specific signature printing. An individual ASN1 method
3828 can now print out signatures instead of the standard hex dump.
3830 More complex signatures (e.g. PSS) can print out more meaningful
3831 information. Include DSA version that prints out the signature
3835 *) Password based recipient info support for CMS library: implementing
3839 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3840 neatly separates the code into cipher and PBE sections and is required
3841 for some algorithms that split PBES2 into separate pieces (such as
3842 password based CMS).
3845 *) Session-handling fixes:
3846 - Fix handling of connections that are resuming with a session ID,
3847 but also support Session Tickets.
3848 - Fix a bug that suppressed issuing of a new ticket if the client
3849 presented a ticket with an expired session.
3850 - Try to set the ticket lifetime hint to something reasonable.
3851 - Make tickets shorter by excluding irrelevant information.
3852 - On the client side, don't ignore renewed tickets.
3853 [Adam Langley, Bodo Moeller (Google)]
3855 *) Fix PSK session representation.
3858 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3860 This work was sponsored by Intel.
3863 *) Add GCM support to TLS library. Some custom code is needed to split
3864 the IV between the fixed (from PRF) and explicit (from TLS record)
3865 portions. This adds all GCM ciphersuites supported by RFC5288 and
3866 RFC5289. Generalise some AES* cipherstrings to include GCM and
3867 add a special AESGCM string for GCM only.
3870 *) Expand range of ctrls for AES GCM. Permit setting invocation
3871 field on decrypt and retrieval of invocation field only on encrypt.
3874 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3875 As required by RFC5289 these ciphersuites cannot be used if for
3876 versions of TLS earlier than 1.2.
3879 *) For FIPS capable OpenSSL interpret a NULL default public key method
3880 as unset and return the appropriate default but do *not* set the default.
3881 This means we can return the appropriate method in applications that
3882 switch between FIPS and non-FIPS modes.
3885 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3886 ENGINE is used then we cannot handle that in the FIPS module so we
3887 keep original code iff non-FIPS operations are allowed.
3890 *) Add -attime option to openssl utilities.
3891 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3893 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3896 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3897 FIPS EC methods unconditionally for now.
3900 *) New build option no-ec2m to disable characteristic 2 code.
3903 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3904 all cases can be covered as some introduce binary incompatibilities.
3907 *) Redirect RSA operations to FIPS module including keygen,
3908 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3911 *) Add similar low level API blocking to ciphers.
3914 *) Low level digest APIs are not approved in FIPS mode: any attempt
3915 to use these will cause a fatal error. Applications that *really* want
3916 to use them can use the private_* version instead.
3919 *) Redirect cipher operations to FIPS module for FIPS builds.
3922 *) Redirect digest operations to FIPS module for FIPS builds.
3925 *) Update build system to add "fips" flag which will link in fipscanister.o
3926 for static and shared library builds embedding a signature if needed.
3929 *) Output TLS supported curves in preference order instead of numerical
3930 order. This is currently hardcoded for the highest order curves first.
3931 This should be configurable so applications can judge speed vs strength.
3934 *) Add TLS v1.2 server support for client authentication.
3937 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3941 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3942 FIPS modules versions.
3945 *) Add TLS v1.2 client side support for client authentication. Keep cache
3946 of handshake records longer as we don't know the hash algorithm to use
3947 until after the certificate request message is received.
3950 *) Initial TLS v1.2 client support. Add a default signature algorithms
3951 extension including all the algorithms we support. Parse new signature
3952 format in client key exchange. Relax some ECC signing restrictions for
3953 TLS v1.2 as indicated in RFC5246.
3956 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3957 to new signature format when needed using client digest preference.
3958 All server ciphersuites should now work correctly in TLS v1.2. No client
3959 support yet and no support for client certificates.
3962 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3963 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3964 ciphersuites. At present only RSA key exchange ciphersuites work with
3965 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3966 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3967 and version checking.
3970 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3971 with this defined it will not be affected by any changes to ssl internal
3972 structures. Add several utility functions to allow openssl application
3973 to work with OPENSSL_NO_SSL_INTERN defined.
3976 *) A long standing patch to add support for SRP from EdelWeb (Peter
3977 Sylvester and Christophe Renou) was integrated.
3978 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3979 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3982 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3985 *) Permit abbreviated handshakes when renegotiating using the function
3986 SSL_renegotiate_abbreviated().
3987 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3989 *) Add call to ENGINE_register_all_complete() to
3990 ENGINE_load_builtin_engines(), so some implementations get used
3991 automatically instead of needing explicit application support.
3994 *) Add support for TLS key exporter as described in RFC5705.
3995 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3997 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3998 a few changes are required:
4000 Add SSL_OP_NO_TLSv1_1 flag.
4001 Add TLSv1_1 methods.
4002 Update version checking logic to handle version 1.1.
4003 Add explicit IV handling (ported from DTLS code).
4004 Add command line options to s_client/s_server.
4007 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
4009 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4010 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4011 content decryption and always return the same error. Note: this attack
4012 needs on average 2^20 messages so it only affects automated senders. The
4013 old behaviour can be re-enabled in the CMS code by setting the
4014 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4015 an MMA defence is not necessary.
4016 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4017 this issue. (CVE-2012-0884)
4020 *) Fix CVE-2011-4619: make sure we really are receiving a
4021 client hello before rejecting multiple SGC restarts. Thanks to
4022 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4025 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
4027 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4028 Thanks to Antonio Martin, Enterprise Secure Access Research and
4029 Development, Cisco Systems, Inc. for discovering this bug and
4030 preparing a fix. (CVE-2012-0050)
4033 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
4035 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
4036 of the Vaudenay padding oracle attack on CBC mode encryption
4037 which enables an efficient plaintext recovery attack against
4038 the OpenSSL implementation of DTLS. Their attack exploits timing
4039 differences arising during decryption processing. A research
4040 paper describing this attack can be found at:
4041 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4042 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4043 Security Group at Royal Holloway, University of London
4044 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4045 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4046 for preparing the fix. (CVE-2011-4108)
4047 [Robin Seggelmann, Michael Tuexen]
4049 *) Clear bytes used for block padding of SSL 3.0 records.
4051 [Adam Langley (Google)]
4053 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4054 Kadianakis <desnacked@gmail.com> for discovering this issue and
4055 Adam Langley for preparing the fix. (CVE-2011-4619)
4056 [Adam Langley (Google)]
4058 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4059 [Andrey Kulikov <amdeich@gmail.com>]
4061 *) Prevent malformed RFC3779 data triggering an assertion failure.
4062 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4063 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4064 [Rob Austein <sra@hactrn.net>]
4066 *) Improved PRNG seeding for VOS.
4067 [Paul Green <Paul.Green@stratus.com>]
4069 *) Fix ssl_ciph.c set-up race.
4070 [Adam Langley (Google)]
4072 *) Fix spurious failures in ecdsatest.c.
4073 [Emilia Käsper (Google)]
4075 *) Fix the BIO_f_buffer() implementation (which was mixing different
4076 interpretations of the '..._len' fields).
4077 [Adam Langley (Google)]
4079 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4080 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4081 threads won't reuse the same blinding coefficients.
4083 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4084 lock to call BN_BLINDING_invert_ex, and avoids one use of
4085 BN_BLINDING_update for each BN_BLINDING structure (previously,
4086 the last update always remained unused).
4087 [Emilia Käsper (Google)]
4089 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4090 [Bob Buckholz (Google)]
4092 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4094 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4095 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4096 [Kaspar Brand <ossl@velox.ch>]
4098 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4099 for multi-threaded use of ECDH. (CVE-2011-3210)
4100 [Adam Langley (Google)]
4102 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4105 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4106 signature public key algorithm by using OID xref utilities instead.
4107 Before this you could only use some ECC ciphersuites with SHA1 only.
4110 *) Add protection against ECDSA timing attacks as mentioned in the paper
4111 by Billy Bob Brumley and Nicola Tuveri, see:
4113 http://eprint.iacr.org/2011/232.pdf
4115 [Billy Bob Brumley and Nicola Tuveri]
4117 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4119 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4120 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4122 *) Fix bug in string printing code: if *any* escaping is enabled we must
4123 escape the escape character (backslash) or the resulting string is
4127 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4129 *) Disable code workaround for ancient and obsolete Netscape browsers
4130 and servers: an attacker can use it in a ciphersuite downgrade attack.
4131 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4134 *) Fixed J-PAKE implementation error, originally discovered by
4135 Sebastien Martini, further info and confirmation from Stefan
4136 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4139 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4141 *) Fix extension code to avoid race conditions which can result in a buffer
4142 overrun vulnerability: resumed sessions must not be modified as they can
4143 be shared by multiple threads. CVE-2010-3864
4146 *) Fix WIN32 build system to correctly link an ENGINE directory into
4150 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4152 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4154 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4156 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4158 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4159 context. The operation can be customised via the ctrl mechanism in
4160 case ENGINEs want to include additional functionality.
4163 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4166 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4167 output hashes compatible with older versions of OpenSSL.
4168 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4170 *) Fix compression algorithm handling: if resuming a session use the
4171 compression algorithm of the resumed session instead of determining
4172 it from client hello again. Don't allow server to change algorithm.
4175 *) Add load_crls() function to apps tidying load_certs() too. Add option
4176 to verify utility to allow additional CRLs to be included.
4179 *) Update OCSP request code to permit adding custom headers to the request:
4180 some responders need this.
4183 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4185 [Julia Lawall <julia@diku.dk>]
4187 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4188 needlessly dereferenced structures, used obsolete functions and
4189 didn't handle all updated verify codes correctly.
4192 *) Disable MD2 in the default configuration.
4195 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4196 indicate the initial BIO being pushed or popped. This makes it possible
4197 to determine whether the BIO is the one explicitly called or as a result
4198 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4199 it handles reference counts correctly and doesn't zero out the I/O bio
4200 when it is not being explicitly popped. WARNING: applications which
4201 included workarounds for the old buggy behaviour will need to be modified
4202 or they could free up already freed BIOs.
4205 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4206 renaming to all platforms (within the 0.9.8 branch, this was
4207 done conditionally on Netware platforms to avoid a name clash).
4208 [Guenter <lists@gknw.net>]
4210 *) Add ECDHE and PSK support to DTLS.
4211 [Michael Tuexen <tuexen@fh-muenster.de>]
4213 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4217 *) Add "missing" function EVP_MD_flags() (without this the only way to
4218 retrieve a digest flags is by accessing the structure directly. Update
4219 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4220 or cipher is registered as in the "from" argument. Print out all
4221 registered digests in the dgst usage message instead of manually
4222 attempting to work them out.
4225 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4226 this allows the use of compression and extensions. Change default cipher
4227 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4228 by default unless an application cipher string requests it.
4231 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4232 key ids to find matching certificates and keys but some PKCS#12 files
4233 don't follow the (somewhat unwritten) rules and this strategy fails.
4234 Now just gather all certificates together and the first private key
4235 then look for the first certificate that matches the key.
4238 *) Support use of registered digest and cipher names for dgst and cipher
4239 commands instead of having to add each one as a special case. So now
4246 openssl dgst -sha256 foo
4248 and this works for ENGINE based algorithms too.
4252 *) Update Gost ENGINE to support parameter files.
4253 [Victor B. Wagner <vitus@cryptocom.ru>]
4255 *) Support GeneralizedTime in ca utility.
4256 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4258 *) Enhance the hash format used for certificate directory links. The new
4259 form uses the canonical encoding (meaning equivalent names will work
4260 even if they aren't identical) and uses SHA1 instead of MD5. This form
4261 is incompatible with the older format and as a result c_rehash should
4262 be used to rebuild symbolic links.
4265 *) Make PKCS#8 the default write format for private keys, replacing the
4266 traditional format. This form is standardised, more secure and doesn't
4267 include an implicit MD5 dependency.
4270 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4271 committed to OpenSSL should pass this lot as a minimum.
4274 *) Add session ticket override functionality for use by EAP-FAST.
4275 [Jouni Malinen <j@w1.fi>]
4277 *) Modify HMAC functions to return a value. Since these can be implemented
4278 in an ENGINE errors can occur.
4281 *) Type-checked OBJ_bsearch_ex.
4284 *) Type-checked OBJ_bsearch. Also some constification necessitated
4285 by type-checking. Still to come: TXT_DB, bsearch(?),
4286 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4290 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4291 seconds to a tm structure directly, instead of going through OS
4292 specific date routines. This avoids any issues with OS routines such
4293 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4294 and X509_time_adj_ex() to cover the extended range. The existing
4295 X509_time_adj() is still usable and will no longer have any date issues.
4298 *) Delta CRL support. New use deltas option which will attempt to locate
4299 and search any appropriate delta CRLs available.
4301 This work was sponsored by Google.
4304 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4305 code and add additional score elements. Validate alternate CRL paths
4306 as part of the CRL checking and indicate a new error "CRL path validation
4307 error" in this case. Applications wanting additional details can use
4308 the verify callback and check the new "parent" field. If this is not
4309 NULL CRL path validation is taking place. Existing applications won't
4310 see this because it requires extended CRL support which is off by
4313 This work was sponsored by Google.
4316 *) Support for freshest CRL extension.
4318 This work was sponsored by Google.
4321 *) Initial indirect CRL support. Currently only supported in the CRLs
4322 passed directly and not via lookup. Process certificate issuer
4323 CRL entry extension and lookup CRL entries by bother issuer name
4324 and serial number. Check and process CRL issuer entry in IDP extension.
4326 This work was sponsored by Google.
4329 *) Add support for distinct certificate and CRL paths. The CRL issuer
4330 certificate is validated separately in this case. Only enabled if
4331 an extended CRL support flag is set: this flag will enable additional
4332 CRL functionality in future.
4334 This work was sponsored by Google.
4337 *) Add support for policy mappings extension.
4339 This work was sponsored by Google.
4342 *) Fixes to pathlength constraint, self issued certificate handling,
4343 policy processing to align with RFC3280 and PKITS tests.
4345 This work was sponsored by Google.
4348 *) Support for name constraints certificate extension. DN, email, DNS
4349 and URI types are currently supported.
4351 This work was sponsored by Google.
4354 *) To cater for systems that provide a pointer-based thread ID rather
4355 than numeric, deprecate the current numeric thread ID mechanism and
4356 replace it with a structure and associated callback type. This
4357 mechanism allows a numeric "hash" to be extracted from a thread ID in
4358 either case, and on platforms where pointers are larger than 'long',
4359 mixing is done to help ensure the numeric 'hash' is usable even if it
4360 can't be guaranteed unique. The default mechanism is to use "&errno"
4361 as a pointer-based thread ID to distinguish between threads.
4363 Applications that want to provide their own thread IDs should now use
4364 CRYPTO_THREADID_set_callback() to register a callback that will call
4365 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4367 Note that ERR_remove_state() is now deprecated, because it is tied
4368 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4369 to free the current thread's error state should be replaced by
4370 ERR_remove_thread_state(NULL).
4372 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4373 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4374 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4375 application was previously providing a numeric thread callback that
4376 was inappropriate for distinguishing threads, then uniqueness might
4377 have been obtained with &errno that happened immediately in the
4378 intermediate development versions of OpenSSL; this is no longer the
4379 case, the numeric thread callback will now override the automatic use
4381 [Geoff Thorpe, with help from Bodo Moeller]
4383 *) Initial support for different CRL issuing certificates. This covers a
4384 simple case where the self issued certificates in the chain exist and
4385 the real CRL issuer is higher in the existing chain.
4387 This work was sponsored by Google.
4390 *) Removed effectively defunct crypto/store from the build.
4393 *) Revamp of STACK to provide stronger type-checking. Still to come:
4394 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4395 ASN1_STRING, CONF_VALUE.
4398 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4399 RAM on SSL connections. This option can save about 34k per idle SSL.
4402 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4403 STACK, TXT_DB, bsearch, qsort.
4406 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4407 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4408 support for data, signedData, compressedData, digestedData and
4409 encryptedData, envelopedData types included. Scripts to check against
4410 RFC4134 examples draft and interop and consistency checks of many
4411 content types and variants.
4414 *) Add options to enc utility to support use of zlib compression BIO.
4417 *) Extend mk1mf to support importing of options and assembly language
4418 files from Configure script, currently only included in VC-WIN32.
4419 The assembly language rules can now optionally generate the source
4420 files from the associated perl scripts.
4423 *) Implement remaining functionality needed to support GOST ciphersuites.
4424 Interop testing has been performed using CryptoPro implementations.
4425 [Victor B. Wagner <vitus@cryptocom.ru>]
4427 *) s390x assembler pack.
4430 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4434 *) Implement Opaque PRF Input TLS extension as specified in
4435 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4436 official specification yet and no extension type assignment by
4437 IANA exists, this extension (for now) will have to be explicitly
4438 enabled when building OpenSSL by providing the extension number
4439 to use. For example, specify an option
4441 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4443 to the "config" or "Configure" script to enable the extension,
4444 assuming extension number 0x9527 (which is a completely arbitrary
4445 and unofficial assignment based on the MD5 hash of the Internet
4446 Draft). Note that by doing so, you potentially lose
4447 interoperability with other TLS implementations since these might
4448 be using the same extension number for other purposes.
4450 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4451 opaque PRF input value to use in the handshake. This will create
4452 an internal copy of the length-'len' string at 'src', and will
4453 return non-zero for success.
4455 To get more control and flexibility, provide a callback function
4458 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4459 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4463 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4466 Callback function 'cb' will be called in handshakes, and is
4467 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4468 Argument 'arg' is for application purposes (the value as given to
4469 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4470 be provided to the callback function). The callback function
4471 has to return non-zero to report success: usually 1 to use opaque
4472 PRF input just if possible, or 2 to enforce use of the opaque PRF
4473 input. In the latter case, the library will abort the handshake
4474 if opaque PRF input is not successfully negotiated.
4476 Arguments 'peerinput' and 'len' given to the callback function
4477 will always be NULL and 0 in the case of a client. A server will
4478 see the client's opaque PRF input through these variables if
4479 available (NULL and 0 otherwise). Note that if the server
4480 provides an opaque PRF input, the length must be the same as the
4481 length of the client's opaque PRF input.
4483 Note that the callback function will only be called when creating
4484 a new session (session resumption can resume whatever was
4485 previously negotiated), and will not be called in SSL 2.0
4486 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4487 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4488 for applications that need to enforce opaque PRF input.
4492 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4495 [Victor B. Wagner <vitus@cryptocom.ru>]
4497 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4498 RFC4507bis. The encrypted ticket format is an encrypted encoded
4499 SSL_SESSION structure, that way new session features are automatically
4502 If a client application caches session in an SSL_SESSION structure
4503 support is transparent because tickets are now stored in the encoded
4506 The SSL_CTX structure automatically generates keys for ticket
4507 protection in servers so again support should be possible
4508 with no application modification.
4510 If a client or server wishes to disable RFC4507 support then the option
4511 SSL_OP_NO_TICKET can be set.
4513 Add a TLS extension debugging callback to allow the contents of any client
4514 or server extensions to be examined.
4516 This work was sponsored by Google.
4519 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4520 OpenSSL should now compile cleanly on gcc 4.2
4521 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4523 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4524 support including streaming MAC support: this is required for GOST
4525 ciphersuite support.
4526 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4528 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4529 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4530 to output in BER and PEM format.
4533 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4534 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4535 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4536 ENGINE support for HMAC keys which are unextractable. New -mac and
4537 -macopt options to dgst utility.
4540 *) New option -sigopt to dgst utility. Update dgst to use
4541 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4542 alternative signing parameters such as X9.31 or PSS in the dgst
4546 *) Change ssl_cipher_apply_rule(), the internal function that does
4547 the work each time a ciphersuite string requests enabling
4548 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4549 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4550 the order of disabled ciphersuites such that those ciphersuites
4551 that most recently went from enabled to disabled not only stay
4552 in order with respect to each other, but also have higher priority
4553 than other disabled ciphersuites the next time ciphersuites are
4556 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4557 the same ciphersuites as with "HIGH" alone, but in a specific
4558 order where the PSK ciphersuites come first (since they are the
4559 most recently disabled ciphersuites when "HIGH" is parsed).
4561 Also, change ssl_create_cipher_list() (using this new
4562 functionality) such that between otherwise identical
4563 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4567 *) Change ssl_create_cipher_list() so that it automatically
4568 arranges the ciphersuites in reasonable order before starting
4569 to process the rule string. Thus, the definition for "DEFAULT"
4570 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4571 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4572 This makes it much easier to arrive at a reasonable default order
4573 in applications for which anonymous ciphers are OK (meaning
4574 that you can't actually use DEFAULT).
4575 [Bodo Moeller; suggested by Victor Duchovni]
4577 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4578 processing) into multiple integers instead of setting
4579 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4580 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4581 (These masks as well as the individual bit definitions are hidden
4582 away into the non-exported interface ssl/ssl_locl.h, so this
4583 change to the definition of the SSL_CIPHER structure shouldn't
4584 affect applications.) This give us more bits for each of these
4585 categories, so there is no longer a need to coagulate AES128 and
4586 AES256 into a single algorithm bit, and to coagulate Camellia128
4587 and Camellia256 into a single algorithm bit, which has led to all
4590 Thus, among other things, the kludge introduced in 0.9.7m and
4591 0.9.8e for masking out AES256 independently of AES128 or masking
4592 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4594 With the change, we also introduce new ciphersuite aliases that
4595 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4599 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4600 Use the leftmost N bytes of the signature input if the input is
4601 larger than the prime q (with N being the size in bytes of q).
4604 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4605 it yet and it is largely untested.
4608 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4611 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4612 some compilers (gcc 4.2 and later) reject their use. Safestack is
4613 reimplemented. Update ASN1 to avoid use of legacy functions.
4616 *) Win32/64 targets are linked with Winsock2.
4619 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4620 to external functions. This can be used to increase CRL handling
4621 efficiency especially when CRLs are very large by (for example) storing
4622 the CRL revoked certificates in a database.
4625 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4626 new CRLs added to a directory can be used. New command line option
4627 -verify_return_error to s_client and s_server. This causes real errors
4628 to be returned by the verify callback instead of carrying on no matter
4629 what. This reflects the way a "real world" verify callback would behave.
4632 *) GOST engine, supporting several GOST algorithms and public key formats.
4633 Kindly donated by Cryptocom.
4636 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4637 partitioned by DP are handled but no indirect CRL or reason partitioning
4638 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4639 selected via a scoring technique which handles IDP and AKID in CRLs.
4642 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4643 will ultimately be used for all verify operations: this will remove the
4644 X509_STORE dependency on certificate verification and allow alternative
4645 lookup methods. X509_STORE based implementations of these two callbacks.
4648 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4649 Modify get_crl() to find a valid (unexpired) CRL if possible.
4652 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4653 this would be called X509_CRL_cmp() but that name is already used by
4654 a function that just compares CRL issuer names. Cache several CRL
4655 extensions in X509_CRL structure and cache CRLDP in X509.
4658 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4659 this maps equivalent X509_NAME structures into a consistent structure.
4660 Name comparison can then be performed rapidly using memcmp().
4663 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4667 *) Allow digests to supply their own micalg string for S/MIME type using
4668 the ctrl EVP_MD_CTRL_MICALG.
4671 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4672 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4673 ctrl. It can then customise the structure before and/or after signing
4677 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4678 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4679 to free up any added signature OIDs.
4682 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4683 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4684 digest and cipher tables. New options added to openssl utility:
4685 list-message-digest-algorithms and list-cipher-algorithms.
4688 *) Change the array representation of binary polynomials: the list
4689 of degrees of non-zero coefficients is now terminated with -1.
4690 Previously it was terminated with 0, which was also part of the
4691 value; thus, the array representation was not applicable to
4692 polynomials where t^0 has coefficient zero. This change makes
4693 the array representation useful in a more general context.
4696 *) Various modifications and fixes to SSL/TLS cipher string
4697 handling. For ECC, the code now distinguishes between fixed ECDH
4698 with RSA certificates on the one hand and with ECDSA certificates
4699 on the other hand, since these are separate ciphersuites. The
4700 unused code for Fortezza ciphersuites has been removed.
4702 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4703 (not "ECDHE"). For consistency with the code for DH
4704 certificates, use of ECDH certificates is now considered ECDH
4705 authentication, not RSA or ECDSA authentication (the latter is
4706 merely the CA's signing algorithm and not actively used in the
4709 The temporary ciphersuite alias "ECCdraft" is no longer
4710 available, and ECC ciphersuites are no longer excluded from "ALL"
4711 and "DEFAULT". The following aliases now exist for RFC 4492
4712 ciphersuites, most of these by analogy with the DH case:
4714 kECDHr - ECDH cert, signed with RSA
4715 kECDHe - ECDH cert, signed with ECDSA
4716 kECDH - ECDH cert (signed with either RSA or ECDSA)
4717 kEECDH - ephemeral ECDH
4718 ECDH - ECDH cert or ephemeral ECDH
4724 AECDH - anonymous ECDH
4725 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4729 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4730 Use correct micalg parameters depending on digest(s) in signed message.
4733 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4734 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4737 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4738 an engine to register a method. Add ENGINE lookups for methods and
4739 functional reference processing.
4742 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4743 EVP_{Sign,Verify}* which allow an application to customise the signature
4747 *) New -resign option to smime utility. This adds one or more signers
4748 to an existing PKCS#7 signedData structure. Also -md option to use an
4749 alternative message digest algorithm for signing.
4752 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4753 create PKCS7 structures containing multiple signers. Update smime
4754 application to support multiple signers.
4757 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4761 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4762 Reorganize PBE internals to lookup from a static table using NIDs,
4763 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4764 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4765 PRF which will be automatically used with PBES2.
4768 *) Replace the algorithm specific calls to generate keys in "req" with the
4772 *) Update PKCS#7 enveloped data routines to use new API. This is now
4773 supported by any public key method supporting the encrypt operation. A
4774 ctrl is added to allow the public key algorithm to examine or modify
4775 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4779 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4780 a default digest type to use. In most cases this will be SHA1 but some
4781 algorithms (such as GOST) need to specify an alternative digest. The
4782 return value indicates how strong the preference is 1 means optional and
4783 2 is mandatory (that is it is the only supported type). Modify
4784 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4785 use the default md. Update openssl utilities to use the default digest
4786 type for signing if it is not explicitly indicated.
4789 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4790 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4791 signing method from the key type. This effectively removes the link
4792 between digests and public key types.
4795 *) Add an OID cross reference table and utility functions. Its purpose is to
4796 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4797 rsaEncryption. This will allow some of the algorithm specific hackery
4798 needed to use the correct OID to be removed.
4801 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4802 structures for PKCS7_sign(). They are now set up by the relevant public
4806 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4809 *) Add support for key derivation (agreement) in the API, DH method and
4813 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4814 public and private key formats. As a side effect these add additional
4815 command line functionality not previously available: DSA signatures can be
4816 generated and verified using pkeyutl and DH key support and generation in
4821 [Oliver Tappe <zooey@hirschkaefer.de>]
4823 *) New make target "install_html_docs" installs HTML renditions of the
4825 [Oliver Tappe <zooey@hirschkaefer.de>]
4827 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4828 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4829 support key and parameter generation and add initial key generation
4830 functionality for RSA.
4833 *) Add functions for main EVP_PKEY_method operations. The undocumented
4834 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4835 EVP_PKEY_{encrypt,decrypt}_old.
4838 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4839 key API, doesn't do much yet.
4842 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4843 public key algorithms. New option to openssl utility:
4844 "list-public-key-algorithms" to print out info.
4847 *) Implement the Supported Elliptic Curves Extension for
4848 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4851 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4852 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4855 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4856 utilities such as rsa, dsa, dsaparam etc except they process any key
4860 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4861 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4862 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4866 *) Initial support for pluggable public key ASN1.
4867 De-spaghettify the public key ASN1 handling. Move public and private
4868 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4869 algorithm specific handling to a single module within the relevant
4870 algorithm directory. Add functions to allow (near) opaque processing
4871 of public and private key structures.
4874 *) Implement the Supported Point Formats Extension for
4875 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4878 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4879 for the psk identity [hint] and the psk callback functions to the
4880 SSL_SESSION, SSL and SSL_CTX structure.
4883 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4887 SSL_CTX_use_psk_identity_hint
4888 SSL_get_psk_identity_hint
4889 SSL_get_psk_identity
4890 SSL_use_psk_identity_hint
4892 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4894 *) Add RFC 3161 compliant time stamp request creation, response generation
4895 and response verification functionality.
4896 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4898 *) Add initial support for TLS extensions, specifically for the server_name
4899 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4900 have new members for a host name. The SSL data structure has an
4901 additional member SSL_CTX *initial_ctx so that new sessions can be
4902 stored in that context to allow for session resumption, even after the
4903 SSL has been switched to a new SSL_CTX in reaction to a client's
4904 server_name extension.
4906 New functions (subject to change):
4908 SSL_get_servername()
4909 SSL_get_servername_type()
4912 New CTRL codes and macros (subject to change):
4914 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4915 - SSL_CTX_set_tlsext_servername_callback()
4916 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4917 - SSL_CTX_set_tlsext_servername_arg()
4918 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4920 openssl s_client has a new '-servername ...' option.
4922 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4923 '-key2 ...', '-servername_fatal' (subject to change). This allows
4924 testing the HostName extension for a specific single host name ('-cert'
4925 and '-key' remain fallbacks for handshakes without HostName
4926 negotiation). If the unrecognized_name alert has to be sent, this by
4927 default is a warning; it becomes fatal with the '-servername_fatal'
4930 [Peter Sylvester, Remy Allais, Christophe Renou]
4932 *) Whirlpool hash implementation is added.
4935 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4936 bn(64,32). Because of instruction set limitations it doesn't have
4937 any negative impact on performance. This was done mostly in order
4938 to make it possible to share assembler modules, such as bn_mul_mont
4939 implementations, between 32- and 64-bit builds without hassle.
4942 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4943 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4947 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4948 dedicated Montgomery multiplication procedure, is introduced.
4949 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4950 "64-bit" performance on certain 32-bit targets.
4953 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4954 in SSL structures. New SSL ctrl to set maximum send fragment size.
4955 Save memory by setting the I/O buffer sizes dynamically instead of
4956 using the maximum available value.
4959 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4960 in addition to the text details.
4963 *) Very, very preliminary EXPERIMENTAL support for printing of general
4964 ASN1 structures. This currently produces rather ugly output and doesn't
4965 handle several customised structures at all.
4968 *) Integrated support for PVK file format and some related formats such
4969 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4970 these in the 'rsa' and 'dsa' utilities.
4973 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4976 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4977 place for the (very old) "NETSCAPE" format certificates which are now
4978 handled using new ASN1 code equivalents.
4981 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4982 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4983 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4986 *) Modify CRL distribution points extension code to print out previously
4987 unsupported fields. Enhance extension setting code to allow setting of
4991 *) Add print and set support for Issuing Distribution Point CRL extension.
4994 *) Change 'Configure' script to enable Camellia by default.
4997 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4999 *) When rejecting SSL/TLS records due to an incorrect version number, never
5000 update s->server with a new major version number. As of
5001 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5002 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5003 the previous behavior could result in a read attempt at NULL when
5004 receiving specific incorrect SSL/TLS records once record payload
5005 protection is active. (CVE-2010-0740)
5006 [Bodo Moeller, Adam Langley <agl@chromium.org>]
5008 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5009 could be crashed if the relevant tables were not present (e.g. chrooted).
5010 [Tomas Hoger <thoger@redhat.com>]
5012 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5014 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5015 [Martin Olsson, Neel Mehta]
5017 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
5018 accommodate for stack sorting, always a write lock!).
5021 *) On some versions of WIN32 Heap32Next is very slow. This can cause
5022 excessive delays in the RAND_poll(): over a minute. As a workaround
5023 include a time check in the inner Heap32Next loop too.
5026 *) The code that handled flushing of data in SSL/TLS originally used the
5027 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5028 the problem outlined in PR#1949. The fix suggested there however can
5029 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5030 of Apache). So instead simplify the code to flush unconditionally.
5031 This should be fine since flushing with no data to flush is a no op.
5034 *) Handle TLS versions 2.0 and later properly and correctly use the
5035 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5036 off ancient servers have a habit of sticking around for a while...
5039 *) Modify compression code so it frees up structures without using the
5040 ex_data callbacks. This works around a problem where some applications
5041 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5042 restarting) then use compression (e.g. SSL with compression) later.
5043 This results in significant per-connection memory leaks and
5044 has caused some security issues including CVE-2008-1678 and
5048 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5049 change when encrypting or decrypting.
5052 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5053 connect and renegotiate with servers which do not support RI.
5054 Until RI is more widely deployed this option is enabled by default.
5057 *) Add "missing" ssl ctrls to clear options and mode.
5060 *) If client attempts to renegotiate and doesn't support RI respond with
5061 a no_renegotiation alert as required by RFC5746. Some renegotiating
5062 TLS clients will continue a connection gracefully when they receive
5063 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5064 waiting for a server hello which it will never receive. Now we treat a
5065 received no_renegotiation alert as a fatal error. This is because
5066 applications requesting a renegotiation might well expect it to succeed
5067 and would have no code in place to handle the server denying it so the
5068 only safe thing to do is to terminate the connection.
5071 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5072 peer supports secure renegotiation and 0 otherwise. Print out peer
5073 renegotiation support in s_client/s_server.
5076 *) Replace the highly broken and deprecated SPKAC certification method with
5077 the updated NID creation version. This should correctly handle UTF8.
5080 *) Implement RFC5746. Re-enable renegotiation but require the extension
5081 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5082 turns out to be a bad idea. It has been replaced by
5083 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5084 SSL_CTX_set_options(). This is really not recommended unless you
5085 know what you are doing.
5086 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5088 *) Fixes to stateless session resumption handling. Use initial_ctx when
5089 issuing and attempting to decrypt tickets in case it has changed during
5090 servername handling. Use a non-zero length session ID when attempting
5091 stateless session resumption: this makes it possible to determine if
5092 a resumption has occurred immediately after receiving server hello
5093 (several places in OpenSSL subtly assume this) instead of later in
5097 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5098 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5099 fixes for a few places where the return code is not checked
5101 [Julia Lawall <julia@diku.dk>]
5103 *) Add --strict-warnings option to Configure script to include devteam
5104 warnings in other configurations.
5107 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5108 makes it possible to install openssl libraries in locations which
5109 have names other than "lib", for example "/usr/lib64" which some
5111 [Steve Henson, based on patch from Jeremy Utley]
5113 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5114 X690 8.9.12 and can produce some misleading textual output of OIDs.
5115 [Steve Henson, reported by Dan Kaminsky]
5117 *) Delete MD2 from algorithm tables. This follows the recommendation in
5118 several standards that it is not used in new applications due to
5119 several cryptographic weaknesses. For binary compatibility reasons
5120 the MD2 API is still compiled in by default.
5123 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5127 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5128 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5130 [Guenter <lists@gknw.net>]
5132 *) Fix the server certificate chain building code to use X509_verify_cert(),
5133 it used to have an ad-hoc builder which was unable to cope with anything
5134 other than a simple chain.
5135 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5137 *) Don't check self signed certificate signatures in X509_verify_cert()
5138 by default (a flag can override this): it just wastes time without
5139 adding any security. As a useful side effect self signed root CAs
5140 with non-FIPS digests are now usable in FIPS mode.
5143 *) In dtls1_process_out_of_seq_message() the check if the current message
5144 is already buffered was missing. For every new message was memory
5145 allocated, allowing an attacker to perform an denial of service attack
5146 with sending out of seq handshake messages until there is no memory
5147 left. Additionally every future message was buffered, even if the
5148 sequence number made no sense and would be part of another handshake.
5149 So only messages with sequence numbers less than 10 in advance will be
5150 buffered. (CVE-2009-1378)
5151 [Robin Seggelmann, discovered by Daniel Mentz]
5153 *) Records are buffered if they arrive with a future epoch to be
5154 processed after finishing the corresponding handshake. There is
5155 currently no limitation to this buffer allowing an attacker to perform
5156 a DOS attack with sending records with future epochs until there is no
5157 memory left. This patch adds the pqueue_size() function to determine
5158 the size of a buffer and limits the record buffer to 100 entries.
5160 [Robin Seggelmann, discovered by Daniel Mentz]
5162 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5163 parent structure is freed. (CVE-2009-1379)
5166 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5167 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5170 [Ilya O. <vrghost@gmail.com>]
5172 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5174 *) Disable renegotiation completely - this fixes a severe security
5175 problem (CVE-2009-3555) at the cost of breaking all
5176 renegotiation. Renegotiation can be re-enabled by setting
5177 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5178 run-time. This is really not recommended unless you know what
5182 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5184 *) Don't set val to NULL when freeing up structures, it is freed up by
5185 underlying code. If sizeof(void *) > sizeof(long) this can result in
5186 zeroing past the valid field. (CVE-2009-0789)
5187 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5189 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5190 checked correctly. This would allow some invalid signed attributes to
5191 appear to verify correctly. (CVE-2009-0591)
5192 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5194 *) Reject UniversalString and BMPString types with invalid lengths. This
5195 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5196 a legal length. (CVE-2009-0590)
5199 *) Set S/MIME signing as the default purpose rather than setting it
5200 unconditionally. This allows applications to override it at the store
5204 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5205 to handle some structures.
5208 *) Improve efficiency of mem_gets: don't search whole buffer each time
5210 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5212 *) New -hex option for openssl rand.
5215 *) Print out UTF8String and NumericString when parsing ASN1.
5218 *) Support NumericString type for name components.
5221 *) Allow CC in the environment to override the automatically chosen
5222 compiler. Note that nothing is done to ensure flags work with the
5226 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5228 *) Properly check EVP_VerifyFinal() and similar return values
5230 [Ben Laurie, Bodo Moeller, Google Security Team]
5232 *) Enable TLS extensions by default.
5235 *) Allow the CHIL engine to be loaded, whether the application is
5236 multithreaded or not. (This does not release the developer from the
5237 obligation to set up the dynamic locking callbacks.)
5238 [Sander Temme <sander@temme.net>]
5240 *) Use correct exit code if there is an error in dgst command.
5241 [Steve Henson; problem pointed out by Roland Dirlewanger]
5243 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5244 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5247 *) Add experimental JPAKE support, including demo authentication in
5248 s_client and s_server.
5251 *) Set the comparison function in v3_addr_canonize().
5252 [Rob Austein <sra@hactrn.net>]
5254 *) Add support for XMPP STARTTLS in s_client.
5255 [Philip Paeps <philip@freebsd.org>]
5257 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5258 to ensure that even with this option, only ciphersuites in the
5259 server's preference list will be accepted. (Note that the option
5260 applies only when resuming a session, so the earlier behavior was
5261 just about the algorithm choice for symmetric cryptography.)
5264 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5266 *) Fix NULL pointer dereference if a DTLS server received
5267 ChangeCipherSpec as first record (CVE-2009-1386).
5270 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5271 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5274 *) The fix in 0.9.8c that supposedly got rid of unsafe
5275 double-checked locking was incomplete for RSA blinding,
5276 addressing just one layer of what turns out to have been
5277 doubly unsafe triple-checked locking.
5279 So now fix this for real by retiring the MONT_HELPER macro
5280 in crypto/rsa/rsa_eay.c.
5282 [Bodo Moeller; problem pointed out by Marius Schilder]
5284 *) Various precautionary measures:
5286 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5288 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5289 (NB: This would require knowledge of the secret session ticket key
5290 to exploit, in which case you'd be SOL either way.)
5292 - Change bn_nist.c so that it will properly handle input BIGNUMs
5293 outside the expected range.
5295 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5298 [Neel Mehta, Bodo Moeller]
5300 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5301 the load fails. Useful for distros.
5302 [Ben Laurie and the FreeBSD team]
5304 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5307 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5310 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5312 This work was sponsored by Logica.
5315 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5316 keystores. Support for SSL/TLS client authentication too.
5317 Not compiled unless enable-capieng specified to Configure.
5319 This work was sponsored by Logica.
5322 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5323 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5324 attribute creation routines such as certificate requests and PKCS#12
5328 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5330 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5331 handshake which could lead to a client crash as found using the
5332 Codenomicon TLS test suite (CVE-2008-1672)
5333 [Steve Henson, Mark Cox]
5335 *) Fix double free in TLS server name extensions which could lead to
5336 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5339 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5341 Clear the error queue to ensure that error entries left from
5342 older function calls do not interfere with the correct operation.
5343 [Lutz Jaenicke, Erik de Castro Lopo]
5345 *) Remove root CA certificates of commercial CAs:
5347 The OpenSSL project does not recommend any specific CA and does not
5348 have any policy with respect to including or excluding any CA.
5349 Therefore it does not make any sense to ship an arbitrary selection
5350 of root CA certificates with the OpenSSL software.
5353 *) RSA OAEP patches to fix two separate invalid memory reads.
5354 The first one involves inputs when 'lzero' is greater than
5355 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5356 before the beginning of from). The second one involves inputs where
5357 the 'db' section contains nothing but zeroes (there is a one-byte
5358 invalid read after the end of 'db').
5359 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5361 *) Partial backport from 0.9.9-dev:
5363 Introduce bn_mul_mont (dedicated Montgomery multiplication
5364 procedure) as a candidate for BIGNUM assembler implementation.
5365 While 0.9.9-dev uses assembler for various architectures, only
5366 x86_64 is available by default here in the 0.9.8 branch, and
5367 32-bit x86 is available through a compile-time setting.
5369 To try the 32-bit x86 assembler implementation, use Configure
5370 option "enable-montasm" (which exists only for this backport).
5372 As "enable-montasm" for 32-bit x86 disclaims code stability
5373 anyway, in this constellation we activate additional code
5374 backported from 0.9.9-dev for further performance improvements,
5375 namely BN_from_montgomery_word. (To enable this otherwise,
5376 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5378 [Andy Polyakov (backport partially by Bodo Moeller)]
5380 *) Add TLS session ticket callback. This allows an application to set
5381 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5382 values. This is useful for key rollover for example where several key
5383 sets may exist with different names.
5386 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5387 This was broken until now in 0.9.8 releases, such that the only way
5388 a registered ENGINE could be used (assuming it initialises
5389 successfully on the host) was to explicitly set it as the default
5390 for the relevant algorithms. This is in contradiction with 0.9.7
5391 behaviour and the documentation. With this fix, when an ENGINE is
5392 registered into a given algorithm's table of implementations, the
5393 'uptodate' flag is reset so that auto-discovery will be used next
5394 time a new context for that algorithm attempts to select an
5396 [Ian Lister (tweaked by Geoff Thorpe)]
5398 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5399 implementation in the following ways:
5401 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5404 Lack of BER streaming support means one pass streaming processing is
5405 only supported if data is detached: setting the streaming flag is
5406 ignored for embedded content.
5408 CMS support is disabled by default and must be explicitly enabled
5409 with the enable-cms configuration option.
5412 *) Update the GMP engine glue to do direct copies between BIGNUM and
5413 mpz_t when openssl and GMP use the same limb size. Otherwise the
5414 existing "conversion via a text string export" trick is still used.
5415 [Paul Sheer <paulsheer@gmail.com>]
5417 *) Zlib compression BIO. This is a filter BIO which compressed and
5418 uncompresses any data passed through it.
5421 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5422 RFC3394 compatible AES key wrapping.
5425 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5426 sets string data without copying. X509_ALGOR_set0() and
5427 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5428 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5429 from an X509_ATTRIBUTE structure optionally checking it occurs only
5430 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5434 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5435 to get the expected BN_FLG_CONSTTIME behavior.
5436 [Bodo Moeller (Google)]
5440 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5441 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5442 - added some more tests to do_tests.pl
5443 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5444 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5445 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5446 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5447 - various changes to netware.pl to enable gcc-cross builds on Win32
5449 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5450 - various changes to fix missing prototype warnings
5451 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5452 - added AES, WHIRLPOOL and CPUID assembler code to build files
5453 - added missing AES assembler make rules to mk1mf.pl
5454 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5455 [Guenter Knauf <eflash@gmx.net>]
5457 *) Implement certificate status request TLS extension defined in RFC3546.
5458 A client can set the appropriate parameters and receive the encoded
5459 OCSP response via a callback. A server can query the supplied parameters
5460 and set the encoded OCSP response in the callback. Add simplified examples
5461 to s_client and s_server.
5464 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5466 *) Fix various bugs:
5467 + Binary incompatibility of ssl_ctx_st structure
5468 + DTLS interoperation with non-compliant servers
5469 + Don't call get_session_cb() without proposed session
5470 + Fix ia64 assembler code
5471 [Andy Polyakov, Steve Henson]
5473 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5475 *) DTLS Handshake overhaul. There were longstanding issues with
5476 OpenSSL DTLS implementation, which were making it impossible for
5477 RFC 4347 compliant client to communicate with OpenSSL server.
5478 Unfortunately just fixing these incompatibilities would "cut off"
5479 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5480 server keeps tolerating non RFC compliant syntax. The opposite is
5481 not true, 0.9.8f client can not communicate with earlier server.
5482 This update even addresses CVE-2007-4995.
5485 *) Changes to avoid need for function casts in OpenSSL: some compilers
5486 (gcc 4.2 and later) reject their use.
5487 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5490 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5491 RFC4507bis. The encrypted ticket format is an encrypted encoded
5492 SSL_SESSION structure, that way new session features are automatically
5495 If a client application caches session in an SSL_SESSION structure
5496 support is transparent because tickets are now stored in the encoded
5499 The SSL_CTX structure automatically generates keys for ticket
5500 protection in servers so again support should be possible
5501 with no application modification.
5503 If a client or server wishes to disable RFC4507 support then the option
5504 SSL_OP_NO_TICKET can be set.
5506 Add a TLS extension debugging callback to allow the contents of any client
5507 or server extensions to be examined.
5509 This work was sponsored by Google.
5512 *) Add initial support for TLS extensions, specifically for the server_name
5513 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5514 have new members for a host name. The SSL data structure has an
5515 additional member SSL_CTX *initial_ctx so that new sessions can be
5516 stored in that context to allow for session resumption, even after the
5517 SSL has been switched to a new SSL_CTX in reaction to a client's
5518 server_name extension.
5520 New functions (subject to change):
5522 SSL_get_servername()
5523 SSL_get_servername_type()
5526 New CTRL codes and macros (subject to change):
5528 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5529 - SSL_CTX_set_tlsext_servername_callback()
5530 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5531 - SSL_CTX_set_tlsext_servername_arg()
5532 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5534 openssl s_client has a new '-servername ...' option.
5536 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5537 '-key2 ...', '-servername_fatal' (subject to change). This allows
5538 testing the HostName extension for a specific single host name ('-cert'
5539 and '-key' remain fallbacks for handshakes without HostName
5540 negotiation). If the unrecognized_name alert has to be sent, this by
5541 default is a warning; it becomes fatal with the '-servername_fatal'
5544 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5546 *) Add AES and SSE2 assembly language support to VC++ build.
5549 *) Mitigate attack on final subtraction in Montgomery reduction.
5552 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5553 (which previously caused an internal error).
5556 *) Squeeze another 10% out of IGE mode when in != out.
5559 *) AES IGE mode speedup.
5560 [Dean Gaudet (Google)]
5562 *) Add the Korean symmetric 128-bit cipher SEED (see
5563 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5564 add SEED ciphersuites from RFC 4162:
5566 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5567 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5568 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5569 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5571 To minimize changes between patchlevels in the OpenSSL 0.9.8
5572 series, SEED remains excluded from compilation unless OpenSSL
5573 is configured with 'enable-seed'.
5574 [KISA, Bodo Moeller]
5576 *) Mitigate branch prediction attacks, which can be practical if a
5577 single processor is shared, allowing a spy process to extract
5578 information. For detailed background information, see
5579 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5580 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5581 and Necessary Software Countermeasures"). The core of the change
5582 are new versions BN_div_no_branch() and
5583 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5584 respectively, which are slower, but avoid the security-relevant
5585 conditional branches. These are automatically called by BN_div()
5586 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5587 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5588 remove a conditional branch.
5590 BN_FLG_CONSTTIME is the new name for the previous
5591 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5592 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5593 in the exponent causes BN_mod_exp_mont() to use the alternative
5594 implementation in BN_mod_exp_mont_consttime().) The old name
5595 remains as a deprecated alias.
5597 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5598 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5599 constant-time implementations for more than just exponentiation.
5600 Here too the old name is kept as a deprecated alias.
5602 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5603 the BN_BLINDING structure gets an independent copy of the
5604 modulus. This means that the previous "BIGNUM *m" argument to
5605 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5606 essentially becomes "const BIGNUM *m", although we can't actually
5607 change this in the header file before 0.9.9. It allows
5608 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5609 enable BN_FLG_CONSTTIME.
5611 [Matthew D Wood (Intel Corp)]
5613 *) In the SSL/TLS server implementation, be strict about session ID
5614 context matching (which matters if an application uses a single
5615 external cache for different purposes). Previously,
5616 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5617 set. This did ensure strict client verification, but meant that,
5618 with applications using a single external cache for quite
5619 different requirements, clients could circumvent ciphersuite
5620 restrictions for a given session ID context by starting a session
5621 in a different context.
5624 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5625 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5626 authentication-only ciphersuites.
5629 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5630 not complete and could lead to a possible single byte overflow
5631 (CVE-2007-5135) [Ben Laurie]
5633 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5635 *) Since AES128 and AES256 (and similarly Camellia128 and
5636 Camellia256) share a single mask bit in the logic of
5637 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5638 kludge to work properly if AES128 is available and AES256 isn't
5639 (or if Camellia128 is available and Camellia256 isn't).
5642 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5643 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5644 When a point or a seed is encoded in a BIT STRING, we need to
5645 prevent the removal of trailing zero bits to get the proper DER
5646 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5647 of a NamedBitList, for which trailing 0 bits need to be removed.)
5650 *) Have SSL/TLS server implementation tolerate "mismatched" record
5651 protocol version while receiving ClientHello even if the
5652 ClientHello is fragmented. (The server can't insist on the
5653 particular protocol version it has chosen before the ServerHello
5654 message has informed the client about his choice.)
5657 *) Add RFC 3779 support.
5658 [Rob Austein for ARIN, Ben Laurie]
5660 *) Load error codes if they are not already present instead of using a
5661 static variable. This allows them to be cleanly unloaded and reloaded.
5662 Improve header file function name parsing.
5665 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5666 or CAPABILITY handshake as required by RFCs.
5669 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5671 *) Introduce limits to prevent malicious keys being able to
5672 cause a denial of service. (CVE-2006-2940)
5673 [Steve Henson, Bodo Moeller]
5675 *) Fix ASN.1 parsing of certain invalid structures that can result
5676 in a denial of service. (CVE-2006-2937) [Steve Henson]
5678 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5679 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5681 *) Fix SSL client code which could crash if connecting to a
5682 malicious SSLv2 server. (CVE-2006-4343)
5683 [Tavis Ormandy and Will Drewry, Google Security Team]
5685 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5686 match only those. Before that, "AES256-SHA" would be interpreted
5687 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5688 the same strength classification in 0.9.7h) as we currently only
5689 have a single AES bit in the ciphersuite description bitmap.
5690 That change, however, also applied to ciphersuite strings such as
5691 "RC4-MD5" that intentionally matched multiple ciphersuites --
5692 namely, SSL 2.0 ciphersuites in addition to the more common ones
5693 from SSL 3.0/TLS 1.0.
5695 So we change the selection algorithm again: Naming an explicit
5696 ciphersuite selects this one ciphersuite, and any other similar
5697 ciphersuite (same bitmap) from *other* protocol versions.
5698 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5699 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5701 Since SSL 2.0 does not have any ciphersuites for which the
5702 128/256 bit distinction would be relevant, this works for now.
5703 The proper fix will be to use different bits for AES128 and
5704 AES256, which would have avoided the problems from the beginning;
5705 however, bits are scarce, so we can only do this in a new release
5706 (not just a patchlevel) when we can change the SSL_CIPHER
5707 definition to split the single 'unsigned long mask' bitmap into
5708 multiple values to extend the available space.
5712 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5714 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5715 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5717 *) Add AES IGE and biIGE modes.
5720 *) Change the Unix randomness entropy gathering to use poll() when
5721 possible instead of select(), since the latter has some
5722 undesirable limitations.
5723 [Darryl Miles via Richard Levitte and Bodo Moeller]
5725 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5726 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5727 cannot be implicitly activated as part of, e.g., the "AES" alias.
5728 However, please upgrade to OpenSSL 0.9.9[-dev] for
5729 non-experimental use of the ECC ciphersuites to get TLS extension
5730 support, which is required for curve and point format negotiation
5731 to avoid potential handshake problems.
5734 *) Disable rogue ciphersuites:
5736 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5737 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5738 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5740 The latter two were purportedly from
5741 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5744 Also deactivate the remaining ciphersuites from
5745 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5746 unofficial, and the ID has long expired.
5749 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5750 dual-core machines) and other potential thread-safety issues.
5753 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5754 versions), which is now available for royalty-free use
5755 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5756 Also, add Camellia TLS ciphersuites from RFC 4132.
5758 To minimize changes between patchlevels in the OpenSSL 0.9.8
5759 series, Camellia remains excluded from compilation unless OpenSSL
5760 is configured with 'enable-camellia'.
5763 *) Disable the padding bug check when compression is in use. The padding
5764 bug check assumes the first packet is of even length, this is not
5765 necessarily true if compression is enabled and can result in false
5766 positives causing handshake failure. The actual bug test is ancient
5767 code so it is hoped that implementations will either have fixed it by
5768 now or any which still have the bug do not support compression.
5771 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5773 *) When applying a cipher rule check to see if string match is an explicit
5774 cipher suite and only match that one cipher suite if it is.
5777 *) Link in manifests for VC++ if needed.
5778 [Austin Ziegler <halostatue@gmail.com>]
5780 *) Update support for ECC-based TLS ciphersuites according to
5781 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5782 TLS extensions, which are supported starting with the 0.9.9
5783 branch, not in the OpenSSL 0.9.8 branch).
5786 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5787 opaque EVP_CIPHER_CTX handling.
5790 *) Fixes and enhancements to zlib compression code. We now only use
5791 "zlib1.dll" and use the default __cdecl calling convention on Win32
5792 to conform with the standards mentioned here:
5793 http://www.zlib.net/DLL_FAQ.txt
5794 Static zlib linking now works on Windows and the new --with-zlib-include
5795 --with-zlib-lib options to Configure can be used to supply the location
5796 of the headers and library. Gracefully handle case where zlib library
5800 *) Several fixes and enhancements to the OID generation code. The old code
5801 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5802 handle numbers larger than ULONG_MAX, truncated printing and had a
5803 non standard OBJ_obj2txt() behaviour.
5806 *) Add support for building of engines under engine/ as shared libraries
5807 under VC++ build system.
5810 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5811 Hopefully, we will not see any false combination of paths any more.
5814 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5816 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5817 (part of SSL_OP_ALL). This option used to disable the
5818 countermeasure against man-in-the-middle protocol-version
5819 rollback in the SSL 2.0 server implementation, which is a bad
5820 idea. (CVE-2005-2969)
5822 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5823 for Information Security, National Institute of Advanced Industrial
5824 Science and Technology [AIST], Japan)]
5826 *) Add two function to clear and return the verify parameter flags.
5829 *) Keep cipherlists sorted in the source instead of sorting them at
5830 runtime, thus removing the need for a lock.
5833 *) Avoid some small subgroup attacks in Diffie-Hellman.
5834 [Nick Mathewson and Ben Laurie]
5836 *) Add functions for well-known primes.
5839 *) Extended Windows CE support.
5840 [Satoshi Nakamura and Andy Polyakov]
5842 *) Initialize SSL_METHOD structures at compile time instead of during
5843 runtime, thus removing the need for a lock.
5846 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5847 attempting to decrypt each encrypted key in turn. Add support to
5851 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5853 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5856 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5859 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5860 key into the same file any more.
5863 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5866 *) Add -utf8 command line and config file option to 'ca'.
5867 [Stefan <stf@udoma.org]
5869 *) Removed the macro des_crypt(), as it seems to conflict with some
5870 libraries. Use DES_crypt().
5873 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5874 involves renaming the source and generated shared-libs for
5875 both. The engines will accept the corrected or legacy ids
5876 ('ncipher' and '4758_cca' respectively) when binding. NB,
5877 this only applies when building 'shared'.
5878 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5880 *) Add attribute functions to EVP_PKEY structure. Modify
5881 PKCS12_create() to recognize a CSP name attribute and
5882 use it. Make -CSP option work again in pkcs12 utility.
5885 *) Add new functionality to the bn blinding code:
5886 - automatic re-creation of the BN_BLINDING parameters after
5887 a fixed number of uses (currently 32)
5888 - add new function for parameter creation
5889 - introduce flags to control the update behaviour of the
5890 BN_BLINDING parameters
5891 - hide BN_BLINDING structure
5892 Add a second BN_BLINDING slot to the RSA structure to improve
5893 performance when a single RSA object is shared among several
5897 *) Add support for DTLS.
5898 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5900 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5901 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5904 *) Remove buggy and incomplete DH cert support from
5905 ssl/ssl_rsa.c and ssl/s3_both.c
5908 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5909 the apps/openssl applications.
5912 *) Compile clean with "-Wall -Wmissing-prototypes
5913 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5914 DEBUG_SAFESTACK must also be set.
5917 *) Change ./Configure so that certain algorithms can be disabled by default.
5918 The new counterpiece to "no-xxx" is "enable-xxx".
5920 The patented RC5 and MDC2 algorithms will now be disabled unless
5921 "enable-rc5" and "enable-mdc2", respectively, are specified.
5923 (IDEA remains enabled despite being patented. This is because IDEA
5924 is frequently required for interoperability, and there is no license
5925 fee for non-commercial use. As before, "no-idea" can be used to
5926 avoid this algorithm.)
5930 *) Add processing of proxy certificates (see RFC 3820). This work was
5931 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5932 EGEE (Enabling Grids for E-science in Europe).
5935 *) RC4 performance overhaul on modern architectures/implementations, such
5936 as Intel P4, IA-64 and AMD64.
5939 *) New utility extract-section.pl. This can be used specify an alternative
5940 section number in a pod file instead of having to treat each file as
5941 a separate case in Makefile. This can be done by adding two lines to the
5944 =for comment openssl_section:XXX
5946 The blank line is mandatory.
5950 *) New arguments -certform, -keyform and -pass for s_client and s_server
5951 to allow alternative format key and certificate files and passphrase
5955 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5956 update associated structures and add various utility functions.
5958 Add new policy related verify parameters, include policy checking in
5959 standard verify code. Enhance 'smime' application with extra parameters
5960 to support policy checking and print out.
5963 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5964 Nehemiah processors. These extensions support AES encryption in hardware
5965 as well as RNG (though RNG support is currently disabled).
5966 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5968 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5971 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5972 [Andy Polyakov and a number of other people]
5974 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5975 implementation contributed by IBM.
5976 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5978 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5979 exponent rather than 'unsigned long'. There is a corresponding change to
5980 the new 'rsa_keygen' element of the RSA_METHOD structure.
5981 [Jelte Jansen, Geoff Thorpe]
5983 *) Functionality for creating the initial serial number file is now
5984 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5986 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5987 number file to 1, which is bound to cause problems. To avoid
5988 the problems while respecting compatibility between different 0.9.7
5989 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5990 CA.pl for serial number initialization. With the new release 0.9.8,
5991 we can fix the problem directly in the 'ca' utility.)
5994 *) Reduced header interdependencies by declaring more opaque objects in
5995 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5996 give fewer recursive includes, which could break lazy source code - so
5997 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5998 developers should define this symbol when building and using openssl to
5999 ensure they track the recommended behaviour, interfaces, [etc], but
6000 backwards-compatible behaviour prevails when this isn't defined.
6003 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
6006 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
6007 This will generate a random key of the appropriate length based on the
6008 cipher context. The EVP_CIPHER can provide its own random key generation
6009 routine to support keys of a specific form. This is used in the des and
6010 3des routines to generate a key of the correct parity. Update S/MIME
6011 code to use new functions and hence generate correct parity DES keys.
6012 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
6013 valid (weak or incorrect parity).
6016 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
6017 as looking them up. This is useful when the verified structure may contain
6018 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
6019 present unless the new PKCS7_NO_CRL flag is asserted.
6022 *) Extend ASN1 oid configuration module. It now additionally accepts the
6025 shortName = some long name, 1.2.3.4
6028 *) Reimplemented the BN_CTX implementation. There is now no more static
6029 limitation on the number of variables it can handle nor the depth of the
6030 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
6031 information can now expand as required, and rather than having a single
6032 static array of bignums, BN_CTX now uses a linked-list of such arrays
6033 allowing it to expand on demand whilst maintaining the usefulness of
6034 BN_CTX's "bundling".
6037 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
6038 to allow all RSA operations to function using a single BN_CTX.
6041 *) Preliminary support for certificate policy evaluation and checking. This
6042 is initially intended to pass the tests outlined in "Conformance Testing
6043 of Relying Party Client Certificate Path Processing Logic" v1.07.
6046 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6047 remained unused and not that useful. A variety of other little bignum
6048 tweaks and fixes have also been made continuing on from the audit (see
6052 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6053 associated ASN1, EVP and SSL functions and old ASN1 macros.
6056 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6057 and this should never fail. So the return value from the use of
6058 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6059 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6062 *) BN_CTX_get() should return zero-valued bignums, providing the same
6063 initialised value as BN_new().
6064 [Geoff Thorpe, suggested by Ulf Möller]
6066 *) Support for inhibitAnyPolicy certificate extension.
6069 *) An audit of the BIGNUM code is underway, for which debugging code is
6070 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6071 is considered valid when processing BIGNUMs, and causes execution to
6072 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6073 further steps are taken to deliberately pollute unused data in BIGNUM
6074 structures to try and expose faulty code further on. For now, openssl will
6075 (in its default mode of operation) continue to tolerate the inconsistent
6076 forms that it has tolerated in the past, but authors and packagers should
6077 consider trying openssl and their own applications when compiled with
6078 these debugging symbols defined. It will help highlight potential bugs in
6079 their own code, and will improve the test coverage for OpenSSL itself. At
6080 some point, these tighter rules will become openssl's default to improve
6081 maintainability, though the assert()s and other overheads will remain only
6082 in debugging configurations. See bn.h for more details.
6083 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6085 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6086 that can only be obtained through BN_CTX_new() (which implicitly
6087 initialises it). The presence of this function only made it possible
6088 to overwrite an existing structure (and cause memory leaks).
6091 *) Because of the callback-based approach for implementing LHASH as a
6092 template type, lh_insert() adds opaque objects to hash-tables and
6093 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6094 to clean up those corresponding objects before destroying the hash table
6095 (and losing the object pointers). So some over-zealous constifications in
6096 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6097 objects as "const" and the lh_doall[_arg] callback wrappers are not
6098 prototyped to have "const" restrictions on the object pointers they are
6099 given (and so aren't required to cast them away any more).
6102 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6103 (speed) prefers to use its own implementation. The two implementations
6104 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6105 its object type properly exposed (MS_TM) instead of casting to/from "char
6106 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6107 aren't necessarily the greatest nomenclatures - but this is what was used
6108 internally to the implementation so I've used that for now.
6111 *) Ensure that deprecated functions do not get compiled when
6112 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6113 the self-tests were still using deprecated key-generation functions so
6114 these have been updated also.
6117 *) Reorganise PKCS#7 code to separate the digest location functionality
6118 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6119 New function PKCS7_set_digest() to set the digest type for PKCS#7
6120 digestedData type. Add additional code to correctly generate the
6121 digestedData type and add support for this type in PKCS7 initialization
6125 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6126 structure of type "other".
6129 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6130 sure the loop does correctly stop and breaking ("division by zero")
6131 modulus operations are not performed. The (pre-generated) prime
6132 table crypto/bn/bn_prime.h was already correct, but it could not be
6133 re-generated on some platforms because of the "division by zero"
6134 situation in the script.
6135 [Ralf S. Engelschall]
6137 *) Update support for ECC-based TLS ciphersuites according to
6138 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6139 SHA-1 now is only used for "small" curves (where the
6140 representation of a field element takes up to 24 bytes); for
6141 larger curves, the field element resulting from ECDH is directly
6142 used as premaster secret.
6143 [Douglas Stebila (Sun Microsystems Laboratories)]
6145 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6146 curve secp160r1 to the tests.
6147 [Douglas Stebila (Sun Microsystems Laboratories)]
6149 *) Add the possibility to load symbols globally with DSO.
6150 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6152 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6153 control of the error stack.
6156 *) Add support for STORE in ENGINE.
6159 *) Add the STORE type. The intention is to provide a common interface
6160 to certificate and key stores, be they simple file-based stores, or
6161 HSM-type store, or LDAP stores, or...
6162 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6165 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6166 pass a list of arguments to any function as well as provide a way
6167 for a function to pass data back to the caller.
6170 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6171 works like BUF_strdup() but can be used to duplicate a portion of
6172 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6176 *) Add the function sk_find_ex() which works like sk_find(), but will
6177 return an index to an element even if an exact match couldn't be
6178 found. The index is guaranteed to point at the element where the
6179 searched-for key would be inserted to preserve sorting order.
6182 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6183 takes an extra flags argument for optional functionality. Currently,
6184 the following flags are defined:
6186 OBJ_BSEARCH_VALUE_ON_NOMATCH
6187 This one gets OBJ_bsearch_ex() to return a pointer to the first
6188 element where the comparing function returns a negative or zero
6191 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6192 This one gets OBJ_bsearch_ex() to return a pointer to the first
6193 element where the comparing function returns zero. This is useful
6194 if there are more than one element where the comparing function
6198 *) Make it possible to create self-signed certificates with 'openssl ca'
6199 in such a way that the self-signed certificate becomes part of the
6200 CA database and uses the same mechanisms for serial number generation
6201 as all other certificate signing. The new flag '-selfsign' enables
6202 this functionality. Adapt CA.sh and CA.pl.in.
6205 *) Add functionality to check the public key of a certificate request
6206 against a given private. This is useful to check that a certificate
6207 request can be signed by that key (self-signing).
6210 *) Make it possible to have multiple active certificates with the same
6211 subject in the CA index file. This is done only if the keyword
6212 'unique_subject' is set to 'no' in the main CA section (default
6213 if 'CA_default') of the configuration file. The value is saved
6214 with the database itself in a separate index attribute file,
6215 named like the index file with '.attr' appended to the name.
6218 *) Generate multi-valued AVAs using '+' notation in config files for
6222 *) Support for nameConstraints certificate extension.
6225 *) Support for policyConstraints certificate extension.
6228 *) Support for policyMappings certificate extension.
6231 *) Make sure the default DSA_METHOD implementation only uses its
6232 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6233 and change its own handlers to be NULL so as to remove unnecessary
6234 indirection. This lets alternative implementations fallback to the
6235 default implementation more easily.
6238 *) Support for directoryName in GeneralName related extensions
6242 *) Make it possible to link applications using Makefile.shared.
6243 Make that possible even when linking against static libraries!
6246 *) Support for single pass processing for S/MIME signing. This now
6247 means that S/MIME signing can be done from a pipe, in addition
6248 cleartext signing (multipart/signed type) is effectively streaming
6249 and the signed data does not need to be all held in memory.
6251 This is done with a new flag PKCS7_STREAM. When this flag is set
6252 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6253 is done after the data is output (and digests calculated) in
6254 SMIME_write_PKCS7().
6257 *) Add full support for -rpath/-R, both in shared libraries and
6258 applications, at least on the platforms where it's known how
6262 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6263 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6264 will now compute a table of multiples of the generator that
6265 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6266 faster (notably in the case of a single point multiplication,
6267 scalar * generator).
6268 [Nils Larsch, Bodo Moeller]
6270 *) IPv6 support for certificate extensions. The various extensions
6271 which use the IP:a.b.c.d can now take IPv6 addresses using the
6272 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6276 *) Added an ENGINE that implements RSA by performing private key
6277 exponentiations with the GMP library. The conversions to and from
6278 GMP's mpz_t format aren't optimised nor are any montgomery forms
6279 cached, and on x86 it appears OpenSSL's own performance has caught up.
6280 However there are likely to be other architectures where GMP could
6281 provide a boost. This ENGINE is not built in by default, but it can be
6282 specified at Configure time and should be accompanied by the necessary
6283 linker additions, eg;
6284 ./config -DOPENSSL_USE_GMP -lgmp
6287 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6288 testing availability of engines with "-t" - the old behaviour is
6289 produced by increasing the feature's verbosity with "-tt".
6292 *) ECDSA routines: under certain error conditions uninitialized BN objects
6293 could be freed. Solution: make sure initialization is performed early
6294 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6298 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6299 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6300 software implementations. For DSA and DH, parameter generation can
6301 also be overridden by providing the appropriate method callbacks.
6304 *) Change the "progress" mechanism used in key-generation and
6305 primality testing to functions that take a new BN_GENCB pointer in
6306 place of callback/argument pairs. The new API functions have "_ex"
6307 postfixes and the older functions are reimplemented as wrappers for
6308 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6309 declarations of the old functions to help (graceful) attempts to
6310 migrate to the new functions. Also, the new key-generation API
6311 functions operate on a caller-supplied key-structure and return
6312 success/failure rather than returning a key or NULL - this is to
6313 help make "keygen" another member function of RSA_METHOD etc.
6315 Example for using the new callback interface:
6317 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6321 BN_GENCB_set(&my_cb, my_callback, my_arg);
6323 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6324 /* For the meaning of a, b in calls to my_callback(), see the
6325 * documentation of the function that calls the callback.
6326 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6327 * my_callback should return 1 if it wants BN_is_prime_ex()
6328 * to continue, or 0 to stop.
6333 *) Change the ZLIB compression method to be stateful, and make it
6334 available to TLS with the number defined in
6335 draft-ietf-tls-compression-04.txt.
6338 *) Add the ASN.1 structures and functions for CertificatePair, which
6339 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6341 CertificatePair ::= SEQUENCE {
6342 forward [0] Certificate OPTIONAL,
6343 reverse [1] Certificate OPTIONAL,
6344 -- at least one of the pair shall be present -- }
6346 Also implement the PEM functions to read and write certificate
6347 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6349 This needed to be defined, mostly for the sake of the LDAP
6350 attribute crossCertificatePair, but may prove useful elsewhere as
6354 *) Make it possible to inhibit symlinking of shared libraries in
6355 Makefile.shared, for Cygwin's sake.
6358 *) Extend the BIGNUM API by creating a function
6359 void BN_set_negative(BIGNUM *a, int neg);
6360 and a macro that behave like
6361 int BN_is_negative(const BIGNUM *a);
6363 to avoid the need to access 'a->neg' directly in applications.
6366 *) Implement fast modular reduction for pseudo-Mersenne primes
6367 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6368 EC_GROUP_new_curve_GFp() will now automatically use this
6370 [Nils Larsch <nla@trustcenter.de>]
6372 *) Add new lock type (CRYPTO_LOCK_BN).
6375 *) Change the ENGINE framework to automatically load engines
6376 dynamically from specific directories unless they could be
6377 found to already be built in or loaded. Move all the
6378 current engines except for the cryptodev one to a new
6380 The engines in engines/ are built as shared libraries if
6381 the "shared" options was given to ./Configure or ./config.
6382 Otherwise, they are inserted in libcrypto.a.
6383 /usr/local/ssl/engines is the default directory for dynamic
6384 engines, but that can be overridden at configure time through
6385 the usual use of --prefix and/or --openssldir, and at run
6386 time with the environment variable OPENSSL_ENGINES.
6387 [Geoff Thorpe and Richard Levitte]
6389 *) Add Makefile.shared, a helper makefile to build shared
6390 libraries. Adapt Makefile.org.
6393 *) Add version info to Win32 DLLs.
6394 [Peter 'Luna' Runestig" <peter@runestig.com>]
6396 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6397 can be added using this API to created arbitrary PKCS#12
6398 files while avoiding the low level API.
6400 New options to PKCS12_create(), key or cert can be NULL and
6401 will then be omitted from the output file. The encryption
6402 algorithm NIDs can be set to -1 for no encryption, the mac
6403 iteration count can be set to 0 to omit the mac.
6405 Enhance pkcs12 utility by making the -nokeys and -nocerts
6406 options work when creating a PKCS#12 file. New option -nomac
6407 to omit the mac, NONE can be set for an encryption algorithm.
6408 New code is modified to use the enhanced PKCS12_create()
6409 instead of the low level API.
6412 *) Extend ASN1 encoder to support indefinite length constructed
6413 encoding. This can output sequences tags and octet strings in
6414 this form. Modify pk7_asn1.c to support indefinite length
6415 encoding. This is experimental and needs additional code to
6416 be useful, such as an ASN1 bio and some enhanced streaming
6419 Extend template encode functionality so that tagging is passed
6420 down to the template encoder.
6423 *) Let 'openssl req' fail if an argument to '-newkey' is not
6424 recognized instead of using RSA as a default.
6427 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6428 As these are not official, they are not included in "ALL";
6429 the "ECCdraft" ciphersuite group alias can be used to select them.
6430 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6432 *) Add ECDH engine support.
6433 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6435 *) Add ECDH in new directory crypto/ecdh/.
6436 [Douglas Stebila (Sun Microsystems Laboratories)]
6438 *) Let BN_rand_range() abort with an error after 100 iterations
6439 without success (which indicates a broken PRNG).
6442 *) Change BN_mod_sqrt() so that it verifies that the input value
6443 is really the square of the return value. (Previously,
6444 BN_mod_sqrt would show GIGO behaviour.)
6447 *) Add named elliptic curves over binary fields from X9.62, SECG,
6448 and WAP/WTLS; add OIDs that were still missing.
6450 [Sheueling Chang Shantz and Douglas Stebila
6451 (Sun Microsystems Laboratories)]
6453 *) Extend the EC library for elliptic curves over binary fields
6454 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6457 EC_GF2m_simple_method
6461 EC_GROUP_new_curve_GF2m
6462 EC_GROUP_set_curve_GF2m
6463 EC_GROUP_get_curve_GF2m
6464 EC_POINT_set_affine_coordinates_GF2m
6465 EC_POINT_get_affine_coordinates_GF2m
6466 EC_POINT_set_compressed_coordinates_GF2m
6468 Point compression for binary fields is disabled by default for
6469 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6472 As binary polynomials are represented as BIGNUMs, various members
6473 of the EC_GROUP and EC_POINT data structures can be shared
6474 between the implementations for prime fields and binary fields;
6475 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6476 are essentially identical to their ..._GFp counterparts.
6477 (For simplicity, the '..._GFp' prefix has been dropped from
6478 various internal method names.)
6480 An internal 'field_div' method (similar to 'field_mul' and
6481 'field_sqr') has been added; this is used only for binary fields.
6483 [Sheueling Chang Shantz and Douglas Stebila
6484 (Sun Microsystems Laboratories)]
6486 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6487 through methods ('mul', 'precompute_mult').
6489 The generic implementations (now internally called 'ec_wNAF_mul'
6490 and 'ec_wNAF_precomputed_mult') remain the default if these
6491 methods are undefined.
6493 [Sheueling Chang Shantz and Douglas Stebila
6494 (Sun Microsystems Laboratories)]
6496 *) New function EC_GROUP_get_degree, which is defined through
6497 EC_METHOD. For curves over prime fields, this returns the bit
6498 length of the modulus.
6500 [Sheueling Chang Shantz and Douglas Stebila
6501 (Sun Microsystems Laboratories)]
6503 *) New functions EC_GROUP_dup, EC_POINT_dup.
6504 (These simply call ..._new and ..._copy).
6506 [Sheueling Chang Shantz and Douglas Stebila
6507 (Sun Microsystems Laboratories)]
6509 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6510 Polynomials are represented as BIGNUMs (where the sign bit is not
6511 used) in the following functions [macros]:
6514 BN_GF2m_sub [= BN_GF2m_add]
6515 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6516 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6517 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6519 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6520 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6521 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6522 BN_GF2m_cmp [= BN_ucmp]
6524 (Note that only the 'mod' functions are actually for fields GF(2^m).
6525 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6527 For some functions, an the irreducible polynomial defining a
6528 field can be given as an 'unsigned int[]' with strictly
6529 decreasing elements giving the indices of those bits that are set;
6530 i.e., p[] represents the polynomial
6531 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6533 p[0] > p[1] > ... > p[k] = 0.
6534 This applies to the following functions:
6539 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6540 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6542 BN_GF2m_mod_sqrt_arr
6543 BN_GF2m_mod_solve_quad_arr
6547 Conversion can be performed by the following functions:
6552 bntest.c has additional tests for binary polynomial arithmetic.
6554 Two implementations for BN_GF2m_mod_div() are available.
6555 The default algorithm simply uses BN_GF2m_mod_inv() and
6556 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6557 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6558 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6560 [Sheueling Chang Shantz and Douglas Stebila
6561 (Sun Microsystems Laboratories)]
6563 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6564 functionality is disabled at compile-time.
6565 [Douglas Stebila <douglas.stebila@sun.com>]
6567 *) Change default behaviour of 'openssl asn1parse' so that more
6568 information is visible when viewing, e.g., a certificate:
6570 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6571 mode the content of non-printable OCTET STRINGs is output in a
6572 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6573 avoid the appearance of a printable string.
6574 [Nils Larsch <nla@trustcenter.de>]
6576 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6578 EC_GROUP_set_asn1_flag()
6579 EC_GROUP_get_asn1_flag()
6580 EC_GROUP_set_point_conversion_form()
6581 EC_GROUP_get_point_conversion_form()
6582 These control ASN1 encoding details:
6583 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6584 has been set to OPENSSL_EC_NAMED_CURVE.
6585 - Points are encoded in uncompressed form by default; options for
6586 asn1_for are as for point2oct, namely
6587 POINT_CONVERSION_COMPRESSED
6588 POINT_CONVERSION_UNCOMPRESSED
6589 POINT_CONVERSION_HYBRID
6591 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6594 EC_GROUP_get0_seed()
6595 EC_GROUP_get_seed_len()
6596 This is used only for ASN1 purposes (so far).
6597 [Nils Larsch <nla@trustcenter.de>]
6599 *) Add 'field_type' member to EC_METHOD, which holds the NID
6600 of the appropriate field type OID. The new function
6601 EC_METHOD_get_field_type() returns this value.
6602 [Nils Larsch <nla@trustcenter.de>]
6607 EC_POINT_point2hex()
6608 EC_POINT_hex2point()
6609 providing useful interfaces to EC_POINT_point2oct() and
6610 EC_POINT_oct2point().
6611 [Nils Larsch <nla@trustcenter.de>]
6613 *) Change internals of the EC library so that the functions
6614 EC_GROUP_set_generator()
6615 EC_GROUP_get_generator()
6616 EC_GROUP_get_order()
6617 EC_GROUP_get_cofactor()
6618 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6619 to methods, which would lead to unnecessary code duplication when
6620 adding different types of curves.
6621 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6623 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6624 arithmetic, and such that modified wNAFs are generated
6625 (which avoid length expansion in many cases).
6628 *) Add a function EC_GROUP_check_discriminant() (defined via
6629 EC_METHOD) that verifies that the curve discriminant is non-zero.
6631 Add a function EC_GROUP_check() that makes some sanity tests
6632 on a EC_GROUP, its generator and order. This includes
6633 EC_GROUP_check_discriminant().
6634 [Nils Larsch <nla@trustcenter.de>]
6636 *) Add ECDSA in new directory crypto/ecdsa/.
6638 Add applications 'openssl ecparam' and 'openssl ecdsa'
6639 (these are based on 'openssl dsaparam' and 'openssl dsa').
6641 ECDSA support is also included in various other files across the
6642 library. Most notably,
6643 - 'openssl req' now has a '-newkey ecdsa:file' option;
6644 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6645 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6646 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6647 them suitable for ECDSA where domain parameters must be
6648 extracted before the specific public key;
6649 - ECDSA engine support has been added.
6650 [Nils Larsch <nla@trustcenter.de>]
6652 *) Include some named elliptic curves, and add OIDs from X9.62,
6653 SECG, and WAP/WTLS. Each curve can be obtained from the new
6655 EC_GROUP_new_by_curve_name(),
6656 and the list of available named curves can be obtained with
6657 EC_get_builtin_curves().
6658 Also add a 'curve_name' member to EC_GROUP objects, which can be
6660 EC_GROUP_set_curve_name()
6661 EC_GROUP_get_curve_name()
6662 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6664 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6665 was actually never needed) and in BN_mul(). The removal in BN_mul()
6666 required a small change in bn_mul_part_recursive() and the addition
6667 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6668 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6669 bn_sub_words() and bn_add_words() except they take arrays with
6673 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6675 *) Cleanse PEM buffers before freeing them since they may contain
6677 [Benjamin Bennett <ben@psc.edu>]
6679 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6680 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6681 authentication-only ciphersuites.
6684 *) Since AES128 and AES256 share a single mask bit in the logic of
6685 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6686 kludge to work properly if AES128 is available and AES256 isn't.
6689 *) Expand security boundary to match 1.1.1 module.
6692 *) Remove redundant features: hash file source, editing of test vectors
6693 modify fipsld to use external fips_premain.c signature.
6696 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6697 run algorithm test programs.
6700 *) Make algorithm test programs more tolerant of whitespace.
6703 *) Have SSL/TLS server implementation tolerate "mismatched" record
6704 protocol version while receiving ClientHello even if the
6705 ClientHello is fragmented. (The server can't insist on the
6706 particular protocol version it has chosen before the ServerHello
6707 message has informed the client about his choice.)
6710 *) Load error codes if they are not already present instead of using a
6711 static variable. This allows them to be cleanly unloaded and reloaded.
6714 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6716 *) Introduce limits to prevent malicious keys being able to
6717 cause a denial of service. (CVE-2006-2940)
6718 [Steve Henson, Bodo Moeller]
6720 *) Fix ASN.1 parsing of certain invalid structures that can result
6721 in a denial of service. (CVE-2006-2937) [Steve Henson]
6723 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6724 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6726 *) Fix SSL client code which could crash if connecting to a
6727 malicious SSLv2 server. (CVE-2006-4343)
6728 [Tavis Ormandy and Will Drewry, Google Security Team]
6730 *) Change ciphersuite string processing so that an explicit
6731 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6732 will no longer include "AES128-SHA"), and any other similar
6733 ciphersuite (same bitmap) from *other* protocol versions (so that
6734 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6735 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6736 changes from 0.9.8b and 0.9.8d.
6739 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6741 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6742 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6744 *) Change the Unix randomness entropy gathering to use poll() when
6745 possible instead of select(), since the latter has some
6746 undesirable limitations.
6747 [Darryl Miles via Richard Levitte and Bodo Moeller]
6749 *) Disable rogue ciphersuites:
6751 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6752 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6753 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6755 The latter two were purportedly from
6756 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6759 Also deactivate the remaining ciphersuites from
6760 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6761 unofficial, and the ID has long expired.
6764 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6765 dual-core machines) and other potential thread-safety issues.
6768 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6770 *) Adapt fipsld and the build system to link against the validated FIPS
6771 module in FIPS mode.
6774 *) Fixes for VC++ 2005 build under Windows.
6777 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6778 from a Windows bash shell such as MSYS. It is autodetected from the
6779 "config" script when run from a VC++ environment. Modify standard VC++
6780 build to use fipscanister.o from the GNU make build.
6783 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6785 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6786 The value now differs depending on if you build for FIPS or not.
6787 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6788 safely run with a non-FIPSed libcrypto, as it may crash because of
6789 the difference induced by this change.
6792 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6794 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6795 (part of SSL_OP_ALL). This option used to disable the
6796 countermeasure against man-in-the-middle protocol-version
6797 rollback in the SSL 2.0 server implementation, which is a bad
6798 idea. (CVE-2005-2969)
6800 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6801 for Information Security, National Institute of Advanced Industrial
6802 Science and Technology [AIST], Japan)]
6804 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6805 mainly for FIPS compliance and not fully integrated at this stage.
6808 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6809 the exponentiation using a fixed-length exponent. (Otherwise,
6810 the information leaked through timing could expose the secret key
6811 after many signatures; cf. Bleichenbacher's attack on DSA with
6815 *) Make a new fixed-window mod_exp implementation the default for
6816 RSA, DSA, and DH private-key operations so that the sequence of
6817 squares and multiplies and the memory access pattern are
6818 independent of the particular secret key. This will mitigate
6819 cache-timing and potential related attacks.
6821 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6822 and this is automatically used by BN_mod_exp_mont() if the new flag
6823 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6824 will use this BN flag for private exponents unless the flag
6825 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6826 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6828 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6830 *) Change the client implementation for SSLv23_method() and
6831 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6832 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6833 (Previously, the SSL 2.0 backwards compatible Client Hello
6834 message format would be used even with SSL_OP_NO_SSLv2.)
6837 *) Add support for smime-type MIME parameter in S/MIME messages which some
6841 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6842 a threadsafe manner. Modify rsa code to use new function and add calls
6843 to dsa and dh code (which had race conditions before).
6846 *) Include the fixed error library code in the C error file definitions
6847 instead of fixing them up at runtime. This keeps the error code
6848 structures constant.
6851 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6853 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6856 *) Fixes for newer kerberos headers. NB: the casts are needed because
6857 the 'length' field is signed on one version and unsigned on another
6858 with no (?) obvious way to tell the difference, without these VC++
6859 complains. Also the "definition" of FAR (blank) is no longer included
6860 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6861 some needed definitions.
6864 *) Undo Cygwin change.
6867 *) Added support for proxy certificates according to RFC 3820.
6868 Because they may be a security thread to unaware applications,
6869 they must be explicitly allowed in run-time. See
6870 docs/HOWTO/proxy_certificates.txt for further information.
6873 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6875 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6876 server and client random values. Previously
6877 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6878 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6880 This change has negligible security impact because:
6882 1. Server and client random values still have 24 bytes of pseudo random
6885 2. Server and client random values are sent in the clear in the initial
6888 3. The master secret is derived using the premaster secret (48 bytes in
6889 size for static RSA ciphersuites) as well as client server and random
6892 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6895 [Stephen Henson, reported by UK NISCC]
6897 *) Use Windows randomness collection on Cygwin.
6900 *) Fix hang in EGD/PRNGD query when communication socket is closed
6901 prematurely by EGD/PRNGD.
6902 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6904 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6907 *) Back-port of selected performance improvements from development
6908 branch, as well as improved support for PowerPC platforms.
6911 *) Add lots of checks for memory allocation failure, error codes to indicate
6912 failure and freeing up memory if a failure occurs.
6913 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6915 *) Add new -passin argument to dgst.
6918 *) Perform some character comparisons of different types in X509_NAME_cmp:
6919 this is needed for some certificates that re-encode DNs into UTF8Strings
6920 (in violation of RFC3280) and can't or won't issue name rollover
6924 *) Make an explicit check during certificate validation to see that
6925 the CA setting in each certificate on the chain is correct. As a
6926 side effect always do the following basic checks on extensions,
6927 not just when there's an associated purpose to the check:
6929 - if there is an unhandled critical extension (unless the user
6930 has chosen to ignore this fault)
6931 - if the path length has been exceeded (if one is set at all)
6932 - that certain extensions fit the associated purpose (if one has
6936 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6938 *) Avoid a race condition when CRLs are checked in a multi threaded
6939 environment. This would happen due to the reordering of the revoked
6940 entries during signature checking and serial number lookup. Now the
6941 encoding is cached and the serial number sort performed under a lock.
6942 Add new STACK function sk_is_sorted().
6945 *) Add Delta CRL to the extension code.
6948 *) Various fixes to s3_pkt.c so alerts are sent properly.
6949 [David Holmes <d.holmes@f5.com>]
6951 *) Reduce the chances of duplicate issuer name and serial numbers (in
6952 violation of RFC3280) using the OpenSSL certificate creation utilities.
6953 This is done by creating a random 64 bit value for the initial serial
6954 number when a serial number file is created or when a self signed
6955 certificate is created using 'openssl req -x509'. The initial serial
6956 number file is created using 'openssl x509 -next_serial' in CA.pl
6957 rather than being initialized to 1.
6960 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6962 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6963 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6964 [Joe Orton, Steve Henson]
6966 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6968 [Joe Orton, Steve Henson]
6970 *) Make it possible to have multiple active certificates with the same
6971 subject in the CA index file. This is done only if the keyword
6972 'unique_subject' is set to 'no' in the main CA section (default
6973 if 'CA_default') of the configuration file. The value is saved
6974 with the database itself in a separate index attribute file,
6975 named like the index file with '.attr' appended to the name.
6978 *) X509 verify fixes. Disable broken certificate workarounds when
6979 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6980 keyUsage extension present. Don't accept CRLs with unhandled critical
6981 extensions: since verify currently doesn't process CRL extensions this
6982 rejects a CRL with *any* critical extensions. Add new verify error codes
6986 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6987 A clarification of RFC2560 will require the use of OCTET STRINGs and
6988 some implementations cannot handle the current raw format. Since OpenSSL
6989 copies and compares OCSP nonces as opaque blobs without any attempt at
6990 parsing them this should not create any compatibility issues.
6993 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6994 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6995 this HMAC (and other) operations are several times slower than OpenSSL
6999 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
7000 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7002 *) Use the correct content when signing type "other".
7005 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
7007 *) Fix various bugs revealed by running the NISCC test suite:
7009 Stop out of bounds reads in the ASN1 code when presented with
7010 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7012 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
7014 If verify callback ignores invalid public key errors don't try to check
7015 certificate signature with the NULL public key.
7019 *) New -ignore_err option in ocsp application to stop the server
7020 exiting on the first error in a request.
7023 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7024 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7028 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7029 extra data after the compression methods not only for TLS 1.0
7030 but also for SSL 3.0 (as required by the specification).
7031 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7033 *) Change X509_certificate_type() to mark the key as exported/exportable
7034 when it's 512 *bits* long, not 512 bytes.
7037 *) Change AES_cbc_encrypt() so it outputs exact multiple of
7038 blocks during encryption.
7041 *) Various fixes to base64 BIO and non blocking I/O. On write
7042 flushes were not handled properly if the BIO retried. On read
7043 data was not being buffered properly and had various logic bugs.
7044 This also affects blocking I/O when the data being decoded is a
7048 *) Various S/MIME bugfixes and compatibility changes:
7049 output correct application/pkcs7 MIME type if
7050 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7051 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7052 of files as .eml work). Correctly handle very long lines in MIME
7056 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7058 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7059 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7060 a protocol version number mismatch like a decryption error
7061 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7064 *) Turn on RSA blinding by default in the default implementation
7065 to avoid a timing attack. Applications that don't want it can call
7066 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7067 They would be ill-advised to do so in most cases.
7068 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7070 *) Change RSA blinding code so that it works when the PRNG is not
7071 seeded (in this case, the secret RSA exponent is abused as
7072 an unpredictable seed -- if it is not unpredictable, there
7073 is no point in blinding anyway). Make RSA blinding thread-safe
7074 by remembering the creator's thread ID in rsa->blinding and
7075 having all other threads use local one-time blinding factors
7076 (this requires more computation than sharing rsa->blinding, but
7077 avoids excessive locking; and if an RSA object is not shared
7078 between threads, blinding will still be very fast).
7081 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7082 ENGINE as defaults for all supported algorithms irrespective of
7083 the 'flags' parameter. 'flags' is now honoured, so applications
7084 should make sure they are passing it correctly.
7087 *) Target "mingw" now allows native Windows code to be generated in
7088 the Cygwin environment as well as with the MinGW compiler.
7091 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7093 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7094 via timing by performing a MAC computation even if incorrect
7095 block cipher padding has been found. This is a countermeasure
7096 against active attacks where the attacker has to distinguish
7097 between bad padding and a MAC verification error. (CVE-2003-0078)
7099 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7100 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7101 Martin Vuagnoux (EPFL, Ilion)]
7103 *) Make the no-err option work as intended. The intention with no-err
7104 is not to have the whole error stack handling routines removed from
7105 libcrypto, it's only intended to remove all the function name and
7106 reason texts, thereby removing some of the footprint that may not
7107 be interesting if those errors aren't displayed anyway.
7109 NOTE: it's still possible for any application or module to have its
7110 own set of error texts inserted. The routines are there, just not
7111 used by default when no-err is given.
7114 *) Add support for FreeBSD on IA64.
7115 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7117 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7118 Kerberos function mit_des_cbc_cksum(). Before this change,
7119 the value returned by DES_cbc_cksum() was like the one from
7120 mit_des_cbc_cksum(), except the bytes were swapped.
7121 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7123 *) Allow an application to disable the automatic SSL chain building.
7124 Before this a rather primitive chain build was always performed in
7125 ssl3_output_cert_chain(): an application had no way to send the
7126 correct chain if the automatic operation produced an incorrect result.
7128 Now the chain builder is disabled if either:
7130 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7132 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7134 The reasoning behind this is that an application would not want the
7135 auto chain building to take place if extra chain certificates are
7136 present and it might also want a means of sending no additional
7137 certificates (for example the chain has two certificates and the
7141 *) Add the possibility to build without the ENGINE framework.
7142 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7144 *) Under Win32 gmtime() can return NULL: check return value in
7145 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7148 *) DSA routines: under certain error conditions uninitialized BN objects
7149 could be freed. Solution: make sure initialization is performed early
7150 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7151 Nils Larsch <nla@trustcenter.de> via PR#459)
7154 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7155 checked on reconnect on the client side, therefore session resumption
7156 could still fail with a "ssl session id is different" error. This
7157 behaviour is masked when SSL_OP_ALL is used due to
7158 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7159 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7160 followup to PR #377.
7163 *) IA-32 assembler support enhancements: unified ELF targets, support
7164 for SCO/Caldera platforms, fix for Cygwin shared build.
7167 *) Add support for FreeBSD on sparc64. As a consequence, support for
7168 FreeBSD on non-x86 processors is separate from x86 processors on
7169 the config script, much like the NetBSD support.
7170 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7172 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7174 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7177 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7178 code (06) was taken as the first octet of the session ID and the last
7179 octet was ignored consequently. As a result SSLv2 client side session
7180 caching could not have worked due to the session ID mismatch between
7182 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7186 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7187 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7191 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7192 seems that in spite of existing for more than a year, many application
7193 author have done nothing to provide the necessary callbacks, which
7194 means that this particular engine will not work properly anywhere.
7195 This is a very unfortunate situation which forces us, in the name
7196 of usability, to give the hw_ncipher.c a static lock, which is part
7198 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7199 appear in 0.9.8 or later. We EXPECT application authors to have
7200 dealt properly with this when 0.9.8 is released (unless we actually
7201 make such changes in the libcrypto locking code that changes will
7202 have to be made anyway).
7205 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7206 octets have been read, EOF or an error occurs. Without this change
7207 some truncated ASN1 structures will not produce an error.
7210 *) Disable Heimdal support, since it hasn't been fully implemented.
7211 Still give the possibility to force the use of Heimdal, but with
7212 warnings and a request that patches get sent to openssl-dev.
7215 *) Add the VC-CE target, introduce the WINCE sysname, and add
7216 INSTALL.WCE and appropriate conditionals to make it build.
7217 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7219 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7220 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7221 edit numbers of the version.
7222 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7224 *) Introduce safe string copy and catenation functions
7225 (BUF_strlcpy() and BUF_strlcat()).
7226 [Ben Laurie (CHATS) and Richard Levitte]
7228 *) Avoid using fixed-size buffers for one-line DNs.
7229 [Ben Laurie (CHATS)]
7231 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7232 resizing buffers containing secrets, and use where appropriate.
7233 [Ben Laurie (CHATS)]
7235 *) Avoid using fixed size buffers for configuration file location.
7236 [Ben Laurie (CHATS)]
7238 *) Avoid filename truncation for various CA files.
7239 [Ben Laurie (CHATS)]
7241 *) Use sizeof in preference to magic numbers.
7242 [Ben Laurie (CHATS)]
7244 *) Avoid filename truncation in cert requests.
7245 [Ben Laurie (CHATS)]
7247 *) Add assertions to check for (supposedly impossible) buffer
7249 [Ben Laurie (CHATS)]
7251 *) Don't cache truncated DNS entries in the local cache (this could
7252 potentially lead to a spoofing attack).
7253 [Ben Laurie (CHATS)]
7255 *) Fix various buffers to be large enough for hex/decimal
7256 representations in a platform independent manner.
7257 [Ben Laurie (CHATS)]
7259 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7260 resizing buffers containing secrets, and use where appropriate.
7261 [Ben Laurie (CHATS)]
7263 *) Add BIO_indent() to avoid much slightly worrying code to do
7265 [Ben Laurie (CHATS)]
7267 *) Convert sprintf()/BIO_puts() to BIO_printf().
7268 [Ben Laurie (CHATS)]
7270 *) buffer_gets() could terminate with the buffer only half
7272 [Ben Laurie (CHATS)]
7274 *) Add assertions to prevent user-supplied crypto functions from
7275 overflowing internal buffers by having large block sizes, etc.
7276 [Ben Laurie (CHATS)]
7278 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7280 [Ben Laurie (CHATS)]
7282 *) Eliminate unused copy of key in RC4.
7283 [Ben Laurie (CHATS)]
7285 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7286 [Ben Laurie (CHATS)]
7288 *) Fix off-by-one error in EGD path.
7289 [Ben Laurie (CHATS)]
7291 *) If RANDFILE path is too long, ignore instead of truncating.
7292 [Ben Laurie (CHATS)]
7294 *) Eliminate unused and incorrectly sized X.509 structure
7296 [Ben Laurie (CHATS)]
7298 *) Eliminate unused and dangerous function knumber().
7299 [Ben Laurie (CHATS)]
7301 *) Eliminate unused and dangerous structure, KSSL_ERR.
7302 [Ben Laurie (CHATS)]
7304 *) Protect against overlong session ID context length in an encoded
7305 session object. Since these are local, this does not appear to be
7307 [Ben Laurie (CHATS)]
7309 *) Change from security patch (see 0.9.6e below) that did not affect
7310 the 0.9.6 release series:
7312 Remote buffer overflow in SSL3 protocol - an attacker could
7313 supply an oversized master key in Kerberos-enabled versions.
7315 [Ben Laurie (CHATS)]
7317 *) Change the SSL kerb5 codes to match RFC 2712.
7320 *) Make -nameopt work fully for req and add -reqopt switch.
7321 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7323 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7324 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7326 *) Make sure tests can be performed even if the corresponding algorithms
7327 have been removed entirely. This was also the last step to make
7328 OpenSSL compilable with DJGPP under all reasonable conditions.
7329 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7331 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7332 to allow version independent disabling of normally unselected ciphers,
7333 which may be activated as a side-effect of selecting a single cipher.
7335 (E.g., cipher list string "RSA" enables ciphersuites that are left
7336 out of "ALL" because they do not provide symmetric encryption.
7337 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7338 [Lutz Jaenicke, Bodo Moeller]
7340 *) Add appropriate support for separate platform-dependent build
7341 directories. The recommended way to make a platform-dependent
7342 build directory is the following (tested on Linux), maybe with
7345 # Place yourself outside of the OpenSSL source tree. In
7346 # this example, the environment variable OPENSSL_SOURCE
7347 # is assumed to contain the absolute OpenSSL source directory.
7348 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7349 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7350 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7351 mkdir -p `dirname $F`
7352 ln -s $OPENSSL_SOURCE/$F $F
7355 To be absolutely sure not to disturb the source tree, a "make clean"
7356 is a good thing. If it isn't successful, don't worry about it,
7357 it probably means the source directory is very clean.
7360 *) Make sure any ENGINE control commands make local copies of string
7361 pointers passed to them whenever necessary. Otherwise it is possible
7362 the caller may have overwritten (or deallocated) the original string
7363 data when a later ENGINE operation tries to use the stored values.
7364 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7366 *) Improve diagnostics in file reading and command-line digests.
7367 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7369 *) Add AES modes CFB and OFB to the object database. Correct an
7370 error in AES-CFB decryption.
7373 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7374 allows existing EVP_CIPHER_CTX structures to be reused after
7375 calling EVP_*Final(). This behaviour is used by encryption
7376 BIOs and some applications. This has the side effect that
7377 applications must explicitly clean up cipher contexts with
7378 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7381 *) Check the values of dna and dnb in bn_mul_recursive before calling
7382 bn_mul_comba (a non zero value means the a or b arrays do not contain
7383 n2 elements) and fallback to bn_mul_normal if either is not zero.
7386 *) Fix escaping of non-ASCII characters when using the -subj option
7387 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7390 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7391 form for "surname", serialNumber has no short form.
7392 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7393 therefore remove "mail" short name for "internet 7".
7394 The OID for unique identifiers in X509 certificates is
7395 x500UniqueIdentifier, not uniqueIdentifier.
7396 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7399 *) Add an "init" command to the ENGINE config module and auto initialize
7400 ENGINEs. Without any "init" command the ENGINE will be initialized
7401 after all ctrl commands have been executed on it. If init=1 the
7402 ENGINE is initialized at that point (ctrls before that point are run
7403 on the uninitialized ENGINE and after on the initialized one). If
7404 init=0 then the ENGINE will not be initialized at all.
7407 *) Fix the 'app_verify_callback' interface so that the user-defined
7408 argument is actually passed to the callback: In the
7409 SSL_CTX_set_cert_verify_callback() prototype, the callback
7410 declaration has been changed from
7413 int (*cb)(X509_STORE_CTX *,void *);
7414 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7415 i=s->ctx->app_verify_callback(&ctx)
7416 has been changed into
7417 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7419 To update applications using SSL_CTX_set_cert_verify_callback(),
7420 a dummy argument can be added to their callback functions.
7421 [D. K. Smetters <smetters@parc.xerox.com>]
7423 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7424 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7426 *) Add and OPENSSL_LOAD_CONF define which will cause
7427 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7428 This allows older applications to transparently support certain
7429 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7430 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7431 load the config file and OPENSSL_add_all_algorithms_conf() which will
7432 always load it have also been added.
7435 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7436 Adjust NIDs and EVP layer.
7437 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7439 *) Config modules support in openssl utility.
7441 Most commands now load modules from the config file,
7442 though in a few (such as version) this isn't done
7443 because it couldn't be used for anything.
7445 In the case of ca and req the config file used is
7446 the same as the utility itself: that is the -config
7447 command line option can be used to specify an
7451 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7452 use "openssl_conf" if filename is NULL use default openssl config file.
7455 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7456 config section name. Add a new flag to tolerate a missing config file
7457 and move code to CONF_modules_load_file().
7460 *) Support for crypto accelerator cards from Accelerated Encryption
7461 Processing, www.aep.ie. (Use engine 'aep')
7462 The support was copied from 0.9.6c [engine] and adapted/corrected
7463 to work with the new engine framework.
7464 [AEP Inc. and Richard Levitte]
7466 *) Support for SureWare crypto accelerator cards from Baltimore
7467 Technologies. (Use engine 'sureware')
7468 The support was copied from 0.9.6c [engine] and adapted
7469 to work with the new engine framework.
7472 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7473 make the newer ENGINE framework commands for the CHIL engine work.
7474 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7476 *) Make it possible to produce shared libraries on ReliantUNIX.
7477 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7479 *) Add the configuration target debug-linux-ppro.
7480 Make 'openssl rsa' use the general key loading routines
7481 implemented in apps.c, and make those routines able to
7482 handle the key format FORMAT_NETSCAPE and the variant
7484 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7486 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7487 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7489 *) Add -keyform to rsautl, and document -engine.
7490 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7492 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7493 BIO_R_NO_SUCH_FILE error code rather than the generic
7494 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7497 *) Add new functions
7499 ERR_peek_last_error_line
7500 ERR_peek_last_error_line_data.
7501 These are similar to
7504 ERR_peek_error_line_data,
7505 but report on the latest error recorded rather than the first one
7506 still in the error queue.
7507 [Ben Laurie, Bodo Moeller]
7509 *) default_algorithms option in ENGINE config module. This allows things
7511 default_algorithms = ALL
7512 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7515 *) Preliminary ENGINE config module.
7518 *) New experimental application configuration code.
7521 *) Change the AES code to follow the same name structure as all other
7522 symmetric ciphers, and behave the same way. Move everything to
7523 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7524 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7526 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7527 [Ben Laurie and Theo de Raadt]
7529 *) Add option to output public keys in req command.
7530 [Massimiliano Pala madwolf@openca.org]
7532 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7533 (up to about 10% better than before for P-192 and P-224).
7536 *) New functions/macros
7538 SSL_CTX_set_msg_callback(ctx, cb)
7539 SSL_CTX_set_msg_callback_arg(ctx, arg)
7540 SSL_set_msg_callback(ssl, cb)
7541 SSL_set_msg_callback_arg(ssl, arg)
7543 to request calling a callback function
7545 void cb(int write_p, int version, int content_type,
7546 const void *buf, size_t len, SSL *ssl, void *arg)
7548 whenever a protocol message has been completely received
7549 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7550 protocol version according to which the SSL library interprets
7551 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7552 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7553 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7554 specification (change_cipher_spec(20), alert(21), handshake(22)).
7555 'buf' and 'len' point to the actual message, 'ssl' to the
7556 SSL object, and 'arg' is the application-defined value set by
7557 SSL[_CTX]_set_msg_callback_arg().
7559 'openssl s_client' and 'openssl s_server' have new '-msg' options
7560 to enable a callback that displays all protocol messages.
7563 *) Change the shared library support so shared libraries are built as
7564 soon as the corresponding static library is finished, and thereby get
7565 openssl and the test programs linked against the shared library.
7566 This still only happens when the keyword "shard" has been given to
7567 the configuration scripts.
7569 NOTE: shared library support is still an experimental thing, and
7570 backward binary compatibility is still not guaranteed.
7571 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7573 *) Add support for Subject Information Access extension.
7574 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7576 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7577 additional bytes when new memory had to be allocated, not just
7578 when reusing an existing buffer.
7581 *) New command line and configuration option 'utf8' for the req command.
7582 This allows field values to be specified as UTF8 strings.
7585 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7586 runs for the former and machine-readable output for the latter.
7589 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7590 of the e-mail address in the DN (i.e., it will go into a certificate
7591 extension only). The new configuration file option 'email_in_dn = no'
7592 has the same effect.
7593 [Massimiliano Pala madwolf@openca.org]
7595 *) Change all functions with names starting with des_ to be starting
7596 with DES_ instead. Add wrappers that are compatible with libdes,
7597 but are named _ossl_old_des_*. Finally, add macros that map the
7598 des_* symbols to the corresponding _ossl_old_des_* if libdes
7599 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7600 desired, the des_* symbols will be mapped to DES_*, with one
7603 Since we provide two compatibility mappings, the user needs to
7604 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7605 compatibility is desired. The default (i.e., when that macro
7606 isn't defined) is OpenSSL 0.9.6c compatibility.
7608 There are also macros that enable and disable the support of old
7609 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7610 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7611 are defined, the default will apply: to support the old des routines.
7613 In either case, one must include openssl/des.h to get the correct
7614 definitions. Do not try to just include openssl/des_old.h, that
7617 NOTE: This is a major break of an old API into a new one. Software
7618 authors are encouraged to switch to the DES_ style functions. Some
7619 time in the future, des_old.h and the libdes compatibility functions
7620 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7621 default), and then completely removed.
7624 *) Test for certificates which contain unsupported critical extensions.
7625 If such a certificate is found during a verify operation it is
7626 rejected by default: this behaviour can be overridden by either
7627 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7628 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7629 X509_supported_extension() has also been added which returns 1 if a
7630 particular extension is supported.
7633 *) Modify the behaviour of EVP cipher functions in similar way to digests
7634 to retain compatibility with existing code.
7637 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7638 compatibility with existing code. In particular the 'ctx' parameter does
7639 not have to be to be initialized before the call to EVP_DigestInit() and
7640 it is tidied up after a call to EVP_DigestFinal(). New function
7641 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7642 EVP_MD_CTX_copy() changed to not require the destination to be
7643 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7644 requires the destination to be valid.
7646 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7647 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7650 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7651 so that complete 'Handshake' protocol structures are kept in memory
7652 instead of overwriting 'msg_type' and 'length' with 'body' data.
7655 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7656 [Massimo Santin via Richard Levitte]
7658 *) Major restructuring to the underlying ENGINE code. This includes
7659 reduction of linker bloat, separation of pure "ENGINE" manipulation
7660 (initialisation, etc) from functionality dealing with implementations
7661 of specific crypto interfaces. This change also introduces integrated
7662 support for symmetric ciphers and digest implementations - so ENGINEs
7663 can now accelerate these by providing EVP_CIPHER and EVP_MD
7664 implementations of their own. This is detailed in crypto/engine/README
7665 as it couldn't be adequately described here. However, there are a few
7666 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7667 were changed in the original introduction of ENGINE code have now
7668 reverted back - the hooking from this code to ENGINE is now a good
7669 deal more passive and at run-time, operations deal directly with
7670 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7671 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7672 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7673 they were not being used by the framework as there is no concept of a
7674 BIGNUM_METHOD and they could not be generalised to the new
7675 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7676 ENGINE_cpy() has been removed as it cannot be consistently defined in
7680 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7683 *) Change mkdef.pl to sort symbols that get the same entry number,
7684 and make sure the automatically generated functions ERR_load_*
7685 become part of libeay.num as well.
7688 *) New function SSL_renegotiate_pending(). This returns true once
7689 renegotiation has been requested (either SSL_renegotiate() call
7690 or HelloRequest/ClientHello received from the peer) and becomes
7691 false once a handshake has been completed.
7692 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7693 sends a HelloRequest, but does not ensure that a handshake takes
7694 place. SSL_renegotiate_pending() is useful for checking if the
7695 client has followed the request.)
7698 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7699 By default, clients may request session resumption even during
7700 renegotiation (if session ID contexts permit); with this option,
7701 session resumption is possible only in the first handshake.
7703 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7704 more bits available for options that should not be part of
7705 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7708 *) Add some demos for certificate and certificate request creation.
7711 *) Make maximum certificate chain size accepted from the peer application
7712 settable (SSL*_get/set_max_cert_list()), as proposed by
7713 "Douglas E. Engert" <deengert@anl.gov>.
7716 *) Add support for shared libraries for Unixware-7
7717 (Boyd Lynn Gerber <gerberb@zenez.com>).
7720 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7721 be done prior to destruction. Use this to unload error strings from
7722 ENGINEs that load their own error strings. NB: This adds two new API
7723 functions to "get" and "set" this destroy handler in an ENGINE.
7726 *) Alter all existing ENGINE implementations (except "openssl" and
7727 "openbsd") to dynamically instantiate their own error strings. This
7728 makes them more flexible to be built both as statically-linked ENGINEs
7729 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7730 Also, add stub code to each that makes building them as self-contained
7731 shared-libraries easier (see README.ENGINE).
7734 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7735 implementations into applications that are completely implemented in
7736 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7737 commands that can be used to configure what shared-library to load and
7738 to control aspects of the way it is handled. Also, made an update to
7739 the README.ENGINE file that brings its information up-to-date and
7740 provides some information and instructions on the "dynamic" ENGINE
7741 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7744 *) Make it possible to unload ranges of ERR strings with a new
7745 "ERR_unload_strings" function.
7748 *) Add a copy() function to EVP_MD.
7751 *) Make EVP_MD routines take a context pointer instead of just the
7752 md_data void pointer.
7755 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7756 that the digest can only process a single chunk of data
7757 (typically because it is provided by a piece of
7758 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7759 is only going to provide a single chunk of data, and hence the
7760 framework needn't accumulate the data for oneshot drivers.
7763 *) As with "ERR", make it possible to replace the underlying "ex_data"
7764 functions. This change also alters the storage and management of global
7765 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7766 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7767 index counters. The API functions that use this state have been changed
7768 to take a "class_index" rather than pointers to the class's local STACK
7769 and counter, and there is now an API function to dynamically create new
7770 classes. This centralisation allows us to (a) plug a lot of the
7771 thread-safety problems that existed, and (b) makes it possible to clean
7772 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7773 such data would previously have always leaked in application code and
7774 workarounds were in place to make the memory debugging turn a blind eye
7775 to it. Application code that doesn't use this new function will still
7776 leak as before, but their memory debugging output will announce it now
7777 rather than letting it slide.
7779 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7780 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7781 has a return value to indicate success or failure.
7784 *) Make it possible to replace the underlying "ERR" functions such that the
7785 global state (2 LHASH tables and 2 locks) is only used by the "default"
7786 implementation. This change also adds two functions to "get" and "set"
7787 the implementation prior to it being automatically set the first time
7788 any other ERR function takes place. Ie. an application can call "get",
7789 pass the return value to a module it has just loaded, and that module
7790 can call its own "set" function using that value. This means the
7791 module's "ERR" operations will use (and modify) the error state in the
7792 application and not in its own statically linked copy of OpenSSL code.
7795 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7796 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7797 the operation, and provides a more encapsulated way for external code
7798 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7799 to use these functions rather than manually incrementing the counts.
7801 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7804 *) Add EVP test program.
7807 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7810 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7811 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7812 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7813 These allow a CRL to be built without having to access X509_CRL fields
7814 directly. Modify 'ca' application to use new functions.
7817 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7818 bug workarounds. Rollback attack detection is a security feature.
7819 The problem will only arise on OpenSSL servers when TLSv1 is not
7820 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7821 Software authors not wanting to support TLSv1 will have special reasons
7822 for their choice and can explicitly enable this option.
7823 [Bodo Moeller, Lutz Jaenicke]
7825 *) Rationalise EVP so it can be extended: don't include a union of
7826 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7827 (similar to those existing for EVP_CIPHER_CTX).
7832 EVP_MD_CTX_init(&md); /* new function call */
7833 EVP_DigestInit(&md, EVP_sha1());
7834 EVP_DigestUpdate(&md, in, len);
7835 EVP_DigestFinal(&md, out, NULL);
7836 EVP_MD_CTX_cleanup(&md); /* new function call */
7840 *) Make DES key schedule conform to the usual scheme, as well as
7841 correcting its structure. This means that calls to DES functions
7842 now have to pass a pointer to a des_key_schedule instead of a
7843 plain des_key_schedule (which was actually always a pointer
7846 des_key_schedule ks;
7848 des_set_key_checked(..., &ks);
7849 des_ncbc_encrypt(..., &ks, ...);
7851 (Note that a later change renames 'des_...' into 'DES_...'.)
7854 *) Initial reduction of linker bloat: the use of some functions, such as
7855 PEM causes large amounts of unused functions to be linked in due to
7856 poor organisation. For example pem_all.c contains every PEM function
7857 which has a knock on effect of linking in large amounts of (unused)
7858 ASN1 code. Grouping together similar functions and splitting unrelated
7859 functions prevents this.
7862 *) Cleanup of EVP macros.
7865 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7866 correct _ecb suffix.
7869 *) Add initial OCSP responder support to ocsp application. The
7870 revocation information is handled using the text based index
7871 use by the ca application. The responder can either handle
7872 requests generated internally, supplied in files (for example
7873 via a CGI script) or using an internal minimal server.
7876 *) Add configuration choices to get zlib compression for TLS.
7879 *) Changes to Kerberos SSL for RFC 2712 compliance:
7880 1. Implemented real KerberosWrapper, instead of just using
7881 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7882 2. Implemented optional authenticator field of KerberosWrapper.
7884 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7885 and authenticator structs; see crypto/krb5/.
7887 Generalized Kerberos calls to support multiple Kerberos libraries.
7888 [Vern Staats <staatsvr@asc.hpc.mil>,
7889 Jeffrey Altman <jaltman@columbia.edu>
7890 via Richard Levitte]
7892 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7893 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7894 values for each of the key sizes rather than having just
7895 parameters (and 'speed' generating keys each time).
7898 *) Speed up EVP routines.
7901 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7902 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7903 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7904 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7906 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7907 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7908 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7911 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7913 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7916 *) Added the OS2-EMX target.
7917 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7919 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7920 to support NCONF routines in extension code. New function CONF_set_nconf()
7921 to allow functions which take an NCONF to also handle the old LHASH
7922 structure: this means that the old CONF compatible routines can be
7923 retained (in particular wrt extensions) without having to duplicate the
7924 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7927 *) Enhance the general user interface with mechanisms for inner control
7928 and with possibilities to have yes/no kind of prompts.
7931 *) Change all calls to low level digest routines in the library and
7932 applications to use EVP. Add missing calls to HMAC_cleanup() and
7933 don't assume HMAC_CTX can be copied using memcpy().
7934 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7936 *) Add the possibility to control engines through control names but with
7937 arbitrary arguments instead of just a string.
7938 Change the key loaders to take a UI_METHOD instead of a callback
7939 function pointer. NOTE: this breaks binary compatibility with earlier
7940 versions of OpenSSL [engine].
7941 Adapt the nCipher code for these new conditions and add a card insertion
7945 *) Enhance the general user interface with mechanisms to better support
7946 dialog box interfaces, application-defined prompts, the possibility
7947 to use defaults (for example default passwords from somewhere else)
7948 and interrupts/cancellations.
7951 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7952 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7955 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7956 tidy up some unnecessarily weird code in 'sk_new()').
7957 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7959 *) Change the key loading routines for ENGINEs to use the same kind
7960 callback (pem_password_cb) as all other routines that need this
7964 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7965 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7966 than this minimum value is recommended.
7969 *) New random seeder for OpenVMS, using the system process statistics
7970 that are easily reachable.
7973 *) Windows apparently can't transparently handle global
7974 variables defined in DLLs. Initialisations such as:
7976 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7978 won't compile. This is used by the any applications that need to
7979 declare their own ASN1 modules. This was fixed by adding the option
7980 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7981 needed for static libraries under Win32.
7984 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7985 setting of purpose and trust fields. New X509_STORE trust and
7986 purpose functions and tidy up setting in other SSL functions.
7989 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7990 structure. These are inherited by X509_STORE_CTX when it is
7991 initialised. This allows various defaults to be set in the
7992 X509_STORE structure (such as flags for CRL checking and custom
7993 purpose or trust settings) for functions which only use X509_STORE_CTX
7994 internally such as S/MIME.
7996 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7997 trust settings if they are not set in X509_STORE. This allows X509_STORE
7998 purposes and trust (in S/MIME for example) to override any set by default.
8000 Add command line options for CRL checking to smime, s_client and s_server
8004 *) Initial CRL based revocation checking. If the CRL checking flag(s)
8005 are set then the CRL is looked up in the X509_STORE structure and
8006 its validity and signature checked, then if the certificate is found
8007 in the CRL the verify fails with a revoked error.
8009 Various new CRL related callbacks added to X509_STORE_CTX structure.
8011 Command line options added to 'verify' application to support this.
8013 This needs some additional work, such as being able to handle multiple
8014 CRLs with different times, extension based lookup (rather than just
8015 by subject name) and ultimately more complete V2 CRL extension
8019 *) Add a general user interface API (crypto/ui/). This is designed
8020 to replace things like des_read_password and friends (backward
8021 compatibility functions using this new API are provided).
8022 The purpose is to remove prompting functions from the DES code
8023 section as well as provide for prompting through dialog boxes in
8024 a window system and the like.
8027 *) Add "ex_data" support to ENGINE so implementations can add state at a
8028 per-structure level rather than having to store it globally.
8031 *) Make it possible for ENGINE structures to be copied when retrieved by
8032 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
8033 This causes the "original" ENGINE structure to act like a template,
8034 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
8035 operational state can be localised to each ENGINE structure, despite the
8036 fact they all share the same "methods". New ENGINE structures returned in
8037 this case have no functional references and the return value is the single
8038 structural reference. This matches the single structural reference returned
8039 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8043 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8044 needs to match any other type at all we need to manually clear the
8048 *) Changes to the "openssl engine" utility to include;
8049 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8050 about an ENGINE's available control commands.
8051 - executing control commands from command line arguments using the
8052 '-pre' and '-post' switches. '-post' is only used if '-t' is
8053 specified and the ENGINE is successfully initialised. The syntax for
8054 the individual commands are colon-separated, for example;
8055 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8058 *) New dynamic control command support for ENGINEs. ENGINEs can now
8059 declare their own commands (numbers), names (strings), descriptions,
8060 and input types for run-time discovery by calling applications. A
8061 subset of these commands are implicitly classed as "executable"
8062 depending on their input type, and only these can be invoked through
8063 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8064 can be based on user input, config files, etc). The distinction is
8065 that "executable" commands cannot return anything other than a boolean
8066 result and can only support numeric or string input, whereas some
8067 discoverable commands may only be for direct use through
8068 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8069 pointers, or other custom uses. The "executable" commands are to
8070 support parameterisations of ENGINE behaviour that can be
8071 unambiguously defined by ENGINEs and used consistently across any
8072 OpenSSL-based application. Commands have been added to all the
8073 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8074 control over shared-library paths without source code alterations.
8077 *) Changed all ENGINE implementations to dynamically allocate their
8078 ENGINEs rather than declaring them statically. Apart from this being
8079 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8080 this also allows the implementations to compile without using the
8081 internal engine_int.h header.
8084 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8085 'const' value. Any code that should be able to modify a RAND_METHOD
8086 should already have non-const pointers to it (ie. they should only
8087 modify their own ones).
8090 *) Made a variety of little tweaks to the ENGINE code.
8091 - "atalla" and "ubsec" string definitions were moved from header files
8092 to C code. "nuron" string definitions were placed in variables
8093 rather than hard-coded - allowing parameterisation of these values
8094 later on via ctrl() commands.
8095 - Removed unused "#if 0"'d code.
8096 - Fixed engine list iteration code so it uses ENGINE_free() to release
8097 structural references.
8098 - Constified the RAND_METHOD element of ENGINE structures.
8099 - Constified various get/set functions as appropriate and added
8100 missing functions (including a catch-all ENGINE_cpy that duplicates
8101 all ENGINE values onto a new ENGINE except reference counts/state).
8102 - Removed NULL parameter checks in get/set functions. Setting a method
8103 or function to NULL is a way of cancelling out a previously set
8104 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8105 and doesn't justify the extra error symbols and code.
8106 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8107 flags from engine_int.h to engine.h.
8108 - Changed prototypes for ENGINE handler functions (init(), finish(),
8109 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8112 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8113 to the algorithm using long division. The binary algorithm can be
8114 used only if the modulus is odd. On 32-bit systems, it is faster
8115 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8116 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8117 up to 450 bits. In 64-bit environments, the binary algorithm
8118 appears to be advantageous for much longer moduli; here we use it
8119 for moduli up to 2048 bits.
8122 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8123 could not support the combine flag in choice fields.
8126 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8127 extensions from a certificate request to the certificate.
8130 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8131 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8132 file: this allows the display of the certificate about to be
8133 signed to be customised, to allow certain fields to be included
8134 or excluded and extension details. The old system didn't display
8135 multicharacter strings properly, omitted fields not in the policy
8136 and couldn't display additional details such as extensions.
8139 *) Function EC_POINTs_mul for multiple scalar multiplication
8140 of an arbitrary number of elliptic curve points
8141 \sum scalars[i]*points[i],
8142 optionally including the generator defined for the EC_GROUP:
8143 scalar*generator + \sum scalars[i]*points[i].
8145 EC_POINT_mul is a simple wrapper function for the typical case
8146 that the point list has just one item (besides the optional
8150 *) First EC_METHODs for curves over GF(p):
8152 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8153 operations and provides various method functions that can also
8154 operate with faster implementations of modular arithmetic.
8156 EC_GFp_mont_method() reuses most functions that are part of
8157 EC_GFp_simple_method, but uses Montgomery arithmetic.
8159 [Bodo Moeller; point addition and point doubling
8160 implementation directly derived from source code provided by
8161 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8163 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8164 crypto/ec/ec_lib.c):
8166 Curves are EC_GROUP objects (with an optional group generator)
8167 based on EC_METHODs that are built into the library.
8169 Points are EC_POINT objects based on EC_GROUP objects.
8171 Most of the framework would be able to handle curves over arbitrary
8172 finite fields, but as there are no obvious types for fields other
8173 than GF(p), some functions are limited to that for now.
8176 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8177 that the file contains a complete HTTP response.
8180 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8181 change the def and num file printf format specifier from "%-40sXXX"
8182 to "%-39s XXX". The latter will always guarantee a space after the
8183 field while the former will cause them to run together if the field
8184 is 40 of more characters long.
8187 *) Constify the cipher and digest 'method' functions and structures
8188 and modify related functions to take constant EVP_MD and EVP_CIPHER
8192 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8193 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8196 *) Modify EVP_Digest*() routines so they now return values. Although the
8197 internal software routines can never fail additional hardware versions
8201 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8203 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8204 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8207 ERR_R_NESTED_ASN1_ERROR
8209 ERR_R_MISSING_ASN1_EOS
8210 were 4 .. 9, conflicting with
8211 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8213 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8214 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8216 Add new error code 'ERR_R_INTERNAL_ERROR'.
8219 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8223 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8224 sets the subject name for a new request or supersedes the
8225 subject name in a given request. Formats that can be parsed are
8226 'CN=Some Name, OU=myOU, C=IT'
8228 'CN=Some Name/OU=myOU/C=IT'.
8230 Add options '-batch' and '-verbose' to 'openssl req'.
8231 [Massimiliano Pala <madwolf@hackmasters.net>]
8233 *) Introduce the possibility to access global variables through
8234 functions on platform were that's the best way to handle exporting
8235 global variables in shared libraries. To enable this functionality,
8236 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8237 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8238 is normally done by Configure or something similar).
8240 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8241 in the source file (foo.c) like this:
8243 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8244 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8246 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8247 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8249 OPENSSL_DECLARE_GLOBAL(int,foo);
8250 #define foo OPENSSL_GLOBAL_REF(foo)
8251 OPENSSL_DECLARE_GLOBAL(double,bar);
8252 #define bar OPENSSL_GLOBAL_REF(bar)
8254 The #defines are very important, and therefore so is including the
8255 header file everywhere where the defined globals are used.
8257 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8258 of ASN.1 items, but that structure is a bit different.
8260 The largest change is in util/mkdef.pl which has been enhanced with
8261 better and easier to understand logic to choose which symbols should
8262 go into the Windows .def files as well as a number of fixes and code
8263 cleanup (among others, algorithm keywords are now sorted
8264 lexicographically to avoid constant rewrites).
8267 *) In BN_div() keep a copy of the sign of 'num' before writing the
8268 result to 'rm' because if rm==num the value will be overwritten
8269 and produce the wrong result if 'num' is negative: this caused
8270 problems with BN_mod() and BN_nnmod().
8273 *) Function OCSP_request_verify(). This checks the signature on an
8274 OCSP request and verifies the signer certificate. The signer
8275 certificate is just checked for a generic purpose and OCSP request
8279 *) Add OCSP_check_validity() function to check the validity of OCSP
8280 responses. OCSP responses are prepared in real time and may only
8281 be a few seconds old. Simply checking that the current time lies
8282 between thisUpdate and nextUpdate max reject otherwise valid responses
8283 caused by either OCSP responder or client clock inaccuracy. Instead
8284 we allow thisUpdate and nextUpdate to fall within a certain period of
8285 the current time. The age of the response can also optionally be
8286 checked. Two new options -validity_period and -status_age added to
8290 *) If signature or public key algorithm is unrecognized print out its
8291 OID rather that just UNKNOWN.
8294 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8295 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8296 ID to be generated from the issuer certificate alone which can then be
8297 passed to OCSP_id_issuer_cmp().
8300 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8301 ASN1 modules to export functions returning ASN1_ITEM pointers
8302 instead of the ASN1_ITEM structures themselves. This adds several
8303 new macros which allow the underlying ASN1 function/structure to
8304 be accessed transparently. As a result code should not use ASN1_ITEM
8305 references directly (such as &X509_it) but instead use the relevant
8306 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8307 use of the new ASN1 code on platforms where exporting structures
8308 is problematical (for example in shared libraries) but exporting
8309 functions returning pointers to structures is not.
8312 *) Add support for overriding the generation of SSL/TLS session IDs.
8313 These callbacks can be registered either in an SSL_CTX or per SSL.
8314 The purpose of this is to allow applications to control, if they wish,
8315 the arbitrary values chosen for use as session IDs, particularly as it
8316 can be useful for session caching in multiple-server environments. A
8317 command-line switch for testing this (and any client code that wishes
8318 to use such a feature) has been added to "s_server".
8319 [Geoff Thorpe, Lutz Jaenicke]
8321 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8322 of the form '#if defined(...) || defined(...) || ...' and
8323 '#if !defined(...) && !defined(...) && ...'. This also avoids
8324 the growing number of special cases it was previously handling.
8327 *) Make all configuration macros available for application by making
8328 sure they are available in opensslconf.h, by giving them names starting
8329 with "OPENSSL_" to avoid conflicts with other packages and by making
8330 sure e_os2.h will cover all platform-specific cases together with
8332 Additionally, it is now possible to define configuration/platform-
8333 specific names (called "system identities"). In the C code, these
8334 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8335 macro with the name beginning with "OPENSSL_SYS_", which is determined
8336 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8340 *) New option -set_serial to 'req' and 'x509' this allows the serial
8341 number to use to be specified on the command line. Previously self
8342 signed certificates were hard coded with serial number 0 and the
8343 CA options of 'x509' had to use a serial number in a file which was
8347 *) New options to 'ca' utility to support V2 CRL entry extensions.
8348 Currently CRL reason, invalidity date and hold instruction are
8349 supported. Add new CRL extensions to V3 code and some new objects.
8352 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8353 disable standard block padding (aka PKCS#5 padding) in the EVP
8354 API, which was previously mandatory. This means that the data is
8355 not padded in any way and so the total length much be a multiple
8356 of the block size, otherwise an error occurs.
8359 *) Initial (incomplete) OCSP SSL support.
8362 *) New function OCSP_parse_url(). This splits up a URL into its host,
8363 port and path components: primarily to parse OCSP URLs. New -url
8364 option to ocsp utility.
8367 *) New nonce behavior. The return value of OCSP_check_nonce() now
8368 reflects the various checks performed. Applications can decide
8369 whether to tolerate certain situations such as an absent nonce
8370 in a response when one was present in a request: the ocsp application
8371 just prints out a warning. New function OCSP_add1_basic_nonce()
8372 this is to allow responders to include a nonce in a response even if
8373 the request is nonce-less.
8376 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8377 skipped when using openssl x509 multiple times on a single input file,
8378 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8381 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8382 set string type: to handle setting ASN1_TIME structures. Fix ca
8383 utility to correctly initialize revocation date of CRLs.
8386 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8387 the clients preferred ciphersuites and rather use its own preferences.
8388 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8389 Internet Explorer by ensuring unchanged hash method during stepup.
8390 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8393 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8394 to aes and add a new 'exist' option to print out symbols that don't
8398 *) Additional options to ocsp utility to allow flags to be set and
8399 additional certificates supplied.
8402 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8403 OCSP client a number of certificate to only verify the response
8407 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8408 handle the new API. Currently only ECB, CBC modes supported. Add new
8411 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8412 Encryption Standard (AES) Ciphersuites for Transport Layer
8413 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8414 not enabled by default and were not part of the "ALL" ciphersuite
8415 alias because they were not yet official; they could be
8416 explicitly requested by specifying the "AESdraft" ciphersuite
8417 group alias. In the final release of OpenSSL 0.9.7, the group
8418 alias is called "AES" and is part of "ALL".)
8419 [Ben Laurie, Steve Henson, Bodo Moeller]
8421 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8422 request to response.
8425 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8426 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8427 extract information from a certificate request. OCSP_response_create()
8428 creates a response and optionally adds a basic response structure.
8429 OCSP_basic_add1_status() adds a complete single response to a basic
8430 response and returns the OCSP_SINGLERESP structure just added (to allow
8431 extensions to be included for example). OCSP_basic_add1_cert() adds a
8432 certificate to a basic response and OCSP_basic_sign() signs a basic
8433 response with various flags. New helper functions ASN1_TIME_check()
8434 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8435 (converts ASN1_TIME to GeneralizedTime).
8438 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8439 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8440 structure from a certificate. X509_pubkey_digest() digests the public_key
8441 contents: this is used in various key identifiers.
8444 *) Make sk_sort() tolerate a NULL argument.
8445 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8447 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8448 passed by the function are trusted implicitly. If any of them signed the
8449 response then it is assumed to be valid and is not verified.
8452 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8453 to data. This was previously part of the PKCS7 ASN1 code. This
8454 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8455 [Steve Henson, reported by Kenneth R. Robinette
8456 <support@securenetterm.com>]
8458 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8459 routines: without these tracing memory leaks is very painful.
8460 Fix leaks in PKCS12 and PKCS7 routines.
8463 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8464 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8465 effectively meant GeneralizedTime would never be used. Now it
8466 is initialised to -1 but X509_time_adj() now has to check the value
8467 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8468 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8469 [Steve Henson, reported by Kenneth R. Robinette
8470 <support@securenetterm.com>]
8472 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8473 result in a zero length in the ASN1_INTEGER structure which was
8474 not consistent with the structure when d2i_ASN1_INTEGER() was used
8475 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8476 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8477 where it did not print out a minus for negative ASN1_INTEGER.
8480 *) Add summary printout to ocsp utility. The various functions which
8481 convert status values to strings have been renamed to:
8482 OCSP_response_status_str(), OCSP_cert_status_str() and
8483 OCSP_crl_reason_str() and are no longer static. New options
8484 to verify nonce values and to disable verification. OCSP response
8485 printout format cleaned up.
8488 *) Add additional OCSP certificate checks. These are those specified
8489 in RFC2560. This consists of two separate checks: the CA of the
8490 certificate being checked must either be the OCSP signer certificate
8491 or the issuer of the OCSP signer certificate. In the latter case the
8492 OCSP signer certificate must contain the OCSP signing extended key
8493 usage. This check is performed by attempting to match the OCSP
8494 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8495 in the OCSP_CERTID structures of the response.
8498 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8499 and related routines. This uses the standard OpenSSL certificate
8500 verify routines to perform initial checks (just CA validity) and
8501 to obtain the certificate chain. Then additional checks will be
8502 performed on the chain. Currently the root CA is checked to see
8503 if it is explicitly trusted for OCSP signing. This is used to set
8504 a root CA as a global signing root: that is any certificate that
8505 chains to that CA is an acceptable OCSP signing certificate.
8508 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8509 extensions from a separate configuration file.
8510 As when reading extensions from the main configuration file,
8511 the '-extensions ...' option may be used for specifying the
8513 [Massimiliano Pala <madwolf@comune.modena.it>]
8515 *) New OCSP utility. Allows OCSP requests to be generated or
8516 read. The request can be sent to a responder and the output
8517 parsed, outputed or printed in text form. Not complete yet:
8518 still needs to check the OCSP response validity.
8521 *) New subcommands for 'openssl ca':
8522 'openssl ca -status <serial>' prints the status of the cert with
8523 the given serial number (according to the index file).
8524 'openssl ca -updatedb' updates the expiry status of certificates
8526 [Massimiliano Pala <madwolf@comune.modena.it>]
8528 *) New '-newreq-nodes' command option to CA.pl. This is like
8529 '-newreq', but calls 'openssl req' with the '-nodes' option
8530 so that the resulting key is not encrypted.
8531 [Damien Miller <djm@mindrot.org>]
8533 *) New configuration for the GNU Hurd.
8534 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8536 *) Initial code to implement OCSP basic response verify. This
8537 is currently incomplete. Currently just finds the signer's
8538 certificate and verifies the signature on the response.
8541 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8542 value of OPENSSLDIR. This is available via the new '-d' option
8543 to 'openssl version', and is also included in 'openssl version -a'.
8546 *) Allowing defining memory allocation callbacks that will be given
8547 file name and line number information in additional arguments
8548 (a const char* and an int). The basic functionality remains, as
8549 well as the original possibility to just replace malloc(),
8550 realloc() and free() by functions that do not know about these
8551 additional arguments. To register and find out the current
8552 settings for extended allocation functions, the following
8553 functions are provided:
8555 CRYPTO_set_mem_ex_functions
8556 CRYPTO_set_locked_mem_ex_functions
8557 CRYPTO_get_mem_ex_functions
8558 CRYPTO_get_locked_mem_ex_functions
8560 These work the same way as CRYPTO_set_mem_functions and friends.
8561 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8562 extended allocation function is enabled.
8563 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8564 a conventional allocation function is enabled.
8565 [Richard Levitte, Bodo Moeller]
8567 *) Finish off removing the remaining LHASH function pointer casts.
8568 There should no longer be any prototype-casting required when using
8569 the LHASH abstraction, and any casts that remain are "bugs". See
8570 the callback types and macros at the head of lhash.h for details
8571 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8574 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8575 If /dev/[u]random devices are not available or do not return enough
8576 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8578 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8579 /etc/entropy will be queried once each in this sequence, querying stops
8580 when enough entropy was collected without querying more sockets.
8583 *) Change the Unix RAND_poll() variant to be able to poll several
8584 random devices, as specified by DEVRANDOM, until a sufficient amount
8585 of data has been collected. We spend at most 10 ms on each file
8586 (select timeout) and read in non-blocking mode. DEVRANDOM now
8587 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8588 (previously it was just the string "/dev/urandom"), so on typical
8589 platforms the 10 ms delay will never occur.
8590 Also separate out the Unix variant to its own file, rand_unix.c.
8591 For VMS, there's a currently-empty rand_vms.c.
8594 *) Move OCSP client related routines to ocsp_cl.c. These
8595 provide utility functions which an application needing
8596 to issue a request to an OCSP responder and analyse the
8597 response will typically need: as opposed to those which an
8598 OCSP responder itself would need which will be added later.
8600 OCSP_request_sign() signs an OCSP request with an API similar
8601 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8602 response. OCSP_response_get1_basic() extracts basic response
8603 from response. OCSP_resp_find_status(): finds and extracts status
8604 information from an OCSP_CERTID structure (which will be created
8605 when the request structure is built). These are built from lower
8606 level functions which work on OCSP_SINGLERESP structures but
8607 won't normally be used unless the application wishes to examine
8608 extensions in the OCSP response for example.
8610 Replace nonce routines with a pair of functions.
8611 OCSP_request_add1_nonce() adds a nonce value and optionally
8612 generates a random value. OCSP_check_nonce() checks the
8613 validity of the nonce in an OCSP response.
8616 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8617 This doesn't copy the supplied OCSP_CERTID and avoids the
8618 need to free up the newly created id. Change return type
8619 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8620 This can then be used to add extensions to the request.
8621 Deleted OCSP_request_new(), since most of its functionality
8622 is now in OCSP_REQUEST_new() (and the case insensitive name
8623 clash) apart from the ability to set the request name which
8624 will be added elsewhere.
8627 *) Update OCSP API. Remove obsolete extensions argument from
8628 various functions. Extensions are now handled using the new
8629 OCSP extension code. New simple OCSP HTTP function which
8630 can be used to send requests and parse the response.
8633 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8634 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8635 uses the special reorder version of SET OF to sort the attributes
8636 and reorder them to match the encoded order. This resolves a long
8637 standing problem: a verify on a PKCS7 structure just after signing
8638 it used to fail because the attribute order did not match the
8639 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8640 it uses the received order. This is necessary to tolerate some broken
8641 software that does not order SET OF. This is handled by encoding
8642 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8643 to produce the required SET OF.
8646 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8647 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8648 files to get correct declarations of the ASN.1 item variables.
8651 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8652 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8653 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8654 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8655 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8656 ASN1_ITEM and no wrapper functions.
8659 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8660 replace the old function pointer based I/O routines. Change most of
8661 the *_d2i_bio() and *_d2i_fp() functions to use these.
8664 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8665 lines, recognize more "algorithms" that can be deselected, and make
8666 it complain about algorithm deselection that isn't recognised.
8669 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8670 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8671 to use new functions. Add NO_ASN1_OLD which can be set to remove
8672 some old style ASN1 functions: this can be used to determine if old
8673 code will still work when these eventually go away.
8676 *) New extension functions for OCSP structures, these follow the
8677 same conventions as certificates and CRLs.
8680 *) New function X509V3_add1_i2d(). This automatically encodes and
8681 adds an extension. Its behaviour can be customised with various
8682 flags to append, replace or delete. Various wrappers added for
8683 certificates and CRLs.
8686 *) Fix to avoid calling the underlying ASN1 print routine when
8687 an extension cannot be parsed. Correct a typo in the
8688 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8691 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8692 entries for variables.
8695 *) Add functionality to apps/openssl.c for detecting locking
8696 problems: As the program is single-threaded, all we have
8697 to do is register a locking callback using an array for
8698 storing which locks are currently held by the program.
8701 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8702 SSL_get_ex_data_X509_STORE_idx(), which is used in
8703 ssl_verify_cert_chain() and thus can be called at any time
8704 during TLS/SSL handshakes so that thread-safety is essential.
8705 Unfortunately, the ex_data design is not at all suited
8706 for multi-threaded use, so it probably should be abolished.
8709 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8710 [Broadcom, tweaked and integrated by Geoff Thorpe]
8712 *) Move common extension printing code to new function
8713 X509V3_print_extensions(). Reorganise OCSP print routines and
8714 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8717 *) New function X509_signature_print() to remove duplication in some
8721 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8722 set (this was treated exactly the same as SET OF previously). This
8723 is used to reorder the STACK representing the structure to match the
8724 encoding. This will be used to get round a problem where a PKCS7
8725 structure which was signed could not be verified because the STACK
8726 order did not reflect the encoded order.
8729 *) Reimplement the OCSP ASN1 module using the new code.
8732 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8733 for its ASN1 operations. The old style function pointers still exist
8734 for now but they will eventually go away.
8737 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8738 completely replaces the old ASN1 functionality with a table driven
8739 encoder and decoder which interprets an ASN1_ITEM structure describing
8740 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8741 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8742 has also been converted to the new form.
8745 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8746 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8747 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8748 for negative moduli.
8751 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8752 of not touching the result's sign bit.
8755 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8759 *) Changed the LHASH code to use prototypes for callbacks, and created
8760 macros to declare and implement thin (optionally static) functions
8761 that provide type-safety and avoid function pointer casting for the
8762 type-specific callbacks.
8765 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8767 [Veers Staats <staatsvr@asc.hpc.mil>,
8768 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8770 *) Reformat the FAQ so the different questions and answers can be divided
8771 in sections depending on the subject.
8774 *) Have the zlib compression code load ZLIB.DLL dynamically under
8778 *) New function BN_mod_sqrt for computing square roots modulo a prime
8779 (using the probabilistic Tonelli-Shanks algorithm unless
8780 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8781 be handled deterministically).
8782 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8784 *) Make BN_mod_inverse faster by explicitly handling small quotients
8785 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8786 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8789 *) New function BN_kronecker.
8792 *) Fix BN_gcd so that it works on negative inputs; the result is
8793 positive unless both parameters are zero.
8794 Previously something reasonably close to an infinite loop was
8795 possible because numbers could be growing instead of shrinking
8796 in the implementation of Euclid's algorithm.
8799 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8800 sign of the number in question.
8802 Fix BN_is_word(a,w) to work correctly for w == 0.
8804 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8805 because its test if the absolute value of 'a' equals 'w'.
8806 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8807 it exists mostly for use in the implementations of BN_is_zero(),
8808 BN_is_one(), and BN_is_word().
8811 *) New function BN_swap.
8814 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8815 the exponentiation functions are more likely to produce reasonable
8816 results on negative inputs.
8819 *) Change BN_mod_mul so that the result is always non-negative.
8820 Previously, it could be negative if one of the factors was negative;
8821 I don't think anyone really wanted that behaviour.
8824 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8825 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8826 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8827 and add new functions:
8836 BN_mod_lshift1_quick
8840 These functions always generate non-negative results.
8842 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8843 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8845 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8846 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8847 be reduced modulo m.
8848 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8851 The following entry accidentally appeared in the CHANGES file
8852 distributed with OpenSSL 0.9.7. The modifications described in
8853 it do *not* apply to OpenSSL 0.9.7.
8855 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8856 was actually never needed) and in BN_mul(). The removal in BN_mul()
8857 required a small change in bn_mul_part_recursive() and the addition
8858 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8859 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8860 bn_sub_words() and bn_add_words() except they take arrays with
8865 *) In 'openssl passwd', verify passwords read from the terminal
8866 unless the '-salt' option is used (which usually means that
8867 verification would just waste user's time since the resulting
8868 hash is going to be compared with some given password hash)
8869 or the new '-noverify' option is used.
8871 This is an incompatible change, but it does not affect
8872 non-interactive use of 'openssl passwd' (passwords on the command
8873 line, '-stdin' option, '-in ...' option) and thus should not
8877 *) Remove all references to RSAref, since there's no more need for it.
8880 *) Make DSO load along a path given through an environment variable
8881 (SHLIB_PATH) with shl_load().
8884 *) Constify the ENGINE code as a result of BIGNUM constification.
8885 Also constify the RSA code and most things related to it. In a
8886 few places, most notable in the depth of the ASN.1 code, ugly
8887 casts back to non-const were required (to be solved at a later
8891 *) Make it so the openssl application has all engines loaded by default.
8894 *) Constify the BIGNUM routines a little more.
8897 *) Add the following functions:
8899 ENGINE_load_cswift()
8901 ENGINE_load_atalla()
8903 ENGINE_load_builtin_engines()
8905 That way, an application can itself choose if external engines that
8906 are built-in in OpenSSL shall ever be used or not. The benefit is
8907 that applications won't have to be linked with libdl or other dso
8908 libraries unless it's really needed.
8910 Changed 'openssl engine' to load all engines on demand.
8911 Changed the engine header files to avoid the duplication of some
8912 declarations (they differed!).
8915 *) 'openssl engine' can now list capabilities.
8918 *) Better error reporting in 'openssl engine'.
8921 *) Never call load_dh_param(NULL) in s_server.
8924 *) Add engine application. It can currently list engines by name and
8925 identity, and test if they are actually available.
8928 *) Improve RPM specification file by forcing symbolic linking and making
8929 sure the installed documentation is also owned by root.root.
8930 [Damien Miller <djm@mindrot.org>]
8932 *) Give the OpenSSL applications more possibilities to make use of
8933 keys (public as well as private) handled by engines.
8936 *) Add OCSP code that comes from CertCo.
8939 *) Add VMS support for the Rijndael code.
8942 *) Added untested support for Nuron crypto accelerator.
8945 *) Add support for external cryptographic devices. This code was
8946 previously distributed separately as the "engine" branch.
8947 [Geoff Thorpe, Richard Levitte]
8949 *) Rework the filename-translation in the DSO code. It is now possible to
8950 have far greater control over how a "name" is turned into a filename
8951 depending on the operating environment and any oddities about the
8952 different shared library filenames on each system.
8955 *) Support threads on FreeBSD-elf in Configure.
8958 *) Fix for SHA1 assembly problem with MASM: it produces
8959 warnings about corrupt line number information when assembling
8960 with debugging information. This is caused by the overlapping
8962 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8965 NCONF_get_number() has no error checking at all. As a replacement,
8966 NCONF_get_number_e() is defined (_e for "error checking") and is
8967 promoted strongly. The old NCONF_get_number is kept around for
8968 binary backward compatibility.
8969 Make it possible for methods to load from something other than a BIO,
8970 by providing a function pointer that is given a name instead of a BIO.
8971 For example, this could be used to load configuration data from an
8975 *) Fix for non blocking accept BIOs. Added new I/O special reason
8976 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8977 with non blocking I/O was not possible because no retry code was
8978 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8982 *) Added the beginnings of Rijndael support.
8985 *) Fix for bug in DirectoryString mask setting. Add support for
8986 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8987 to allow certificate printing to more controllable, additional
8988 'certopt' option to 'x509' to allow new printing options to be
8992 *) Clean old EAY MD5 hack from e_os.h.
8995 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8997 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8998 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8999 [Joe Orton, Steve Henson]
9001 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
9003 *) Fix additional bug revealed by the NISCC test suite:
9005 Stop bug triggering large recursion when presented with
9006 certain ASN.1 tags (CVE-2003-0851)
9009 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
9011 *) Fix various bugs revealed by running the NISCC test suite:
9013 Stop out of bounds reads in the ASN1 code when presented with
9014 invalid tags (CVE-2003-0543 and CVE-2003-0544).
9016 If verify callback ignores invalid public key errors don't try to check
9017 certificate signature with the NULL public key.
9021 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
9022 if the server requested one: as stated in TLS 1.0 and SSL 3.0
9026 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
9027 extra data after the compression methods not only for TLS 1.0
9028 but also for SSL 3.0 (as required by the specification).
9029 [Bodo Moeller; problem pointed out by Matthias Loepfe]
9031 *) Change X509_certificate_type() to mark the key as exported/exportable
9032 when it's 512 *bits* long, not 512 bytes.
9035 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
9037 *) Countermeasure against the Klima-Pokorny-Rosa extension of
9038 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
9039 a protocol version number mismatch like a decryption error
9040 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9043 *) Turn on RSA blinding by default in the default implementation
9044 to avoid a timing attack. Applications that don't want it can call
9045 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9046 They would be ill-advised to do so in most cases.
9047 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9049 *) Change RSA blinding code so that it works when the PRNG is not
9050 seeded (in this case, the secret RSA exponent is abused as
9051 an unpredictable seed -- if it is not unpredictable, there
9052 is no point in blinding anyway). Make RSA blinding thread-safe
9053 by remembering the creator's thread ID in rsa->blinding and
9054 having all other threads use local one-time blinding factors
9055 (this requires more computation than sharing rsa->blinding, but
9056 avoids excessive locking; and if an RSA object is not shared
9057 between threads, blinding will still be very fast).
9060 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9062 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9063 via timing by performing a MAC computation even if incorrect
9064 block cipher padding has been found. This is a countermeasure
9065 against active attacks where the attacker has to distinguish
9066 between bad padding and a MAC verification error. (CVE-2003-0078)
9068 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9069 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9070 Martin Vuagnoux (EPFL, Ilion)]
9072 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9074 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9075 memory from its contents. This is done with a counter that will
9076 place alternating values in each byte. This can be used to solve
9077 two issues: 1) the removal of calls to memset() by highly optimizing
9078 compilers, and 2) cleansing with other values than 0, since those can
9079 be read through on certain media, for example a swap space on disk.
9082 *) Bugfix: client side session caching did not work with external caching,
9083 because the session->cipher setting was not restored when reloading
9084 from the external cache. This problem was masked, when
9085 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9086 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9089 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9090 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9091 [Zeev Lieber <zeev-l@yahoo.com>]
9093 *) Undo an undocumented change introduced in 0.9.6e which caused
9094 repeated calls to OpenSSL_add_all_ciphers() and
9095 OpenSSL_add_all_digests() to be ignored, even after calling
9099 *) Change the default configuration reader to deal with last line not
9100 being properly terminated.
9103 *) Change X509_NAME_cmp() so it applies the special rules on handling
9104 DN values that are of type PrintableString, as well as RDNs of type
9105 emailAddress where the value has the type ia5String.
9106 [stefank@valicert.com via Richard Levitte]
9108 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9109 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9110 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9111 the bitwise-OR of the two for use by the majority of applications
9112 wanting this behaviour, and update the docs. The documented
9113 behaviour and actual behaviour were inconsistent and had been
9114 changing anyway, so this is more a bug-fix than a behavioural
9116 [Geoff Thorpe, diagnosed by Nadav Har'El]
9118 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9119 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9122 *) Fix initialization code race conditions in
9123 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9124 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9125 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9126 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9127 ssl2_get_cipher_by_char(),
9128 ssl3_get_cipher_by_char().
9129 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9131 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9132 the cached sessions are flushed, as the remove_cb() might use ex_data
9133 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9134 (see [openssl.org #212]).
9135 [Geoff Thorpe, Lutz Jaenicke]
9137 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9138 length, instead of the encoding length to d2i_ASN1_OBJECT.
9141 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9143 *) [In 0.9.6g-engine release:]
9144 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9145 [Lynn Gazis <lgazis@rainbow.com>]
9147 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9149 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9150 and get fix the header length calculation.
9151 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9152 Alon Kantor <alonk@checkpoint.com> (and others),
9155 *) Use proper error handling instead of 'assertions' in buffer
9156 overflow checks added in 0.9.6e. This prevents DoS (the
9157 assertions could call abort()).
9158 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9160 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9162 *) Add various sanity checks to asn1_get_length() to reject
9163 the ASN1 length bytes if they exceed sizeof(long), will appear
9164 negative or the content length exceeds the length of the
9166 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9168 *) Fix cipher selection routines: ciphers without encryption had no flags
9169 for the cipher strength set and where therefore not handled correctly
9170 by the selection routines (PR #130).
9173 *) Fix EVP_dsa_sha macro.
9177 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9178 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9179 that was added in OpenSSL 0.9.6d.
9181 As the countermeasure turned out to be incompatible with some
9182 broken SSL implementations, the new option is part of SSL_OP_ALL.
9183 SSL_OP_ALL is usually employed when compatibility with weird SSL
9184 implementations is desired (e.g. '-bugs' option to 's_client' and
9185 's_server'), so the new option is automatically set in many
9189 *) Changes in security patch:
9191 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9192 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9193 Air Force Materiel Command, USAF, under agreement number
9196 *) Add various sanity checks to asn1_get_length() to reject
9197 the ASN1 length bytes if they exceed sizeof(long), will appear
9198 negative or the content length exceeds the length of the
9199 supplied buffer. (CVE-2002-0659)
9200 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9202 *) Assertions for various potential buffer overflows, not known to
9204 [Ben Laurie (CHATS)]
9206 *) Various temporary buffers to hold ASCII versions of integers were
9207 too small for 64 bit platforms. (CVE-2002-0655)
9208 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9210 *) Remote buffer overflow in SSL3 protocol - an attacker could
9211 supply an oversized session ID to a client. (CVE-2002-0656)
9212 [Ben Laurie (CHATS)]
9214 *) Remote buffer overflow in SSL2 protocol - an attacker could
9215 supply an oversized client master key. (CVE-2002-0656)
9216 [Ben Laurie (CHATS)]
9218 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9220 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9221 encoded as NULL) with id-dsa-with-sha1.
9222 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9224 *) Check various X509_...() return values in apps/req.c.
9225 [Nils Larsch <nla@trustcenter.de>]
9227 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9228 an end-of-file condition would erroneously be flagged, when the CRLF
9229 was just at the end of a processed block. The bug was discovered when
9230 processing data through a buffering memory BIO handing the data to a
9231 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9232 <ptsekov@syntrex.com> and Nedelcho Stanev.
9235 *) Implement a countermeasure against a vulnerability recently found
9236 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9237 before application data chunks to avoid the use of known IVs
9238 with data potentially chosen by the attacker.
9241 *) Fix length checks in ssl3_get_client_hello().
9244 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9245 to prevent ssl3_read_internal() from incorrectly assuming that
9246 ssl3_read_bytes() found application data while handshake
9247 processing was enabled when in fact s->s3->in_read_app_data was
9248 merely automatically cleared during the initial handshake.
9249 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9251 *) Fix object definitions for Private and Enterprise: they were not
9252 recognized in their shortname (=lowercase) representation. Extend
9253 obj_dat.pl to issue an error when using undefined keywords instead
9254 of silently ignoring the problem (Svenning Sorensen
9255 <sss@sss.dnsalias.net>).
9258 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9259 generators, i.e. generators other than 2 and 5. (Previously, the
9260 code did not properly initialise the 'add' and 'rem' values to
9261 BN_generate_prime().)
9263 In the new general case, we do not insist that 'generator' is
9264 actually a primitive root: This requirement is rather pointless;
9265 a generator of the order-q subgroup is just as good, if not
9269 *) Map new X509 verification errors to alerts. Discovered and submitted by
9270 Tom Wu <tom@arcot.com>.
9273 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9274 returning non-zero before the data has been completely received
9275 when using non-blocking I/O.
9276 [Bodo Moeller; problem pointed out by John Hughes]
9278 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9279 [Ben Laurie, Lutz Jaenicke]
9281 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9282 Yoram Zahavi <YoramZ@gilian.com>).
9285 *) Add information about CygWin 1.3 and on, and preserve proper
9286 configuration for the versions before that.
9287 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9289 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9290 check whether we deal with a copy of a session and do not delete from
9291 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9292 <izhar@checkpoint.com>.
9295 *) Do not store session data into the internal session cache, if it
9296 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9297 flag is set). Proposed by Aslam <aslam@funk.com>.
9300 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9304 *) [In 0.9.6d-engine release:]
9305 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9306 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9308 *) Add the configuration target linux-s390x.
9309 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9311 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9312 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9313 variable as an indication that a ClientHello message has been
9314 received. As the flag value will be lost between multiple
9315 invocations of ssl3_accept when using non-blocking I/O, the
9316 function may not be aware that a handshake has actually taken
9317 place, thus preventing a new session from being added to the
9320 To avoid this problem, we now set s->new_session to 2 instead of
9321 using a local variable.
9322 [Lutz Jaenicke, Bodo Moeller]
9324 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9325 if the SSL_R_LENGTH_MISMATCH error is detected.
9326 [Geoff Thorpe, Bodo Moeller]
9328 *) New 'shared_ldflag' column in Configure platform table.
9331 *) Fix EVP_CIPHER_mode macro.
9332 ["Dan S. Camper" <dan@bti.net>]
9334 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9335 type, we must throw them away by setting rr->length to 0.
9336 [D P Chang <dpc@qualys.com>]
9338 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9340 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9341 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9342 worked incorrectly for those cases where range = 10..._2 and
9343 3*range is two bits longer than range.)
9346 *) Only add signing time to PKCS7 structures if it is not already
9350 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9351 OBJ_ld_ce should be OBJ_id_ce.
9352 Also some ip-pda OIDs in crypto/objects/objects.txt were
9353 incorrect (cf. RFC 3039).
9354 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9356 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9357 returns early because it has nothing to do.
9358 [Andy Schneider <andy.schneider@bjss.co.uk>]
9360 *) [In 0.9.6c-engine release:]
9361 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9362 [Andy Schneider <andy.schneider@bjss.co.uk>]
9364 *) [In 0.9.6c-engine release:]
9365 Add support for Cryptographic Appliance's keyserver technology.
9366 (Use engine 'keyclient')
9367 [Cryptographic Appliances and Geoff Thorpe]
9369 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9370 is called via tools/c89.sh because arguments have to be
9371 rearranged (all '-L' options must appear before the first object
9373 [Richard Shapiro <rshapiro@abinitio.com>]
9375 *) [In 0.9.6c-engine release:]
9376 Add support for Broadcom crypto accelerator cards, backported
9378 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9380 *) [In 0.9.6c-engine release:]
9381 Add support for SureWare crypto accelerator cards from
9382 Baltimore Technologies. (Use engine 'sureware')
9383 [Baltimore Technologies and Mark Cox]
9385 *) [In 0.9.6c-engine release:]
9386 Add support for crypto accelerator cards from Accelerated
9387 Encryption Processing, www.aep.ie. (Use engine 'aep')
9388 [AEP Inc. and Mark Cox]
9390 *) Add a configuration entry for gcc on UnixWare.
9391 [Gary Benson <gbenson@redhat.com>]
9393 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9394 messages are stored in a single piece (fixed-length part and
9395 variable-length part combined) and fix various bugs found on the way.
9398 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9399 instead. BIO_gethostbyname() does not know what timeouts are
9400 appropriate, so entries would stay in cache even when they have
9402 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9404 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9405 faced with a pathologically small ClientHello fragment that does
9406 not contain client_version: Instead of aborting with an error,
9407 simply choose the highest available protocol version (i.e.,
9408 TLS 1.0 unless it is disabled). In practice, ClientHello
9409 messages are never sent like this, but this change gives us
9410 strictly correct behaviour at least for TLS.
9413 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9414 never resets s->method to s->ctx->method when called from within
9415 one of the SSL handshake functions.
9416 [Bodo Moeller; problem pointed out by Niko Baric]
9418 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9419 (sent using the client's version number) if client_version is
9420 smaller than the protocol version in use. Also change
9421 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9422 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9423 the client will at least see that alert.
9426 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9430 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9431 client receives HelloRequest while in a handshake.
9432 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9434 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9435 should end in 'break', not 'goto end' which circumvents various
9436 cleanups done in state SSL_ST_OK. But session related stuff
9437 must be disabled for SSL_ST_OK in the case that we just sent a
9440 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9441 before just sending a HelloRequest.
9442 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9444 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9445 reveal whether illegal block cipher padding was found or a MAC
9446 verification error occurred. (Neither SSLerr() codes nor alerts
9447 are directly visible to potential attackers, but the information
9448 may leak via logfiles.)
9450 Similar changes are not required for the SSL 2.0 implementation
9451 because the number of padding bytes is sent in clear for SSL 2.0,
9452 and the extra bytes are just ignored. However ssl/s2_pkt.c
9453 failed to verify that the purported number of padding bytes is in
9457 *) Add OpenUNIX-8 support including shared libraries
9458 (Boyd Lynn Gerber <gerberb@zenez.com>).
9461 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9462 'wristwatch attack' using huge encoding parameters (cf.
9463 James H. Manger's CRYPTO 2001 paper). Note that the
9464 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9465 encoding parameters and hence was not vulnerable.
9468 *) BN_sqr() bug fix.
9469 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9471 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9472 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9473 followed by modular reduction.
9474 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9476 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9477 equivalent based on BN_pseudo_rand() instead of BN_rand().
9480 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9481 This function was broken, as the check for a new client hello message
9482 to handle SGC did not allow these large messages.
9483 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9486 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9489 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9490 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9493 *) Rework the configuration and shared library support for Tru64 Unix.
9494 The configuration part makes use of modern compiler features and
9495 still retains old compiler behavior for those that run older versions
9496 of the OS. The shared library support part includes a variant that
9497 uses the RPATH feature, and is available through the special
9498 configuration target "alpha-cc-rpath", which will never be selected
9500 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9502 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9503 with the same message size as in ssl3_get_certificate_request().
9504 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9505 messages might inadvertently be reject as too long.
9506 [Petr Lampa <lampa@fee.vutbr.cz>]
9508 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9511 *) Modified SSL library such that the verify_callback that has been set
9512 specificly for an SSL object with SSL_set_verify() is actually being
9513 used. Before the change, a verify_callback set with this function was
9514 ignored and the verify_callback() set in the SSL_CTX at the time of
9515 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9516 to allow the necessary settings.
9519 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9520 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9521 done automatically (in contradiction to the requirements of the C
9522 standard). This made problems when used from OpenSSH.
9525 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9526 dh->length and always used
9528 BN_rand_range(priv_key, dh->p).
9530 BN_rand_range() is not necessary for Diffie-Hellman, and this
9531 specific range makes Diffie-Hellman unnecessarily inefficient if
9532 dh->length (recommended exponent length) is much smaller than the
9533 length of dh->p. We could use BN_rand_range() if the order of
9534 the subgroup was stored in the DH structure, but we only have
9539 BN_rand(priv_key, l, ...)
9541 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9547 RSA_eay_public_encrypt
9548 RSA_eay_private_decrypt
9549 RSA_eay_private_encrypt (signing)
9550 RSA_eay_public_decrypt (signature verification)
9552 (default implementations for RSA_public_encrypt,
9553 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9554 always reject numbers >= n.
9557 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9558 to synchronize access to 'locking_thread'. This is necessary on
9559 systems where access to 'locking_thread' (an 'unsigned long'
9560 variable) is not atomic.
9563 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9564 *before* setting the 'crypto_lock_rand' flag. The previous code had
9565 a race condition if 0 is a valid thread ID.
9566 [Travis Vitek <vitek@roguewave.com>]
9568 *) Add support for shared libraries under Irix.
9569 [Albert Chin-A-Young <china@thewrittenword.com>]
9571 *) Add configuration option to build on Linux on both big-endian and
9573 [Ralf Baechle <ralf@uni-koblenz.de>]
9575 *) Add the possibility to create shared libraries on HP-UX.
9578 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9580 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9581 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9582 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9583 PRNG state recovery was possible based on the output of
9584 one PRNG request appropriately sized to gain knowledge on
9585 'md' followed by enough consecutive 1-byte PRNG requests
9586 to traverse all of 'state'.
9588 1. When updating 'md_local' (the current thread's copy of 'md')
9589 during PRNG output generation, hash all of the previous
9590 'md_local' value, not just the half used for PRNG output.
9592 2. Make the number of bytes from 'state' included into the hash
9593 independent from the number of PRNG bytes requested.
9595 The first measure alone would be sufficient to avoid
9596 Markku-Juhani's attack. (Actually it had never occurred
9597 to me that the half of 'md_local' used for chaining was the
9598 half from which PRNG output bytes were taken -- I had always
9599 assumed that the secret half would be used.) The second
9600 measure makes sure that additional data from 'state' is never
9601 mixed into 'md_local' in small portions; this heuristically
9602 further strengthens the PRNG.
9605 *) Fix crypto/bn/asm/mips3.s.
9608 *) When only the key is given to "enc", the IV is undefined. Print out
9609 an error message in this case.
9612 *) Handle special case when X509_NAME is empty in X509 printing routines.
9615 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9616 positive and less than q.
9619 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9620 used: it isn't thread safe and the add_lock_callback should handle
9622 [Paul Rose <Paul.Rose@bridge.com>]
9624 *) Verify that incoming data obeys the block size in
9625 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9629 [Ulf Möller, Bodo Möller]
9631 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9632 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9633 when fixing the server behaviour for backwards-compatible 'client
9634 hello' messages. (Note that the attack is impractical against
9635 SSL 3.0 and TLS 1.0 anyway because length and version checking
9636 means that the probability of guessing a valid ciphertext is
9637 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9640 Before 0.9.5, the countermeasure (hide the error by generating a
9641 random 'decryption result') did not work properly because
9642 ERR_clear_error() was missing, meaning that SSL_get_error() would
9643 detect the supposedly ignored error.
9645 Both problems are now fixed.
9648 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9649 (previously it was 1024).
9652 *) Fix for compatibility mode trust settings: ignore trust settings
9653 unless some valid trust or reject settings are present.
9656 *) Fix for blowfish EVP: its a variable length cipher.
9659 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9660 parameters in DSA public key structures and return an error in the
9661 DSA routines if parameters are absent.
9664 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9665 in the current directory if neither $RANDFILE nor $HOME was set.
9666 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9667 caused some confusion to Windows users who haven't defined $HOME.
9668 Thus RAND_file_name() is changed again: e_os.h can define a
9669 DEFAULT_HOME, which will be used if $HOME is not set.
9670 For Windows, we use "C:"; on other platforms, we still require
9671 environment variables.
9673 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9674 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9675 having multiple threads call RAND_poll() concurrently.
9678 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9679 combination of a flag and a thread ID variable.
9680 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9681 flag), *other* threads can enter ssleay_add_bytes without obeying
9682 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9683 that they do not hold after the first thread unsets add_do_not_lock).
9686 *) Change bctest again: '-x' expressions are not available in all
9690 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9692 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9693 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9695 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9696 the default extension for executables, if any. Also, make the perl
9697 scripts that use symlink() to test if it really exists and use "cp"
9698 if it doesn't. All this made OpenSSL compilable and installable in
9702 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9703 If SEQUENCE is length is indefinite just set c->slen to the total
9704 amount of data available.
9705 [Steve Henson, reported by shige@FreeBSD.org]
9706 [This change does not apply to 0.9.7.]
9708 *) Change bctest to avoid here-documents inside command substitution
9709 (workaround for FreeBSD /bin/sh bug).
9710 For compatibility with Ultrix, avoid shell functions (introduced
9711 in the bctest version that searches along $PATH).
9714 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9715 with des_encrypt() defined on some operating systems, like Solaris
9719 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9720 On the Importance of Eliminating Errors in Cryptographic
9721 Computations, J. Cryptology 14 (2001) 2, 101-119,
9722 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9725 *) MIPS assembler BIGNUM division bug fix.
9728 *) Disabled incorrect Alpha assembler code.
9731 *) Fix PKCS#7 decode routines so they correctly update the length
9732 after reading an EOC for the EXPLICIT tag.
9734 [This change does not apply to 0.9.7.]
9736 *) Fix bug in PKCS#12 key generation routines. This was triggered
9737 if a 3DES key was generated with a 0 initial byte. Include
9738 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9739 (but broken) behaviour.
9742 *) Enhance bctest to search for a working bc along $PATH and print
9744 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9746 *) Fix memory leaks in err.c: free err_data string if necessary;
9747 don't write to the wrong index in ERR_set_error_data.
9750 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9754 *) Replace rdtsc with _emit statements for VC++ version 5.
9755 [Jeremy Cooper <jeremy@baymoo.org>]
9757 *) Make it possible to reuse SSLv2 sessions.
9760 *) In copy_email() check for >= 0 as a return value for
9761 X509_NAME_get_index_by_NID() since 0 is a valid index.
9762 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9764 *) Avoid coredump with unsupported or invalid public keys by checking if
9765 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9766 PKCS7_verify() fails with non detached data.
9769 *) Don't use getenv in library functions when run as setuid/setgid.
9770 New function OPENSSL_issetugid().
9773 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9774 due to incorrect handling of multi-threading:
9776 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9778 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9780 3. Count how many times MemCheck_off() has been called so that
9781 nested use can be treated correctly. This also avoids
9782 inband-signalling in the previous code (which relied on the
9783 assumption that thread ID 0 is impossible).
9786 *) Add "-rand" option also to s_client and s_server.
9789 *) Fix CPU detection on Irix 6.x.
9790 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9791 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9793 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9796 [This change does not apply to 0.9.7.]
9798 *) Use the cached encoding of an X509_NAME structure rather than
9799 copying it. This is apparently the reason for the libsafe "errors"
9800 but the code is actually correct.
9803 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9804 Bleichenbacher's DSA attack.
9805 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9806 to be set and top=0 forces the highest bit to be set; top=-1 is new
9807 and leaves the highest bit random.
9808 [Ulf Moeller, Bodo Moeller]
9810 *) In the NCONF_...-based implementations for CONF_... queries
9811 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9812 a temporary CONF structure with the data component set to NULL
9813 (which gives segmentation faults in lh_retrieve).
9814 Instead, use NULL for the CONF pointer in CONF_get_string and
9815 CONF_get_number (which may use environment variables) and directly
9816 return NULL from CONF_get_section.
9819 *) Fix potential buffer overrun for EBCDIC.
9822 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9823 keyUsage if basicConstraints absent for a CA.
9826 *) Make SMIME_write_PKCS7() write mail header values with a format that
9827 is more generally accepted (no spaces before the semicolon), since
9828 some programs can't parse those values properly otherwise. Also make
9829 sure BIO's that break lines after each write do not create invalid
9833 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9834 macros previously used would not encode an empty SEQUENCE OF
9835 and break the signature.
9837 [This change does not apply to 0.9.7.]
9839 *) Zero the premaster secret after deriving the master secret in
9843 *) Add some EVP_add_digest_alias registrations (as found in
9844 OpenSSL_add_all_digests()) to SSL_library_init()
9845 aka OpenSSL_add_ssl_algorithms(). This provides improved
9846 compatibility with peers using X.509 certificates
9847 with unconventional AlgorithmIdentifier OIDs.
9850 *) Fix for Irix with NO_ASM.
9851 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9853 *) ./config script fixes.
9854 [Ulf Moeller, Richard Levitte]
9856 *) Fix 'openssl passwd -1'.
9859 *) Change PKCS12_key_gen_asc() so it can cope with non null
9860 terminated strings whose length is passed in the passlen
9861 parameter, for example from PEM callbacks. This was done
9862 by adding an extra length parameter to asc2uni().
9863 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9865 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9866 call failed, free the DSA structure.
9869 *) Fix to uni2asc() to cope with zero length Unicode strings.
9870 These are present in some PKCS#12 files.
9873 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9874 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9875 when writing a 32767 byte record.
9876 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9878 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9879 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9881 (RSA objects have a reference count access to which is protected
9882 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9883 so they are meant to be shared between threads.)
9884 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9885 "Reddie, Steven" <Steven.Reddie@ca.com>]
9887 *) Fix a deadlock in CRYPTO_mem_leaks().
9890 *) Use better test patterns in bntest.
9893 *) rand_win.c fix for Borland C.
9896 *) BN_rshift bugfix for n == 0.
9899 *) Add a 'bctest' script that checks for some known 'bc' bugs
9900 so that 'make test' does not abort just because 'bc' is broken.
9903 *) Store verify_result within SSL_SESSION also for client side to
9904 avoid potential security hole. (Re-used sessions on the client side
9905 always resulted in verify_result==X509_V_OK, not using the original
9906 result of the server certificate verification.)
9909 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9910 SSL3_RT_APPLICATION_DATA, return 0.
9911 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9915 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9916 releases, have been re-implemented by renaming the previous
9917 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9918 and ssl3_read_internal, respectively, and adding 'peek' parameters
9919 to them. The new ssl[23]_{read,peek} functions are calls to
9920 ssl[23]_read_internal with the 'peek' flag set appropriately.
9921 A 'peek' parameter has also been added to ssl3_read_bytes, which
9922 does the actual work for ssl3_read_internal.
9925 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9926 the method-specific "init()" handler. Also clean up ex_data after
9927 calling the method-specific "finish()" handler. Previously, this was
9928 happening the other way round.
9931 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9932 The previous value, 12, was not always sufficient for BN_mod_exp().
9935 *) Make sure that shared libraries get the internal name engine with
9936 the full version number and not just 0. This should mark the
9937 shared libraries as not backward compatible. Of course, this should
9938 be changed again when we can guarantee backward binary compatibility.
9941 *) Fix typo in get_cert_by_subject() in by_dir.c
9942 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9944 *) Rework the system to generate shared libraries:
9946 - Make note of the expected extension for the shared libraries and
9947 if there is a need for symbolic links from for example libcrypto.so.0
9948 to libcrypto.so.0.9.7. There is extended info in Configure for
9951 - Make as few rebuilds of the shared libraries as possible.
9953 - Still avoid linking the OpenSSL programs with the shared libraries.
9955 - When installing, install the shared libraries separately from the
9959 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9961 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9962 and not in SSL_clear because the latter is also used by the
9963 accept/connect functions; previously, the settings made by
9964 SSL_set_read_ahead would be lost during the handshake.
9965 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9967 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9968 Previously, it would create entries for disabled algorithms no
9972 *) Added several new manual pages for SSL_* function.
9975 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9977 *) In ssl23_get_client_hello, generate an error message when faced
9978 with an initial SSL 3.0/TLS record that is too small to contain the
9979 first two bytes of the ClientHello message, i.e. client_version.
9980 (Note that this is a pathologic case that probably has never happened
9981 in real life.) The previous approach was to use the version number
9982 from the record header as a substitute; but our protocol choice
9983 should not depend on that one because it is not authenticated
9984 by the Finished messages.
9987 *) More robust randomness gathering functions for Windows.
9988 [Jeffrey Altman <jaltman@columbia.edu>]
9990 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9991 not set then we don't setup the error code for issuer check errors
9992 to avoid possibly overwriting other errors which the callback does
9993 handle. If an application does set the flag then we assume it knows
9994 what it is doing and can handle the new informational codes
9998 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9999 a general "ANY" type, as such it should be able to decode anything
10000 including tagged types. However it didn't check the class so it would
10001 wrongly interpret tagged types in the same way as their universal
10002 counterpart and unknown types were just rejected. Changed so that the
10003 tagged and unknown types are handled in the same way as a SEQUENCE:
10004 that is the encoding is stored intact. There is also a new type
10005 "V_ASN1_OTHER" which is used when the class is not universal, in this
10006 case we have no idea what the actual type is so we just lump them all
10010 *) On VMS, stdout may very well lead to a file that is written to
10011 in a record-oriented fashion. That means that every write() will
10012 write a separate record, which will be read separately by the
10013 programs trying to read from it. This can be very confusing.
10015 The solution is to put a BIO filter in the way that will buffer
10016 text until a linefeed is reached, and then write everything a
10017 line at a time, so every record written will be an actual line,
10018 not chunks of lines and not (usually doesn't happen, but I've
10019 seen it once) several lines in one record. BIO_f_linebuffer() is
10022 Currently, it's a VMS-only method, because that's where it has
10023 been tested well enough.
10026 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
10027 it can return incorrect results.
10028 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
10029 but it was in 0.9.6-beta[12].)
10032 *) Disable the check for content being present when verifying detached
10033 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
10034 include zero length content when signing messages.
10037 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
10038 BIO_ctrl (for BIO pairs).
10041 *) Add DSO method for VMS.
10044 *) Bug fix: Montgomery multiplication could produce results with the
10048 *) Add RPM specification openssl.spec and modify it to build three
10049 packages. The default package contains applications, application
10050 documentation and run-time libraries. The devel package contains
10051 include files, static libraries and function documentation. The
10052 doc package contains the contents of the doc directory. The original
10053 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10056 *) Add a large number of documentation files for many SSL routines.
10057 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10059 *) Add a configuration entry for Sony News 4.
10060 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10062 *) Don't set the two most significant bits to one when generating a
10063 random number < q in the DSA library.
10066 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10067 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10068 the underlying transport is blocking) if a handshake took place.
10069 (The default behaviour is needed by applications such as s_client
10070 and s_server that use select() to determine when to use SSL_read;
10071 but for applications that know in advance when to expect data, it
10072 just makes things more complicated.)
10075 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10079 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10080 work better on such systems.
10081 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10083 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10084 Update PKCS12_parse() so it copies the friendlyName and the
10085 keyid to the certificates aux info.
10088 *) Fix bug in PKCS7_verify() which caused an infinite loop
10089 if there was more than one signature.
10090 [Sven Uszpelkat <su@celocom.de>]
10092 *) Major change in util/mkdef.pl to include extra information
10093 about each symbol, as well as presenting variables as well
10094 as functions. This change means that there's n more need
10095 to rebuild the .num files when some algorithms are excluded.
10098 *) Allow the verify time to be set by an application,
10099 rather than always using the current time.
10102 *) Phase 2 verify code reorganisation. The certificate
10103 verify code now looks up an issuer certificate by a
10104 number of criteria: subject name, authority key id
10105 and key usage. It also verifies self signed certificates
10106 by the same criteria. The main comparison function is
10107 X509_check_issued() which performs these checks.
10109 Lot of changes were necessary in order to support this
10110 without completely rewriting the lookup code.
10112 Authority and subject key identifier are now cached.
10114 The LHASH 'certs' is X509_STORE has now been replaced
10115 by a STACK_OF(X509_OBJECT). This is mainly because an
10116 LHASH can't store or retrieve multiple objects with
10117 the same hash value.
10119 As a result various functions (which were all internal
10120 use only) have changed to handle the new X509_STORE
10121 structure. This will break anything that messed round
10122 with X509_STORE internally.
10124 The functions X509_STORE_add_cert() now checks for an
10125 exact match, rather than just subject name.
10127 The X509_STORE API doesn't directly support the retrieval
10128 of multiple certificates matching a given criteria, however
10129 this can be worked round by performing a lookup first
10130 (which will fill the cache with candidate certificates)
10131 and then examining the cache for matches. This is probably
10132 the best we can do without throwing out X509_LOOKUP
10133 entirely (maybe later...).
10135 The X509_VERIFY_CTX structure has been enhanced considerably.
10137 All certificate lookup operations now go via a get_issuer()
10138 callback. Although this currently uses an X509_STORE it
10139 can be replaced by custom lookups. This is a simple way
10140 to bypass the X509_STORE hackery necessary to make this
10141 work and makes it possible to use more efficient techniques
10142 in future. A very simple version which uses a simple
10143 STACK for its trusted certificate store is also provided
10144 using X509_STORE_CTX_trusted_stack().
10146 The verify_cb() and verify() callbacks now have equivalents
10147 in the X509_STORE_CTX structure.
10149 X509_STORE_CTX also has a 'flags' field which can be used
10150 to customise the verify behaviour.
10153 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10154 excludes S/MIME capabilities.
10157 *) When a certificate request is read in keep a copy of the
10158 original encoding of the signed data and use it when outputting
10159 again. Signatures then use the original encoding rather than
10160 a decoded, encoded version which may cause problems if the
10161 request is improperly encoded.
10164 *) For consistency with other BIO_puts implementations, call
10165 buffer_write(b, ...) directly in buffer_puts instead of calling
10168 In BIO_puts, increment b->num_write as in BIO_write.
10169 [Peter.Sylvester@EdelWeb.fr]
10171 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10172 BN_zero, we may not return a BIGNUM with an array consisting of
10173 words set to zero.)
10176 *) Avoid calling abort() from within the library when problems are
10177 detected, except if preprocessor symbols have been defined
10178 (such as REF_CHECK, BN_DEBUG etc.).
10181 *) New openssl application 'rsautl'. This utility can be
10182 used for low level RSA operations. DER public key
10183 BIO/fp routines also added.
10186 *) New Configure entry and patches for compiling on QNX 4.
10187 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10189 *) A demo state-machine implementation was sponsored by
10190 Nuron (http://www.nuron.com/) and is now available in
10191 demos/state_machine.
10194 *) New options added to the 'dgst' utility for signature
10195 generation and verification.
10198 *) Unrecognized PKCS#7 content types are now handled via a
10199 catch all ASN1_TYPE structure. This allows unsupported
10200 types to be stored as a "blob" and an application can
10201 encode and decode it manually.
10204 *) Fix various signed/unsigned issues to make a_strex.c
10205 compile under VC++.
10206 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10208 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10209 length if passed a buffer. ASN1_INTEGER_to_BN failed
10210 if passed a NULL BN and its argument was negative.
10211 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10213 *) Modification to PKCS#7 encoding routines to output definite
10214 length encoding. Since currently the whole structures are in
10215 memory there's not real point in using indefinite length
10216 constructed encoding. However if OpenSSL is compiled with
10217 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10220 *) Added BIO_vprintf() and BIO_vsnprintf().
10223 *) Added more prefixes to parse for in the strings written
10224 through a logging bio, to cover all the levels that are available
10225 through syslog. The prefixes are now:
10227 PANIC, EMERG, EMR => LOG_EMERG
10228 ALERT, ALR => LOG_ALERT
10229 CRIT, CRI => LOG_CRIT
10230 ERROR, ERR => LOG_ERR
10231 WARNING, WARN, WAR => LOG_WARNING
10232 NOTICE, NOTE, NOT => LOG_NOTICE
10233 INFO, INF => LOG_INFO
10234 DEBUG, DBG => LOG_DEBUG
10236 and as before, if none of those prefixes are present at the
10237 beginning of the string, LOG_ERR is chosen.
10239 On Win32, the LOG_* levels are mapped according to this:
10241 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10242 LOG_WARNING => EVENTLOG_WARNING_TYPE
10243 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10247 *) Made it possible to reconfigure with just the configuration
10248 argument "reconf" or "reconfigure". The command line arguments
10249 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10250 and are retrieved from there when reconfiguring.
10253 *) MD4 implemented.
10254 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10256 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10259 *) The obj_dat.pl script was messing up the sorting of object
10260 names. The reason was that it compared the quoted version
10261 of strings as a result "OCSP" > "OCSP Signing" because
10262 " > SPACE. Changed script to store unquoted versions of
10263 names and add quotes on output. It was also omitting some
10264 names from the lookup table if they were given a default
10265 value (that is if SN is missing it is given the same
10266 value as LN and vice versa), these are now added on the
10267 grounds that if an object has a name we should be able to
10268 look it up. Finally added warning output when duplicate
10269 short or long names are found.
10272 *) Changes needed for Tandem NSK.
10273 [Scott Uroff <scott@xypro.com>]
10275 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10276 RSA_padding_check_SSLv23(), special padding was never detected
10277 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10278 version rollback attacks was not effective.
10280 In s23_clnt.c, don't use special rollback-attack detection padding
10281 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10282 client; similarly, in s23_srvr.c, don't do the rollback check if
10283 SSL 2.0 is the only protocol enabled in the server.
10286 *) Make it possible to get hexdumps of unprintable data with 'openssl
10287 asn1parse'. By implication, the functions ASN1_parse_dump() and
10288 BIO_dump_indent() are added.
10291 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10292 these print out strings and name structures based on various
10293 flags including RFC2253 support and proper handling of
10294 multibyte characters. Added options to the 'x509' utility
10295 to allow the various flags to be set.
10298 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10299 Also change the functions X509_cmp_current_time() and
10300 X509_gmtime_adj() work with an ASN1_TIME structure,
10301 this will enable certificates using GeneralizedTime in validity
10302 dates to be checked.
10305 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10306 negative public key encodings) on by default,
10307 NO_NEG_PUBKEY_BUG can be set to disable it.
10310 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10311 content octets. An i2c_ASN1_OBJECT is unnecessary because
10312 the encoding can be trivially obtained from the structure.
10315 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10316 not read locks (CRYPTO_r_[un]lock).
10319 *) A first attempt at creating official support for shared
10320 libraries through configuration. I've kept it so the
10321 default is static libraries only, and the OpenSSL programs
10322 are always statically linked for now, but there are
10323 preparations for dynamic linking in place.
10324 This has been tested on Linux and Tru64.
10327 *) Randomness polling function for Win9x, as described in:
10328 Peter Gutmann, Software Generation of Practically Strong
10332 *) Fix so PRNG is seeded in req if using an already existing
10336 *) New options to smime application. -inform and -outform
10337 allow alternative formats for the S/MIME message including
10338 PEM and DER. The -content option allows the content to be
10339 specified separately. This should allow things like Netscape
10340 form signing output easier to verify.
10343 *) Fix the ASN1 encoding of tags using the 'long form'.
10346 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10347 STRING types. These convert content octets to and from the
10348 underlying type. The actual tag and length octets are
10349 already assumed to have been read in and checked. These
10350 are needed because all other string types have virtually
10351 identical handling apart from the tag. By having versions
10352 of the ASN1 functions that just operate on content octets
10353 IMPLICIT tagging can be handled properly. It also allows
10354 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10355 and ASN1_INTEGER are identical apart from the tag.
10358 *) Change the handling of OID objects as follows:
10360 - New object identifiers are inserted in objects.txt, following
10361 the syntax given in objects.README.
10362 - objects.pl is used to process obj_mac.num and create a new
10364 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10367 This is currently kind of a hack, and the perl code in objects.pl
10368 isn't very elegant, but it works as I intended. The simplest way
10369 to check that it worked correctly is to look in obj_dat.h and
10370 check the array nid_objs and make sure the objects haven't moved
10371 around (this is important!). Additions are OK, as well as
10372 consistent name changes.
10375 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10378 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10379 The given file adds to whatever has already been seeded into the
10380 random pool through the RANDFILE configuration file option or
10381 environment variable, or the default random state file.
10384 *) mkstack.pl now sorts each macro group into lexical order.
10385 Previously the output order depended on the order the files
10386 appeared in the directory, resulting in needless rewriting
10390 *) Patches to make OpenSSL compile under Win32 again. Mostly
10391 work arounds for the VC++ problem that it treats func() as
10392 func(void). Also stripped out the parts of mkdef.pl that
10393 added extra typesafe functions: these no longer exist.
10396 *) Reorganisation of the stack code. The macros are now all
10397 collected in safestack.h . Each macro is defined in terms of
10398 a "stack macro" of the form SKM_<name>(type, a, b). The
10399 DEBUG_SAFESTACK is now handled in terms of function casts,
10400 this has the advantage of retaining type safety without the
10401 use of additional functions. If DEBUG_SAFESTACK is not defined
10402 then the non typesafe macros are used instead. Also modified the
10403 mkstack.pl script to handle the new form. Needs testing to see
10404 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10405 the default if no major problems. Similar behaviour for ASN1_SET_OF
10406 and PKCS12_STACK_OF.
10409 *) When some versions of IIS use the 'NET' form of private key the
10410 key derivation algorithm is different. Normally MD5(password) is
10411 used as a 128 bit RC4 key. In the modified case
10412 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10413 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10414 as the old Netscape_RSA functions except they have an additional
10415 'sgckey' parameter which uses the modified algorithm. Also added
10416 an -sgckey command line option to the rsa utility. Thanks to
10417 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10418 algorithm to openssl-dev.
10421 *) The evp_local.h macros were using 'c.##kname' which resulted in
10422 invalid expansion on some systems (SCO 5.0.5 for example).
10423 Corrected to 'c.kname'.
10424 [Phillip Porch <root@theporch.com>]
10426 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10427 a STACK of email addresses from a certificate or request, these look
10428 in the subject name and the subject alternative name extensions and
10429 omit any duplicate addresses.
10432 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10433 This makes DSA verification about 2 % faster.
10436 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10437 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10438 plus overhead for 1024 bit moduli).
10439 This makes exponentiations about 0.5 % faster for 1024 bit
10440 exponents (as measured by "openssl speed rsa2048").
10443 *) Rename memory handling macros to avoid conflicts with other
10445 Malloc => OPENSSL_malloc
10446 Malloc_locked => OPENSSL_malloc_locked
10447 Realloc => OPENSSL_realloc
10448 Free => OPENSSL_free
10451 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10452 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10455 *) CygWin32 support.
10456 [John Jarvie <jjarvie@newsguy.com>]
10458 *) The type-safe stack code has been rejigged. It is now only compiled
10459 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10460 by default all type-specific stack functions are "#define"d back to
10461 standard stack functions. This results in more streamlined output
10462 but retains the type-safety checking possibilities of the original
10466 *) The STACK code has been cleaned up, and certain type declarations
10467 that didn't make a lot of sense have been brought in line. This has
10468 also involved a cleanup of sorts in safestack.h to more correctly
10469 map type-safe stack functions onto their plain stack counterparts.
10470 This work has also resulted in a variety of "const"ifications of
10471 lots of the code, especially "_cmp" operations which should normally
10472 be prototyped with "const" parameters anyway.
10475 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10476 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10477 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10478 where all of 'md' is used each time the PRNG is used, but 'state'
10479 is used only indexed by a cyclic counter. As entropy may not be
10480 well distributed from the beginning, 'md' is important as a
10481 chaining variable. However, the output function chains only half
10482 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10483 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10484 in all of 'state' being rewritten, with the new values depending
10485 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10488 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10489 the handshake is continued after ssl_verify_cert_chain();
10490 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10491 can lead to 'unexplainable' connection aborts later.
10492 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10494 *) Major EVP API cipher revision.
10495 Add hooks for extra EVP features. This allows various cipher
10496 parameters to be set in the EVP interface. Support added for variable
10497 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10498 setting of RC2 and RC5 parameters.
10500 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10503 Remove lots of duplicated code from the EVP library. For example *every*
10504 cipher init() function handles the 'iv' in the same way according to the
10505 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10506 for CFB and OFB modes they zero ctx->num.
10508 New functionality allows removal of S/MIME code RC2 hack.
10510 Most of the routines have the same form and so can be declared in terms
10513 By shifting this to the top level EVP_CipherInit() it can be removed from
10514 all individual ciphers. If the cipher wants to handle IVs or keys
10515 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10518 Change lots of functions like EVP_EncryptUpdate() to now return a
10519 value: although software versions of the algorithms cannot fail
10520 any installed hardware versions can.
10523 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10524 this option is set, tolerate broken clients that send the negotiated
10525 protocol version number instead of the requested protocol version
10529 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10530 i.e. non-zero for export ciphersuites, zero otherwise.
10531 Previous versions had this flag inverted, inconsistent with
10532 rsa_tmp_cb (..._TMP_RSA_CB).
10533 [Bodo Moeller; problem reported by Amit Chopra]
10535 *) Add missing DSA library text string. Work around for some IIS
10536 key files with invalid SEQUENCE encoding.
10539 *) Add a document (doc/standards.txt) that list all kinds of standards
10540 and so on that are implemented in OpenSSL.
10543 *) Enhance c_rehash script. Old version would mishandle certificates
10544 with the same subject name hash and wouldn't handle CRLs at all.
10545 Added -fingerprint option to crl utility, to support new c_rehash
10549 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10552 *) Fix for SSL server purpose checking. Server checking was
10553 rejecting certificates which had extended key usage present
10554 but no ssl client purpose.
10555 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10557 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10558 is a little unclear about how a blank password is handled.
10559 Since the password in encoded as a BMPString with terminating
10560 double NULL a zero length password would end up as just the
10561 double NULL. However no password at all is different and is
10562 handled differently in the PKCS#12 key generation code. NS
10563 treats a blank password as zero length. MSIE treats it as no
10564 password on export: but it will try both on import. We now do
10565 the same: PKCS12_parse() tries zero length and no password if
10566 the password is set to "" or NULL (NULL is now a valid password:
10567 it wasn't before) as does the pkcs12 application.
10570 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10571 perror when PEM_read_bio_X509_REQ fails, the error message must
10572 be obtained from the error queue.
10575 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10576 it in ERR_remove_state if appropriate, and change ERR_get_state
10577 accordingly to avoid race conditions (this is necessary because
10578 thread_hash is no longer constant once set).
10581 *) Bugfix for linux-elf makefile.one.
10584 *) RSA_get_default_method() will now cause a default
10585 RSA_METHOD to be chosen if one doesn't exist already.
10586 Previously this was only set during a call to RSA_new()
10587 or RSA_new_method(NULL) meaning it was possible for
10588 RSA_get_default_method() to return NULL.
10591 *) Added native name translation to the existing DSO code
10592 that will convert (if the flag to do so is set) filenames
10593 that are sufficiently small and have no path information
10594 into a canonical native form. Eg. "blah" converted to
10595 "libblah.so" or "blah.dll" etc.
10598 *) New function ERR_error_string_n(e, buf, len) which is like
10599 ERR_error_string(e, buf), but writes at most 'len' bytes
10600 including the 0 terminator. For ERR_error_string_n, 'buf'
10602 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10604 *) CONF library reworked to become more general. A new CONF
10605 configuration file reader "class" is implemented as well as a
10606 new functions (NCONF_*, for "New CONF") to handle it. The now
10607 old CONF_* functions are still there, but are reimplemented to
10608 work in terms of the new functions. Also, a set of functions
10609 to handle the internal storage of the configuration data is
10610 provided to make it easier to write new configuration file
10611 reader "classes" (I can definitely see something reading a
10612 configuration file in XML format, for example), called _CONF_*,
10613 or "the configuration storage API"...
10615 The new configuration file reading functions are:
10617 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10618 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10620 NCONF_default, NCONF_WIN32
10622 NCONF_dump_fp, NCONF_dump_bio
10624 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10625 NCONF_new creates a new CONF object. This works in the same way
10626 as other interfaces in OpenSSL, like the BIO interface.
10627 NCONF_dump_* dump the internal storage of the configuration file,
10628 which is useful for debugging. All other functions take the same
10629 arguments as the old CONF_* functions wth the exception of the
10630 first that must be a `CONF *' instead of a `LHASH *'.
10632 To make it easer to use the new classes with the old CONF_* functions,
10633 the function CONF_set_default_method is provided.
10636 *) Add '-tls1' option to 'openssl ciphers', which was already
10637 mentioned in the documentation but had not been implemented.
10638 (This option is not yet really useful because even the additional
10639 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10642 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10643 OpenSSL-based applications) load shared libraries and bind to
10644 them in a portable way.
10645 [Geoff Thorpe, with contributions from Richard Levitte]
10647 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10649 *) Make sure _lrotl and _lrotr are only used with MSVC.
10651 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10652 (the default implementation of RAND_status).
10654 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10655 to '-clrext' (= clear extensions), as intended and documented.
10656 [Bodo Moeller; inconsistency pointed out by Michael Attili
10657 <attili@amaxo.com>]
10659 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10660 was larger than the MD block size.
10661 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10663 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10664 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10665 using the passed key: if the passed key was a private key the result
10666 of X509_print(), for example, would be to print out all the private key
10670 *) des_quad_cksum() byte order bug fix.
10671 [Ulf Möller, using the problem description in krb4-0.9.7, where
10672 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10674 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10676 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10678 *) For easily testing in shell scripts whether some command
10679 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10680 returns with exit code 0 iff no command of the given name is available.
10681 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10682 the output goes to stdout and nothing is printed to stderr.
10683 Additional arguments are always ignored.
10685 Since for each cipher there is a command of the same name,
10686 the 'no-cipher' compilation switches can be tested this way.
10688 ('openssl no-XXX' is not able to detect pseudo-commands such
10689 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10692 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10695 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10696 is set; it will be thrown away anyway because each handshake creates
10698 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10699 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10700 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10701 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10704 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10705 'Q' and 'R' lose their special meanings (quit/renegotiate).
10706 This is part of what -quiet does; unlike -quiet, -ign_eof
10707 does not suppress any output.
10710 *) Add compatibility options to the purpose and trust code. The
10711 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10712 accepts a certificate or CA, this was the previous behaviour,
10713 with all the associated security issues.
10715 X509_TRUST_COMPAT is the old trust behaviour: only and
10716 automatically trust self signed roots in certificate store. A
10717 new trust setting X509_TRUST_DEFAULT is used to specify that
10718 a purpose has no associated trust setting and it should instead
10719 use the value in the default purpose.
10722 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10723 and fix a memory leak.
10726 *) In util/mkerr.pl (which implements 'make errors'), preserve
10727 reason strings from the previous version of the .c file, as
10728 the default to have only downcase letters (and digits) in
10729 automatically generated reasons codes is not always appropriate.
10732 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10733 using strerror. Previously, ERR_reason_error_string() returned
10734 library names as reason strings for SYSerr; but SYSerr is a special
10735 case where small numbers are errno values, not library numbers.
10738 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10739 converts DSA parameters into DH parameters. (When creating parameters,
10740 DSA_generate_parameters is used.)
10743 *) Include 'length' (recommended exponent length) in C code generated
10744 by 'openssl dhparam -C'.
10747 *) The second argument to set_label in perlasm was already being used
10748 so couldn't be used as a "file scope" flag. Moved to third argument
10752 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10753 instead of RAND_bytes for encryption IVs and salts.
10756 *) Include RAND_status() into RAND_METHOD instead of implementing
10757 it only for md_rand.c Otherwise replacing the PRNG by calling
10758 RAND_set_rand_method would be impossible.
10761 *) Don't let DSA_generate_key() enter an infinite loop if the random
10762 number generation fails.
10765 *) New 'rand' application for creating pseudo-random output.
10768 *) Added configuration support for Linux/IA64
10769 [Rolf Haberrecker <rolf@suse.de>]
10771 *) Assembler module support for Mingw32.
10774 *) Shared library support for HPUX (in shlib/).
10775 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10777 *) Shared library support for Solaris gcc.
10778 [Lutz Behnke <behnke@trustcenter.de>]
10780 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10782 *) PKCS7_encrypt() was adding text MIME headers twice because they
10783 were added manually and by SMIME_crlf_copy().
10786 *) In bntest.c don't call BN_rand with zero bits argument.
10787 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10789 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10790 case was implemented. This caused BN_div_recp() to fail occasionally.
10793 *) Add an optional second argument to the set_label() in the perl
10794 assembly language builder. If this argument exists and is set
10795 to 1 it signals that the assembler should use a symbol whose
10796 scope is the entire file, not just the current function. This
10797 is needed with MASM which uses the format label:: for this scope.
10798 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10800 *) Change the ASN1 types so they are typedefs by default. Before
10801 almost all types were #define'd to ASN1_STRING which was causing
10802 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10806 *) Change names of new functions to the new get1/get0 naming
10807 convention: After 'get1', the caller owns a reference count
10808 and has to call ..._free; 'get0' returns a pointer to some
10809 data structure without incrementing reference counters.
10810 (Some of the existing 'get' functions increment a reference
10811 counter, some don't.)
10812 Similarly, 'set1' and 'add1' functions increase reference
10813 counters or duplicate objects.
10816 *) Allow for the possibility of temp RSA key generation failure:
10817 the code used to assume it always worked and crashed on failure.
10820 *) Fix potential buffer overrun problem in BIO_printf().
10821 [Ulf Möller, using public domain code by Patrick Powell; problem
10822 pointed out by David Sacerdote <das33@cornell.edu>]
10824 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10825 RAND_egd() and RAND_status(). In the command line application,
10826 the EGD socket can be specified like a seed file using RANDFILE
10830 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10831 Some CAs (e.g. Verisign) distribute certificates in this form.
10834 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10835 list to exclude them. This means that no special compilation option
10836 is needed to use anonymous DH: it just needs to be included in the
10840 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10841 EVP_MD_type. The old functionality is available in a new macro called
10842 EVP_MD_md(). Change code that uses it and update docs.
10845 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10846 where the 'void *' argument is replaced by a function pointer argument.
10847 Previously 'void *' was abused to point to functions, which works on
10848 many platforms, but is not correct. As these functions are usually
10849 called by macros defined in OpenSSL header files, most source code
10850 should work without changes.
10853 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10854 sections with information on -D... compiler switches used for
10855 compiling the library so that applications can see them. To enable
10856 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10857 must be defined. E.g.,
10858 #define OPENSSL_ALGORITHM_DEFINES
10859 #include <openssl/opensslconf.h>
10860 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10861 [Richard Levitte, Ulf and Bodo Möller]
10863 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10867 *) Change the 'other' type in certificate aux info to a STACK_OF
10868 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10869 the required ASN1 format: arbitrary types determined by an OID.
10872 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10873 argument to 'req'. This is not because the function is newer or
10874 better than others it just uses the work 'NEW' in the certificate
10875 request header lines. Some software needs this.
10878 *) Reorganise password command line arguments: now passwords can be
10879 obtained from various sources. Delete the PEM_cb function and make
10880 it the default behaviour: i.e. if the callback is NULL and the
10881 usrdata argument is not NULL interpret it as a null terminated pass
10882 phrase. If usrdata and the callback are NULL then the pass phrase
10883 is prompted for as usual.
10886 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10887 the support is automatically enabled. The resulting binaries will
10888 autodetect the card and use it if present.
10889 [Ben Laurie and Compaq Inc.]
10891 *) Work around for Netscape hang bug. This sends certificate request
10892 and server done in one record. Since this is perfectly legal in the
10893 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10894 the bugs/SSLv3 entry for more info.
10897 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10900 *) Add -rand argument to smime and pkcs12 applications and read/write
10904 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10907 *) Add command line password options to the remaining applications.
10910 *) Bug fix for BN_div_recp() for numerators with an even number of
10914 *) More tests in bntest.c, and changed test_bn output.
10917 *) ./config recognizes MacOS X now.
10920 *) Bug fix for BN_div() when the first words of num and divisor are
10921 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10924 *) Add support for various broken PKCS#8 formats, and command line
10925 options to produce them.
10928 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10929 get temporary BIGNUMs from a BN_CTX.
10932 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10936 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10937 include a #define from the old name to the new. The original intent
10938 was that statically linked binaries could for example just call
10939 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10940 link with digests. This never worked because SSLeay_add_all_digests()
10941 and SSLeay_add_all_ciphers() were in the same source file so calling
10942 one would link with the other. They are now in separate source files.
10945 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10948 *) Use a less unusual form of the Miller-Rabin primality test (it used
10949 a binary algorithm for exponentiation integrated into the Miller-Rabin
10950 loop, our standard modexp algorithms are faster).
10953 *) Support for the EBCDIC character set completed.
10954 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10956 *) Source code cleanups: use const where appropriate, eliminate casts,
10957 use void * instead of char * in lhash.
10960 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10961 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10962 this the server could overwrite ephemeral keys that the client
10966 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10967 using 50 iterations of the Rabin-Miller test.
10969 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10970 iterations of the Rabin-Miller test as required by the appendix
10971 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10972 As BN_is_prime_fasttest includes trial division, DSA parameter
10973 generation becomes much faster.
10975 This implies a change for the callback functions in DSA_is_prime
10976 and DSA_generate_parameters: The callback function is called once
10977 for each positive witness in the Rabin-Miller test, not just
10978 occasionally in the inner loop; and the parameters to the
10979 callback function now provide an iteration count for the outer
10980 loop rather than for the current invocation of the inner loop.
10981 DSA_generate_parameters additionally can call the callback
10982 function with an 'iteration count' of -1, meaning that a
10983 candidate has passed the trial division test (when q is generated
10984 from an application-provided seed, trial division is skipped).
10987 *) New function BN_is_prime_fasttest that optionally does trial
10988 division before starting the Rabin-Miller test and has
10989 an additional BN_CTX * argument (whereas BN_is_prime always
10990 has to allocate at least one BN_CTX).
10991 'callback(1, -1, cb_arg)' is called when a number has passed the
10992 trial division stage.
10995 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10999 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
11002 *) New function BN_pseudo_rand().
11005 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
11006 bignum version of BN_from_montgomery() with the working code from
11007 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
11011 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
11012 made it impossible to use the same SSL_SESSION data structure in
11013 SSL2 clients in multiple threads.
11016 *) The return value of RAND_load_file() no longer counts bytes obtained
11017 by stat(). RAND_load_file(..., -1) is new and uses the complete file
11018 to seed the PRNG (previously an explicit byte count was required).
11019 [Ulf Möller, Bodo Möller]
11021 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
11022 used (char *) instead of (void *) and had casts all over the place.
11025 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
11028 *) Retain source code compatibility for BN_prime_checks macro:
11029 BN_is_prime(..., BN_prime_checks, ...) now uses
11030 BN_prime_checks_for_size to determine the appropriate number of
11031 Rabin-Miller iterations.
11034 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
11035 DH_CHECK_P_NOT_SAFE_PRIME.
11036 (Check if this is true? OpenPGP calls them "strong".)
11039 *) Merge the functionality of "dh" and "gendh" programs into a new program
11040 "dhparam". The old programs are retained for now but will handle DH keys
11041 (instead of parameters) in future.
11044 *) Make the ciphers, s_server and s_client programs check the return values
11045 when a new cipher list is set.
11048 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11049 ciphers. Before when the 56bit ciphers were enabled the sorting was
11052 The syntax for the cipher sorting has been extended to support sorting by
11053 cipher-strength (using the strength_bits hard coded in the tables).
11054 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11056 Fix a bug in the cipher-command parser: when supplying a cipher command
11057 string with an "undefined" symbol (neither command nor alphanumeric
11058 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11059 an error is flagged.
11061 Due to the strength-sorting extension, the code of the
11062 ssl_create_cipher_list() function was completely rearranged. I hope that
11063 the readability was also increased :-)
11064 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11066 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11067 for the first serial number and places 2 in the serial number file. This
11068 avoids problems when the root CA is created with serial number zero and
11069 the first user certificate has the same issuer name and serial number
11073 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11074 the new code. Add documentation for this stuff.
11077 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11078 X509_*() to X509at_*() on the grounds that they don't handle X509
11079 structures and behave in an analogous way to the X509v3 functions:
11080 they shouldn't be called directly but wrapper functions should be used
11083 So we also now have some wrapper functions that call the X509at functions
11084 when passed certificate requests. (TO DO: similar things can be done with
11085 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11086 things. Some of these need some d2i or i2d and print functionality
11087 because they handle more complex structures.)
11090 *) Add missing #ifndefs that caused missing symbols when building libssl
11091 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11092 NO_RSA in ssl/s2*.c.
11093 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11095 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11096 has a return value which indicates the quality of the random data
11097 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11098 error queue. New function RAND_pseudo_bytes() generates output that is
11099 guaranteed to be unique but not unpredictable. RAND_add is like
11100 RAND_seed, but takes an extra argument for an entropy estimate
11101 (RAND_seed always assumes full entropy).
11104 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11105 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11106 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11107 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11108 false-positive rate of at most 2^-80 for random input.
11111 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11114 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11115 in the 0.9.5 release), this returns the chain
11116 from an X509_CTX structure with a dup of the stack and all
11117 the X509 reference counts upped: so the stack will exist
11118 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11121 Also make SSL_SESSION_print() print out the verify return
11125 *) Add manpage for the pkcs12 command. Also change the default
11126 behaviour so MAC iteration counts are used unless the new
11127 -nomaciter option is used. This improves file security and
11128 only older versions of MSIE (4.0 for example) need it.
11131 *) Honor the no-xxx Configure options when creating .DEF files.
11134 *) Add PKCS#10 attributes to field table: challengePassword,
11135 unstructuredName and unstructuredAddress. These are taken from
11136 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11137 international characters are used.
11139 More changes to X509_ATTRIBUTE code: allow the setting of types
11140 based on strings. Remove the 'loc' parameter when adding
11141 attributes because these will be a SET OF encoding which is sorted
11145 *) Initial changes to the 'req' utility to allow request generation
11146 automation. This will allow an application to just generate a template
11147 file containing all the field values and have req construct the
11150 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11151 used all over the place including certificate requests and PKCS#7
11152 structures. They are currently handled manually where necessary with
11153 some primitive wrappers for PKCS#7. The new functions behave in a
11154 manner analogous to the X509 extension functions: they allow
11155 attributes to be looked up by NID and added.
11157 Later something similar to the X509V3 code would be desirable to
11158 automatically handle the encoding, decoding and printing of the
11159 more complex types. The string types like challengePassword can
11160 be handled by the string table functions.
11162 Also modified the multi byte string table handling. Now there is
11163 a 'global mask' which masks out certain types. The table itself
11164 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11165 is useful when for example there is only one permissible type
11166 (as in countryName) and using the mask might result in no valid
11170 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11171 SSL_get_peer_finished to allow applications to obtain the latest
11172 Finished messages sent to the peer or expected from the peer,
11173 respectively. (SSL_get_peer_finished is usually the Finished message
11174 actually received from the peer, otherwise the protocol will be aborted.)
11176 As the Finished message are message digests of the complete handshake
11177 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11178 be used for external authentication procedures when the authentication
11179 provided by SSL/TLS is not desired or is not enough.
11182 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11183 the host supports BWX extension and if Compaq C is present on the
11184 $PATH. Just exploiting of the BWX extension results in 20-30%
11185 performance kick for some algorithms, e.g. DES and RC4 to mention
11186 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11190 *) Add support for MS "fast SGC". This is arguably a violation of the
11191 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11192 weak crypto and after checking the certificate is SGC a second one
11193 with strong crypto. MS SGC stops the first handshake after receiving
11194 the server certificate message and sends a second client hello. Since
11195 a server will typically do all the time consuming operations before
11196 expecting any further messages from the client (server key exchange
11197 is the most expensive) there is little difference between the two.
11199 To get OpenSSL to support MS SGC we have to permit a second client
11200 hello message after we have sent server done. In addition we have to
11201 reset the MAC if we do get this second client hello.
11204 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11205 if a DER encoded private key is RSA or DSA traditional format. Changed
11206 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11207 format DER encoded private key. Newer code should use PKCS#8 format which
11208 has the key type encoded in the ASN1 structure. Added DER private key
11209 support to pkcs8 application.
11212 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11213 ciphersuites has been selected (as required by the SSL 3/TLS 1
11214 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11215 is set, we interpret this as a request to violate the specification
11216 (the worst that can happen is a handshake failure, and 'correct'
11217 behaviour would result in a handshake failure anyway).
11220 *) In SSL_CTX_add_session, take into account that there might be multiple
11221 SSL_SESSION structures with the same session ID (e.g. when two threads
11222 concurrently obtain them from an external cache).
11223 The internal cache can handle only one SSL_SESSION with a given ID,
11224 so if there's a conflict, we now throw out the old one to achieve
11228 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11229 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11230 some routines that use cipher OIDs: some ciphers do not have OIDs
11231 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11235 *) Simplify the trust setting structure and code. Now we just have
11236 two sequences of OIDs for trusted and rejected settings. These will
11237 typically have values the same as the extended key usage extension
11238 and any application specific purposes.
11240 The trust checking code now has a default behaviour: it will just
11241 check for an object with the same NID as the passed id. Functions can
11242 be provided to override either the default behaviour or the behaviour
11243 for a given id. SSL client, server and email already have functions
11244 in place for compatibility: they check the NID and also return "trusted"
11245 if the certificate is self signed.
11248 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11249 traditional format into an EVP_PKEY structure.
11252 *) Add a password callback function PEM_cb() which either prompts for
11253 a password if usr_data is NULL or otherwise assumes it is a null
11254 terminated password. Allow passwords to be passed on command line
11255 environment or config files in a few more utilities.
11258 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11259 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11260 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11261 Update documentation.
11264 *) Support for ASN1 "NULL" type. This could be handled before by using
11265 ASN1_TYPE but there wasn't any function that would try to read a NULL
11266 and produce an error if it couldn't. For compatibility we also have
11267 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11268 don't allocate anything because they don't need to.
11271 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11273 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11275 *) Rebuild of the memory allocation routines used by OpenSSL code and
11276 possibly others as well. The purpose is to make an interface that
11277 provide hooks so anyone can build a separate set of allocation and
11278 deallocation routines to be used by OpenSSL, for example memory
11279 pool implementations, or something else, which was previously hard
11280 since Malloc(), Realloc() and Free() were defined as macros having
11281 the values malloc, realloc and free, respectively (except for Win32
11282 compilations). The same is provided for memory debugging code.
11283 OpenSSL already comes with functionality to find memory leaks, but
11284 this gives people a chance to debug other memory problems.
11286 With these changes, a new set of functions and macros have appeared:
11288 CRYPTO_set_mem_debug_functions() [F]
11289 CRYPTO_get_mem_debug_functions() [F]
11290 CRYPTO_dbg_set_options() [F]
11291 CRYPTO_dbg_get_options() [F]
11292 CRYPTO_malloc_debug_init() [M]
11294 The memory debug functions are NULL by default, unless the library
11295 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11296 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11297 gives the standard debugging functions that come with OpenSSL) or
11298 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11299 provided by the library user) must be used. When the standard
11300 debugging functions are used, CRYPTO_dbg_set_options can be used to
11301 request additional information:
11302 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11303 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11305 Also, things like CRYPTO_set_mem_functions will always give the
11306 expected result (the new set of functions is used for allocation
11307 and deallocation) at all times, regardless of platform and compiler
11310 To finish it up, some functions that were never use in any other
11311 way than through macros have a new API and new semantic:
11313 CRYPTO_dbg_malloc()
11314 CRYPTO_dbg_realloc()
11317 All macros of value have retained their old syntax.
11318 [Richard Levitte and Bodo Moeller]
11320 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11321 ordering of SMIMECapabilities wasn't in "strength order" and there
11322 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11326 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11327 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11328 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11330 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11331 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11332 functionality to handle multipart/signed properly) and a utility
11333 called 'smime' to call all this stuff. This is based on code I
11334 originally wrote for Celo who have kindly allowed it to be
11335 included in OpenSSL.
11338 *) Add variants des_set_key_checked and des_set_key_unchecked of
11339 des_set_key (aka des_key_sched). Global variable des_check_key
11340 decides which of these is called by des_set_key; this way
11341 des_check_key behaves as it always did, but applications and
11342 the library itself, which was buggy for des_check_key == 1,
11343 have a cleaner way to pick the version they need.
11346 *) New function PKCS12_newpass() which changes the password of a
11350 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11351 dynamic mix. In both cases the ids can be used as an index into the
11352 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11353 functions so they accept a list of the field values and the
11354 application doesn't need to directly manipulate the X509_TRUST
11358 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11362 *) Modify the way the V3 extension code looks up extensions. This now
11363 works in a similar way to the object code: we have some "standard"
11364 extensions in a static table which is searched with OBJ_bsearch()
11365 and the application can add dynamic ones if needed. The file
11366 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11367 updated whenever a new extension is added to the core code and kept
11368 in ext_nid order. There is a simple program 'tabtest.c' which checks
11369 this. New extensions are not added too often so this file can readily
11370 be maintained manually.
11372 There are two big advantages in doing things this way. The extensions
11373 can be looked up immediately and no longer need to be "added" using
11374 X509V3_add_standard_extensions(): this function now does nothing.
11375 [Side note: I get *lots* of email saying the extension code doesn't
11376 work because people forget to call this function]
11377 Also no dynamic allocation is done unless new extensions are added:
11378 so if we don't add custom extensions there is no need to call
11379 X509V3_EXT_cleanup().
11382 *) Modify enc utility's salting as follows: make salting the default. Add a
11383 magic header, so unsalted files fail gracefully instead of just decrypting
11384 to garbage. This is because not salting is a big security hole, so people
11385 should be discouraged from doing it.
11388 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11389 digest to be passed on the command line but it only used this
11390 parameter when signing a certificate. Modified so all relevant
11391 operations are affected by the digest parameter including the
11392 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11393 DSA key was used because it didn't fix the digest.
11396 *) Initial certificate chain verify code. Currently tests the untrusted
11397 certificates for consistency with the verify purpose (which is set
11398 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11400 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11401 this is because it will reject chains with invalid extensions whereas
11402 every previous version of OpenSSL and SSLeay made no checks at all.
11404 Trust code: checks the root CA for the relevant trust settings. Trust
11405 settings have an initial value consistent with the verify purpose: e.g.
11406 if the verify purpose is for SSL client use it expects the CA to be
11407 trusted for SSL client use. However the default value can be changed to
11408 permit custom trust settings: one example of this would be to only trust
11409 certificates from a specific "secure" set of CAs.
11411 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11412 which should be used for version portability: especially since the
11413 verify structure is likely to change more often now.
11415 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11416 to set them. If not set then assume SSL clients will verify SSL servers
11419 Two new options to the verify program: -untrusted allows a set of
11420 untrusted certificates to be passed in and -purpose which sets the
11421 intended purpose of the certificate. If a purpose is set then the
11422 new chain verify code is used to check extension consistency.
11425 *) Support for the authority information access extension.
11428 *) Modify RSA and DSA PEM read routines to transparently handle
11429 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11430 public keys in a format compatible with certificate
11431 SubjectPublicKeyInfo structures. Unfortunately there were already
11432 functions called *_PublicKey_* which used various odd formats so
11433 these are retained for compatibility: however the DSA variants were
11434 never in a public release so they have been deleted. Changed dsa/rsa
11435 utilities to handle the new format: note no releases ever handled public
11436 keys so we should be OK.
11438 The primary motivation for this change is to avoid the same fiasco
11439 that dogs private keys: there are several incompatible private key
11440 formats some of which are standard and some OpenSSL specific and
11441 require various evil hacks to allow partial transparent handling and
11442 even then it doesn't work with DER formats. Given the option anything
11443 other than PKCS#8 should be dumped: but the other formats have to
11444 stay in the name of compatibility.
11446 With public keys and the benefit of hindsight one standard format
11447 is used which works with EVP_PKEY, RSA or DSA structures: though
11448 it clearly returns an error if you try to read the wrong kind of key.
11450 Added a -pubkey option to the 'x509' utility to output the public key.
11451 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11452 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11453 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11454 that do the same as the EVP_PKEY_assign_*() except they up the
11455 reference count of the added key (they don't "swallow" the
11459 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11460 CRLs would fail if the file contained no certificates or no CRLs:
11461 added a new function to read in both types and return the number
11462 read: this means that if none are read it will be an error. The
11463 DER versions of the certificate and CRL reader would always fail
11464 because it isn't possible to mix certificates and CRLs in DER format
11465 without choking one or the other routine. Changed this to just read
11466 a certificate: this is the best we can do. Also modified the code
11467 in apps/verify.c to take notice of return codes: it was previously
11468 attempting to read in certificates from NULL pointers and ignoring
11469 any errors: this is one reason why the cert and CRL reader seemed
11470 to work. It doesn't check return codes from the default certificate
11471 routines: these may well fail if the certificates aren't installed.
11474 *) Code to support otherName option in GeneralName.
11477 *) First update to verify code. Change the verify utility
11478 so it warns if it is passed a self signed certificate:
11479 for consistency with the normal behaviour. X509_verify
11480 has been modified to it will now verify a self signed
11481 certificate if *exactly* the same certificate appears
11482 in the store: it was previously impossible to trust a
11483 single self signed certificate. This means that:
11484 openssl verify ss.pem
11485 now gives a warning about a self signed certificate but
11486 openssl verify -CAfile ss.pem ss.pem
11490 *) For servers, store verify_result in SSL_SESSION data structure
11491 (and add it to external session representation).
11492 This is needed when client certificate verifications fails,
11493 but an application-provided verification callback (set by
11494 SSL_CTX_set_cert_verify_callback) allows accepting the session
11495 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11496 but returns 1): When the session is reused, we have to set
11497 ssl->verify_result to the appropriate error code to avoid
11499 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11501 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11502 case in PKCS7_dataInit() where the signed PKCS7 structure
11503 didn't contain any existing data because it was being created.
11504 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11506 *) Add a salt to the key derivation routines in enc.c. This
11507 forms the first 8 bytes of the encrypted file. Also add a
11508 -S option to allow a salt to be input on the command line.
11511 *) New function X509_cmp(). Oddly enough there wasn't a function
11512 to compare two certificates. We do this by working out the SHA1
11513 hash and comparing that. X509_cmp() will be needed by the trust
11517 *) SSL_get1_session() is like SSL_get_session(), but increments
11518 the reference count in the SSL_SESSION returned.
11519 [Geoff Thorpe <geoff@eu.c2.net>]
11521 *) Fix for 'req': it was adding a null to request attributes.
11522 Also change the X509_LOOKUP and X509_INFO code to handle
11523 certificate auxiliary information.
11526 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11530 *) Add the possibility to add extra information to the memory leak
11531 detecting output, to form tracebacks, showing from where each
11532 allocation was originated: CRYPTO_push_info("constant string") adds
11533 the string plus current file name and line number to a per-thread
11534 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11535 is like calling CYRPTO_pop_info() until the stack is empty.
11536 Also updated memory leak detection code to be multi-thread-safe.
11539 *) Add options -text and -noout to pkcs7 utility and delete the
11540 encryption options which never did anything. Update docs.
11543 *) Add options to some of the utilities to allow the pass phrase
11544 to be included on either the command line (not recommended on
11545 OSes like Unix) or read from the environment. Update the
11546 manpages and fix a few bugs.
11549 *) Add a few manpages for some of the openssl commands.
11552 *) Fix the -revoke option in ca. It was freeing up memory twice,
11553 leaking and not finding already revoked certificates.
11556 *) Extensive changes to support certificate auxiliary information.
11557 This involves the use of X509_CERT_AUX structure and X509_AUX
11558 functions. An X509_AUX function such as PEM_read_X509_AUX()
11559 can still read in a certificate file in the usual way but it
11560 will also read in any additional "auxiliary information". By
11561 doing things this way a fair degree of compatibility can be
11562 retained: existing certificates can have this information added
11563 using the new 'x509' options.
11565 Current auxiliary information includes an "alias" and some trust
11566 settings. The trust settings will ultimately be used in enhanced
11567 certificate chain verification routines: currently a certificate
11568 can only be trusted if it is self signed and then it is trusted
11572 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11573 The problem was that one of the replacement routines had not been working
11574 since SSLeay releases. For now the offending routine has been replaced
11575 with non-optimised assembler. Even so, this now gives around 95%
11576 performance improvement for 1024 bit RSA signs.
11579 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11580 handling. Most clients have the effective key size in bits equal to
11581 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11582 A few however don't do this and instead use the size of the decrypted key
11583 to determine the RC2 key length and the AlgorithmIdentifier to determine
11584 the effective key length. In this case the effective key length can still
11585 be 40 bits but the key length can be 168 bits for example. This is fixed
11586 by manually forcing an RC2 key into the EVP_PKEY structure because the
11587 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11588 the key length and effective key length are equal.
11591 *) Add a bunch of functions that should simplify the creation of
11592 X509_NAME structures. Now you should be able to do:
11593 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11594 and have it automatically work out the correct field type and fill in
11595 the structures. The more adventurous can try:
11596 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11597 and it will (hopefully) work out the correct multibyte encoding.
11600 *) Change the 'req' utility to use the new field handling and multibyte
11601 copy routines. Before the DN field creation was handled in an ad hoc
11602 way in req, ca, and x509 which was rather broken and didn't support
11603 BMPStrings or UTF8Strings. Since some software doesn't implement
11604 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11605 using the dirstring_type option. See the new comment in the default
11606 openssl.cnf for more info.
11609 *) Make crypto/rand/md_rand.c more robust:
11610 - Assure unique random numbers after fork().
11611 - Make sure that concurrent threads access the global counter and
11612 md serializably so that we never lose entropy in them
11613 or use exactly the same state in multiple threads.
11614 Access to the large state is not always serializable because
11615 the additional locking could be a performance killer, and
11616 md should be large enough anyway.
11619 *) New file apps/app_rand.c with commonly needed functionality
11620 for handling the random seed file.
11622 Use the random seed file in some applications that previously did not:
11624 dsaparam -genkey (which also ignored its '-rand' option),
11627 x509 (when signing).
11628 Except on systems with /dev/urandom, it is crucial to have a random
11629 seed file at least for key creation, DSA signing, and for DH exchanges;
11630 for RSA signatures we could do without one.
11632 gendh and gendsa (unlike genrsa) used to read only the first byte
11633 of each file listed in the '-rand' option. The function as previously
11634 found in genrsa is now in app_rand.c and is used by all programs
11635 that support '-rand'.
11638 *) In RAND_write_file, use mode 0600 for creating files;
11639 don't just chmod when it may be too late.
11642 *) Report an error from X509_STORE_load_locations
11643 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11646 *) New function ASN1_mbstring_copy() this copies a string in either
11647 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11648 into an ASN1_STRING type. A mask of permissible types is passed
11649 and it chooses the "minimal" type to use or an error if not type
11653 *) Add function equivalents to the various macros in asn1.h. The old
11654 macros are retained with an M_ prefix. Code inside the library can
11655 use the M_ macros. External code (including the openssl utility)
11656 should *NOT* in order to be "shared library friendly".
11659 *) Add various functions that can check a certificate's extensions
11660 to see if it usable for various purposes such as SSL client,
11661 server or S/MIME and CAs of these types. This is currently
11662 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11663 verification. Also added a -purpose flag to x509 utility to
11664 print out all the purposes.
11667 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11671 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11672 for, obtain and decode and extension and obtain its critical flag.
11673 This allows all the necessary extension code to be handled in a
11674 single function call.
11677 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11678 platforms. See crypto/rc4/rc4_enc.c for further details.
11681 *) New -noout option to asn1parse. This causes no output to be produced
11682 its main use is when combined with -strparse and -out to extract data
11683 from a file (which may not be in ASN.1 format).
11686 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11687 when producing the local key id.
11688 [Richard Levitte <levitte@stacken.kth.se>]
11690 *) New option -dhparam in s_server. This allows a DH parameter file to be
11691 stated explicitly. If it is not stated then it tries the first server
11692 certificate file. The previous behaviour hard coded the filename
11696 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11697 a public key to be input or output. For example:
11698 openssl rsa -in key.pem -pubout -out pubkey.pem
11699 Also added necessary DSA public key functions to handle this.
11702 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11703 in the message. This was handled by allowing
11704 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11705 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11707 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11708 to the end of the strings whereas this didn't. This would cause problems
11709 if strings read with d2i_ASN1_bytes() were later modified.
11710 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11712 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11713 data and it contains EOF it will end up returning an error. This is
11714 caused by input 46 bytes long. The cause is due to the way base64
11715 BIOs find the start of base64 encoded data. They do this by trying a
11716 trial decode on each line until they find one that works. When they
11717 do a flag is set and it starts again knowing it can pass all the
11718 data directly through the decoder. Unfortunately it doesn't reset
11719 the context it uses. This means that if EOF is reached an attempt
11720 is made to pass two EOFs through the context and this causes the
11721 resulting error. This can also cause other problems as well. As is
11722 usual with these problems it takes *ages* to find and the fix is
11723 trivial: move one line.
11724 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11726 *) Ugly workaround to get s_client and s_server working under Windows. The
11727 old code wouldn't work because it needed to select() on sockets and the
11728 tty (for keypresses and to see if data could be written). Win32 only
11729 supports select() on sockets so we select() with a 1s timeout on the
11730 sockets and then see if any characters are waiting to be read, if none
11731 are present then we retry, we also assume we can always write data to
11732 the tty. This isn't nice because the code then blocks until we've
11733 received a complete line of data and it is effectively polling the
11734 keyboard at 1s intervals: however it's quite a bit better than not
11735 working at all :-) A dedicated Windows application might handle this
11736 with an event loop for example.
11739 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11740 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11741 will be called when RSA_sign() and RSA_verify() are used. This is useful
11742 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11743 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11744 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11745 This necessitated the support of an extra signature type NID_md5_sha1
11746 for SSL signatures and modifications to the SSL library to use it instead
11747 of calling RSA_public_decrypt() and RSA_private_encrypt().
11750 *) Add new -verify -CAfile and -CApath options to the crl program, these
11751 will lookup a CRL issuers certificate and verify the signature in a
11752 similar way to the verify program. Tidy up the crl program so it
11753 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11754 less strict. It will now permit CRL extensions even if it is not
11755 a V2 CRL: this will allow it to tolerate some broken CRLs.
11758 *) Initialize all non-automatic variables each time one of the openssl
11759 sub-programs is started (this is necessary as they may be started
11760 multiple times from the "OpenSSL>" prompt).
11761 [Lennart Bang, Bodo Moeller]
11763 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11764 removing all other RSA functionality (this is what NO_RSA does). This
11765 is so (for example) those in the US can disable those operations covered
11766 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11770 *) Non-copying interface to BIO pairs.
11771 (still largely untested)
11774 *) New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
11775 ASCII string. This was handled independently in various places before.
11778 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11779 UTF8 strings a character at a time.
11782 *) Use client_version from client hello to select the protocol
11783 (s23_srvr.c) and for RSA client key exchange verification
11784 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11787 *) Add various utility functions to handle SPKACs, these were previously
11788 handled by poking round in the structure internals. Added new function
11789 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11790 print, verify and generate SPKACs. Based on an original idea from
11791 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11794 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11797 *) Allow the config file extension section to be overwritten on the
11798 command line. Based on an original idea from Massimiliano Pala
11799 <madwolf@comune.modena.it>. The new option is called -extensions
11800 and can be applied to ca, req and x509. Also -reqexts to override
11801 the request extensions in req and -crlexts to override the crl extensions
11805 *) Add new feature to the SPKAC handling in ca. Now you can include
11806 the same field multiple times by preceding it by "XXXX." for example:
11809 this is the same syntax as used in the req config file.
11812 *) Allow certificate extensions to be added to certificate requests. These
11813 are specified in a 'req_extensions' option of the req section of the
11814 config file. They can be printed out with the -text option to req but
11815 are otherwise ignored at present.
11818 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11819 data read consists of only the final block it would not decrypted because
11820 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11821 A misplaced 'break' also meant the decrypted final block might not be
11822 copied until the next read.
11825 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11826 a few extra parameters to the DH structure: these will be useful if
11827 for example we want the value of 'q' or implement X9.42 DH.
11830 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11831 provides hooks that allow the default DSA functions or functions on a
11832 "per key" basis to be replaced. This allows hardware acceleration and
11833 hardware key storage to be handled without major modification to the
11834 library. Also added low level modexp hooks and CRYPTO_EX structure and
11835 associated functions.
11838 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11839 as "read only": it can't be written to and the buffer it points to will
11840 not be freed. Reading from a read only BIO is much more efficient than
11841 a normal memory BIO. This was added because there are several times when
11842 an area of memory needs to be read from a BIO. The previous method was
11843 to create a memory BIO and write the data to it, this results in two
11844 copies of the data and an O(n^2) reading algorithm. There is a new
11845 function BIO_new_mem_buf() which creates a read only memory BIO from
11846 an area of memory. Also modified the PKCS#7 routines to use read only
11850 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11851 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11852 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11853 but a retry condition occurred while trying to read the rest.
11856 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11857 NID_pkcs7_encrypted by default: this was wrong since this should almost
11858 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11859 the encrypted data type: this is a more sensible place to put it and it
11860 allows the PKCS#12 code to be tidied up that duplicated this
11864 *) Changed obj_dat.pl script so it takes its input and output files on
11865 the command line. This should avoid shell escape redirection problems
11869 *) Initial support for certificate extension requests, these are included
11870 in things like Xenroll certificate requests. Included functions to allow
11871 extensions to be obtained and added.
11874 *) -crlf option to s_client and s_server for sending newlines as
11875 CRLF (as required by many protocols).
11878 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11880 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11881 [Ralf S. Engelschall]
11883 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11884 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11886 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11890 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11891 DH parameters/keys (q is lost during that conversion, but the resulting
11892 DH parameters contain its length).
11894 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11895 much faster than DH_generate_parameters (which creates parameters
11896 where p = 2*q + 1), and also the smaller q makes DH computations
11897 much more efficient (160-bit exponentiation instead of 1024-bit
11898 exponentiation); so this provides a convenient way to support DHE
11899 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11900 utter importance to use
11901 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11903 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11904 when such DH parameters are used, because otherwise small subgroup
11905 attacks may become possible!
11908 *) Avoid memory leak in i2d_DHparams.
11911 *) Allow the -k option to be used more than once in the enc program:
11912 this allows the same encrypted message to be read by multiple recipients.
11915 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11916 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11917 it will always use the numerical form of the OID, even if it has a short
11921 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11922 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11923 otherwise bn_mod_exp was called. In the case of hardware keys for example
11924 no private key components need be present and it might store extra data
11925 in the RSA structure, which cannot be accessed from bn_mod_exp.
11926 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11927 private key operations.
11930 *) Added support for SPARC Linux.
11933 *) pem_password_cb function type incompatibly changed from
11934 typedef int pem_password_cb(char *buf, int size, int rwflag);
11936 ....(char *buf, int size, int rwflag, void *userdata);
11937 so that applications can pass data to their callbacks:
11938 The PEM[_ASN1]_{read,write}... functions and macros now take an
11939 additional void * argument, which is just handed through whenever
11940 the password callback is called.
11941 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11943 New function SSL_CTX_set_default_passwd_cb_userdata.
11945 Compatibility note: As many C implementations push function arguments
11946 onto the stack in reverse order, the new library version is likely to
11947 interoperate with programs that have been compiled with the old
11948 pem_password_cb definition (PEM_whatever takes some data that
11949 happens to be on the stack as its last argument, and the callback
11950 just ignores this garbage); but there is no guarantee whatsoever that
11953 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11954 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11955 problems not only on Windows, but also on some Unix platforms.
11956 To avoid problematic command lines, these definitions are now in an
11957 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11958 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11961 *) MIPS III/IV assembler module is reimplemented.
11964 *) More DES library cleanups: remove references to srand/rand and
11965 delete an unused file.
11968 *) Add support for the free Netwide assembler (NASM) under Win32,
11969 since not many people have MASM (ml) and it can be hard to obtain.
11970 This is currently experimental but it seems to work OK and pass all
11971 the tests. Check out INSTALL.W32 for info.
11974 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11975 without temporary keys kept an extra copy of the server key,
11976 and connections with temporary keys did not free everything in case
11980 *) New function RSA_check_key and new openssl rsa option -check
11981 for verifying the consistency of RSA keys.
11982 [Ulf Moeller, Bodo Moeller]
11984 *) Various changes to make Win32 compile work:
11985 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11986 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11987 comparison" warnings.
11988 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11991 *) Add a debugging option to PKCS#5 v2 key generation function: when
11992 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11993 derived keys are printed to stderr.
11996 *) Copy the flags in ASN1_STRING_dup().
11997 [Roman E. Pavlov <pre@mo.msk.ru>]
11999 *) The x509 application mishandled signing requests containing DSA
12000 keys when the signing key was also DSA and the parameters didn't match.
12002 It was supposed to omit the parameters when they matched the signing key:
12003 the verifying software was then supposed to automatically use the CA's
12004 parameters if they were absent from the end user certificate.
12006 Omitting parameters is no longer recommended. The test was also
12007 the wrong way round! This was probably due to unusual behaviour in
12008 EVP_cmp_parameters() which returns 1 if the parameters match.
12009 This meant that parameters were omitted when they *didn't* match and
12010 the certificate was useless. Certificates signed with 'ca' didn't have
12012 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
12014 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
12015 The interface is as follows:
12016 Applications can use
12017 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
12018 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
12019 "off" is now the default.
12020 The library internally uses
12021 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
12022 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
12023 to disable memory-checking temporarily.
12025 Some inconsistent states that previously were possible (and were
12026 even the default) are now avoided.
12028 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
12029 with each memory chunk allocated; this is occasionally more helpful
12030 than just having a counter.
12032 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
12034 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
12038 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
12039 which largely parallels "options", but is for changing API behaviour,
12040 whereas "options" are about protocol behaviour.
12041 Initial "mode" flags are:
12043 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12044 a single record has been written.
12045 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12046 retries use the same buffer location.
12047 (But all of the contents must be
12051 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12054 *) Fix problems with no-hmac etc.
12055 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12057 *) New functions RSA_get_default_method(), RSA_set_method() and
12058 RSA_get_method(). These allows replacement of RSA_METHODs without having
12059 to mess around with the internals of an RSA structure.
12062 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12063 Also really enable memory leak checks in openssl.c and in some
12065 [Chad C. Mulligan, Bodo Moeller]
12067 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12068 up the length of negative integers. This has now been simplified to just
12069 store the length when it is first determined and use it later, rather
12070 than trying to keep track of where data is copied and updating it to
12072 [Steve Henson, reported by Brien Wheeler
12073 <bwheeler@authentica-security.com>]
12075 *) Add a new function PKCS7_signatureVerify. This allows the verification
12076 of a PKCS#7 signature but with the signing certificate passed to the
12077 function itself. This contrasts with PKCS7_dataVerify which assumes the
12078 certificate is present in the PKCS#7 structure. This isn't always the
12079 case: certificates can be omitted from a PKCS#7 structure and be
12080 distributed by "out of band" means (such as a certificate database).
12083 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12084 function prototypes in pem.h, also change util/mkdef.pl to add the
12085 necessary function names.
12088 *) mk1mf.pl (used by Windows builds) did not properly read the
12089 options set by Configure in the top level Makefile, and Configure
12090 was not even able to write more than one option correctly.
12091 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12094 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12095 file to be loaded from a BIO or FILE pointer. The BIO version will
12096 for example allow memory BIOs to contain config info.
12099 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12100 Whoever hopes to achieve shared-library compatibility across versions
12101 must use this, not the compile-time macro.
12102 (Exercise 0.9.4: Which is the minimum library version required by
12104 Note: All this applies only to multi-threaded programs, others don't
12108 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12109 through a BIO pair triggered the default case, i.e.
12110 SSLerr(...,SSL_R_UNKNOWN_STATE).
12113 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12114 can use the SSL library even if none of the specific BIOs is
12118 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12119 for the encoded length.
12120 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12122 *) Add initial documentation of the X509V3 functions.
12125 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12126 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12127 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12128 secure PKCS#8 private key format with a high iteration count.
12131 *) Fix determination of Perl interpreter: A perl or perl5
12132 _directory_ in $PATH was also accepted as the interpreter.
12133 [Ralf S. Engelschall]
12135 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12136 wrong with it but it was very old and did things like calling
12137 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12138 unusual formatting.
12141 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12142 to use the new extension code.
12145 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12146 with macros. This should make it easier to change their form, add extra
12147 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12151 *) Add to configuration table a new entry that can specify an alternative
12152 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12153 according to Mark Crispin <MRC@Panda.COM>.
12157 *) DES CBC did not update the IV. Weird.
12160 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12161 Changing the behaviour of the former might break existing programs --
12162 where IV updating is needed, des_ncbc_encrypt can be used.
12165 *) When bntest is run from "make test" it drives bc to check its
12166 calculations, as well as internally checking them. If an internal check
12167 fails, it needs to cause bc to give a non-zero result or make test carries
12168 on without noticing the failure. Fixed.
12171 *) DES library cleanups.
12174 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12175 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12176 ciphers. NOTE: although the key derivation function has been verified
12177 against some published test vectors it has not been extensively tested
12178 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12182 *) Instead of "mkdir -p", which is not fully portable, use new
12183 Perl script "util/mkdir-p.pl".
12186 *) Rewrite the way password based encryption (PBE) is handled. It used to
12187 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12188 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12189 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12190 the 'parameter' field of the AlgorithmIdentifier is passed to the
12191 underlying key generation function so it must do its own ASN1 parsing.
12192 This has also changed the EVP_PBE_CipherInit() function which now has a
12193 'parameter' argument instead of literal salt and iteration count values
12194 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12197 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12198 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12199 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12200 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12201 value was just used as a "magic string" and not used directly its
12202 value doesn't matter.
12205 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12209 *) "linux-sparc64" configuration (ultrapenguin).
12210 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12211 "linux-sparc" configuration.
12212 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12214 *) config now generates no-xxx options for missing ciphers.
12217 *) Support the EBCDIC character set (work in progress).
12218 File ebcdic.c not yet included because it has a different license.
12219 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12221 *) Support BS2000/OSD-POSIX.
12222 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12224 *) Make callbacks for key generation use void * instead of char *.
12227 *) Make S/MIME samples compile (not yet tested).
12230 *) Additional typesafe stacks.
12233 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12237 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12239 *) New configuration variant "sco5-gcc".
12241 *) Updated some demos.
12242 [Sean O Riordain, Wade Scholine]
12244 *) Add missing BIO_free at exit of pkcs12 application.
12247 *) Fix memory leak in conf.c.
12250 *) Updates for Win32 to assembler version of MD5.
12253 *) Set #! path to perl in apps/der_chop to where we found it
12254 instead of using a fixed path.
12257 *) SHA library changes for irix64-mips4-cc.
12260 *) Improvements for VMS support.
12264 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12266 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12267 This also avoids the problems with SC4.2 and unpatched SC5.
12268 [Andy Polyakov <appro@fy.chalmers.se>]
12270 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12271 These are required because of the typesafe stack would otherwise break
12272 existing code. If old code used a structure member which used to be STACK
12273 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12274 sk_num or sk_value it would produce an error because the num, data members
12275 are not present in STACK_OF. Now it just produces a warning. sk_set
12276 replaces the old method of assigning a value to sk_value
12277 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12278 that does this will no longer work (and should use sk_set instead) but
12279 this could be regarded as a "questionable" behaviour anyway.
12282 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12283 correctly handle encrypted S/MIME data.
12286 *) Change type of various DES function arguments from des_cblock
12287 (which means, in function argument declarations, pointer to char)
12288 to des_cblock * (meaning pointer to array with 8 char elements),
12289 which allows the compiler to do more typechecking; it was like
12290 that back in SSLeay, but with lots of ugly casts.
12292 Introduce new type const_des_cblock.
12295 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12296 problems: find RecipientInfo structure that matches recipient certificate
12297 and initialise the ASN1 structures properly based on passed cipher.
12300 *) Belatedly make the BN tests actually check the results.
12303 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12304 to and from BNs: it was completely broken. New compilation option
12305 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12306 key elements as negative integers.
12309 *) Reorganize and speed up MD5.
12310 [Andy Polyakov <appro@fy.chalmers.se>]
12313 [Richard Levitte <richard@levitte.org>]
12315 *) New option -out to asn1parse to allow the parsed structure to be
12316 output to a file. This is most useful when combined with the -strparse
12317 option to examine the output of things like OCTET STRINGS.
12320 *) Make SSL library a little more fool-proof by not requiring any longer
12321 that SSL_set_{accept,connect}_state be called before
12322 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12323 in many applications because usually everything *appeared* to work as
12324 intended anyway -- now it really works as intended).
12327 *) Move openssl.cnf out of lib/.
12330 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12331 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12332 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12333 [Ralf S. Engelschall]
12335 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12336 handle PKCS#7 enveloped data properly.
12337 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12339 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12340 copying pointers. The cert_st handling is changed by this in
12341 various ways (and thus what used to be known as ctx->default_cert
12342 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12343 any longer when s->cert does not give us what we need).
12344 ssl_cert_instantiate becomes obsolete by this change.
12345 As soon as we've got the new code right (possibly it already is?),
12346 we have solved a couple of bugs of the earlier code where s->cert
12347 was used as if it could not have been shared with other SSL structures.
12349 Note that using the SSL API in certain dirty ways now will result
12350 in different behaviour than observed with earlier library versions:
12351 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12352 does not influence s as it used to.
12354 In order to clean up things more thoroughly, inside SSL_SESSION
12355 we don't use CERT any longer, but a new structure SESS_CERT
12356 that holds per-session data (if available); currently, this is
12357 the peer's certificate chain and, for clients, the server's certificate
12358 and temporary key. CERT holds only those values that can have
12359 meaningful defaults in an SSL_CTX.
12362 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12363 from the internal representation. Various PKCS#7 fixes: remove some
12364 evil casts and set the enc_dig_alg field properly based on the signing
12368 *) Allow PKCS#12 password to be set from the command line or the
12369 environment. Let 'ca' get its config file name from the environment
12370 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12374 *) Allow certificate policies extension to use an IA5STRING for the
12375 organization field. This is contrary to the PKIX definition but
12376 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12380 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12381 without disallowing inline assembler and the like for non-pedantic builds.
12384 *) Support Borland C++ builder.
12385 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12387 *) Support Mingw32.
12390 *) SHA-1 cleanups and performance enhancements.
12391 [Andy Polyakov <appro@fy.chalmers.se>]
12393 *) Sparc v8plus assembler for the bignum library.
12394 [Andy Polyakov <appro@fy.chalmers.se>]
12396 *) Accept any -xxx and +xxx compiler options in Configure.
12399 *) Update HPUX configuration.
12402 *) Add missing sk_<type>_unshift() function to safestack.h
12403 [Ralf S. Engelschall]
12405 *) New function SSL_CTX_use_certificate_chain_file that sets the
12406 "extra_cert"s in addition to the certificate. (This makes sense
12407 only for "PEM" format files, as chains as a whole are not
12411 *) Support verify_depth from the SSL API.
12412 x509_vfy.c had what can be considered an off-by-one-error:
12413 Its depth (which was not part of the external interface)
12414 was actually counting the number of certificates in a chain;
12415 now it really counts the depth.
12418 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12419 instead of X509err, which often resulted in confusing error
12420 messages since the error codes are not globally unique
12421 (e.g. an alleged error in ssl3_accept when a certificate
12422 didn't match the private key).
12424 *) New function SSL_CTX_set_session_id_context that allows to set a default
12425 value (so that you don't need SSL_set_session_id_context for each
12426 connection using the SSL_CTX).
12429 *) OAEP decoding bug fix.
12432 *) Support INSTALL_PREFIX for package builders, as proposed by
12436 *) New Configure options "threads" and "no-threads". For systems
12437 where the proper compiler options are known (currently Solaris
12438 and Linux), "threads" is the default.
12441 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12444 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12445 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12446 such as /usr/local/bin.
12449 *) "make linux-shared" to build shared libraries.
12450 [Niels Poppe <niels@netbox.org>]
12452 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12455 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12456 extension adding in x509 utility.
12459 *) Remove NOPROTO sections and error code comments.
12462 *) Partial rewrite of the DEF file generator to now parse the ANSI
12466 *) New Configure options --prefix=DIR and --openssldir=DIR.
12469 *) Complete rewrite of the error code script(s). It is all now handled
12470 by one script at the top level which handles error code gathering,
12471 header rewriting and C source file generation. It should be much better
12472 than the old method: it now uses a modified version of Ulf's parser to
12473 read the ANSI prototypes in all header files (thus the old K&R definitions
12474 aren't needed for error creation any more) and do a better job of
12475 translating function codes into names. The old 'ASN1 error code imbedded
12476 in a comment' is no longer necessary and it doesn't use .err files which
12477 have now been deleted. Also the error code call doesn't have to appear all
12478 on one line (which resulted in some large lines...).
12481 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12484 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12485 0 (which usually indicates a closed connection), but continue reading.
12488 *) Fix some race conditions.
12491 *) Add support for CRL distribution points extension. Add Certificate
12492 Policies and CRL distribution points documentation.
12495 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12498 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12499 8 of keying material. Merlin has also confirmed interop with this fix
12500 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12501 [Merlin Hughes <merlin@baltimore.ie>]
12503 *) Fix lots of warnings.
12504 [Richard Levitte <levitte@stacken.kth.se>]
12506 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12507 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12508 [Richard Levitte <levitte@stacken.kth.se>]
12510 *) Fix problems with sizeof(long) == 8.
12511 [Andy Polyakov <appro@fy.chalmers.se>]
12513 *) Change functions to ANSI C.
12516 *) Fix typos in error codes.
12517 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12519 *) Remove defunct assembler files from Configure.
12522 *) SPARC v8 assembler BIGNUM implementation.
12523 [Andy Polyakov <appro@fy.chalmers.se>]
12525 *) Support for Certificate Policies extension: both print and set.
12526 Various additions to support the r2i method this uses.
12529 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12530 return a const string when you are expecting an allocated buffer.
12533 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12534 types DirectoryString and DisplayText.
12537 *) Add code to allow r2i extensions to access the configuration database,
12538 add an LHASH database driver and add several ctx helper functions.
12541 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12542 fail when they extended the size of a BIGNUM.
12545 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12546 support typesafe stack.
12549 *) Fix typo in SSL_[gs]et_options().
12550 [Nils Frostberg <nils@medcom.se>]
12552 *) Delete various functions and files that belonged to the (now obsolete)
12553 old X509V3 handling code.
12556 *) New Configure option "rsaref".
12559 *) Don't auto-generate pem.h.
12562 *) Introduce type-safe ASN.1 SETs.
12565 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12566 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12568 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12569 that links with OpenSSL (well at least cause lots of warnings), but fear
12570 not: the conversion is trivial, and it eliminates loads of evil casts. A
12571 few STACKed things have been converted already. Feel free to convert more.
12572 In the fullness of time, I'll do away with the STACK type altogether.
12575 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12576 specified in <certfile> by updating the entry in the index.txt file.
12577 This way one no longer has to edit the index.txt file manually for
12578 revoking a certificate. The -revoke option does the gory details now.
12579 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12581 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12582 `-text' option at all and this way the `-noout -text' combination was
12583 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12584 [Ralf S. Engelschall]
12586 *) Make sure a corresponding plain text error message exists for the
12587 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12588 verify callback function determined that a certificate was revoked.
12589 [Ralf S. Engelschall]
12591 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12592 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12593 all available ciphers including rc5, which was forgotten until now.
12594 In order to let the testing shell script know which algorithms
12595 are available, a new (up to now undocumented) command
12596 "openssl list-cipher-commands" is used.
12599 *) Bugfix: s_client occasionally would sleep in select() when
12600 it should have checked SSL_pending() first.
12603 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12604 the raw DSA values prior to ASN.1 encoding.
12607 *) Tweaks to Configure
12608 [Niels Poppe <niels@netbox.org>]
12610 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12614 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12617 *) New config option to avoid instructions that are illegal on the 80386.
12618 The default code is faster, but requires at least a 486.
12621 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12622 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12623 same as SSL2_VERSION anyway.
12626 *) New "-showcerts" option for s_client.
12629 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12630 application. Various cleanups and fixes.
12633 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12634 modify error routines to work internally. Add error codes and PBE init
12635 to library startup routines.
12638 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12639 packing functions to asn1 and evp. Changed function names and error
12640 codes along the way.
12643 *) PKCS12 integration: and so it begins... First of several patches to
12644 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12645 objects to objects.h
12648 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12649 and display support for Thawte strong extranet extension.
12652 *) Add LinuxPPC support.
12653 [Jeff Dubrule <igor@pobox.org>]
12655 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12656 bn_div_words in alpha.s.
12657 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12659 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12660 OAEP isn't supported when OpenSSL is built with RSAref.
12661 [Ulf Moeller <ulf@fitug.de>]
12663 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12664 so they no longer are missing under -DNOPROTO.
12665 [Soren S. Jorvang <soren@t.dk>]
12668 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12670 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12671 doesn't work when the session is reused. Coming soon!
12674 *) Fix a security hole, that allows sessions to be reused in the wrong
12675 context thus bypassing client cert protection! All software that uses
12676 client certs and session caches in multiple contexts NEEDS PATCHING to
12677 allow session reuse! A fuller solution is in the works.
12678 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12680 *) Some more source tree cleanups (removed obsolete files
12681 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12682 permission on "config" script to be executable) and a fix for the INSTALL
12684 [Ulf Moeller <ulf@fitug.de>]
12686 *) Remove some legacy and erroneous uses of malloc, free instead of
12688 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12690 *) Make rsa_oaep_test return non-zero on error.
12691 [Ulf Moeller <ulf@fitug.de>]
12693 *) Add support for native Solaris shared libraries. Configure
12694 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12695 if someone would make that last step automatic.
12696 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12698 *) ctx_size was not built with the right compiler during "make links". Fixed.
12701 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12702 except NULL ciphers". This means the default cipher list will no longer
12703 enable NULL ciphers. They need to be specifically enabled e.g. with
12704 the string "DEFAULT:eNULL".
12707 *) Fix to RSA private encryption routines: if p < q then it would
12708 occasionally produce an invalid result. This will only happen with
12709 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12712 *) Be less restrictive and allow also `perl util/perlpath.pl
12713 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12714 because this way one can also use an interpreter named `perl5' (which is
12715 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12716 installed as `perl').
12717 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12719 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12720 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12722 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12723 advapi32.lib to Win32 build and change the pem test comparison
12724 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12725 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12726 and crypto/des/ede_cbcm_enc.c.
12729 *) DES quad checksum was broken on big-endian architectures. Fixed.
12732 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12733 Win32 test batch file so it (might) work again. The Win32 test batch file
12734 is horrible: I feel ill....
12737 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12738 in e_os.h. Audit of header files to check ANSI and non ANSI
12739 sections: 10 functions were absent from non ANSI section and not exported
12740 from Windows DLLs. Fixed up libeay.num for new functions.
12743 *) Make `openssl version' output lines consistent.
12744 [Ralf S. Engelschall]
12746 *) Fix Win32 symbol export lists for BIO functions: Added
12747 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12748 to ms/libeay{16,32}.def.
12749 [Ralf S. Engelschall]
12751 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12752 fine under Unix and passes some trivial tests I've now added. But the
12753 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12754 added to make sure no one expects that this stuff really works in the
12755 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12756 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12758 [Ralf S. Engelschall]
12760 *) Fix the generation of two part addresses in perl.
12761 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12763 *) Add config entry for Linux on MIPS.
12764 [John Tobey <jtobey@channel1.com>]
12766 *) Make links whenever Configure is run, unless we are on Windoze.
12769 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12770 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12774 *) Add a useful kludge to allow package maintainers to specify compiler and
12775 other platforms details on the command line without having to patch the
12776 Configure script everytime: One now can use ``perl Configure
12777 <id>:<details>'', i.e. platform ids are allowed to have details appended
12778 to them (separated by colons). This is treated as there would be a static
12779 pre-configured entry in Configure's %table under key <id> with value
12780 <details> and ``perl Configure <id>'' is called. So, when you want to
12781 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12782 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12783 now, which overrides the FreeBSD-elf entry on-the-fly.
12784 [Ralf S. Engelschall]
12786 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12789 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12790 on the `perl Configure ...' command line. This way one can compile
12791 OpenSSL libraries with Position Independent Code (PIC) which is needed
12792 for linking it into DSOs.
12793 [Ralf S. Engelschall]
12795 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12799 *) Cleaned up the LICENSE document: The official contact for any license
12800 questions now is the OpenSSL core team under openssl-core@openssl.org.
12801 And add a paragraph about the dual-license situation to make sure people
12802 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12803 to the OpenSSL toolkit.
12804 [Ralf S. Engelschall]
12806 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12807 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12808 Additionally cleaned up the `make links' target: Remove unnecessary
12809 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12810 to speed processing and no longer clutter the display with confusing
12811 stuff. Instead only the actually done links are displayed.
12812 [Ralf S. Engelschall]
12814 *) Permit null encryption ciphersuites, used for authentication only. It used
12815 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12816 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12820 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12821 signed attributes when verifying signatures (this would break them),
12822 the detached data encoding was wrong and public keys obtained using
12823 X509_get_pubkey() weren't freed.
12826 *) Add text documentation for the BUFFER functions. Also added a work around
12827 to a Win95 console bug. This was triggered by the password read stuff: the
12828 last character typed gets carried over to the next fread(). If you were
12829 generating a new cert request using 'req' for example then the last
12830 character of the passphrase would be CR which would then enter the first
12834 *) Added the new `Includes OpenSSL Cryptography Software' button as
12835 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12836 button and can be used by applications based on OpenSSL to show the
12837 relationship to the OpenSSL project.
12838 [Ralf S. Engelschall]
12840 *) Remove confusing variables in function signatures in files
12841 ssl/ssl_lib.c and ssl/ssl.h.
12842 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12844 *) Don't install bss_file.c under PREFIX/include/
12845 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12847 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12848 functions that return function pointers and has support for NT specific
12849 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12850 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12851 unsigned to signed types: this was killing the Win32 compile.
12854 *) Add new certificate file to stack functions,
12855 SSL_add_dir_cert_subjects_to_stack() and
12856 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12857 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12858 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12859 This means that Apache-SSL and similar packages don't have to mess around
12860 to add as many CAs as they want to the preferred list.
12863 *) Experiment with doxygen documentation. Currently only partially applied to
12865 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12866 openssl.doxy as the configuration file.
12869 *) Get rid of remaining C++-style comments which strict C compilers hate.
12870 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12872 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12873 compiled in by default: it has problems with large keys.
12876 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12877 DH private keys and/or callback functions which directly correspond to
12878 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12879 is needed for applications which have to configure certificates on a
12880 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12882 For the RSA certificate situation is makes no difference, but
12883 for the DSA certificate situation this fixes the "no shared cipher"
12884 problem where the OpenSSL cipher selection procedure failed because the
12885 temporary keys were not overtaken from the context and the API provided
12886 no way to reconfigure them.
12887 The new functions now let applications reconfigure the stuff and they
12888 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12889 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12890 non-public-API function ssl_cert_instantiate() is used as a helper
12891 function and also to reduce code redundancy inside ssl_rsa.c.
12892 [Ralf S. Engelschall]
12894 *) Move s_server -dcert and -dkey options out of the undocumented feature
12895 area because they are useful for the DSA situation and should be
12896 recognized by the users.
12897 [Ralf S. Engelschall]
12899 *) Fix the cipher decision scheme for export ciphers: the export bits are
12900 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12901 SSL_EXP_MASK. So, the original variable has to be used instead of the
12902 already masked variable.
12903 [Richard Levitte <levitte@stacken.kth.se>]
12905 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12906 [Richard Levitte <levitte@stacken.kth.se>]
12908 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12909 from `int' to `unsigned int' because it's a length and initialized by
12910 EVP_DigestFinal() which expects an `unsigned int *'.
12911 [Richard Levitte <levitte@stacken.kth.se>]
12913 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12914 script. Instead use the usual Shell->Perl transition trick.
12915 [Ralf S. Engelschall]
12917 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12918 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12919 -noout -modulus' as it's already the case for `openssl rsa -noout
12920 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12921 currently the public key is printed (a decision which was already done by
12922 `openssl dsa -modulus' in the past) which serves a similar purpose.
12923 Additionally the NO_RSA no longer completely removes the whole -modulus
12924 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12926 [Ralf S. Engelschall]
12928 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12929 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12930 [Arne Ansper <arne@ats.cyber.ee>]
12932 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12933 to be added. Now both 'req' and 'ca' can use new objects defined in the
12937 *) Add cool BIO that does syslog (or event log on NT).
12938 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12940 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12941 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12942 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12943 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12946 *) Add preliminary config info for new extension code.
12949 *) Make RSA_NO_PADDING really use no padding.
12950 [Ulf Moeller <ulf@fitug.de>]
12952 *) Generate errors when private/public key check is done.
12955 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12956 for some CRL extensions and new objects added.
12959 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12960 key usage extension and fuller support for authority key id.
12963 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12964 padding method for RSA, which is recommended for new applications in PKCS
12965 #1 v2.0 (RFC 2437, October 1998).
12966 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12967 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12968 against Bleichbacher's attack on RSA.
12969 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12972 *) Updates to the new SSL compression code
12973 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12975 *) Fix so that the version number in the master secret, when passed
12976 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12977 (because the server will not accept higher), that the version number
12978 is 0x03,0x01, not 0x03,0x00
12979 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12981 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12982 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12983 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12986 *) Support for RAW extensions where an arbitrary extension can be
12987 created by including its DER encoding. See apps/openssl.cnf for
12991 *) Make sure latest Perl versions don't interpret some generated C array
12992 code as Perl array code in the crypto/err/err_genc.pl script.
12993 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12995 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12996 not many people have the assembler. Various Win32 compilation fixes and
12997 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12998 build instructions.
13001 *) Modify configure script 'Configure' to automatically create crypto/date.h
13002 file under Win32 and also build pem.h from pem.org. New script
13003 util/mkfiles.pl to create the MINFO file on environments that can't do a
13004 'make files': perl util/mkfiles.pl >MINFO should work.
13007 *) Major rework of DES function declarations, in the pursuit of correctness
13008 and purity. As a result, many evil casts evaporated, and some weirdness,
13009 too. You may find this causes warnings in your code. Zapping your evil
13010 casts will probably fix them. Mostly.
13013 *) Fix for a typo in asn1.h. Bug fix to object creation script
13014 obj_dat.pl. It considered a zero in an object definition to mean
13015 "end of object": none of the objects in objects.h have any zeros
13016 so it wasn't spotted.
13017 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
13019 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
13020 Masking (CBCM). In the absence of test vectors, the best I have been able
13021 to do is check that the decrypt undoes the encrypt, so far. Send me test
13022 vectors if you have them.
13025 *) Correct calculation of key length for export ciphers (too much space was
13026 allocated for null ciphers). This has not been tested!
13029 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
13030 message is now correct (it understands "crypto" and "ssl" on its
13031 command line). There is also now an "update" option. This will update
13032 the util/ssleay.num and util/libeay.num files with any new functions.
13034 perl util/mkdef.pl crypto ssl update
13035 it will update them.
13038 *) Overhauled the Perl interface (perl/*):
13039 - ported BN stuff to OpenSSL's different BN library
13040 - made the perl/ source tree CVS-aware
13041 - renamed the package from SSLeay to OpenSSL (the files still contain
13042 their history because I've copied them in the repository)
13043 - removed obsolete files (the test scripts will be replaced
13044 by better Test::Harness variants in the future)
13045 [Ralf S. Engelschall]
13047 *) First cut for a very conservative source tree cleanup:
13048 1. merge various obsolete readme texts into doc/ssleay.txt
13049 where we collect the old documents and readme texts.
13050 2. remove the first part of files where I'm already sure that we no
13051 longer need them because of three reasons: either they are just temporary
13052 files which were left by Eric or they are preserved original files where
13053 I've verified that the diff is also available in the CVS via "cvs diff
13054 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13055 the crypto/md/ stuff).
13056 [Ralf S. Engelschall]
13058 *) More extension code. Incomplete support for subject and issuer alt
13059 name, issuer and authority key id. Change the i2v function parameters
13060 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13061 what that's for :-) Fix to ASN1 macro which messed up
13062 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13065 *) Preliminary support for ENUMERATED type. This is largely copied from the
13069 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13070 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13072 *) Make sure `make rehash' target really finds the `openssl' program.
13073 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13075 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13076 like to hear about it if this slows down other processors.
13079 *) Add CygWin32 platform information to Configure script.
13080 [Alan Batie <batie@aahz.jf.intel.com>]
13082 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13083 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13085 *) New program nseq to manipulate netscape certificate sequences
13088 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13092 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13093 but the BN code had some problems that would cause failures when
13094 doing certificate verification and some other functions.
13095 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13097 *) Add ASN1 and PEM code to support netscape certificate sequences.
13100 *) Add ASN1 and PEM code to support netscape certificate sequences.
13103 *) Add several PKIX and private extended key usage OIDs.
13106 *) Modify the 'ca' program to handle the new extension code. Modify
13107 openssl.cnf for new extension format, add comments.
13110 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13111 and add a sample to openssl.cnf so req -x509 now adds appropriate
13115 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13116 error code, add initial support to X509_print() and x509 application.
13119 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13120 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13121 stuff is currently isolated and isn't even compiled yet.
13124 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13125 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13126 Removed the versions check from X509 routines when loading extensions:
13127 this allows certain broken certificates that don't set the version
13128 properly to be processed.
13131 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13132 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13133 can still be regenerated with "make depend".
13136 *) Spelling mistake in C version of CAST-128.
13137 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13139 *) Changes to the error generation code. The perl script err-code.pl
13140 now reads in the old error codes and retains the old numbers, only
13141 adding new ones if necessary. It also only changes the .err files if new
13142 codes are added. The makefiles have been modified to only insert errors
13143 when needed (to avoid needlessly modifying header files). This is done
13144 by only inserting errors if the .err file is newer than the auto generated
13145 C file. To rebuild all the error codes from scratch (the old behaviour)
13146 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13147 or delete all the .err files.
13150 *) CAST-128 was incorrectly implemented for short keys. The C version has
13151 been fixed, but is untested. The assembler versions are also fixed, but
13152 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13153 to regenerate it if needed.
13154 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13155 Hagino <itojun@kame.net>]
13157 *) File was opened incorrectly in randfile.c.
13158 [Ulf Möller <ulf@fitug.de>]
13160 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13161 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13162 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13163 al: it's just almost always a UTCTime. Note this patch adds new error
13164 codes so do a "make errors" if there are problems.
13167 *) Correct Linux 1 recognition in config.
13168 [Ulf Möller <ulf@fitug.de>]
13170 *) Remove pointless MD5 hash when using DSA keys in ca.
13171 [Anonymous <nobody@replay.com>]
13173 *) Generate an error if given an empty string as a cert directory. Also
13174 generate an error if handed NULL (previously returned 0 to indicate an
13175 error, but didn't set one).
13176 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13178 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13181 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13182 parameters. This was causing a warning which killed off the Win32 compile.
13185 *) Remove C++ style comments from crypto/bn/bn_local.h.
13186 [Neil Costigan <neil.costigan@celocom.com>]
13188 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13189 based on a text string, looking up short and long names and finally
13190 "dot" format. The "dot" format stuff didn't work. Added new function
13191 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13192 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13193 OID is not part of the table.
13196 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13197 X509_LOOKUP_by_alias().
13200 *) Sort openssl functions by name.
13203 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13204 encryption from sample DSA keys (in case anyone is interested the password
13208 *) Make _all_ *_free functions accept a NULL pointer.
13209 [Frans Heymans <fheymans@isaserver.be>]
13211 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13213 [Anonymous <nobody@replay.com>]
13215 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13216 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13218 *) Don't blow it for numeric -newkey arguments to apps/req.
13219 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13221 *) Temp key "for export" tests were wrong in s3_srvr.c.
13222 [Anonymous <nobody@replay.com>]
13224 *) Add prototype for temp key callback functions
13225 SSL_CTX_set_tmp_{rsa,dh}_callback().
13228 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13229 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13232 *) X509_name_add_entry() freed the wrong thing after an error.
13233 [Arne Ansper <arne@ats.cyber.ee>]
13235 *) rsa_eay.c would attempt to free a NULL context.
13236 [Arne Ansper <arne@ats.cyber.ee>]
13238 *) BIO_s_socket() had a broken should_retry() on Windoze.
13239 [Arne Ansper <arne@ats.cyber.ee>]
13241 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13242 [Arne Ansper <arne@ats.cyber.ee>]
13244 *) Make sure the already existing X509_STORE->depth variable is initialized
13245 in X509_STORE_new(), but document the fact that this variable is still
13246 unused in the certificate verification process.
13247 [Ralf S. Engelschall]
13249 *) Fix the various library and apps files to free up pkeys obtained from
13250 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13253 *) Fix reference counting in X509_PUBKEY_get(). This makes
13254 demos/maurice/example2.c work, amongst others, probably.
13255 [Steve Henson and Ben Laurie]
13257 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13258 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13259 are no longer created. This way we have a single and consistent command
13260 line interface `openssl <command>', similar to `cvs <command>'.
13261 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13263 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13264 BIT STRING wrapper always have zero unused bits.
13267 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13270 *) Make the top-level INSTALL documentation easier to understand.
13273 *) Makefiles updated to exit if an error occurs in a sub-directory
13274 make (including if user presses ^C) [Paul Sutton]
13276 *) Make Montgomery context stuff explicit in RSA data structure.
13279 *) Fix build order of pem and err to allow for generated pem.h.
13282 *) Fix renumbering bug in X509_NAME_delete_entry().
13285 *) Enhanced the err-ins.pl script so it makes the error library number
13286 global and can add a library name. This is needed for external ASN1 and
13287 other error libraries.
13290 *) Fixed sk_insert which never worked properly.
13293 *) Fix ASN1 macros so they can handle indefinite length constructed
13294 EXPLICIT tags. Some non standard certificates use these: they can now
13298 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13299 into a single doc/ssleay.txt bundle. This way the information is still
13300 preserved but no longer messes up this directory. Now it's new room for
13301 the new set of documentation files.
13302 [Ralf S. Engelschall]
13304 *) SETs were incorrectly DER encoded. This was a major pain, because they
13305 shared code with SEQUENCEs, which aren't coded the same. This means that
13306 almost everything to do with SETs or SEQUENCEs has either changed name or
13307 number of arguments.
13308 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13310 *) Fix test data to work with the above.
13313 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13314 was already fixed by Eric for 0.9.1 it seems.
13315 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13317 *) Autodetect FreeBSD3.
13320 *) Fix various bugs in Configure. This affects the following platforms:
13324 unixware-2.0-pentium
13328 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13329 before they are needed.
13332 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13336 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13338 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13339 changed SSLeay to OpenSSL in version strings.
13340 [Ralf S. Engelschall]
13342 *) Some fixups to the top-level documents.
13345 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13346 because the symlink to include/ was missing.
13347 [Ralf S. Engelschall]
13349 *) Incorporated the popular no-RSA/DSA-only patches
13350 which allow to compile a RSA-free SSLeay.
13351 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13353 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13354 when "ssleay" is still not found.
13355 [Ralf S. Engelschall]
13357 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13358 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13360 *) Updated the README file.
13361 [Ralf S. Engelschall]
13363 *) Added various .cvsignore files in the CVS repository subdirs
13364 to make a "cvs update" really silent.
13365 [Ralf S. Engelschall]
13367 *) Recompiled the error-definition header files and added
13368 missing symbols to the Win32 linker tables.
13369 [Ralf S. Engelschall]
13371 *) Cleaned up the top-level documents;
13372 o new files: CHANGES and LICENSE
13373 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13374 o merged COPYRIGHT into LICENSE
13375 o removed obsolete TODO file
13376 o renamed MICROSOFT to INSTALL.W32
13377 [Ralf S. Engelschall]
13379 *) Removed dummy files from the 0.9.1b source tree:
13380 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13381 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13382 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13383 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13384 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13385 [Ralf S. Engelschall]
13387 *) Added various platform portability fixes.
13390 *) The Genesis of the OpenSSL rpject:
13391 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13392 Young and Tim J. Hudson created while they were working for C2Net until
13394 [The OpenSSL Project]
13397 Changes between 0.9.0b and 0.9.1b [not released]
13399 *) Updated a few CA certificates under certs/
13402 *) Changed some BIGNUM api stuff.
13405 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13406 DGUX x86, Linux Alpha, etc.
13409 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13410 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13414 *) Add -strparse option to asn1pars program which parses nested
13416 [Dr Stephen Henson <shenson@bigfoot.com>]
13418 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13421 *) DSA fix for "ca" program.
13424 *) Added "-genkey" option to "dsaparam" program.
13427 *) Added RIPE MD160 (rmd160) message digest.
13430 *) Added -a (all) option to "ssleay version" command.
13433 *) Added PLATFORM define which is the id given to Configure.
13436 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13439 *) Extended the ASN.1 parser routines.
13442 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13445 *) Added a BN_CTX to the BN library.
13448 *) Fixed the weak key values in DES library
13451 *) Changed API in EVP library for cipher aliases.
13454 *) Added support for RC2/64bit cipher.
13457 *) Converted the lhash library to the crypto/mem.c functions.
13460 *) Added more recognized ASN.1 object ids.
13463 *) Added more RSA padding checks for SSL/TLS.
13466 *) Added BIO proxy/filter functionality.
13469 *) Added extra_certs to SSL_CTX which can be used
13470 send extra CA certificates to the client in the CA cert chain sending
13471 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13474 *) Now Fortezza is denied in the authentication phase because
13475 this is key exchange mechanism is not supported by SSLeay at all.
13478 *) Additional PKCS1 checks.
13481 *) Support the string "TLSv1" for all TLS v1 ciphers.
13484 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13485 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13488 *) Fixed a few memory leaks.
13491 *) Fixed various code and comment typos.
13494 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13495 bytes sent in the client random.
13496 [Edward Bishop <ebishop@spyglass.com>]