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 *) Introduced the new functions EVP_DigestSignInit_ex() and
13 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
14 EVP_DigestVerifyUpdate() have been converted to functions. See the man
15 pages for further details.
18 *) s390x assembly pack: add hardware-support for P-256, P-384, P-521,
19 X25519, X448, Ed25519 and Ed448.
22 *) Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
26 *) Deprecated the public definition of ERR_STATE as well as the function
27 ERR_get_state(). This is done in preparation of making ERR_STATE an
31 *) Added ERR functionality to give callers access to the stored function
32 names that have replaced the older function code based functions.
34 New functions are ERR_get_error_func(), ERR_peek_error_func(),
35 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
36 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
37 and ERR_peek_last_error_all().
39 These functions have become deprecated: ERR_get_error_line_data(),
40 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
41 ERR_func_error_string().
44 *) Extended testing to be verbose for failing tests only. The make variables
45 VERBOSE_FAILURE or VF can be used to enable this:
47 $ make VF=1 test # Unix
48 $ mms /macro=(VF=1) test ! OpenVMS
49 $ nmake VF=1 test # Windows
53 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
54 used even when parsing explicit parameters, when loading a serialized key
55 or calling `EC_GROUP_new_from_ecpkparameters()`/
56 `EC_GROUP_new_from_ecparameters()`.
57 This prevents bypass of security hardening and performance gains,
58 especially for curves with specialized EC_METHODs.
59 By default, if a key encoded with explicit parameters is loaded and later
60 serialized, the output is still encoded with explicit parameters, even if
61 internally a "named" EC_GROUP is used for computation.
64 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
65 this change, EC_GROUP_set_generator would accept order and/or cofactor as
66 NULL. After this change, only the cofactor parameter can be NULL. It also
67 does some minimal sanity checks on the passed order.
71 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
72 An attack is simple, if the first CMS_recipientInfo is valid but the
73 second CMS_recipientInfo is chosen ciphertext. If the second
74 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
75 encryption key will be replaced by garbage, and the message cannot be
76 decoded, but if the RSA decryption fails, the correct encryption key is
77 used and the recipient will not notice the attack.
78 As a work around for this potential attack the length of the decrypted
79 key must be equal to the cipher default key length, in case the
80 certifiate is not given and all recipientInfo are tried out.
81 The old behaviour can be re-enabled in the CMS code by setting the
82 CMS_DEBUG_DECRYPT flag.
85 *) Early start up entropy quality from the DEVRANDOM seed source has been
86 improved for older Linux systems. The RAND subsystem will wait for
87 /dev/random to be producing output before seeding from /dev/urandom.
88 The seeded state is stored for future library initialisations using
89 a system global shared memory segment. The shared memory identifier
90 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
91 the desired value. The default identifier is 114.
94 *) Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
95 when primes for RSA keys are computed.
96 Since we previously always generated primes == 2 (mod 3) for RSA keys,
97 the 2-prime and 3-prime RSA modules were easy to distinguish, since
98 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
99 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
100 This avoids possible fingerprinting of newly generated RSA modules.
103 *) Correct the extended master secret constant on EBCDIC systems. Without this
104 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
105 negotiate EMS will fail. Unfortunately this also means that TLS connections
106 between EBCDIC systems with this fix, and EBCDIC systems without this
107 fix will fail if they negotiate EMS.
110 *) Changed the library initialisation so that the config file is now loaded
111 by default. This was already the case for libssl. It now occurs for both
112 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
113 OPENSSL_init_crypto() to suppress automatic loading of a config file.
116 *) Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
117 where the former acts as a replacement for ERR_put_error(), and the
118 latter replaces the combination ERR_put_error()+ERR_add_error_data().
119 ERR_raise_data() adds more flexibility by taking a format string and
120 an arbitrary number of arguments following it, to be processed with
124 *) Introduced a new function, OSSL_PROVIDER_available(), which can be used
125 to check if a named provider is loaded and available. When called, it
126 will also activate all fallback providers if such are still present.
129 *) Enforce a minimum DH modulus size of 512 bits.
132 *) Changed DH parameters to generate the order q subgroup instead of 2q.
133 Previously generated DH parameters are still accepted by DH_check
134 but DH_generate_key works around that by clearing bit 0 of the
135 private key for those. This avoids leaking bit 0 of the private key.
138 *) Significantly reduce secure memory usage by the randomness pools.
141 *) {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
145 *) A new type, EVP_KEYEXCH, has been introduced to represent key exchange
146 algorithms. An implementation of a key exchange algorithm can be obtained
147 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
148 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
149 the older EVP_PKEY_derive_init() function. See the man pages for the new
150 functions for further details.
153 *) The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
156 *) Removed the function names from error messages and deprecated the
159 *) Removed NextStep support and the macro OPENSSL_UNISTD
162 *) Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
163 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
164 Also removed "export var as function" capability; we do not export
165 variables, only functions.
168 *) RC5_32_set_key has been changed to return an int type, with 0 indicating
169 an error and 1 indicating success. In previous versions of OpenSSL this
170 was a void type. If a key was set longer than the maximum possible this
174 *) Support SM2 signing and verification schemes with X509 certificate.
177 *) Use SHA256 as the default digest for TS query in the ts app.
180 *) Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
181 This checks that the salt length is at least 128 bits, the derived key
182 length is at least 112 bits, and that the iteration count is at least 1000.
183 For backwards compatibility these checks are disabled by default in the
184 default provider, but are enabled by default in the fips provider.
185 To enable or disable these checks use the control
186 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
189 *) Default cipher lists/suites are now available via a function, the
190 #defines are deprecated.
193 *) Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
194 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
195 for Windows Store apps easier. Also, the "no-uplink" option has been added.
198 *) Join the directories crypto/x509 and crypto/x509v3
201 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
202 This changes the size when using the genpkey app when no size is given. It
203 fixes an omission in earlier changes that changed all RSA, DSA and DH
204 generation apps to use 2048 bits by default.
207 *) Added command 'openssl kdf' that uses the EVP_KDF API.
210 *) Added command 'openssl mac' that uses the EVP_MAC API.
213 *) Added OPENSSL_info() to get diverse built-in OpenSSL data, such
214 as default directories. Also added the command 'openssl info'
215 for scripting purposes.
218 *) The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
219 deprecated. These undocumented functions were never integrated into the EVP
220 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
221 Bi-directional IGE mode. These modes were never formally standardised and
222 usage of these functions is believed to be very small. In particular
223 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
224 is ever used. The security implications are believed to be minimal, but
225 this issue was never fixed for backwards compatibility reasons. New code
226 should not use these modes.
229 *) Add prediction resistance to the DRBG reseeding process.
232 *) Limit the number of blocks in a data unit for AES-XTS to 2^20 as
233 mandated by IEEE Std 1619-2018.
236 *) Added newline escaping functionality to a filename when using openssl dgst.
237 This output format is to replicate the output format found in the '*sum'
238 checksum programs. This aims to preserve backward compatibility.
239 [Matt Eaton, Richard Levitte, and Paul Dale]
241 *) Removed the heartbeat message in DTLS feature, as it has very
242 little usage and doesn't seem to fulfill a valuable purpose.
243 The configuration option is now deprecated.
246 *) Changed the output of 'openssl {digestname} < file' to display the
247 digest name in its output.
250 *) Added a new generic trace API which provides support for enabling
251 instrumentation through trace output. This feature is mainly intended
252 as an aid for developers and is disabled by default. To utilize it,
253 OpenSSL needs to be configured with the `enable-trace` option.
255 If the tracing API is enabled, the application can activate trace output
256 by registering BIOs as trace channels for a number of tracing and debugging
259 The 'openssl' application has been expanded to enable any of the types
260 available via environment variables defined by the user, and serves as
261 one possible example on how to use this functionality.
262 [Richard Levitte & Matthias St. Pierre]
264 *) Added build tests for C++. These are generated files that only do one
265 thing, to include one public OpenSSL head file each. This tests that
266 the public header files can be usefully included in a C++ application.
268 This test isn't enabled by default. It can be enabled with the option
269 'enable-buildtest-c++'.
272 *) Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
275 *) Add KMAC to EVP_MAC.
278 *) Added property based algorithm implementation selection framework to
282 *) Added SCA hardening for modular field inversion in EC_GROUP through
283 a new dedicated field_inv() pointer in EC_METHOD.
284 This also addresses a leakage affecting conversions from projective
285 to affine coordinates.
286 [Billy Bob Brumley, Nicola Tuveri]
288 *) Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
289 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
290 those algorithms that were already supported through the EVP_PKEY API
291 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
292 and scrypt are now wrappers that call EVP_KDF.
295 *) Build devcrypto engine as a dynamic engine.
298 *) Add keyed BLAKE2 to EVP_MAC.
301 *) Fix a bug in the computation of the endpoint-pair shared secret used
302 by DTLS over SCTP. This breaks interoperability with older versions
303 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
304 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
305 interoperability with such broken implementations. However, enabling
306 this switch breaks interoperability with correct implementations.
308 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
309 re-used X509_PUBKEY object if the second PUBKEY is malformed.
312 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
315 *) Change the license to the Apache License v2.0.
318 *) Change the possible version information given with OPENSSL_API_COMPAT.
319 It may be a pre-3.0.0 style numerical version number as it was defined
320 in 1.1.0, and it may also simply take the major version number.
322 Because of the version numbering of pre-3.0.0 releases, the values 0,
323 1 and 2 are equivalent to 0x00908000L (0.9.8), 0x10000000L (1.0.0) and
324 0x10100000L (1.1.0), respectively.
327 *) Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
329 o Major releases (indicated by incrementing the MAJOR release number)
330 may introduce incompatible API/ABI changes.
331 o Minor releases (indicated by incrementing the MINOR release number)
332 may introduce new features but retain API/ABI compatibility.
333 o Patch releases (indicated by incrementing the PATCH number)
334 are intended for bug fixes and other improvements of existing
335 features only (like improving performance or adding documentation)
336 and retain API/ABI compatibility.
339 *) Add support for RFC5297 SIV mode (siv128), including AES-SIV.
342 *) Remove the 'dist' target and add a tarball building script. The
343 'dist' target has fallen out of use, and it shouldn't be
344 necessary to configure just to create a source distribution.
347 *) Recreate the OS390-Unix config target. It no longer relies on a
348 special script like it did for OpenSSL pre-1.1.0.
351 *) Instead of having the source directories listed in Configure, add
352 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
356 *) Add GMAC to EVP_MAC.
359 *) Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
362 *) Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
363 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
364 to facilitate the continued use of MACs through raw private keys in
365 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
368 *) Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
369 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
372 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
373 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
374 are retained for backwards compatibility.
377 *) AES-XTS mode now enforces that its two keys are different to mitigate
378 the attacked described in "Efficient Instantiations of Tweakable
379 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
380 Details of this attack can be obtained from:
381 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
384 *) Rename the object files, i.e. give them other names than in previous
385 versions. Their names now include the name of the final product, as
386 well as its type mnemonic (bin, lib, shlib).
389 *) Added new option for 'openssl list', '-objects', which will display the
390 list of built in objects, i.e. OIDs with names.
393 *) Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
394 improves application performance by removing data copies and providing
395 applications with zero-copy system calls such as sendfile and splice.
398 Changes between 1.1.1a and 1.1.1b [xx XXX xxxx]
400 *) Change the info callback signals for the start and end of a post-handshake
401 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
402 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
403 confused by this and assume that a TLSv1.2 renegotiation has started. This
404 can break KeyUpdate handling. Instead we no longer signal the start and end
405 of a post handshake message exchange (although the messages themselves are
406 still signalled). This could break some applications that were expecting
407 the old signals. However without this KeyUpdate is not usable for many
411 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
413 *) Timing vulnerability in DSA signature generation
415 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
416 timing side channel attack. An attacker could use variations in the signing
417 algorithm to recover the private key.
419 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
423 *) Timing vulnerability in ECDSA signature generation
425 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
426 timing side channel attack. An attacker could use variations in the signing
427 algorithm to recover the private key.
429 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
433 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
434 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
435 of two gigabytes and the error handling improved.
437 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
438 categorized as a normal bug, not a security issue, because the DRBG reseeds
439 automatically and is fully functional even without additional randomness
440 provided by the application.
442 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
444 *) Add a new ClientHello callback. Provides a callback interface that gives
445 the application the ability to adjust the nascent SSL object at the
446 earliest stage of ClientHello processing, immediately after extensions have
447 been collected but before they have been processed. In particular, this
448 callback can adjust the supported TLS versions in response to the contents
452 *) Add SM2 base algorithm support.
455 *) s390x assembly pack: add (improved) hardware-support for the following
456 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
457 aes-cfb/cfb8, aes-ecb.
460 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
461 parameter is no longer accepted, as it leads to a corrupt table. NULL
462 pem_str is reserved for alias entries only.
465 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
466 step for prime curves. The new implementation is based on formulae from
467 differential addition-and-doubling in homogeneous projective coordinates
468 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
469 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
470 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
471 to work in projective coordinates.
472 [Billy Bob Brumley, Nicola Tuveri]
474 *) Change generating and checking of primes so that the error rate of not
475 being prime depends on the intended use based on the size of the input.
476 For larger primes this will result in more rounds of Miller-Rabin.
477 The maximal error rate for primes with more than 1080 bits is lowered
479 [Kurt Roeckx, Annie Yousar]
481 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
484 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
485 moving between systems, and to avoid confusion when a Windows build is
486 done with mingw vs with MSVC. For POSIX installs, there's still a
487 symlink or copy named 'tsget' to avoid that confusion as well.
490 *) Revert blinding in ECDSA sign and instead make problematic addition
491 length-invariant. Switch even to fixed-length Montgomery multiplication.
494 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
495 step for binary curves. The new implementation is based on formulae from
496 differential addition-and-doubling in mixed Lopez-Dahab projective
497 coordinates, modified to independently blind the operands.
498 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
500 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
501 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
502 EC_METHODs to implement their own specialized "ladder step", to take
503 advantage of more favorable coordinate systems or more efficient
504 differential addition-and-doubling algorithms.
505 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
507 *) Modified the random device based seed sources to keep the relevant
508 file descriptors open rather than reopening them on each access.
509 This allows such sources to operate in a chroot() jail without
510 the associated device nodes being available. This behaviour can be
511 controlled using RAND_keep_random_devices_open().
514 *) Numerous side-channel attack mitigations have been applied. This may have
515 performance impacts for some algorithms for the benefit of improved
516 security. Specific changes are noted in this change log by their respective
520 *) AIX shared library support overhaul. Switch to AIX "natural" way of
521 handling shared libraries, which means collecting shared objects of
522 different versions and bitnesses in one common archive. This allows to
523 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
524 doesn't affect the way 3rd party applications are linked, only how
525 multi-version installation is managed.
528 *) Make ec_group_do_inverse_ord() more robust and available to other
529 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
530 mitigations are applied to the fallback BN_mod_inverse().
531 When using this function rather than BN_mod_inverse() directly, new
532 EC cryptosystem implementations are then safer-by-default.
535 *) Add coordinate blinding for EC_POINT and implement projective
536 coordinate blinding for generic prime curves as a countermeasure to
537 chosen point SCA attacks.
538 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
540 *) Add blinding to ECDSA and DSA signatures to protect against side channel
541 attacks discovered by Keegan Ryan (NCC Group).
544 *) Enforce checking in the pkeyutl command line app to ensure that the input
545 length does not exceed the maximum supported digest length when performing
546 a sign, verify or verifyrecover operation.
549 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
550 I/O in combination with something like select() or poll() will hang. This
551 can be turned off again using SSL_CTX_clear_mode().
552 Many applications do not properly handle non-application data records, and
553 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
554 around the problems in those applications, but can also break some.
555 It's recommended to read the manpages about SSL_read(), SSL_write(),
556 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
557 SSL_CTX_set_read_ahead() again.
560 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
561 now allow empty (zero character) pass phrases.
564 *) Apply blinding to binary field modular inversion and remove patent
565 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
568 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
569 binary and prime elliptic curves.
572 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
573 constant time fixed point multiplication.
576 *) Revise elliptic curve scalar multiplication with timing attack
577 defenses: ec_wNAF_mul redirects to a constant time implementation
578 when computing fixed point and variable point multiplication (which
579 in OpenSSL are mostly used with secret scalars in keygen, sign,
580 ECDH derive operations).
581 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
584 *) Updated CONTRIBUTING
587 *) Updated DRBG / RAND to request nonce and additional low entropy
588 randomness from the system.
589 [Matthias St. Pierre]
591 *) Updated 'openssl rehash' to use OpenSSL consistent default.
594 *) Moved the load of the ssl_conf module to libcrypto, which helps
595 loading engines that libssl uses before libssl is initialised.
598 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
601 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
602 [Ingo Schwarze, Rich Salz]
604 *) Added output of accepting IP address and port for 'openssl s_server'
607 *) Added a new API for TLSv1.3 ciphersuites:
608 SSL_CTX_set_ciphersuites()
609 SSL_set_ciphersuites()
612 *) Memory allocation failures consistently add an error to the error
616 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
617 in libcrypto when run as setuid/setgid.
620 *) Load any config file by default when libssl is used.
623 *) Added new public header file <openssl/rand_drbg.h> and documentation
624 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
625 [Matthias St. Pierre]
627 *) QNX support removed (cannot find contributors to get their approval
628 for the license change).
631 *) TLSv1.3 replay protection for early data has been implemented. See the
632 SSL_read_early_data() man page for further details.
635 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
636 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
637 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
638 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
639 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
640 configuration has been separated out. See the ciphers man page or the
641 SSL_CTX_set_ciphersuites() man page for more information.
644 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
645 in responder mode now supports the new "-multi" option, which
646 spawns the specified number of child processes to handle OCSP
647 requests. The "-timeout" option now also limits the OCSP
648 responder's patience to wait to receive the full client request
649 on a newly accepted connection. Child processes are respawned
650 as needed, and the CA index file is automatically reloaded
651 when changed. This makes it possible to run the "ocsp" responder
652 as a long-running service, making the OpenSSL CA somewhat more
653 feature-complete. In this mode, most diagnostic messages logged
654 after entering the event loop are logged via syslog(3) rather than
658 *) Added support for X448 and Ed448. Heavily based on original work by
662 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
663 objects loaded. This adds the functions OSSL_STORE_expect() and
664 OSSL_STORE_find() as well as needed tools to construct searches and
665 get the search data out of them.
668 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
669 version of OpenSSL should review their configuration settings to ensure
670 that they are still appropriate for TLSv1.3. For further information see:
671 https://wiki.openssl.org/index.php/TLS1.3
674 *) Grand redesign of the OpenSSL random generator
676 The default RAND method now utilizes an AES-CTR DRBG according to
677 NIST standard SP 800-90Ar1. The new random generator is essentially
678 a port of the default random generator from the OpenSSL FIPS 2.0
679 object module. It is a hybrid deterministic random bit generator
680 using an AES-CTR bit stream and which seeds and reseeds itself
681 automatically using trusted system entropy sources.
683 Some of its new features are:
684 o Support for multiple DRBG instances with seed chaining.
685 o The default RAND method makes use of a DRBG.
686 o There is a public and private DRBG instance.
687 o The DRBG instances are fork-safe.
688 o Keep all global DRBG instances on the secure heap if it is enabled.
689 o The public and private DRBG instance are per thread for lock free
691 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
693 *) Changed Configure so it only says what it does and doesn't dump
694 so much data. Instead, ./configdata.pm should be used as a script
695 to display all sorts of configuration data.
698 *) Added processing of "make variables" to Configure.
701 *) Added SHA512/224 and SHA512/256 algorithm support.
704 *) The last traces of Netware support, first removed in 1.1.0, have
708 *) Get rid of Makefile.shared, and in the process, make the processing
709 of certain files (rc.obj, or the .def/.map/.opt files produced from
710 the ordinal files) more visible and hopefully easier to trace and
711 debug (or make silent).
714 *) Make it possible to have environment variable assignments as
715 arguments to config / Configure.
718 *) Add multi-prime RSA (RFC 8017) support.
721 *) Add SM3 implemented according to GB/T 32905-2016
722 [ Jack Lloyd <jack.lloyd@ribose.com>,
723 Ronald Tse <ronald.tse@ribose.com>,
724 Erick Borsboom <erick.borsboom@ribose.com> ]
726 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
727 as documented in RFC6066.
728 Based on a patch from Tomasz Moń
729 [Filipe Raimundo da Silva]
731 *) Add SM4 implemented according to GB/T 32907-2016.
732 [ Jack Lloyd <jack.lloyd@ribose.com>,
733 Ronald Tse <ronald.tse@ribose.com>,
734 Erick Borsboom <erick.borsboom@ribose.com> ]
736 *) Reimplement -newreq-nodes and ERR_error_string_n; the
737 original author does not agree with the license change.
740 *) Add ARIA AEAD TLS support.
743 *) Some macro definitions to support VS6 have been removed. Visual
744 Studio 6 has not worked since 1.1.0
747 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
748 without clearing the errors.
751 *) Add "atfork" functions. If building on a system that without
752 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
753 requirements. The RAND facility now uses/requires this.
759 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
760 not possible to disable entirely. However, it's still possible to
761 disable the console reading UI method, UI_OpenSSL() (use UI_null()
764 To disable, configure with 'no-ui-console'. 'no-ui' is still
765 possible to use as an alias. Check at compile time with the
766 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
767 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
770 *) Add a STORE module, which implements a uniform and URI based reader of
771 stores that can contain keys, certificates, CRLs and numerous other
772 objects. The main API is loosely based on a few stdio functions,
773 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
774 OSSL_STORE_error and OSSL_STORE_close.
775 The implementation uses backends called "loaders" to implement arbitrary
776 URI schemes. There is one built in "loader" for the 'file' scheme.
779 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
780 then adjusted to work on FreeBSD 8.4 as well.
781 Enable by configuring with 'enable-devcryptoeng'. This is done by default
782 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
785 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
786 util/mkerr.pl, which is adapted to allow those prefixes, leading to
787 error code calls like this:
789 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
791 With this change, we claim the namespaces OSSL and OPENSSL in a manner
792 that can be encoded in C. For the foreseeable future, this will only
794 [Richard Levitte and Tim Hudson]
796 *) Removed BSD cryptodev engine.
799 *) Add a build target 'build_all_generated', to build all generated files
800 and only that. This can be used to prepare everything that requires
801 things like perl for a system that lacks perl and then move everything
802 to that system and do the rest of the build there.
805 *) In the UI interface, make it possible to duplicate the user data. This
806 can be used by engines that need to retain the data for a longer time
807 than just the call where this user data is passed.
810 *) Ignore the '-named_curve auto' value for compatibility of applications
812 [Tomas Mraz <tmraz@fedoraproject.org>]
814 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
815 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
816 alerts across multiple records (some of which could be empty). In practice
817 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
818 prohibts this altogether and other libraries (BoringSSL, NSS) do not
819 support this at all. Supporting it adds significant complexity to the
820 record layer, and its removal is unlikely to cause inter-operability
824 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
825 with Z. These are meant to replace LONG and ZLONG and to be size safe.
826 The use of LONG and ZLONG is discouraged and scheduled for deprecation
830 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
831 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
832 [Richard Levitte, Andy Polyakov]
834 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
838 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
839 platform rather than 'mingw'.
842 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
843 success if they are asked to add an object which already exists
844 in the store. This change cascades to other functions which load
845 certificates and CRLs.
848 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
849 facilitate stack unwinding even from assembly subroutines.
852 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
853 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
856 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
857 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
858 which is the minimum version we support.
861 *) Certificate time validation (X509_cmp_time) enforces stricter
862 compliance with RFC 5280. Fractional seconds and timezone offsets
863 are no longer allowed.
866 *) Add support for ARIA
869 *) s_client will now send the Server Name Indication (SNI) extension by
870 default unless the new "-noservername" option is used. The server name is
871 based on the host provided to the "-connect" option unless overridden by
875 *) Add support for SipHash
878 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
879 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
880 prevent issues where no progress is being made and the peer continually
881 sends unrecognised record types, using up resources processing them.
884 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
885 using the algorithm defined in
886 https://www.akkadia.org/drepper/SHA-crypt.txt
889 *) Heartbeat support has been removed; the ABI is changed for now.
890 [Richard Levitte, Rich Salz]
892 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
895 *) The RSA "null" method, which was partially supported to avoid patent
896 issues, has been replaced to always returns NULL.
900 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
902 *) Client DoS due to large DH parameter
904 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
905 malicious server can send a very large prime value to the client. This will
906 cause the client to spend an unreasonably long period of time generating a
907 key for this prime resulting in a hang until the client has finished. This
908 could be exploited in a Denial Of Service attack.
910 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
914 *) Cache timing vulnerability in RSA Key Generation
916 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
917 a cache timing side channel attack. An attacker with sufficient access to
918 mount cache timing attacks during the RSA key generation process could
919 recover the private key.
921 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
922 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
926 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
927 parameter is no longer accepted, as it leads to a corrupt table. NULL
928 pem_str is reserved for alias entries only.
931 *) Revert blinding in ECDSA sign and instead make problematic addition
932 length-invariant. Switch even to fixed-length Montgomery multiplication.
935 *) Change generating and checking of primes so that the error rate of not
936 being prime depends on the intended use based on the size of the input.
937 For larger primes this will result in more rounds of Miller-Rabin.
938 The maximal error rate for primes with more than 1080 bits is lowered
940 [Kurt Roeckx, Annie Yousar]
942 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
945 *) Add blinding to ECDSA and DSA signatures to protect against side channel
946 attacks discovered by Keegan Ryan (NCC Group).
949 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
950 now allow empty (zero character) pass phrases.
953 *) Certificate time validation (X509_cmp_time) enforces stricter
954 compliance with RFC 5280. Fractional seconds and timezone offsets
955 are no longer allowed.
958 *) Fixed a text canonicalisation bug in CMS
960 Where a CMS detached signature is used with text content the text goes
961 through a canonicalisation process first prior to signing or verifying a
962 signature. This process strips trailing space at the end of lines, converts
963 line terminators to CRLF and removes additional trailing line terminators
964 at the end of a file. A bug in the canonicalisation process meant that
965 some characters, such as form-feed, were incorrectly treated as whitespace
966 and removed. This is contrary to the specification (RFC5485). This fix
967 could mean that detached text data signed with an earlier version of
968 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
969 signed with a fixed OpenSSL may fail to verify with an earlier version of
970 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
971 and use the "-binary" flag (for the "cms" command line application) or set
972 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
975 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
977 *) Constructed ASN.1 types with a recursive definition could exceed the stack
979 Constructed ASN.1 types with a recursive definition (such as can be found
980 in PKCS7) could eventually exceed the stack given malicious input with
981 excessive recursion. This could result in a Denial Of Service attack. There
982 are no such structures used within SSL/TLS that come from untrusted sources
983 so this is considered safe.
985 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
990 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
992 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
993 effectively reduced to only comparing the least significant bit of each
994 byte. This allows an attacker to forge messages that would be considered as
995 authenticated in an amount of tries lower than that guaranteed by the
996 security claims of the scheme. The module can only be compiled by the
997 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
999 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1004 *) Add a build target 'build_all_generated', to build all generated files
1005 and only that. This can be used to prepare everything that requires
1006 things like perl for a system that lacks perl and then move everything
1007 to that system and do the rest of the build there.
1010 *) Backport SSL_OP_NO_RENGOTIATION
1012 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1013 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1014 changes this is no longer possible in 1.1.0. Therefore the new
1015 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1016 1.1.0 to provide equivalent functionality.
1018 Note that if an application built against 1.1.0h headers (or above) is run
1019 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1020 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1023 *) Removed the OS390-Unix config target. It relied on a script that doesn't
1027 *) rsaz_1024_mul_avx2 overflow bug on x86_64
1029 There is an overflow bug in the AVX2 Montgomery multiplication procedure
1030 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
1031 Analysis suggests that attacks against RSA and DSA as a result of this
1032 defect would be very difficult to perform and are not believed likely.
1033 Attacks against DH1024 are considered just feasible, because most of the
1034 work necessary to deduce information about a private key may be performed
1035 offline. The amount of resources required for such an attack would be
1036 significant. However, for an attack on TLS to be meaningful, the server
1037 would have to share the DH1024 private key among multiple clients, which is
1038 no longer an option since CVE-2016-0701.
1040 This only affects processors that support the AVX2 but not ADX extensions
1041 like Intel Haswell (4th generation).
1043 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1044 was originally found via the OSS-Fuzz project.
1048 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
1050 *) bn_sqrx8x_internal carry bug on x86_64
1052 There is a carry propagating bug in the x86_64 Montgomery squaring
1053 procedure. No EC algorithms are affected. Analysis suggests that attacks
1054 against RSA and DSA as a result of this defect would be very difficult to
1055 perform and are not believed likely. Attacks against DH are considered just
1056 feasible (although very difficult) because most of the work necessary to
1057 deduce information about a private key may be performed offline. The amount
1058 of resources required for such an attack would be very significant and
1059 likely only accessible to a limited number of attackers. An attacker would
1060 additionally need online access to an unpatched system using the target
1061 private key in a scenario with persistent DH parameters and a private
1062 key that is shared between multiple clients.
1064 This only affects processors that support the BMI1, BMI2 and ADX extensions
1065 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1067 This issue was reported to OpenSSL by the OSS-Fuzz project.
1071 *) Malformed X.509 IPAddressFamily could cause OOB read
1073 If an X.509 certificate has a malformed IPAddressFamily extension,
1074 OpenSSL could do a one-byte buffer overread. The most likely result
1075 would be an erroneous display of the certificate in text format.
1077 This issue was reported to OpenSSL by the OSS-Fuzz project.
1081 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1083 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1084 platform rather than 'mingw'.
1087 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1088 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1089 which is the minimum version we support.
1092 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1094 *) Encrypt-Then-Mac renegotiation crash
1096 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1097 negotiated where it was not in the original handshake (or vice-versa) then
1098 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1099 and servers are affected.
1101 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1105 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1107 *) Truncated packet could crash via OOB read
1109 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1110 cipher is being used, then a truncated packet can cause that host to
1111 perform an out-of-bounds read, usually resulting in a crash.
1113 This issue was reported to OpenSSL by Robert Święcki of Google.
1117 *) Bad (EC)DHE parameters cause a client crash
1119 If a malicious server supplies bad parameters for a DHE or ECDHE key
1120 exchange then this can result in the client attempting to dereference a
1121 NULL pointer leading to a client crash. This could be exploited in a Denial
1124 This issue was reported to OpenSSL by Guido Vranken.
1128 *) BN_mod_exp may produce incorrect results on x86_64
1130 There is a carry propagating bug in the x86_64 Montgomery squaring
1131 procedure. No EC algorithms are affected. Analysis suggests that attacks
1132 against RSA and DSA as a result of this defect would be very difficult to
1133 perform and are not believed likely. Attacks against DH are considered just
1134 feasible (although very difficult) because most of the work necessary to
1135 deduce information about a private key may be performed offline. The amount
1136 of resources required for such an attack would be very significant and
1137 likely only accessible to a limited number of attackers. An attacker would
1138 additionally need online access to an unpatched system using the target
1139 private key in a scenario with persistent DH parameters and a private
1140 key that is shared between multiple clients. For example this can occur by
1141 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1142 similar to CVE-2015-3193 but must be treated as a separate problem.
1144 This issue was reported to OpenSSL by the OSS-Fuzz project.
1148 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1150 *) ChaCha20/Poly1305 heap-buffer-overflow
1152 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1153 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1154 crash. This issue is not considered to be exploitable beyond a DoS.
1156 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1160 *) CMS Null dereference
1162 Applications parsing invalid CMS structures can crash with a NULL pointer
1163 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1164 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1165 structure callback if an attempt is made to free certain invalid encodings.
1166 Only CHOICE structures using a callback which do not handle NULL value are
1169 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1173 *) Montgomery multiplication may produce incorrect results
1175 There is a carry propagating bug in the Broadwell-specific Montgomery
1176 multiplication procedure that handles input lengths divisible by, but
1177 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1178 and DH private keys are impossible. This is because the subroutine in
1179 question is not used in operations with the private key itself and an input
1180 of the attacker's direct choice. Otherwise the bug can manifest itself as
1181 transient authentication and key negotiation failures or reproducible
1182 erroneous outcome of public-key operations with specially crafted input.
1183 Among EC algorithms only Brainpool P-512 curves are affected and one
1184 presumably can attack ECDH key negotiation. Impact was not analyzed in
1185 detail, because pre-requisites for attack are considered unlikely. Namely
1186 multiple clients have to choose the curve in question and the server has to
1187 share the private key among them, neither of which is default behaviour.
1188 Even then only clients that chose the curve will be affected.
1190 This issue was publicly reported as transient failures and was not
1191 initially recognized as a security issue. Thanks to Richard Morgan for
1192 providing reproducible case.
1196 *) Removed automatic addition of RPATH in shared libraries and executables,
1197 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1200 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1202 *) Fix Use After Free for large message sizes
1204 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1205 message larger than approx 16k is received then the underlying buffer to
1206 store the incoming message is reallocated and moved. Unfortunately a
1207 dangling pointer to the old location is left which results in an attempt to
1208 write to the previously freed location. This is likely to result in a
1209 crash, however it could potentially lead to execution of arbitrary code.
1211 This issue only affects OpenSSL 1.1.0a.
1213 This issue was reported to OpenSSL by Robert Święcki.
1217 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1219 *) OCSP Status Request extension unbounded memory growth
1221 A malicious client can send an excessively large OCSP Status Request
1222 extension. If that client continually requests renegotiation, sending a
1223 large OCSP Status Request extension each time, then there will be unbounded
1224 memory growth on the server. This will eventually lead to a Denial Of
1225 Service attack through memory exhaustion. Servers with a default
1226 configuration are vulnerable even if they do not support OCSP. Builds using
1227 the "no-ocsp" build time option are not affected.
1229 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1233 *) SSL_peek() hang on empty record
1235 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1236 sends an empty record. This could be exploited by a malicious peer in a
1237 Denial Of Service attack.
1239 This issue was reported to OpenSSL by Alex Gaynor.
1243 *) Excessive allocation of memory in tls_get_message_header() and
1244 dtls1_preprocess_fragment()
1246 A (D)TLS message includes 3 bytes for its length in the header for the
1247 message. This would allow for messages up to 16Mb in length. Messages of
1248 this length are excessive and OpenSSL includes a check to ensure that a
1249 peer is sending reasonably sized messages in order to avoid too much memory
1250 being consumed to service a connection. A flaw in the logic of version
1251 1.1.0 means that memory for the message is allocated too early, prior to
1252 the excessive message length check. Due to way memory is allocated in
1253 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1254 to service a connection. This could lead to a Denial of Service through
1255 memory exhaustion. However, the excessive message length check still takes
1256 place, and this would cause the connection to immediately fail. Assuming
1257 that the application calls SSL_free() on the failed connection in a timely
1258 manner then the 21Mb of allocated memory will then be immediately freed
1259 again. Therefore the excessive memory allocation will be transitory in
1260 nature. This then means that there is only a security impact if:
1262 1) The application does not call SSL_free() in a timely manner in the event
1263 that the connection fails
1265 2) The application is working in a constrained environment where there is
1266 very little free memory
1268 3) The attacker initiates multiple connection attempts such that there are
1269 multiple connections in a state where memory has been allocated for the
1270 connection; SSL_free() has not yet been called; and there is insufficient
1271 memory to service the multiple requests.
1273 Except in the instance of (1) above any Denial Of Service is likely to be
1274 transitory because as soon as the connection fails the memory is
1275 subsequently freed again in the SSL_free() call. However there is an
1276 increased risk during this period of application crashes due to the lack of
1277 memory - which would then mean a more serious Denial of Service.
1279 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1280 (CVE-2016-6307 and CVE-2016-6308)
1283 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1284 had to be removed. Primary reason is that vendor assembler can't
1285 assemble our modules with -KPIC flag. As result it, assembly
1286 support, was not even available as option. But its lack means
1287 lack of side-channel resistant code, which is incompatible with
1288 security by todays standards. Fortunately gcc is readily available
1289 prepackaged option, which we firmly point at...
1292 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1294 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1295 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1296 (to any value) allows Windows user to access PKCS#12 file generated
1297 with Windows CryptoAPI and protected with non-ASCII password, as well
1298 as files generated under UTF-8 locale on Linux also protected with
1302 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1303 have been disabled by default and removed from DEFAULT, just like RC4.
1304 See the RC4 item below to re-enable both.
1307 *) The method for finding the storage location for the Windows RAND seed file
1308 has changed. First we check %RANDFILE%. If that is not set then we check
1309 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1310 all else fails we fall back to C:\.
1313 *) The EVP_EncryptUpdate() function has had its return type changed from void
1314 to int. A return of 0 indicates and error while a return of 1 indicates
1318 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1319 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1320 off the constant time implementation for RSA, DSA and DH have been made
1321 no-ops and deprecated.
1324 *) Windows RAND implementation was simplified to only get entropy by
1325 calling CryptGenRandom(). Various other RAND-related tickets
1327 [Joseph Wylie Yandle, Rich Salz]
1329 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1330 and OPENSSL_LH_, respectively. The old names are available
1331 with API compatibility. They new names are now completely documented.
1334 *) Unify TYPE_up_ref(obj) methods signature.
1335 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1336 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1337 int (instead of void) like all others TYPE_up_ref() methods.
1338 So now these methods also check the return value of CRYPTO_atomic_add(),
1339 and the validity of object reference counter.
1340 [fdasilvayy@gmail.com]
1342 *) With Windows Visual Studio builds, the .pdb files are installed
1343 alongside the installed libraries and executables. For a static
1344 library installation, ossl_static.pdb is the associate compiler
1345 generated .pdb file to be used when linking programs.
1348 *) Remove openssl.spec. Packaging files belong with the packagers.
1351 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1352 recognise x86_64 architectures automatically. You can still decide
1353 to build for a different bitness with the environment variable
1354 KERNEL_BITS (can be 32 or 64), for example:
1356 KERNEL_BITS=32 ./config
1360 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1361 256 bit AES and HMAC with SHA256.
1364 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1367 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1370 *) To enable users to have their own config files and build file templates,
1371 Configure looks in the directory indicated by the environment variable
1372 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1373 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1374 name and is used as is.
1377 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1378 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1379 X509_CERT_FILE_CTX was removed.
1382 *) "shared" builds are now the default. To create only static libraries use
1383 the "no-shared" Configure option.
1386 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1387 All of these option have not worked for some while and are fundamental
1391 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1392 global cleanup functions are no longer required because they are handled
1393 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1394 Explicitly de-initing can cause problems (e.g. where a library that uses
1395 OpenSSL de-inits, but an application is still using it). The affected
1396 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1397 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1398 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1399 COMP_zlib_cleanup().
1402 *) --strict-warnings no longer enables runtime debugging options
1403 such as REF_DEBUG. Instead, debug options are automatically
1404 enabled with '--debug' builds.
1405 [Andy Polyakov, Emilia Käsper]
1407 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1408 have been moved out of the public header files. New functions for managing
1409 these have been added.
1412 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1413 objects have been moved out of the public header files. New
1414 functions for managing these have been added.
1417 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1418 have been moved out of the public header files. New functions for managing
1419 these have been added.
1422 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1423 moved out of the public header files. New functions for managing these
1427 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1430 *) Removed the mk1mf build scripts.
1433 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1434 it is always safe to #include a header now.
1437 *) Removed the aged BC-32 config and all its supporting scripts
1440 *) Removed support for Ultrix, Netware, and OS/2.
1443 *) Add support for HKDF.
1444 [Alessandro Ghedini]
1446 *) Add support for blake2b and blake2s
1449 *) Added support for "pipelining". Ciphers that have the
1450 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1451 encryptions/decryptions simultaneously. There are currently no built-in
1452 ciphers with this property but the expectation is that engines will be able
1453 to offer it to significantly improve throughput. Support has been extended
1454 into libssl so that multiple records for a single connection can be
1455 processed in one go (for >=TLS 1.1).
1458 *) Added the AFALG engine. This is an async capable engine which is able to
1459 offload work to the Linux kernel. In this initial version it only supports
1460 AES128-CBC. The kernel must be version 4.1.0 or greater.
1463 *) OpenSSL now uses a new threading API. It is no longer necessary to
1464 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1465 are two supported threading models: pthreads and windows threads. It is
1466 also possible to configure OpenSSL at compile time for "no-threads". The
1467 old threading API should no longer be used. The functions have been
1468 replaced with "no-op" compatibility macros.
1469 [Alessandro Ghedini, Matt Caswell]
1471 *) Modify behavior of ALPN to invoke callback after SNI/servername
1472 callback, such that updates to the SSL_CTX affect ALPN.
1475 *) Add SSL_CIPHER queries for authentication and key-exchange.
1478 *) Changes to the DEFAULT cipherlist:
1479 - Prefer (EC)DHE handshakes over plain RSA.
1480 - Prefer AEAD ciphers over legacy ciphers.
1481 - Prefer ECDSA over RSA when both certificates are available.
1482 - Prefer TLSv1.2 ciphers/PRF.
1483 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1487 *) Change the ECC default curve list to be this, in order: x25519,
1488 secp256r1, secp521r1, secp384r1.
1491 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1492 disabled by default. They can be re-enabled using the
1493 enable-weak-ssl-ciphers option to Configure.
1496 *) If the server has ALPN configured, but supports no protocols that the
1497 client advertises, send a fatal "no_application_protocol" alert.
1498 This behaviour is SHALL in RFC 7301, though it isn't universally
1499 implemented by other servers.
1502 *) Add X25519 support.
1503 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1504 for public and private key encoding using the format documented in
1505 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1506 key generation and key derivation.
1508 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1512 *) Deprecate SRP_VBASE_get_by_user.
1513 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1514 In order to fix an unavoidable memory leak (CVE-2016-0798),
1515 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1516 seed, even if the seed is configured.
1518 Users should use SRP_VBASE_get1_by_user instead. Note that in
1519 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1520 also that even though configuring the SRP seed attempts to hide
1521 invalid usernames by continuing the handshake with fake
1522 credentials, this behaviour is not constant time and no strong
1523 guarantees are made that the handshake is indistinguishable from
1524 that of a valid user.
1527 *) Configuration change; it's now possible to build dynamic engines
1528 without having to build shared libraries and vice versa. This
1529 only applies to the engines in engines/, those in crypto/engine/
1530 will always be built into libcrypto (i.e. "static").
1532 Building dynamic engines is enabled by default; to disable, use
1533 the configuration option "disable-dynamic-engine".
1535 The only requirements for building dynamic engines are the
1536 presence of the DSO module and building with position independent
1537 code, so they will also automatically be disabled if configuring
1538 with "disable-dso" or "disable-pic".
1540 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1541 are also taken away from openssl/opensslconf.h, as they are
1545 *) Configuration change; if there is a known flag to compile
1546 position independent code, it will always be applied on the
1547 libcrypto and libssl object files, and never on the application
1548 object files. This means other libraries that use routines from
1549 libcrypto / libssl can be made into shared libraries regardless
1550 of how OpenSSL was configured.
1552 If this isn't desirable, the configuration options "disable-pic"
1553 or "no-pic" can be used to disable the use of PIC. This will
1554 also disable building shared libraries and dynamic engines.
1557 *) Removed JPAKE code. It was experimental and has no wide use.
1560 *) The INSTALL_PREFIX Makefile variable has been renamed to
1561 DESTDIR. That makes for less confusion on what this variable
1562 is for. Also, the configuration option --install_prefix is
1566 *) Heartbeat for TLS has been removed and is disabled by default
1567 for DTLS; configure with enable-heartbeats. Code that uses the
1568 old #define's might need to be updated.
1569 [Emilia Käsper, Rich Salz]
1571 *) Rename REF_CHECK to REF_DEBUG.
1574 *) New "unified" build system
1576 The "unified" build system is aimed to be a common system for all
1577 platforms we support. With it comes new support for VMS.
1579 This system builds supports building in a different directory tree
1580 than the source tree. It produces one Makefile (for unix family
1581 or lookalikes), or one descrip.mms (for VMS).
1583 The source of information to make the Makefile / descrip.mms is
1584 small files called 'build.info', holding the necessary
1585 information for each directory with source to compile, and a
1586 template in Configurations, like unix-Makefile.tmpl or
1589 With this change, the library names were also renamed on Windows
1590 and on VMS. They now have names that are closer to the standard
1591 on Unix, and include the major version number, and in certain
1592 cases, the architecture they are built for. See "Notes on shared
1593 libraries" in INSTALL.
1595 We rely heavily on the perl module Text::Template.
1598 *) Added support for auto-initialisation and de-initialisation of the library.
1599 OpenSSL no longer requires explicit init or deinit routines to be called,
1600 except in certain circumstances. See the OPENSSL_init_crypto() and
1601 OPENSSL_init_ssl() man pages for further information.
1604 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1605 "peer" argument is now expected to be a BIO_ADDR object.
1607 *) Rewrite of BIO networking library. The BIO library lacked consistent
1608 support of IPv6, and adding it required some more extensive
1609 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1610 which hold all types of addresses and chains of address information.
1611 It also introduces a new API, with functions like BIO_socket,
1612 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1613 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1614 have been adapted accordingly.
1617 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1621 *) CRIME protection: disable compression by default, even if OpenSSL is
1622 compiled with zlib enabled. Applications can still enable compression
1623 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1624 using the SSL_CONF library to configure compression.
1627 *) The signature of the session callback configured with
1628 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1629 was explicitly marked as 'const unsigned char*' instead of
1633 *) Always DPURIFY. Remove the use of uninitialized memory in the
1634 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1637 *) Removed many obsolete configuration items, including
1638 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1639 MD2_CHAR, MD2_INT, MD2_LONG
1641 IDEA_SHORT, IDEA_LONG
1642 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1643 [Rich Salz, with advice from Andy Polyakov]
1645 *) Many BN internals have been moved to an internal header file.
1646 [Rich Salz with help from Andy Polyakov]
1648 *) Configuration and writing out the results from it has changed.
1649 Files such as Makefile include/openssl/opensslconf.h and are now
1650 produced through general templates, such as Makefile.in and
1651 crypto/opensslconf.h.in and some help from the perl module
1654 Also, the center of configuration information is no longer
1655 Makefile. Instead, Configure produces a perl module in
1656 configdata.pm which holds most of the config data (in the hash
1657 table %config), the target data that comes from the target
1658 configuration in one of the Configurations/*.conf files (in
1662 *) To clarify their intended purposes, the Configure options
1663 --prefix and --openssldir change their semantics, and become more
1664 straightforward and less interdependent.
1666 --prefix shall be used exclusively to give the location INSTALLTOP
1667 where programs, scripts, libraries, include files and manuals are
1668 going to be installed. The default is now /usr/local.
1670 --openssldir shall be used exclusively to give the default
1671 location OPENSSLDIR where certificates, private keys, CRLs are
1672 managed. This is also where the default openssl.cnf gets
1674 If the directory given with this option is a relative path, the
1675 values of both the --prefix value and the --openssldir value will
1676 be combined to become OPENSSLDIR.
1677 The default for --openssldir is INSTALLTOP/ssl.
1679 Anyone who uses --openssldir to specify where OpenSSL is to be
1680 installed MUST change to use --prefix instead.
1683 *) The GOST engine was out of date and therefore it has been removed. An up
1684 to date GOST engine is now being maintained in an external repository.
1685 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1686 support for GOST ciphersuites (these are only activated if a GOST engine
1690 *) EGD is no longer supported by default; use enable-egd when
1692 [Ben Kaduk and Rich Salz]
1694 *) The distribution now has Makefile.in files, which are used to
1695 create Makefile's when Configure is run. *Configure must be run
1696 before trying to build now.*
1699 *) The return value for SSL_CIPHER_description() for error conditions
1703 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1705 Obtaining and performing DNSSEC validation of TLSA records is
1706 the application's responsibility. The application provides
1707 the TLSA records of its choice to OpenSSL, and these are then
1708 used to authenticate the peer.
1710 The TLSA records need not even come from DNS. They can, for
1711 example, be used to implement local end-entity certificate or
1712 trust-anchor "pinning", where the "pin" data takes the form
1713 of TLSA records, which can augment or replace verification
1714 based on the usual WebPKI public certification authorities.
1717 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1718 continues to support deprecated interfaces in default builds.
1719 However, applications are strongly advised to compile their
1720 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1721 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1722 or the 1.1.0 releases.
1724 In environments in which all applications have been ported to
1725 not use any deprecated interfaces OpenSSL's Configure script
1726 should be used with the --api=1.1.0 option to entirely remove
1727 support for the deprecated features from the library and
1728 unconditionally disable them in the installed headers.
1729 Essentially the same effect can be achieved with the "no-deprecated"
1730 argument to Configure, except that this will always restrict
1731 the build to just the latest API, rather than a fixed API
1734 As applications are ported to future revisions of the API,
1735 they should update their compile-time OPENSSL_API_COMPAT define
1736 accordingly, but in most cases should be able to continue to
1737 compile with later releases.
1739 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1740 0x10000000L and 0x00908000L, respectively. However those
1741 versions did not support the OPENSSL_API_COMPAT feature, and
1742 so applications are not typically tested for explicit support
1743 of just the undeprecated features of either release.
1746 *) Add support for setting the minimum and maximum supported protocol.
1747 It can bet set via the SSL_set_min_proto_version() and
1748 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1749 MaxProtocol. It's recommended to use the new APIs to disable
1750 protocols instead of disabling individual protocols using
1751 SSL_set_options() or SSL_CONF's Protocol. This change also
1752 removes support for disabling TLS 1.2 in the OpenSSL TLS
1753 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1756 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1759 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1760 and integrates ECDSA and ECDH functionality into EC. Implementations can
1761 now redirect key generation and no longer need to convert to or from
1764 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1765 include the ec.h header file instead.
1768 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1769 ciphers who are no longer supported and drops support the ephemeral RSA key
1770 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1773 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1774 opaque. For HMAC_CTX, the following constructors and destructors
1777 HMAC_CTX *HMAC_CTX_new(void);
1778 void HMAC_CTX_free(HMAC_CTX *ctx);
1780 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1781 destroy such methods has been added. See EVP_MD_meth_new(3) and
1782 EVP_CIPHER_meth_new(3) for documentation.
1785 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1786 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1787 EVP_MD_CTX_reset() should be called instead to reinitialise
1788 an already created structure.
1789 2) For consistency with the majority of our object creators and
1790 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1791 EVP_MD_CTX_(new|free). The old names are retained as macros
1792 for deprecated builds.
1795 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1796 cryptographic operations to be performed asynchronously as long as an
1797 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1798 further details. Libssl has also had this capability integrated with the
1799 introduction of the new mode SSL_MODE_ASYNC and associated error
1800 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1801 pages. This work was developed in partnership with Intel Corp.
1804 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1805 always enabled now. If you want to disable the support you should
1806 exclude it using the list of supported ciphers. This also means that the
1807 "-no_ecdhe" option has been removed from s_server.
1810 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1811 SSL_{CTX_}set1_curves() which can set a list.
1814 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1815 curve you want to support using SSL_{CTX_}set1_curves().
1818 *) State machine rewrite. The state machine code has been significantly
1819 refactored in order to remove much duplication of code and solve issues
1820 with the old code (see ssl/statem/README for further details). This change
1821 does have some associated API changes. Notably the SSL_state() function
1822 has been removed and replaced by SSL_get_state which now returns an
1823 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1824 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1828 *) All instances of the string "ssleay" in the public API were replaced
1829 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1830 Some error codes related to internal RSA_eay API's were renamed.
1833 *) The demo files in crypto/threads were moved to demo/threads.
1836 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1838 [Matt Caswell, Rich Salz]
1840 *) New ASN.1 embed macro.
1842 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1843 structure is not allocated: it is part of the parent. That is instead of
1851 This reduces memory fragmentation and make it impossible to accidentally
1852 set a mandatory field to NULL.
1854 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1855 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1856 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1860 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1863 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1864 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1865 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1866 DES and RC4 ciphersuites.
1869 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1870 This changes the decoding behaviour for some invalid messages,
1871 though the change is mostly in the more lenient direction, and
1872 legacy behaviour is preserved as much as possible.
1875 *) Fix no-stdio build.
1876 [ David Woodhouse <David.Woodhouse@intel.com> and also
1877 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1879 *) New testing framework
1880 The testing framework has been largely rewritten and is now using
1881 perl and the perl modules Test::Harness and an extended variant of
1882 Test::More called OpenSSL::Test to do its work. All test scripts in
1883 test/ have been rewritten into test recipes, and all direct calls to
1884 executables in test/Makefile have become individual recipes using the
1885 simplified testing OpenSSL::Test::Simple.
1887 For documentation on our testing modules, do:
1889 perldoc test/testlib/OpenSSL/Test/Simple.pm
1890 perldoc test/testlib/OpenSSL/Test.pm
1894 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1895 are used; the latter aborts on memory leaks (usually checked on exit).
1896 Some undocumented "set malloc, etc., hooks" functions were removed
1897 and others were changed. All are now documented.
1900 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1902 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1904 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1905 from RFC4279, RFC4785, RFC5487, RFC5489.
1907 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1908 original RSA_PSK patch.
1911 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1912 era flag was never set throughout the codebase (only read). Also removed
1913 SSL3_FLAGS_POP_BUFFER which was only used if
1914 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1917 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1918 to be "oneline" instead of "compat".
1921 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1922 not aware of clients that still exhibit this bug, and the workaround
1923 hasn't been working properly for a while.
1926 *) The return type of BIO_number_read() and BIO_number_written() as well as
1927 the corresponding num_read and num_write members in the BIO structure has
1928 changed from unsigned long to uint64_t. On platforms where an unsigned
1929 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1933 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1934 OpenSSL without support for them. It also means that maintaining
1935 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1936 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1939 *) Removed support for the two export grade static DH ciphersuites
1940 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1941 were newly added (along with a number of other static DH ciphersuites) to
1942 1.0.2. However the two export ones have *never* worked since they were
1943 introduced. It seems strange in any case to be adding new export
1944 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1947 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1948 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1949 and turned into macros which simply call the new preferred function names
1950 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1951 should use the new names instead. Also as part of this change the ssl23.h
1952 header file has been removed.
1955 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1956 code and the associated standard is no longer considered fit-for-purpose.
1959 *) RT2547 was closed. When generating a private key, try to make the
1960 output file readable only by the owner. This behavior change might
1961 be noticeable when interacting with other software.
1963 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1967 *) Added HTTP GET support to the ocsp command.
1970 *) Changed default digest for the dgst and enc commands from MD5 to
1974 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1977 *) Added support for TLS extended master secret from
1978 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1979 initial patch which was a great help during development.
1982 *) All libssl internal structures have been removed from the public header
1983 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1984 now redundant). Users should not attempt to access internal structures
1985 directly. Instead they should use the provided API functions.
1988 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1989 Access to deprecated functions can be re-enabled by running config with
1990 "enable-deprecated". In addition applications wishing to use deprecated
1991 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1992 will, by default, disable some transitive includes that previously existed
1993 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1996 *) Added support for OCB mode. OpenSSL has been granted a patent license
1997 compatible with the OpenSSL license for use of OCB. Details are available
1998 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1999 for OCB can be removed by calling config with no-ocb.
2002 *) SSLv2 support has been removed. It still supports receiving a SSLv2
2003 compatible client hello.
2006 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
2007 done while fixing the error code for the key-too-small case.
2008 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
2010 *) CA.sh has been removed; use CA.pl instead.
2013 *) Removed old DES API.
2016 *) Remove various unsupported platforms:
2022 Sinix/ReliantUNIX RM400
2027 16-bit platforms such as WIN16
2030 *) Clean up OPENSSL_NO_xxx #define's
2031 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
2032 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
2033 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
2034 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
2035 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
2036 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
2037 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
2038 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
2039 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
2040 Remove MS_STATIC; it's a relic from platforms <32 bits.
2043 *) Cleaned up dead code
2044 Remove all but one '#ifdef undef' which is to be looked at.
2047 *) Clean up calling of xxx_free routines.
2048 Just like free(), fix most of the xxx_free routines to accept
2049 NULL. Remove the non-null checks from callers. Save much code.
2052 *) Add secure heap for storage of private keys (when possible).
2053 Add BIO_s_secmem(), CBIGNUM, etc.
2054 Contributed by Akamai Technologies under our Corporate CLA.
2057 *) Experimental support for a new, fast, unbiased prime candidate generator,
2058 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2059 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
2061 *) New output format NSS in the sess_id command line tool. This allows
2062 exporting the session id and the master key in NSS keylog format.
2063 [Martin Kaiser <martin@kaiser.cx>]
2065 *) Harmonize version and its documentation. -f flag is used to display
2067 [mancha <mancha1@zoho.com>]
2069 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2070 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2071 [mancha <mancha1@zoho.com>]
2073 *) Fix some double frees. These are not thought to be exploitable.
2074 [mancha <mancha1@zoho.com>]
2076 *) A missing bounds check in the handling of the TLS heartbeat extension
2077 can be used to reveal up to 64k of memory to a connected client or
2080 Thanks for Neel Mehta of Google Security for discovering this bug and to
2081 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2082 preparing the fix (CVE-2014-0160)
2083 [Adam Langley, Bodo Moeller]
2085 *) Fix for the attack described in the paper "Recovering OpenSSL
2086 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2087 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2088 http://eprint.iacr.org/2014/140
2090 Thanks to Yuval Yarom and Naomi Benger for discovering this
2091 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2092 [Yuval Yarom and Naomi Benger]
2094 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2095 this fixes a limitation in previous versions of OpenSSL.
2098 *) Experimental encrypt-then-mac support.
2100 Experimental support for encrypt then mac from
2101 draft-gutmann-tls-encrypt-then-mac-02.txt
2103 To enable it set the appropriate extension number (0x42 for the test
2104 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2106 For non-compliant peers (i.e. just about everything) this should have no
2109 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2113 *) Add EVP support for key wrapping algorithms, to avoid problems with
2114 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2115 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2116 algorithms and include tests cases.
2119 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2123 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2124 MGF1 digest and OAEP label.
2127 *) Make openssl verify return errors.
2128 [Chris Palmer <palmer@google.com> and Ben Laurie]
2130 *) New function ASN1_TIME_diff to calculate the difference between two
2131 ASN1_TIME structures or one structure and the current time.
2134 *) Update fips_test_suite to support multiple command line options. New
2135 test to induce all self test errors in sequence and check expected
2139 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2140 sign or verify all in one operation.
2143 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2144 test programs and fips_test_suite. Includes functionality to parse
2145 the minimal script output of fipsalgest.pl directly.
2148 *) Add authorisation parameter to FIPS_module_mode_set().
2151 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2154 *) Use separate DRBG fields for internal and external flags. New function
2155 FIPS_drbg_health_check() to perform on demand health checking. Add
2156 generation tests to fips_test_suite with reduced health check interval to
2157 demonstrate periodic health checking. Add "nodh" option to
2158 fips_test_suite to skip very slow DH test.
2161 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2165 *) More extensive health check for DRBG checking many more failure modes.
2166 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2167 combination: call this in fips_test_suite.
2170 *) Add support for canonical generation of DSA parameter 'g'. See
2173 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2174 POST to handle HMAC cases.
2177 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2178 to return numerical and string versions of the FIPS module number.
2181 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2182 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2183 outside the validated module in the FIPS capable OpenSSL.
2186 *) Minor change to DRBG entropy callback semantics. In some cases
2187 there is no multiple of the block length between min_len and
2188 max_len. Allow the callback to return more than max_len bytes
2189 of entropy but discard any extra: it is the callback's responsibility
2190 to ensure that the extra data discarded does not impact the
2191 requested amount of entropy.
2194 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2195 information in FIPS186-3, SP800-57 and SP800-131A.
2198 *) CCM support via EVP. Interface is very similar to GCM case except we
2199 must supply all data in one chunk (i.e. no update, final) and the
2200 message length must be supplied if AAD is used. Add algorithm test
2204 *) Initial version of POST overhaul. Add POST callback to allow the status
2205 of POST to be monitored and/or failures induced. Modify fips_test_suite
2206 to use callback. Always run all selftests even if one fails.
2209 *) XTS support including algorithm test driver in the fips_gcmtest program.
2210 Note: this does increase the maximum key length from 32 to 64 bytes but
2211 there should be no binary compatibility issues as existing applications
2212 will never use XTS mode.
2215 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2216 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2217 performs algorithm blocking for unapproved PRNG types. Also do not
2218 set PRNG type in FIPS_mode_set(): leave this to the application.
2219 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2220 the standard OpenSSL PRNG: set additional data to a date time vector.
2223 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2224 This shouldn't present any incompatibility problems because applications
2225 shouldn't be using these directly and any that are will need to rethink
2226 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2229 *) Extensive self tests and health checking required by SP800-90 DRBG.
2230 Remove strength parameter from FIPS_drbg_instantiate and always
2231 instantiate at maximum supported strength.
2234 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2237 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2240 *) New function DH_compute_key_padded() to compute a DH key and pad with
2241 leading zeroes if needed: this complies with SP800-56A et al.
2244 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2245 anything, incomplete, subject to change and largely untested at present.
2248 *) Modify fipscanisteronly build option to only build the necessary object
2249 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2252 *) Add experimental option FIPSSYMS to give all symbols in
2253 fipscanister.o and FIPS or fips prefix. This will avoid
2254 conflicts with future versions of OpenSSL. Add perl script
2255 util/fipsas.pl to preprocess assembly language source files
2256 and rename any affected symbols.
2259 *) Add selftest checks and algorithm block of non-fips algorithms in
2260 FIPS mode. Remove DES2 from selftests.
2263 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2264 return internal method without any ENGINE dependencies. Add new
2265 tiny fips sign and verify functions.
2268 *) New build option no-ec2m to disable characteristic 2 code.
2271 *) New build option "fipscanisteronly". This only builds fipscanister.o
2272 and (currently) associated fips utilities. Uses the file Makefile.fips
2273 instead of Makefile.org as the prototype.
2276 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2277 Update fips_gcmtest to use IV generator.
2280 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2281 setting output buffer to NULL. The *Final function must be
2282 called although it will not retrieve any additional data. The tag
2283 can be set or retrieved with a ctrl. The IV length is by default 12
2284 bytes (96 bits) but can be set to an alternative value. If the IV
2285 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2289 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2290 underlying do_cipher function handles all cipher semantics itself
2291 including padding and finalisation. This is useful if (for example)
2292 an ENGINE cipher handles block padding itself. The behaviour of
2293 do_cipher is subtly changed if this flag is set: the return value
2294 is the number of characters written to the output buffer (zero is
2295 no longer an error code) or a negative error code. Also if the
2296 input buffer is NULL and length 0 finalisation should be performed.
2299 *) If a candidate issuer certificate is already part of the constructed
2300 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2303 *) Improve forward-security support: add functions
2305 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2306 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2308 for use by SSL/TLS servers; the callback function will be called whenever a
2309 new session is created, and gets to decide whether the session may be
2310 cached to make it resumable (return 0) or not (return 1). (As by the
2311 SSL/TLS protocol specifications, the session_id sent by the server will be
2312 empty to indicate that the session is not resumable; also, the server will
2313 not generate RFC 4507 (RFC 5077) session tickets.)
2315 A simple reasonable callback implementation is to return is_forward_secure.
2316 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2317 by the SSL/TLS server library, indicating whether it can provide forward
2319 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2321 *) New -verify_name option in command line utilities to set verification
2325 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2326 Add CMAC pkey methods.
2329 *) Experimental renegotiation in s_server -www mode. If the client
2330 browses /reneg connection is renegotiated. If /renegcert it is
2331 renegotiated requesting a certificate.
2334 *) Add an "external" session cache for debugging purposes to s_server. This
2335 should help trace issues which normally are only apparent in deployed
2336 multi-process servers.
2339 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2340 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2341 BIO_set_cipher() and some obscure PEM functions were changed so they
2342 can now return an error. The RAND changes required a change to the
2343 RAND_METHOD structure.
2346 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2347 a gcc attribute to warn if the result of a function is ignored. This
2348 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2349 whose return value is often ignored.
2352 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2353 These allow SCTs (signed certificate timestamps) to be requested and
2354 validated when establishing a connection.
2355 [Rob Percival <robpercival@google.com>]
2357 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2359 *) Prevent padding oracle in AES-NI CBC MAC check
2361 A MITM attacker can use a padding oracle attack to decrypt traffic
2362 when the connection uses an AES CBC cipher and the server support
2365 This issue was introduced as part of the fix for Lucky 13 padding
2366 attack (CVE-2013-0169). The padding check was rewritten to be in
2367 constant time by making sure that always the same bytes are read and
2368 compared against either the MAC or padding bytes. But it no longer
2369 checked that there was enough data to have both the MAC and padding
2372 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2376 *) Fix EVP_EncodeUpdate overflow
2378 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2379 Base64 encoding of binary data. If an attacker is able to supply very large
2380 amounts of input data then a length check can overflow resulting in a heap
2383 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2384 the PEM_write_bio* family of functions. These are mainly used within the
2385 OpenSSL command line applications, so any application which processes data
2386 from an untrusted source and outputs it as a PEM file should be considered
2387 vulnerable to this issue. User applications that call these APIs directly
2388 with large amounts of untrusted data may also be vulnerable.
2390 This issue was reported by Guido Vranken.
2394 *) Fix EVP_EncryptUpdate overflow
2396 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2397 is able to supply very large amounts of input data after a previous call to
2398 EVP_EncryptUpdate() with a partial block then a length check can overflow
2399 resulting in a heap corruption. Following an analysis of all OpenSSL
2400 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2401 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2402 the first called function after an EVP_EncryptInit(), and therefore that
2403 specific call must be safe. The second form is where the length passed to
2404 EVP_EncryptUpdate() can be seen from the code to be some small value and
2405 therefore there is no possibility of an overflow. Since all instances are
2406 one of these two forms, it is believed that there can be no overflows in
2407 internal code due to this problem. It should be noted that
2408 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2409 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2410 of these calls have also been analysed too and it is believed there are no
2411 instances in internal usage where an overflow could occur.
2413 This issue was reported by Guido Vranken.
2417 *) Prevent ASN.1 BIO excessive memory allocation
2419 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2420 a short invalid encoding can cause allocation of large amounts of memory
2421 potentially consuming excessive resources or exhausting memory.
2423 Any application parsing untrusted data through d2i BIO functions is
2424 affected. The memory based functions such as d2i_X509() are *not* affected.
2425 Since the memory based functions are used by the TLS library, TLS
2426 applications are not affected.
2428 This issue was reported by Brian Carpenter.
2434 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2435 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2436 in arbitrary stack data being returned in the buffer.
2438 This issue was reported by Guido Vranken.
2442 *) Modify behavior of ALPN to invoke callback after SNI/servername
2443 callback, such that updates to the SSL_CTX affect ALPN.
2446 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2450 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2451 methods are enabled and ssl2 is disabled the methods return NULL.
2454 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2456 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2457 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2458 provide any "EXPORT" or "LOW" strength ciphers.
2461 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2462 is by default disabled at build-time. Builds that are not configured with
2463 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2464 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2465 will need to explicitly call either of:
2467 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2469 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2471 as appropriate. Even if either of those is used, or the application
2472 explicitly uses the version-specific SSLv2_method() or its client and
2473 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2474 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2475 ciphers, and SSLv2 56-bit DES are no longer available.
2479 *) Fix a double-free in DSA code
2481 A double free bug was discovered when OpenSSL parses malformed DSA private
2482 keys and could lead to a DoS attack or memory corruption for applications
2483 that receive DSA private keys from untrusted sources. This scenario is
2486 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2491 *) Disable SRP fake user seed to address a server memory leak.
2493 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2495 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2496 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2497 was changed to ignore the "fake user" SRP seed, even if the seed
2500 Users should use SRP_VBASE_get1_by_user instead. Note that in
2501 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2502 also that even though configuring the SRP seed attempts to hide
2503 invalid usernames by continuing the handshake with fake
2504 credentials, this behaviour is not constant time and no strong
2505 guarantees are made that the handshake is indistinguishable from
2506 that of a valid user.
2510 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2512 In the BN_hex2bn function the number of hex digits is calculated using an
2513 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2514 large values of |i| this can result in |bn_expand| not allocating any
2515 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2516 field as NULL leading to a subsequent NULL ptr deref. For very large values
2517 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2518 In this case memory is allocated to the internal BIGNUM data field, but it
2519 is insufficiently sized leading to heap corruption. A similar issue exists
2520 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2521 is ever called by user applications with very large untrusted hex/dec data.
2522 This is anticipated to be a rare occurrence.
2524 All OpenSSL internal usage of these functions use data that is not expected
2525 to be untrusted, e.g. config file data or application command line
2526 arguments. If user developed applications generate config file data based
2527 on untrusted data then it is possible that this could also lead to security
2528 consequences. This is also anticipated to be rare.
2530 This issue was reported to OpenSSL by Guido Vranken.
2534 *) Fix memory issues in BIO_*printf functions
2536 The internal |fmtstr| function used in processing a "%s" format string in
2537 the BIO_*printf functions could overflow while calculating the length of a
2538 string and cause an OOB read when printing very long strings.
2540 Additionally the internal |doapr_outch| function can attempt to write to an
2541 OOB memory location (at an offset from the NULL pointer) in the event of a
2542 memory allocation failure. In 1.0.2 and below this could be caused where
2543 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2544 could be in processing a very long "%s" format string. Memory leaks can
2547 The first issue may mask the second issue dependent on compiler behaviour.
2548 These problems could enable attacks where large amounts of untrusted data
2549 is passed to the BIO_*printf functions. If applications use these functions
2550 in this way then they could be vulnerable. OpenSSL itself uses these
2551 functions when printing out human-readable dumps of ASN.1 data. Therefore
2552 applications that print this data could be vulnerable if the data is from
2553 untrusted sources. OpenSSL command line applications could also be
2554 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2555 as command line arguments.
2557 Libssl is not considered directly vulnerable. Additionally certificates etc
2558 received via remote connections via libssl are also unlikely to be able to
2559 trigger these issues because of message size limits enforced within libssl.
2561 This issue was reported to OpenSSL Guido Vranken.
2565 *) Side channel attack on modular exponentiation
2567 A side-channel attack was found which makes use of cache-bank conflicts on
2568 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2569 of RSA keys. The ability to exploit this issue is limited as it relies on
2570 an attacker who has control of code in a thread running on the same
2571 hyper-threaded core as the victim thread which is performing decryptions.
2573 This issue was reported to OpenSSL by Yuval Yarom, The University of
2574 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2575 Nadia Heninger, University of Pennsylvania with more information at
2576 http://cachebleed.info.
2580 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2581 if no keysize is specified with default_bits. This fixes an
2582 omission in an earlier change that changed all RSA/DSA key generation
2583 apps to use 2048 bits by default.
2586 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2587 *) DH small subgroups
2589 Historically OpenSSL only ever generated DH parameters based on "safe"
2590 primes. More recently (in version 1.0.2) support was provided for
2591 generating X9.42 style parameter files such as those required for RFC 5114
2592 support. The primes used in such files may not be "safe". Where an
2593 application is using DH configured with parameters based on primes that are
2594 not "safe" then an attacker could use this fact to find a peer's private
2595 DH exponent. This attack requires that the attacker complete multiple
2596 handshakes in which the peer uses the same private DH exponent. For example
2597 this could be used to discover a TLS server's private DH exponent if it's
2598 reusing the private DH exponent or it's using a static DH ciphersuite.
2600 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2601 TLS. It is not on by default. If the option is not set then the server
2602 reuses the same private DH exponent for the life of the server process and
2603 would be vulnerable to this attack. It is believed that many popular
2604 applications do set this option and would therefore not be at risk.
2606 The fix for this issue adds an additional check where a "q" parameter is
2607 available (as is the case in X9.42 based parameters). This detects the
2608 only known attack, and is the only possible defense for static DH
2609 ciphersuites. This could have some performance impact.
2611 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2612 default and cannot be disabled. This could have some performance impact.
2614 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2618 *) SSLv2 doesn't block disabled ciphers
2620 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2621 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2622 been disabled, provided that the SSLv2 protocol was not also disabled via
2625 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2626 and Sebastian Schinzel.
2630 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2632 *) BN_mod_exp may produce incorrect results on x86_64
2634 There is a carry propagating bug in the x86_64 Montgomery squaring
2635 procedure. No EC algorithms are affected. Analysis suggests that attacks
2636 against RSA and DSA as a result of this defect would be very difficult to
2637 perform and are not believed likely. Attacks against DH are considered just
2638 feasible (although very difficult) because most of the work necessary to
2639 deduce information about a private key may be performed offline. The amount
2640 of resources required for such an attack would be very significant and
2641 likely only accessible to a limited number of attackers. An attacker would
2642 additionally need online access to an unpatched system using the target
2643 private key in a scenario with persistent DH parameters and a private
2644 key that is shared between multiple clients. For example this can occur by
2645 default in OpenSSL DHE based SSL/TLS ciphersuites.
2647 This issue was reported to OpenSSL by Hanno Böck.
2651 *) Certificate verify crash with missing PSS parameter
2653 The signature verification routines will crash with a NULL pointer
2654 dereference if presented with an ASN.1 signature using the RSA PSS
2655 algorithm and absent mask generation function parameter. Since these
2656 routines are used to verify certificate signature algorithms this can be
2657 used to crash any certificate verification operation and exploited in a
2658 DoS attack. Any application which performs certificate verification is
2659 vulnerable including OpenSSL clients and servers which enable client
2662 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2666 *) X509_ATTRIBUTE memory leak
2668 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2669 memory. This structure is used by the PKCS#7 and CMS routines so any
2670 application which reads PKCS#7 or CMS data from untrusted sources is
2671 affected. SSL/TLS is not affected.
2673 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2678 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2679 This changes the decoding behaviour for some invalid messages,
2680 though the change is mostly in the more lenient direction, and
2681 legacy behaviour is preserved as much as possible.
2684 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2686 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2688 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2690 *) Alternate chains certificate forgery
2692 During certificate verification, OpenSSL will attempt to find an
2693 alternative certificate chain if the first attempt to build such a chain
2694 fails. An error in the implementation of this logic can mean that an
2695 attacker could cause certain checks on untrusted certificates to be
2696 bypassed, such as the CA flag, enabling them to use a valid leaf
2697 certificate to act as a CA and "issue" an invalid certificate.
2699 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2703 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2705 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2706 incompatibility in the handling of HMAC. The previous ABI has now been
2710 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2712 *) Malformed ECParameters causes infinite loop
2714 When processing an ECParameters structure OpenSSL enters an infinite loop
2715 if the curve specified is over a specially malformed binary polynomial
2718 This can be used to perform denial of service against any
2719 system which processes public keys, certificate requests or
2720 certificates. This includes TLS clients and TLS servers with
2721 client authentication enabled.
2723 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2727 *) Exploitable out-of-bounds read in X509_cmp_time
2729 X509_cmp_time does not properly check the length of the ASN1_TIME
2730 string and can read a few bytes out of bounds. In addition,
2731 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2734 An attacker can use this to craft malformed certificates and CRLs of
2735 various sizes and potentially cause a segmentation fault, resulting in
2736 a DoS on applications that verify certificates or CRLs. TLS clients
2737 that verify CRLs are affected. TLS clients and servers with client
2738 authentication enabled may be affected if they use custom verification
2741 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2742 independently by Hanno Böck.
2746 *) PKCS7 crash with missing EnvelopedContent
2748 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2749 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2750 with missing content and trigger a NULL pointer dereference on parsing.
2752 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2753 structures from untrusted sources are affected. OpenSSL clients and
2754 servers are not affected.
2756 This issue was reported to OpenSSL by Michal Zalewski (Google).
2760 *) CMS verify infinite loop with unknown hash function
2762 When verifying a signedData message the CMS code can enter an infinite loop
2763 if presented with an unknown hash function OID. This can be used to perform
2764 denial of service against any system which verifies signedData messages using
2766 This issue was reported to OpenSSL by Johannes Bauer.
2770 *) Race condition handling NewSessionTicket
2772 If a NewSessionTicket is received by a multi-threaded client when attempting to
2773 reuse a previous ticket then a race condition can occur potentially leading to
2774 a double free of the ticket data.
2778 *) Only support 256-bit or stronger elliptic curves with the
2779 'ecdh_auto' setting (server) or by default (client). Of supported
2780 curves, prefer P-256 (both).
2783 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2785 *) ClientHello sigalgs DoS fix
2787 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2788 invalid signature algorithms extension a NULL pointer dereference will
2789 occur. This can be exploited in a DoS attack against the server.
2791 This issue was was reported to OpenSSL by David Ramos of Stanford
2794 [Stephen Henson and Matt Caswell]
2796 *) Multiblock corrupted pointer fix
2798 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2799 feature only applies on 64 bit x86 architecture platforms that support AES
2800 NI instructions. A defect in the implementation of "multiblock" can cause
2801 OpenSSL's internal write buffer to become incorrectly set to NULL when
2802 using non-blocking IO. Typically, when the user application is using a
2803 socket BIO for writing, this will only result in a failed connection.
2804 However if some other BIO is used then it is likely that a segmentation
2805 fault will be triggered, thus enabling a potential DoS attack.
2807 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2811 *) Segmentation fault in DTLSv1_listen fix
2813 The DTLSv1_listen function is intended to be stateless and processes the
2814 initial ClientHello from many peers. It is common for user code to loop
2815 over the call to DTLSv1_listen until a valid ClientHello is received with
2816 an associated cookie. A defect in the implementation of DTLSv1_listen means
2817 that state is preserved in the SSL object from one invocation to the next
2818 that can lead to a segmentation fault. Errors processing the initial
2819 ClientHello can trigger this scenario. An example of such an error could be
2820 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2823 This issue was reported to OpenSSL by Per Allansson.
2827 *) Segmentation fault in ASN1_TYPE_cmp fix
2829 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2830 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2831 certificate signature algorithm consistency this can be used to crash any
2832 certificate verification operation and exploited in a DoS attack. Any
2833 application which performs certificate verification is vulnerable including
2834 OpenSSL clients and servers which enable client authentication.
2838 *) Segmentation fault for invalid PSS parameters fix
2840 The signature verification routines will crash with a NULL pointer
2841 dereference if presented with an ASN.1 signature using the RSA PSS
2842 algorithm and invalid parameters. Since these routines are used to verify
2843 certificate signature algorithms this can be used to crash any
2844 certificate verification operation and exploited in a DoS attack. Any
2845 application which performs certificate verification is vulnerable including
2846 OpenSSL clients and servers which enable client authentication.
2848 This issue was was reported to OpenSSL by Brian Carpenter.
2852 *) ASN.1 structure reuse memory corruption fix
2854 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2855 memory corruption via an invalid write. Such reuse is and has been
2856 strongly discouraged and is believed to be rare.
2858 Applications that parse structures containing CHOICE or ANY DEFINED BY
2859 components may be affected. Certificate parsing (d2i_X509 and related
2860 functions) are however not affected. OpenSSL clients and servers are
2865 *) PKCS7 NULL pointer dereferences fix
2867 The PKCS#7 parsing code does not handle missing outer ContentInfo
2868 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2869 missing content and trigger a NULL pointer dereference on parsing.
2871 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2872 otherwise parse PKCS#7 structures from untrusted sources are
2873 affected. OpenSSL clients and servers are not affected.
2875 This issue was reported to OpenSSL by Michal Zalewski (Google).
2879 *) DoS via reachable assert in SSLv2 servers fix
2881 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2882 servers that both support SSLv2 and enable export cipher suites by sending
2883 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2885 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2886 (OpenSSL development team).
2890 *) Empty CKE with client auth and DHE fix
2892 If client auth is used then a server can seg fault in the event of a DHE
2893 ciphersuite being selected and a zero length ClientKeyExchange message
2894 being sent by the client. This could be exploited in a DoS attack.
2898 *) Handshake with unseeded PRNG fix
2900 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2901 with an unseeded PRNG. The conditions are:
2902 - The client is on a platform where the PRNG has not been seeded
2903 automatically, and the user has not seeded manually
2904 - A protocol specific client method version has been used (i.e. not
2905 SSL_client_methodv23)
2906 - A ciphersuite is used that does not require additional random data from
2907 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2909 If the handshake succeeds then the client random that has been used will
2910 have been generated from a PRNG with insufficient entropy and therefore the
2911 output may be predictable.
2913 For example using the following command with an unseeded openssl will
2914 succeed on an unpatched platform:
2916 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2920 *) Use After Free following d2i_ECPrivatekey error fix
2922 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2923 could cause a use after free condition. This, in turn, could cause a double
2924 free in several private key parsing functions (such as d2i_PrivateKey
2925 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2926 for applications that receive EC private keys from untrusted
2927 sources. This scenario is considered rare.
2929 This issue was discovered by the BoringSSL project and fixed in their
2934 *) X509_to_X509_REQ NULL pointer deref fix
2936 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2937 the certificate key is invalid. This function is rarely used in practice.
2939 This issue was discovered by Brian Carpenter.
2943 *) Removed the export ciphers from the DEFAULT ciphers
2946 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2948 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2949 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2950 So far those who have to target multiple platforms would compromise
2951 and argue that binary targeting say ARMv5 would still execute on
2952 ARMv8. "Universal" build resolves this compromise by providing
2953 near-optimal performance even on newer platforms.
2956 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2957 (other platforms pending).
2958 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2960 *) Add support for the SignedCertificateTimestampList certificate and
2961 OCSP response extensions from RFC6962.
2964 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2965 for corner cases. (Certain input points at infinity could lead to
2966 bogus results, with non-infinity inputs mapped to infinity too.)
2969 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2970 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2971 common cases are optimized and there still is room for further
2972 improvements. Vector Permutation AES for Altivec is also added.
2975 *) Add support for little-endian ppc64 Linux target.
2976 [Marcelo Cerri (IBM)]
2978 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2979 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2980 are optimized and there still is room for further improvements.
2981 Both 32- and 64-bit modes are supported.
2982 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2984 *) Improved ARMv7 NEON support.
2987 *) Support for SPARC Architecture 2011 crypto extensions, first
2988 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2989 SHA256/512, MD5, GHASH and modular exponentiation.
2990 [Andy Polyakov, David Miller]
2992 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2994 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2996 *) Support for new and upcoming Intel processors, including AVX2,
2997 BMI and SHA ISA extensions. This includes additional "stitched"
2998 implementations, AESNI-SHA256 and GCM, and multi-buffer support
3001 This work was sponsored by Intel Corp.
3004 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
3005 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
3006 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
3009 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3010 this fixes a limitation in previous versions of OpenSSL.
3013 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3014 MGF1 digest and OAEP label.
3017 *) Add EVP support for key wrapping algorithms, to avoid problems with
3018 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3019 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3020 algorithms and include tests cases.
3023 *) Add functions to allocate and set the fields of an ECDSA_METHOD
3025 [Douglas E. Engert, Steve Henson]
3027 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
3028 difference in days and seconds between two tm or ASN1_TIME structures.
3031 *) Add -rev test option to s_server to just reverse order of characters
3032 received by client and send back to server. Also prints an abbreviated
3033 summary of the connection parameters.
3036 *) New option -brief for s_client and s_server to print out a brief summary
3037 of connection parameters.
3040 *) Add callbacks for arbitrary TLS extensions.
3041 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
3043 *) New option -crl_download in several openssl utilities to download CRLs
3044 from CRLDP extension in certificates.
3047 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
3050 *) New function X509_CRL_diff to generate a delta CRL from the difference
3051 of two full CRLs. Add support to "crl" utility.
3054 *) New functions to set lookup_crls function and to retrieve
3055 X509_STORE from X509_STORE_CTX.
3058 *) Print out deprecated issuer and subject unique ID fields in
3062 *) Extend OCSP I/O functions so they can be used for simple general purpose
3063 HTTP as well as OCSP. New wrapper function which can be used to download
3064 CRLs using the OCSP API.
3067 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3070 *) SSL_CONF* functions. These provide a common framework for application
3071 configuration using configuration files or command lines.
3074 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
3075 message callback and prints the results. Needs compile time option
3076 "enable-ssl-trace". New options to s_client and s_server to enable
3080 *) New ctrl and macro to retrieve supported points extensions.
3081 Print out extension in s_server and s_client.
3084 *) New functions to retrieve certificate signature and signature
3088 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3092 *) New Suite B modes for TLS code. These use and enforce the requirements
3093 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3094 only use Suite B curves. The Suite B modes can be set by using the
3095 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3098 *) New chain verification flags for Suite B levels of security. Check
3099 algorithms are acceptable when flags are set in X509_verify_cert.
3102 *) Make tls1_check_chain return a set of flags indicating checks passed
3103 by a certificate chain. Add additional tests to handle client
3104 certificates: checks for matching certificate type and issuer name
3108 *) If an attempt is made to use a signature algorithm not in the peer
3109 preference list abort the handshake. If client has no suitable
3110 signature algorithms in response to a certificate request do not
3111 use the certificate.
3114 *) If server EC tmp key is not in client preference list abort handshake.
3117 *) Add support for certificate stores in CERT structure. This makes it
3118 possible to have different stores per SSL structure or one store in
3119 the parent SSL_CTX. Include distinct stores for certificate chain
3120 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3121 to build and store a certificate chain in CERT structure: returning
3122 an error if the chain cannot be built: this will allow applications
3123 to test if a chain is correctly configured.
3125 Note: if the CERT based stores are not set then the parent SSL_CTX
3126 store is used to retain compatibility with existing behaviour.
3130 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3131 mask based on the current session, check mask when sending client
3132 hello and checking the requested ciphersuite.
3135 *) New ctrls to retrieve and set certificate types in a certificate
3136 request message. Print out received values in s_client. If certificate
3137 types is not set with custom values set sensible values based on
3138 supported signature algorithms.
3141 *) Support for distinct client and server supported signature algorithms.
3144 *) Add certificate callback. If set this is called whenever a certificate
3145 is required by client or server. An application can decide which
3146 certificate chain to present based on arbitrary criteria: for example
3147 supported signature algorithms. Add very simple example to s_server.
3148 This fixes many of the problems and restrictions of the existing client
3149 certificate callback: for example you can now clear an existing
3150 certificate and specify the whole chain.
3153 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3154 the certificate can be used for (if anything). Set valid_flags field
3155 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3156 to have similar checks in it.
3158 Add new "cert_flags" field to CERT structure and include a "strict mode".
3159 This enforces some TLS certificate requirements (such as only permitting
3160 certificate signature algorithms contained in the supported algorithms
3161 extension) which some implementations ignore: this option should be used
3162 with caution as it could cause interoperability issues.
3165 *) Update and tidy signature algorithm extension processing. Work out
3166 shared signature algorithms based on preferences and peer algorithms
3167 and print them out in s_client and s_server. Abort handshake if no
3168 shared signature algorithms.
3171 *) Add new functions to allow customised supported signature algorithms
3172 for SSL and SSL_CTX structures. Add options to s_client and s_server
3176 *) New function SSL_certs_clear() to delete all references to certificates
3177 from an SSL structure. Before this once a certificate had been added
3178 it couldn't be removed.
3181 *) Integrate hostname, email address and IP address checking with certificate
3182 verification. New verify options supporting checking in openssl utility.
3185 *) Fixes and wildcard matching support to hostname and email checking
3186 functions. Add manual page.
3187 [Florian Weimer (Red Hat Product Security Team)]
3189 *) New functions to check a hostname email or IP address against a
3190 certificate. Add options x509 utility to print results of checks against
3194 *) Fix OCSP checking.
3195 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3197 *) Initial experimental support for explicitly trusted non-root CAs.
3198 OpenSSL still tries to build a complete chain to a root but if an
3199 intermediate CA has a trust setting included that is used. The first
3200 setting is used: whether to trust (e.g., -addtrust option to the x509
3204 *) Add -trusted_first option which attempts to find certificates in the
3205 trusted store even if an untrusted chain is also supplied.
3208 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3209 platform support for Linux and Android.
3212 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3215 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3216 When in FIPS mode the approved implementations are used as normal,
3217 when not in FIPS mode the internal unapproved versions are used instead.
3218 This means that the FIPS capable OpenSSL isn't forced to use the
3219 (often lower performance) FIPS implementations outside FIPS mode.
3222 *) Transparently support X9.42 DH parameters when calling
3223 PEM_read_bio_DHparameters. This means existing applications can handle
3224 the new parameter format automatically.
3227 *) Initial experimental support for X9.42 DH parameter format: mainly
3228 to support use of 'q' parameter for RFC5114 parameters.
3231 *) Add DH parameters from RFC5114 including test data to dhtest.
3234 *) Support for automatic EC temporary key parameter selection. If enabled
3235 the most preferred EC parameters are automatically used instead of
3236 hardcoded fixed parameters. Now a server just has to call:
3237 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3238 support ECDH and use the most appropriate parameters.
3241 *) Enhance and tidy EC curve and point format TLS extension code. Use
3242 static structures instead of allocation if default values are used.
3243 New ctrls to set curves we wish to support and to retrieve shared curves.
3244 Print out shared curves in s_server. New options to s_server and s_client
3245 to set list of supported curves.
3248 *) New ctrls to retrieve supported signature algorithms and
3249 supported curve values as an array of NIDs. Extend openssl utility
3250 to print out received values.
3253 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3254 between NIDs and the more common NIST names such as "P-256". Enhance
3255 ecparam utility and ECC method to recognise the NIST names for curves.
3258 *) Enhance SSL/TLS certificate chain handling to support different
3259 chains for each certificate instead of one chain in the parent SSL_CTX.
3262 *) Support for fixed DH ciphersuite client authentication: where both
3263 server and client use DH certificates with common parameters.
3266 *) Support for fixed DH ciphersuites: those requiring DH server
3270 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3272 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3273 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3274 X509_CINF_get_signature were reverted post internal team review.
3276 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3278 *) Build fixes for the Windows and OpenVMS platforms
3279 [Matt Caswell and Richard Levitte]
3281 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3283 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3284 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3285 dereference. This could lead to a Denial Of Service attack. Thanks to
3286 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3290 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3291 dtls1_buffer_record function under certain conditions. In particular this
3292 could occur if an attacker sent repeated DTLS records with the same
3293 sequence number but for the next epoch. The memory leak could be exploited
3294 by an attacker in a Denial of Service attack through memory exhaustion.
3295 Thanks to Chris Mueller for reporting this issue.
3299 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3300 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3301 method would be set to NULL which could later result in a NULL pointer
3302 dereference. Thanks to Frank Schmirler for reporting this issue.
3306 *) Abort handshake if server key exchange message is omitted for ephemeral
3309 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3310 reporting this issue.
3314 *) Remove non-export ephemeral RSA code on client and server. This code
3315 violated the TLS standard by allowing the use of temporary RSA keys in
3316 non-export ciphersuites and could be used by a server to effectively
3317 downgrade the RSA key length used to a value smaller than the server
3318 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3319 INRIA or reporting this issue.
3323 *) Fixed issue where DH client certificates are accepted without verification.
3324 An OpenSSL server will accept a DH certificate for client authentication
3325 without the certificate verify message. This effectively allows a client to
3326 authenticate without the use of a private key. This only affects servers
3327 which trust a client certificate authority which issues certificates
3328 containing DH keys: these are extremely rare and hardly ever encountered.
3329 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3334 *) Ensure that the session ID context of an SSL is updated when its
3335 SSL_CTX is updated via SSL_set_SSL_CTX.
3337 The session ID context is typically set from the parent SSL_CTX,
3338 and can vary with the CTX.
3341 *) Fix various certificate fingerprint issues.
3343 By using non-DER or invalid encodings outside the signed portion of a
3344 certificate the fingerprint can be changed without breaking the signature.
3345 Although no details of the signed portion of the certificate can be changed
3346 this can cause problems with some applications: e.g. those using the
3347 certificate fingerprint for blacklists.
3349 1. Reject signatures with non zero unused bits.
3351 If the BIT STRING containing the signature has non zero unused bits reject
3352 the signature. All current signature algorithms require zero unused bits.
3354 2. Check certificate algorithm consistency.
3356 Check the AlgorithmIdentifier inside TBS matches the one in the
3357 certificate signature. NB: this will result in signature failure
3358 errors for some broken certificates.
3360 Thanks to Konrad Kraszewski from Google for reporting this issue.
3362 3. Check DSA/ECDSA signatures use DER.
3364 Re-encode DSA/ECDSA signatures and compare with the original received
3365 signature. Return an error if there is a mismatch.
3367 This will reject various cases including garbage after signature
3368 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3369 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3370 (negative or with leading zeroes).
3372 Further analysis was conducted and fixes were developed by Stephen Henson
3373 of the OpenSSL core team.
3378 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3379 results on some platforms, including x86_64. This bug occurs at random
3380 with a very low probability, and is not known to be exploitable in any
3381 way, though its exact impact is difficult to determine. Thanks to Pieter
3382 Wuille (Blockstream) who reported this issue and also suggested an initial
3383 fix. Further analysis was conducted by the OpenSSL development team and
3384 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3385 the OpenSSL core team.
3389 *) Do not resume sessions on the server if the negotiated protocol
3390 version does not match the session's version. Resuming with a different
3391 version, while not strictly forbidden by the RFC, is of questionable
3392 sanity and breaks all known clients.
3393 [David Benjamin, Emilia Käsper]
3395 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3396 early CCS messages during renegotiation. (Note that because
3397 renegotiation is encrypted, this early CCS was not exploitable.)
3400 *) Tighten client-side session ticket handling during renegotiation:
3401 ensure that the client only accepts a session ticket if the server sends
3402 the extension anew in the ServerHello. Previously, a TLS client would
3403 reuse the old extension state and thus accept a session ticket if one was
3404 announced in the initial ServerHello.
3406 Similarly, ensure that the client requires a session ticket if one
3407 was advertised in the ServerHello. Previously, a TLS client would
3408 ignore a missing NewSessionTicket message.
3411 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3413 *) SRTP Memory Leak.
3415 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3416 sends a carefully crafted handshake message, to cause OpenSSL to fail
3417 to free up to 64k of memory causing a memory leak. This could be
3418 exploited in a Denial Of Service attack. This issue affects OpenSSL
3419 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3420 whether SRTP is used or configured. Implementations of OpenSSL that
3421 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3423 The fix was developed by the OpenSSL team.
3427 *) Session Ticket Memory Leak.
3429 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3430 integrity of that ticket is first verified. In the event of a session
3431 ticket integrity check failing, OpenSSL will fail to free memory
3432 causing a memory leak. By sending a large number of invalid session
3433 tickets an attacker could exploit this issue in a Denial Of Service
3438 *) Build option no-ssl3 is incomplete.
3440 When OpenSSL is configured with "no-ssl3" as a build option, servers
3441 could accept and complete a SSL 3.0 handshake, and clients could be
3442 configured to send them.
3444 [Akamai and the OpenSSL team]
3446 *) Add support for TLS_FALLBACK_SCSV.
3447 Client applications doing fallback retries should call
3448 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3450 [Adam Langley, Bodo Moeller]
3452 *) Add additional DigestInfo checks.
3454 Re-encode DigestInto in DER and check against the original when
3455 verifying RSA signature: this will reject any improperly encoded
3456 DigestInfo structures.
3458 Note: this is a precautionary measure and no attacks are currently known.
3462 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3464 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3465 SRP code can be overrun an internal buffer. Add sanity check that
3466 g, A, B < N to SRP code.
3468 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3469 Group for discovering this issue.
3473 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3474 TLS 1.0 instead of higher protocol versions when the ClientHello message
3475 is badly fragmented. This allows a man-in-the-middle attacker to force a
3476 downgrade to TLS 1.0 even if both the server and the client support a
3477 higher protocol version, by modifying the client's TLS records.
3479 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3480 researching this issue.
3484 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3485 to a denial of service attack. A malicious server can crash the client
3486 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3487 ciphersuite and sending carefully crafted handshake messages.
3489 Thanks to Felix Gröbert (Google) for discovering and researching this
3494 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3495 to leak memory. This can be exploited through a Denial of Service attack.
3496 Thanks to Adam Langley for discovering and researching this issue.
3500 *) An attacker can force openssl to consume large amounts of memory whilst
3501 processing DTLS handshake messages. This can be exploited through a
3502 Denial of Service attack.
3503 Thanks to Adam Langley for discovering and researching this issue.
3507 *) An attacker can force an error condition which causes openssl to crash
3508 whilst processing DTLS packets due to memory being freed twice. This
3509 can be exploited through a Denial of Service attack.
3510 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3515 *) If a multithreaded client connects to a malicious server using a resumed
3516 session and the server sends an ec point format extension it could write
3517 up to 255 bytes to freed memory.
3519 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3524 *) A malicious server can crash an OpenSSL client with a null pointer
3525 dereference (read) by specifying an SRP ciphersuite even though it was not
3526 properly negotiated with the client. This can be exploited through a
3527 Denial of Service attack.
3529 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3530 discovering and researching this issue.
3534 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3535 X509_name_oneline, X509_name_print_ex et al. to leak some information
3536 from the stack. Applications may be affected if they echo pretty printing
3537 output to the attacker.
3539 Thanks to Ivan Fratric (Google) for discovering this issue.
3541 [Emilia Käsper, and Steve Henson]
3543 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3544 for corner cases. (Certain input points at infinity could lead to
3545 bogus results, with non-infinity inputs mapped to infinity too.)
3548 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3550 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3551 handshake can force the use of weak keying material in OpenSSL
3552 SSL/TLS clients and servers.
3554 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3555 researching this issue. (CVE-2014-0224)
3556 [KIKUCHI Masashi, Steve Henson]
3558 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3559 OpenSSL DTLS client the code can be made to recurse eventually crashing
3562 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3564 [Imre Rad, Steve Henson]
3566 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3567 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3568 client or server. This is potentially exploitable to run arbitrary
3569 code on a vulnerable client or server.
3571 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3572 [Jüri Aedla, Steve Henson]
3574 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3575 are subject to a denial of service attack.
3577 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3578 this issue. (CVE-2014-3470)
3579 [Felix Gröbert, Ivan Fratric, Steve Henson]
3581 *) Harmonize version and its documentation. -f flag is used to display
3583 [mancha <mancha1@zoho.com>]
3585 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3586 in i2d_ECPrivateKey.
3587 [mancha <mancha1@zoho.com>]
3589 *) Fix some double frees. These are not thought to be exploitable.
3590 [mancha <mancha1@zoho.com>]
3592 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3594 *) A missing bounds check in the handling of the TLS heartbeat extension
3595 can be used to reveal up to 64k of memory to a connected client or
3598 Thanks for Neel Mehta of Google Security for discovering this bug and to
3599 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3600 preparing the fix (CVE-2014-0160)
3601 [Adam Langley, Bodo Moeller]
3603 *) Fix for the attack described in the paper "Recovering OpenSSL
3604 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3605 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3606 http://eprint.iacr.org/2014/140
3608 Thanks to Yuval Yarom and Naomi Benger for discovering this
3609 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3610 [Yuval Yarom and Naomi Benger]
3612 *) TLS pad extension: draft-agl-tls-padding-03
3614 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3615 TLS client Hello record length value would otherwise be > 255 and
3616 less that 512 pad with a dummy extension containing zeroes so it
3617 is at least 512 bytes long.
3619 [Adam Langley, Steve Henson]
3621 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3623 *) Fix for TLS record tampering bug. A carefully crafted invalid
3624 handshake could crash OpenSSL with a NULL pointer exception.
3625 Thanks to Anton Johansson for reporting this issues.
3628 *) Keep original DTLS digest and encryption contexts in retransmission
3629 structures so we can use the previous session parameters if they need
3630 to be resent. (CVE-2013-6450)
3633 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3634 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3635 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3636 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3637 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3638 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3639 [Rob Stradling, Adam Langley]
3641 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3643 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3644 supporting platforms or when small records were transferred.
3645 [Andy Polyakov, Steve Henson]
3647 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3649 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3651 This addresses the flaw in CBC record processing discovered by
3652 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3653 at: http://www.isg.rhul.ac.uk/tls/
3655 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3656 Security Group at Royal Holloway, University of London
3657 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3658 Emilia Käsper for the initial patch.
3660 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3662 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3663 ciphersuites which can be exploited in a denial of service attack.
3664 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3665 and detecting this bug and to Wolfgang Ettlinger
3666 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3670 *) Return an error when checking OCSP signatures when key is NULL.
3671 This fixes a DoS attack. (CVE-2013-0166)
3674 *) Make openssl verify return errors.
3675 [Chris Palmer <palmer@google.com> and Ben Laurie]
3677 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3678 the right response is stapled. Also change SSL_get_certificate()
3679 so it returns the certificate actually sent.
3680 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3681 [Rob Stradling <rob.stradling@comodo.com>]
3683 *) Fix possible deadlock when decoding public keys.
3686 *) Don't use TLS 1.0 record version number in initial client hello
3690 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3692 *) Sanity check record length before skipping explicit IV in TLS
3693 1.2, 1.1 and DTLS to fix DoS attack.
3695 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3696 fuzzing as a service testing platform.
3700 *) Initialise tkeylen properly when encrypting CMS messages.
3701 Thanks to Solar Designer of Openwall for reporting this issue.
3704 *) In FIPS mode don't try to use composite ciphers as they are not
3708 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3710 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3711 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3712 mean any application compiled against OpenSSL 1.0.0 headers setting
3713 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3714 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3715 0x10000000L Any application which was previously compiled against
3716 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3717 will need to be recompiled as a result. Letting be results in
3718 inability to disable specifically TLS 1.1 and in client context,
3719 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3722 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3723 disable just protocol X, but all protocols above X *if* there are
3724 protocols *below* X still enabled. In more practical terms it means
3725 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3726 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3727 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3731 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3733 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3734 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3735 in CRYPTO_realloc_clean.
3737 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3738 issue and to Adam Langley <agl@chromium.org> for fixing it.
3740 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3742 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3745 *) Workarounds for some broken servers that "hang" if a client hello
3746 record length exceeds 255 bytes.
3748 1. Do not use record version number > TLS 1.0 in initial client
3749 hello: some (but not all) hanging servers will now work.
3750 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3751 the number of ciphers sent in the client hello. This should be
3752 set to an even number, such as 50, for example by passing:
3753 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3754 Most broken servers should now work.
3755 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3756 TLS 1.2 client support entirely.
3759 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3762 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3764 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3765 STRING form instead of a DigestInfo.
3768 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3769 and the RSA_sign/RSA_verify functions. This was made more apparent when
3770 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3771 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3772 the correct format in RSA_verify so both forms transparently work.
3775 *) Some servers which support TLS 1.0 can choke if we initially indicate
3776 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3777 encrypted premaster secret. As a workaround use the maximum permitted
3778 client version in client hello, this should keep such servers happy
3779 and still work with previous versions of OpenSSL.
3782 *) Add support for TLS/DTLS heartbeats.
3783 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3785 *) Add support for SCTP.
3786 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3788 *) Improved PRNG seeding for VOS.
3789 [Paul Green <Paul.Green@stratus.com>]
3791 *) Extensive assembler packs updates, most notably:
3793 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3794 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3795 - x86_64: bit-sliced AES implementation;
3796 - ARM: NEON support, contemporary platforms optimizations;
3797 - s390x: z196 support;
3798 - *: GHASH and GF(2^m) multiplication implementations;
3802 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3803 (removal of unnecessary code)
3804 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3806 *) Add TLS key material exporter from RFC 5705.
3809 *) Add DTLS-SRTP negotiation from RFC 5764.
3812 *) Add Next Protocol Negotiation,
3813 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3814 disabled with a no-npn flag to config or Configure. Code donated
3816 [Adam Langley <agl@google.com> and Ben Laurie]
3818 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3819 NIST-P256, NIST-P521, with constant-time single point multiplication on
3820 typical inputs. Compiler support for the nonstandard type __uint128_t is
3821 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3822 Code made available under Apache License version 2.0.
3824 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3825 line to include this in your build of OpenSSL, and run "make depend" (or
3826 "make update"). This enables the following EC_METHODs:
3828 EC_GFp_nistp224_method()
3829 EC_GFp_nistp256_method()
3830 EC_GFp_nistp521_method()
3832 EC_GROUP_new_by_curve_name() will automatically use these (while
3833 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3835 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3837 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3838 all platforms. Move ssize_t definition from e_os.h to the public
3839 header file e_os2.h as it now appears in public header file cms.h
3842 *) New -sigopt option to the ca, req and x509 utilities. Additional
3843 signature parameters can be passed using this option and in
3847 *) Add RSA PSS signing function. This will generate and set the
3848 appropriate AlgorithmIdentifiers for PSS based on those in the
3849 corresponding EVP_MD_CTX structure. No application support yet.
3852 *) Support for companion algorithm specific ASN1 signing routines.
3853 New function ASN1_item_sign_ctx() signs a pre-initialised
3854 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3855 the appropriate parameters.
3858 *) Add new algorithm specific ASN1 verification initialisation function
3859 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3860 handling will be the same no matter what EVP_PKEY_METHOD is used.
3861 Add a PSS handler to support verification of PSS signatures: checked
3862 against a number of sample certificates.
3865 *) Add signature printing for PSS. Add PSS OIDs.
3866 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3868 *) Add algorithm specific signature printing. An individual ASN1 method
3869 can now print out signatures instead of the standard hex dump.
3871 More complex signatures (e.g. PSS) can print out more meaningful
3872 information. Include DSA version that prints out the signature
3876 *) Password based recipient info support for CMS library: implementing
3880 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3881 neatly separates the code into cipher and PBE sections and is required
3882 for some algorithms that split PBES2 into separate pieces (such as
3883 password based CMS).
3886 *) Session-handling fixes:
3887 - Fix handling of connections that are resuming with a session ID,
3888 but also support Session Tickets.
3889 - Fix a bug that suppressed issuing of a new ticket if the client
3890 presented a ticket with an expired session.
3891 - Try to set the ticket lifetime hint to something reasonable.
3892 - Make tickets shorter by excluding irrelevant information.
3893 - On the client side, don't ignore renewed tickets.
3894 [Adam Langley, Bodo Moeller (Google)]
3896 *) Fix PSK session representation.
3899 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3901 This work was sponsored by Intel.
3904 *) Add GCM support to TLS library. Some custom code is needed to split
3905 the IV between the fixed (from PRF) and explicit (from TLS record)
3906 portions. This adds all GCM ciphersuites supported by RFC5288 and
3907 RFC5289. Generalise some AES* cipherstrings to include GCM and
3908 add a special AESGCM string for GCM only.
3911 *) Expand range of ctrls for AES GCM. Permit setting invocation
3912 field on decrypt and retrieval of invocation field only on encrypt.
3915 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3916 As required by RFC5289 these ciphersuites cannot be used if for
3917 versions of TLS earlier than 1.2.
3920 *) For FIPS capable OpenSSL interpret a NULL default public key method
3921 as unset and return the appropriate default but do *not* set the default.
3922 This means we can return the appropriate method in applications that
3923 switch between FIPS and non-FIPS modes.
3926 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3927 ENGINE is used then we cannot handle that in the FIPS module so we
3928 keep original code iff non-FIPS operations are allowed.
3931 *) Add -attime option to openssl utilities.
3932 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3934 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3937 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3938 FIPS EC methods unconditionally for now.
3941 *) New build option no-ec2m to disable characteristic 2 code.
3944 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3945 all cases can be covered as some introduce binary incompatibilities.
3948 *) Redirect RSA operations to FIPS module including keygen,
3949 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3952 *) Add similar low level API blocking to ciphers.
3955 *) Low level digest APIs are not approved in FIPS mode: any attempt
3956 to use these will cause a fatal error. Applications that *really* want
3957 to use them can use the private_* version instead.
3960 *) Redirect cipher operations to FIPS module for FIPS builds.
3963 *) Redirect digest operations to FIPS module for FIPS builds.
3966 *) Update build system to add "fips" flag which will link in fipscanister.o
3967 for static and shared library builds embedding a signature if needed.
3970 *) Output TLS supported curves in preference order instead of numerical
3971 order. This is currently hardcoded for the highest order curves first.
3972 This should be configurable so applications can judge speed vs strength.
3975 *) Add TLS v1.2 server support for client authentication.
3978 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3982 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3983 FIPS modules versions.
3986 *) Add TLS v1.2 client side support for client authentication. Keep cache
3987 of handshake records longer as we don't know the hash algorithm to use
3988 until after the certificate request message is received.
3991 *) Initial TLS v1.2 client support. Add a default signature algorithms
3992 extension including all the algorithms we support. Parse new signature
3993 format in client key exchange. Relax some ECC signing restrictions for
3994 TLS v1.2 as indicated in RFC5246.
3997 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3998 to new signature format when needed using client digest preference.
3999 All server ciphersuites should now work correctly in TLS v1.2. No client
4000 support yet and no support for client certificates.
4003 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
4004 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
4005 ciphersuites. At present only RSA key exchange ciphersuites work with
4006 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
4007 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
4008 and version checking.
4011 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
4012 with this defined it will not be affected by any changes to ssl internal
4013 structures. Add several utility functions to allow openssl application
4014 to work with OPENSSL_NO_SSL_INTERN defined.
4017 *) A long standing patch to add support for SRP from EdelWeb (Peter
4018 Sylvester and Christophe Renou) was integrated.
4019 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
4020 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
4023 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
4026 *) Permit abbreviated handshakes when renegotiating using the function
4027 SSL_renegotiate_abbreviated().
4028 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4030 *) Add call to ENGINE_register_all_complete() to
4031 ENGINE_load_builtin_engines(), so some implementations get used
4032 automatically instead of needing explicit application support.
4035 *) Add support for TLS key exporter as described in RFC5705.
4036 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
4038 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
4039 a few changes are required:
4041 Add SSL_OP_NO_TLSv1_1 flag.
4042 Add TLSv1_1 methods.
4043 Update version checking logic to handle version 1.1.
4044 Add explicit IV handling (ported from DTLS code).
4045 Add command line options to s_client/s_server.
4048 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
4050 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4051 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4052 content decryption and always return the same error. Note: this attack
4053 needs on average 2^20 messages so it only affects automated senders. The
4054 old behaviour can be re-enabled in the CMS code by setting the
4055 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4056 an MMA defence is not necessary.
4057 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4058 this issue. (CVE-2012-0884)
4061 *) Fix CVE-2011-4619: make sure we really are receiving a
4062 client hello before rejecting multiple SGC restarts. Thanks to
4063 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4066 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
4068 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4069 Thanks to Antonio Martin, Enterprise Secure Access Research and
4070 Development, Cisco Systems, Inc. for discovering this bug and
4071 preparing a fix. (CVE-2012-0050)
4074 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
4076 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
4077 of the Vaudenay padding oracle attack on CBC mode encryption
4078 which enables an efficient plaintext recovery attack against
4079 the OpenSSL implementation of DTLS. Their attack exploits timing
4080 differences arising during decryption processing. A research
4081 paper describing this attack can be found at:
4082 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4083 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4084 Security Group at Royal Holloway, University of London
4085 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4086 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4087 for preparing the fix. (CVE-2011-4108)
4088 [Robin Seggelmann, Michael Tuexen]
4090 *) Clear bytes used for block padding of SSL 3.0 records.
4092 [Adam Langley (Google)]
4094 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4095 Kadianakis <desnacked@gmail.com> for discovering this issue and
4096 Adam Langley for preparing the fix. (CVE-2011-4619)
4097 [Adam Langley (Google)]
4099 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4100 [Andrey Kulikov <amdeich@gmail.com>]
4102 *) Prevent malformed RFC3779 data triggering an assertion failure.
4103 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4104 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4105 [Rob Austein <sra@hactrn.net>]
4107 *) Improved PRNG seeding for VOS.
4108 [Paul Green <Paul.Green@stratus.com>]
4110 *) Fix ssl_ciph.c set-up race.
4111 [Adam Langley (Google)]
4113 *) Fix spurious failures in ecdsatest.c.
4114 [Emilia Käsper (Google)]
4116 *) Fix the BIO_f_buffer() implementation (which was mixing different
4117 interpretations of the '..._len' fields).
4118 [Adam Langley (Google)]
4120 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4121 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4122 threads won't reuse the same blinding coefficients.
4124 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4125 lock to call BN_BLINDING_invert_ex, and avoids one use of
4126 BN_BLINDING_update for each BN_BLINDING structure (previously,
4127 the last update always remained unused).
4128 [Emilia Käsper (Google)]
4130 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4131 [Bob Buckholz (Google)]
4133 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4135 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4136 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4137 [Kaspar Brand <ossl@velox.ch>]
4139 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4140 for multi-threaded use of ECDH. (CVE-2011-3210)
4141 [Adam Langley (Google)]
4143 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4146 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4147 signature public key algorithm by using OID xref utilities instead.
4148 Before this you could only use some ECC ciphersuites with SHA1 only.
4151 *) Add protection against ECDSA timing attacks as mentioned in the paper
4152 by Billy Bob Brumley and Nicola Tuveri, see:
4154 http://eprint.iacr.org/2011/232.pdf
4156 [Billy Bob Brumley and Nicola Tuveri]
4158 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4160 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4161 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4163 *) Fix bug in string printing code: if *any* escaping is enabled we must
4164 escape the escape character (backslash) or the resulting string is
4168 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4170 *) Disable code workaround for ancient and obsolete Netscape browsers
4171 and servers: an attacker can use it in a ciphersuite downgrade attack.
4172 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4175 *) Fixed J-PAKE implementation error, originally discovered by
4176 Sebastien Martini, further info and confirmation from Stefan
4177 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4180 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4182 *) Fix extension code to avoid race conditions which can result in a buffer
4183 overrun vulnerability: resumed sessions must not be modified as they can
4184 be shared by multiple threads. CVE-2010-3864
4187 *) Fix WIN32 build system to correctly link an ENGINE directory into
4191 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4193 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4195 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4197 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4199 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4200 context. The operation can be customised via the ctrl mechanism in
4201 case ENGINEs want to include additional functionality.
4204 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4207 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4208 output hashes compatible with older versions of OpenSSL.
4209 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4211 *) Fix compression algorithm handling: if resuming a session use the
4212 compression algorithm of the resumed session instead of determining
4213 it from client hello again. Don't allow server to change algorithm.
4216 *) Add load_crls() function to apps tidying load_certs() too. Add option
4217 to verify utility to allow additional CRLs to be included.
4220 *) Update OCSP request code to permit adding custom headers to the request:
4221 some responders need this.
4224 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4226 [Julia Lawall <julia@diku.dk>]
4228 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4229 needlessly dereferenced structures, used obsolete functions and
4230 didn't handle all updated verify codes correctly.
4233 *) Disable MD2 in the default configuration.
4236 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4237 indicate the initial BIO being pushed or popped. This makes it possible
4238 to determine whether the BIO is the one explicitly called or as a result
4239 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4240 it handles reference counts correctly and doesn't zero out the I/O bio
4241 when it is not being explicitly popped. WARNING: applications which
4242 included workarounds for the old buggy behaviour will need to be modified
4243 or they could free up already freed BIOs.
4246 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4247 renaming to all platforms (within the 0.9.8 branch, this was
4248 done conditionally on Netware platforms to avoid a name clash).
4249 [Guenter <lists@gknw.net>]
4251 *) Add ECDHE and PSK support to DTLS.
4252 [Michael Tuexen <tuexen@fh-muenster.de>]
4254 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4258 *) Add "missing" function EVP_MD_flags() (without this the only way to
4259 retrieve a digest flags is by accessing the structure directly. Update
4260 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4261 or cipher is registered as in the "from" argument. Print out all
4262 registered digests in the dgst usage message instead of manually
4263 attempting to work them out.
4266 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4267 this allows the use of compression and extensions. Change default cipher
4268 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4269 by default unless an application cipher string requests it.
4272 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4273 key ids to find matching certificates and keys but some PKCS#12 files
4274 don't follow the (somewhat unwritten) rules and this strategy fails.
4275 Now just gather all certificates together and the first private key
4276 then look for the first certificate that matches the key.
4279 *) Support use of registered digest and cipher names for dgst and cipher
4280 commands instead of having to add each one as a special case. So now
4287 openssl dgst -sha256 foo
4289 and this works for ENGINE based algorithms too.
4293 *) Update Gost ENGINE to support parameter files.
4294 [Victor B. Wagner <vitus@cryptocom.ru>]
4296 *) Support GeneralizedTime in ca utility.
4297 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4299 *) Enhance the hash format used for certificate directory links. The new
4300 form uses the canonical encoding (meaning equivalent names will work
4301 even if they aren't identical) and uses SHA1 instead of MD5. This form
4302 is incompatible with the older format and as a result c_rehash should
4303 be used to rebuild symbolic links.
4306 *) Make PKCS#8 the default write format for private keys, replacing the
4307 traditional format. This form is standardised, more secure and doesn't
4308 include an implicit MD5 dependency.
4311 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4312 committed to OpenSSL should pass this lot as a minimum.
4315 *) Add session ticket override functionality for use by EAP-FAST.
4316 [Jouni Malinen <j@w1.fi>]
4318 *) Modify HMAC functions to return a value. Since these can be implemented
4319 in an ENGINE errors can occur.
4322 *) Type-checked OBJ_bsearch_ex.
4325 *) Type-checked OBJ_bsearch. Also some constification necessitated
4326 by type-checking. Still to come: TXT_DB, bsearch(?),
4327 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4331 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4332 seconds to a tm structure directly, instead of going through OS
4333 specific date routines. This avoids any issues with OS routines such
4334 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4335 and X509_time_adj_ex() to cover the extended range. The existing
4336 X509_time_adj() is still usable and will no longer have any date issues.
4339 *) Delta CRL support. New use deltas option which will attempt to locate
4340 and search any appropriate delta CRLs available.
4342 This work was sponsored by Google.
4345 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4346 code and add additional score elements. Validate alternate CRL paths
4347 as part of the CRL checking and indicate a new error "CRL path validation
4348 error" in this case. Applications wanting additional details can use
4349 the verify callback and check the new "parent" field. If this is not
4350 NULL CRL path validation is taking place. Existing applications won't
4351 see this because it requires extended CRL support which is off by
4354 This work was sponsored by Google.
4357 *) Support for freshest CRL extension.
4359 This work was sponsored by Google.
4362 *) Initial indirect CRL support. Currently only supported in the CRLs
4363 passed directly and not via lookup. Process certificate issuer
4364 CRL entry extension and lookup CRL entries by bother issuer name
4365 and serial number. Check and process CRL issuer entry in IDP extension.
4367 This work was sponsored by Google.
4370 *) Add support for distinct certificate and CRL paths. The CRL issuer
4371 certificate is validated separately in this case. Only enabled if
4372 an extended CRL support flag is set: this flag will enable additional
4373 CRL functionality in future.
4375 This work was sponsored by Google.
4378 *) Add support for policy mappings extension.
4380 This work was sponsored by Google.
4383 *) Fixes to pathlength constraint, self issued certificate handling,
4384 policy processing to align with RFC3280 and PKITS tests.
4386 This work was sponsored by Google.
4389 *) Support for name constraints certificate extension. DN, email, DNS
4390 and URI types are currently supported.
4392 This work was sponsored by Google.
4395 *) To cater for systems that provide a pointer-based thread ID rather
4396 than numeric, deprecate the current numeric thread ID mechanism and
4397 replace it with a structure and associated callback type. This
4398 mechanism allows a numeric "hash" to be extracted from a thread ID in
4399 either case, and on platforms where pointers are larger than 'long',
4400 mixing is done to help ensure the numeric 'hash' is usable even if it
4401 can't be guaranteed unique. The default mechanism is to use "&errno"
4402 as a pointer-based thread ID to distinguish between threads.
4404 Applications that want to provide their own thread IDs should now use
4405 CRYPTO_THREADID_set_callback() to register a callback that will call
4406 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4408 Note that ERR_remove_state() is now deprecated, because it is tied
4409 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4410 to free the current thread's error state should be replaced by
4411 ERR_remove_thread_state(NULL).
4413 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4414 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4415 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4416 application was previously providing a numeric thread callback that
4417 was inappropriate for distinguishing threads, then uniqueness might
4418 have been obtained with &errno that happened immediately in the
4419 intermediate development versions of OpenSSL; this is no longer the
4420 case, the numeric thread callback will now override the automatic use
4422 [Geoff Thorpe, with help from Bodo Moeller]
4424 *) Initial support for different CRL issuing certificates. This covers a
4425 simple case where the self issued certificates in the chain exist and
4426 the real CRL issuer is higher in the existing chain.
4428 This work was sponsored by Google.
4431 *) Removed effectively defunct crypto/store from the build.
4434 *) Revamp of STACK to provide stronger type-checking. Still to come:
4435 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4436 ASN1_STRING, CONF_VALUE.
4439 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4440 RAM on SSL connections. This option can save about 34k per idle SSL.
4443 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4444 STACK, TXT_DB, bsearch, qsort.
4447 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4448 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4449 support for data, signedData, compressedData, digestedData and
4450 encryptedData, envelopedData types included. Scripts to check against
4451 RFC4134 examples draft and interop and consistency checks of many
4452 content types and variants.
4455 *) Add options to enc utility to support use of zlib compression BIO.
4458 *) Extend mk1mf to support importing of options and assembly language
4459 files from Configure script, currently only included in VC-WIN32.
4460 The assembly language rules can now optionally generate the source
4461 files from the associated perl scripts.
4464 *) Implement remaining functionality needed to support GOST ciphersuites.
4465 Interop testing has been performed using CryptoPro implementations.
4466 [Victor B. Wagner <vitus@cryptocom.ru>]
4468 *) s390x assembler pack.
4471 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4475 *) Implement Opaque PRF Input TLS extension as specified in
4476 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4477 official specification yet and no extension type assignment by
4478 IANA exists, this extension (for now) will have to be explicitly
4479 enabled when building OpenSSL by providing the extension number
4480 to use. For example, specify an option
4482 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4484 to the "config" or "Configure" script to enable the extension,
4485 assuming extension number 0x9527 (which is a completely arbitrary
4486 and unofficial assignment based on the MD5 hash of the Internet
4487 Draft). Note that by doing so, you potentially lose
4488 interoperability with other TLS implementations since these might
4489 be using the same extension number for other purposes.
4491 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4492 opaque PRF input value to use in the handshake. This will create
4493 an internal copy of the length-'len' string at 'src', and will
4494 return non-zero for success.
4496 To get more control and flexibility, provide a callback function
4499 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4500 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4504 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4507 Callback function 'cb' will be called in handshakes, and is
4508 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4509 Argument 'arg' is for application purposes (the value as given to
4510 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4511 be provided to the callback function). The callback function
4512 has to return non-zero to report success: usually 1 to use opaque
4513 PRF input just if possible, or 2 to enforce use of the opaque PRF
4514 input. In the latter case, the library will abort the handshake
4515 if opaque PRF input is not successfully negotiated.
4517 Arguments 'peerinput' and 'len' given to the callback function
4518 will always be NULL and 0 in the case of a client. A server will
4519 see the client's opaque PRF input through these variables if
4520 available (NULL and 0 otherwise). Note that if the server
4521 provides an opaque PRF input, the length must be the same as the
4522 length of the client's opaque PRF input.
4524 Note that the callback function will only be called when creating
4525 a new session (session resumption can resume whatever was
4526 previously negotiated), and will not be called in SSL 2.0
4527 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4528 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4529 for applications that need to enforce opaque PRF input.
4533 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4536 [Victor B. Wagner <vitus@cryptocom.ru>]
4538 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4539 RFC4507bis. The encrypted ticket format is an encrypted encoded
4540 SSL_SESSION structure, that way new session features are automatically
4543 If a client application caches session in an SSL_SESSION structure
4544 support is transparent because tickets are now stored in the encoded
4547 The SSL_CTX structure automatically generates keys for ticket
4548 protection in servers so again support should be possible
4549 with no application modification.
4551 If a client or server wishes to disable RFC4507 support then the option
4552 SSL_OP_NO_TICKET can be set.
4554 Add a TLS extension debugging callback to allow the contents of any client
4555 or server extensions to be examined.
4557 This work was sponsored by Google.
4560 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4561 OpenSSL should now compile cleanly on gcc 4.2
4562 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4564 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4565 support including streaming MAC support: this is required for GOST
4566 ciphersuite support.
4567 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4569 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4570 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4571 to output in BER and PEM format.
4574 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4575 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4576 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4577 ENGINE support for HMAC keys which are unextractable. New -mac and
4578 -macopt options to dgst utility.
4581 *) New option -sigopt to dgst utility. Update dgst to use
4582 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4583 alternative signing parameters such as X9.31 or PSS in the dgst
4587 *) Change ssl_cipher_apply_rule(), the internal function that does
4588 the work each time a ciphersuite string requests enabling
4589 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4590 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4591 the order of disabled ciphersuites such that those ciphersuites
4592 that most recently went from enabled to disabled not only stay
4593 in order with respect to each other, but also have higher priority
4594 than other disabled ciphersuites the next time ciphersuites are
4597 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4598 the same ciphersuites as with "HIGH" alone, but in a specific
4599 order where the PSK ciphersuites come first (since they are the
4600 most recently disabled ciphersuites when "HIGH" is parsed).
4602 Also, change ssl_create_cipher_list() (using this new
4603 functionality) such that between otherwise identical
4604 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4608 *) Change ssl_create_cipher_list() so that it automatically
4609 arranges the ciphersuites in reasonable order before starting
4610 to process the rule string. Thus, the definition for "DEFAULT"
4611 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4612 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4613 This makes it much easier to arrive at a reasonable default order
4614 in applications for which anonymous ciphers are OK (meaning
4615 that you can't actually use DEFAULT).
4616 [Bodo Moeller; suggested by Victor Duchovni]
4618 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4619 processing) into multiple integers instead of setting
4620 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4621 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4622 (These masks as well as the individual bit definitions are hidden
4623 away into the non-exported interface ssl/ssl_locl.h, so this
4624 change to the definition of the SSL_CIPHER structure shouldn't
4625 affect applications.) This give us more bits for each of these
4626 categories, so there is no longer a need to coagulate AES128 and
4627 AES256 into a single algorithm bit, and to coagulate Camellia128
4628 and Camellia256 into a single algorithm bit, which has led to all
4631 Thus, among other things, the kludge introduced in 0.9.7m and
4632 0.9.8e for masking out AES256 independently of AES128 or masking
4633 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4635 With the change, we also introduce new ciphersuite aliases that
4636 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4640 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4641 Use the leftmost N bytes of the signature input if the input is
4642 larger than the prime q (with N being the size in bytes of q).
4645 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4646 it yet and it is largely untested.
4649 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4652 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4653 some compilers (gcc 4.2 and later) reject their use. Safestack is
4654 reimplemented. Update ASN1 to avoid use of legacy functions.
4657 *) Win32/64 targets are linked with Winsock2.
4660 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4661 to external functions. This can be used to increase CRL handling
4662 efficiency especially when CRLs are very large by (for example) storing
4663 the CRL revoked certificates in a database.
4666 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4667 new CRLs added to a directory can be used. New command line option
4668 -verify_return_error to s_client and s_server. This causes real errors
4669 to be returned by the verify callback instead of carrying on no matter
4670 what. This reflects the way a "real world" verify callback would behave.
4673 *) GOST engine, supporting several GOST algorithms and public key formats.
4674 Kindly donated by Cryptocom.
4677 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4678 partitioned by DP are handled but no indirect CRL or reason partitioning
4679 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4680 selected via a scoring technique which handles IDP and AKID in CRLs.
4683 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4684 will ultimately be used for all verify operations: this will remove the
4685 X509_STORE dependency on certificate verification and allow alternative
4686 lookup methods. X509_STORE based implementations of these two callbacks.
4689 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4690 Modify get_crl() to find a valid (unexpired) CRL if possible.
4693 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4694 this would be called X509_CRL_cmp() but that name is already used by
4695 a function that just compares CRL issuer names. Cache several CRL
4696 extensions in X509_CRL structure and cache CRLDP in X509.
4699 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4700 this maps equivalent X509_NAME structures into a consistent structure.
4701 Name comparison can then be performed rapidly using memcmp().
4704 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4708 *) Allow digests to supply their own micalg string for S/MIME type using
4709 the ctrl EVP_MD_CTRL_MICALG.
4712 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4713 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4714 ctrl. It can then customise the structure before and/or after signing
4718 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4719 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4720 to free up any added signature OIDs.
4723 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4724 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4725 digest and cipher tables. New options added to openssl utility:
4726 list-message-digest-algorithms and list-cipher-algorithms.
4729 *) Change the array representation of binary polynomials: the list
4730 of degrees of non-zero coefficients is now terminated with -1.
4731 Previously it was terminated with 0, which was also part of the
4732 value; thus, the array representation was not applicable to
4733 polynomials where t^0 has coefficient zero. This change makes
4734 the array representation useful in a more general context.
4737 *) Various modifications and fixes to SSL/TLS cipher string
4738 handling. For ECC, the code now distinguishes between fixed ECDH
4739 with RSA certificates on the one hand and with ECDSA certificates
4740 on the other hand, since these are separate ciphersuites. The
4741 unused code for Fortezza ciphersuites has been removed.
4743 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4744 (not "ECDHE"). For consistency with the code for DH
4745 certificates, use of ECDH certificates is now considered ECDH
4746 authentication, not RSA or ECDSA authentication (the latter is
4747 merely the CA's signing algorithm and not actively used in the
4750 The temporary ciphersuite alias "ECCdraft" is no longer
4751 available, and ECC ciphersuites are no longer excluded from "ALL"
4752 and "DEFAULT". The following aliases now exist for RFC 4492
4753 ciphersuites, most of these by analogy with the DH case:
4755 kECDHr - ECDH cert, signed with RSA
4756 kECDHe - ECDH cert, signed with ECDSA
4757 kECDH - ECDH cert (signed with either RSA or ECDSA)
4758 kEECDH - ephemeral ECDH
4759 ECDH - ECDH cert or ephemeral ECDH
4765 AECDH - anonymous ECDH
4766 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4770 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4771 Use correct micalg parameters depending on digest(s) in signed message.
4774 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4775 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4778 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4779 an engine to register a method. Add ENGINE lookups for methods and
4780 functional reference processing.
4783 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4784 EVP_{Sign,Verify}* which allow an application to customise the signature
4788 *) New -resign option to smime utility. This adds one or more signers
4789 to an existing PKCS#7 signedData structure. Also -md option to use an
4790 alternative message digest algorithm for signing.
4793 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4794 create PKCS7 structures containing multiple signers. Update smime
4795 application to support multiple signers.
4798 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4802 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4803 Reorganize PBE internals to lookup from a static table using NIDs,
4804 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4805 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4806 PRF which will be automatically used with PBES2.
4809 *) Replace the algorithm specific calls to generate keys in "req" with the
4813 *) Update PKCS#7 enveloped data routines to use new API. This is now
4814 supported by any public key method supporting the encrypt operation. A
4815 ctrl is added to allow the public key algorithm to examine or modify
4816 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4820 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4821 a default digest type to use. In most cases this will be SHA1 but some
4822 algorithms (such as GOST) need to specify an alternative digest. The
4823 return value indicates how strong the preference is 1 means optional and
4824 2 is mandatory (that is it is the only supported type). Modify
4825 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4826 use the default md. Update openssl utilities to use the default digest
4827 type for signing if it is not explicitly indicated.
4830 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4831 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4832 signing method from the key type. This effectively removes the link
4833 between digests and public key types.
4836 *) Add an OID cross reference table and utility functions. Its purpose is to
4837 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4838 rsaEncryption. This will allow some of the algorithm specific hackery
4839 needed to use the correct OID to be removed.
4842 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4843 structures for PKCS7_sign(). They are now set up by the relevant public
4847 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4850 *) Add support for key derivation (agreement) in the API, DH method and
4854 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4855 public and private key formats. As a side effect these add additional
4856 command line functionality not previously available: DSA signatures can be
4857 generated and verified using pkeyutl and DH key support and generation in
4862 [Oliver Tappe <zooey@hirschkaefer.de>]
4864 *) New make target "install_html_docs" installs HTML renditions of the
4866 [Oliver Tappe <zooey@hirschkaefer.de>]
4868 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4869 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4870 support key and parameter generation and add initial key generation
4871 functionality for RSA.
4874 *) Add functions for main EVP_PKEY_method operations. The undocumented
4875 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4876 EVP_PKEY_{encrypt,decrypt}_old.
4879 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4880 key API, doesn't do much yet.
4883 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4884 public key algorithms. New option to openssl utility:
4885 "list-public-key-algorithms" to print out info.
4888 *) Implement the Supported Elliptic Curves Extension for
4889 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4892 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4893 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4896 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4897 utilities such as rsa, dsa, dsaparam etc except they process any key
4901 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4902 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4903 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4907 *) Initial support for pluggable public key ASN1.
4908 De-spaghettify the public key ASN1 handling. Move public and private
4909 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4910 algorithm specific handling to a single module within the relevant
4911 algorithm directory. Add functions to allow (near) opaque processing
4912 of public and private key structures.
4915 *) Implement the Supported Point Formats Extension for
4916 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4919 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4920 for the psk identity [hint] and the psk callback functions to the
4921 SSL_SESSION, SSL and SSL_CTX structure.
4924 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4928 SSL_CTX_use_psk_identity_hint
4929 SSL_get_psk_identity_hint
4930 SSL_get_psk_identity
4931 SSL_use_psk_identity_hint
4933 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4935 *) Add RFC 3161 compliant time stamp request creation, response generation
4936 and response verification functionality.
4937 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4939 *) Add initial support for TLS extensions, specifically for the server_name
4940 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4941 have new members for a host name. The SSL data structure has an
4942 additional member SSL_CTX *initial_ctx so that new sessions can be
4943 stored in that context to allow for session resumption, even after the
4944 SSL has been switched to a new SSL_CTX in reaction to a client's
4945 server_name extension.
4947 New functions (subject to change):
4949 SSL_get_servername()
4950 SSL_get_servername_type()
4953 New CTRL codes and macros (subject to change):
4955 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4956 - SSL_CTX_set_tlsext_servername_callback()
4957 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4958 - SSL_CTX_set_tlsext_servername_arg()
4959 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4961 openssl s_client has a new '-servername ...' option.
4963 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4964 '-key2 ...', '-servername_fatal' (subject to change). This allows
4965 testing the HostName extension for a specific single host name ('-cert'
4966 and '-key' remain fallbacks for handshakes without HostName
4967 negotiation). If the unrecognized_name alert has to be sent, this by
4968 default is a warning; it becomes fatal with the '-servername_fatal'
4971 [Peter Sylvester, Remy Allais, Christophe Renou]
4973 *) Whirlpool hash implementation is added.
4976 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4977 bn(64,32). Because of instruction set limitations it doesn't have
4978 any negative impact on performance. This was done mostly in order
4979 to make it possible to share assembler modules, such as bn_mul_mont
4980 implementations, between 32- and 64-bit builds without hassle.
4983 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4984 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4988 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4989 dedicated Montgomery multiplication procedure, is introduced.
4990 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4991 "64-bit" performance on certain 32-bit targets.
4994 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4995 in SSL structures. New SSL ctrl to set maximum send fragment size.
4996 Save memory by setting the I/O buffer sizes dynamically instead of
4997 using the maximum available value.
5000 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
5001 in addition to the text details.
5004 *) Very, very preliminary EXPERIMENTAL support for printing of general
5005 ASN1 structures. This currently produces rather ugly output and doesn't
5006 handle several customised structures at all.
5009 *) Integrated support for PVK file format and some related formats such
5010 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
5011 these in the 'rsa' and 'dsa' utilities.
5014 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
5017 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
5018 place for the (very old) "NETSCAPE" format certificates which are now
5019 handled using new ASN1 code equivalents.
5022 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
5023 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
5024 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
5027 *) Modify CRL distribution points extension code to print out previously
5028 unsupported fields. Enhance extension setting code to allow setting of
5032 *) Add print and set support for Issuing Distribution Point CRL extension.
5035 *) Change 'Configure' script to enable Camellia by default.
5038 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
5040 *) When rejecting SSL/TLS records due to an incorrect version number, never
5041 update s->server with a new major version number. As of
5042 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5043 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5044 the previous behavior could result in a read attempt at NULL when
5045 receiving specific incorrect SSL/TLS records once record payload
5046 protection is active. (CVE-2010-0740)
5047 [Bodo Moeller, Adam Langley <agl@chromium.org>]
5049 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5050 could be crashed if the relevant tables were not present (e.g. chrooted).
5051 [Tomas Hoger <thoger@redhat.com>]
5053 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5055 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5056 [Martin Olsson, Neel Mehta]
5058 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
5059 accommodate for stack sorting, always a write lock!).
5062 *) On some versions of WIN32 Heap32Next is very slow. This can cause
5063 excessive delays in the RAND_poll(): over a minute. As a workaround
5064 include a time check in the inner Heap32Next loop too.
5067 *) The code that handled flushing of data in SSL/TLS originally used the
5068 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5069 the problem outlined in PR#1949. The fix suggested there however can
5070 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5071 of Apache). So instead simplify the code to flush unconditionally.
5072 This should be fine since flushing with no data to flush is a no op.
5075 *) Handle TLS versions 2.0 and later properly and correctly use the
5076 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5077 off ancient servers have a habit of sticking around for a while...
5080 *) Modify compression code so it frees up structures without using the
5081 ex_data callbacks. This works around a problem where some applications
5082 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5083 restarting) then use compression (e.g. SSL with compression) later.
5084 This results in significant per-connection memory leaks and
5085 has caused some security issues including CVE-2008-1678 and
5089 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5090 change when encrypting or decrypting.
5093 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5094 connect and renegotiate with servers which do not support RI.
5095 Until RI is more widely deployed this option is enabled by default.
5098 *) Add "missing" ssl ctrls to clear options and mode.
5101 *) If client attempts to renegotiate and doesn't support RI respond with
5102 a no_renegotiation alert as required by RFC5746. Some renegotiating
5103 TLS clients will continue a connection gracefully when they receive
5104 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5105 waiting for a server hello which it will never receive. Now we treat a
5106 received no_renegotiation alert as a fatal error. This is because
5107 applications requesting a renegotiation might well expect it to succeed
5108 and would have no code in place to handle the server denying it so the
5109 only safe thing to do is to terminate the connection.
5112 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5113 peer supports secure renegotiation and 0 otherwise. Print out peer
5114 renegotiation support in s_client/s_server.
5117 *) Replace the highly broken and deprecated SPKAC certification method with
5118 the updated NID creation version. This should correctly handle UTF8.
5121 *) Implement RFC5746. Re-enable renegotiation but require the extension
5122 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5123 turns out to be a bad idea. It has been replaced by
5124 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5125 SSL_CTX_set_options(). This is really not recommended unless you
5126 know what you are doing.
5127 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5129 *) Fixes to stateless session resumption handling. Use initial_ctx when
5130 issuing and attempting to decrypt tickets in case it has changed during
5131 servername handling. Use a non-zero length session ID when attempting
5132 stateless session resumption: this makes it possible to determine if
5133 a resumption has occurred immediately after receiving server hello
5134 (several places in OpenSSL subtly assume this) instead of later in
5138 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5139 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5140 fixes for a few places where the return code is not checked
5142 [Julia Lawall <julia@diku.dk>]
5144 *) Add --strict-warnings option to Configure script to include devteam
5145 warnings in other configurations.
5148 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5149 makes it possible to install openssl libraries in locations which
5150 have names other than "lib", for example "/usr/lib64" which some
5152 [Steve Henson, based on patch from Jeremy Utley]
5154 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5155 X690 8.9.12 and can produce some misleading textual output of OIDs.
5156 [Steve Henson, reported by Dan Kaminsky]
5158 *) Delete MD2 from algorithm tables. This follows the recommendation in
5159 several standards that it is not used in new applications due to
5160 several cryptographic weaknesses. For binary compatibility reasons
5161 the MD2 API is still compiled in by default.
5164 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5168 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5169 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5171 [Guenter <lists@gknw.net>]
5173 *) Fix the server certificate chain building code to use X509_verify_cert(),
5174 it used to have an ad-hoc builder which was unable to cope with anything
5175 other than a simple chain.
5176 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5178 *) Don't check self signed certificate signatures in X509_verify_cert()
5179 by default (a flag can override this): it just wastes time without
5180 adding any security. As a useful side effect self signed root CAs
5181 with non-FIPS digests are now usable in FIPS mode.
5184 *) In dtls1_process_out_of_seq_message() the check if the current message
5185 is already buffered was missing. For every new message was memory
5186 allocated, allowing an attacker to perform an denial of service attack
5187 with sending out of seq handshake messages until there is no memory
5188 left. Additionally every future message was buffered, even if the
5189 sequence number made no sense and would be part of another handshake.
5190 So only messages with sequence numbers less than 10 in advance will be
5191 buffered. (CVE-2009-1378)
5192 [Robin Seggelmann, discovered by Daniel Mentz]
5194 *) Records are buffered if they arrive with a future epoch to be
5195 processed after finishing the corresponding handshake. There is
5196 currently no limitation to this buffer allowing an attacker to perform
5197 a DOS attack with sending records with future epochs until there is no
5198 memory left. This patch adds the pqueue_size() function to determine
5199 the size of a buffer and limits the record buffer to 100 entries.
5201 [Robin Seggelmann, discovered by Daniel Mentz]
5203 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5204 parent structure is freed. (CVE-2009-1379)
5207 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5208 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5211 [Ilya O. <vrghost@gmail.com>]
5213 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5215 *) Disable renegotiation completely - this fixes a severe security
5216 problem (CVE-2009-3555) at the cost of breaking all
5217 renegotiation. Renegotiation can be re-enabled by setting
5218 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5219 run-time. This is really not recommended unless you know what
5223 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5225 *) Don't set val to NULL when freeing up structures, it is freed up by
5226 underlying code. If sizeof(void *) > sizeof(long) this can result in
5227 zeroing past the valid field. (CVE-2009-0789)
5228 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5230 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5231 checked correctly. This would allow some invalid signed attributes to
5232 appear to verify correctly. (CVE-2009-0591)
5233 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5235 *) Reject UniversalString and BMPString types with invalid lengths. This
5236 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5237 a legal length. (CVE-2009-0590)
5240 *) Set S/MIME signing as the default purpose rather than setting it
5241 unconditionally. This allows applications to override it at the store
5245 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5246 to handle some structures.
5249 *) Improve efficiency of mem_gets: don't search whole buffer each time
5251 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5253 *) New -hex option for openssl rand.
5256 *) Print out UTF8String and NumericString when parsing ASN1.
5259 *) Support NumericString type for name components.
5262 *) Allow CC in the environment to override the automatically chosen
5263 compiler. Note that nothing is done to ensure flags work with the
5267 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5269 *) Properly check EVP_VerifyFinal() and similar return values
5271 [Ben Laurie, Bodo Moeller, Google Security Team]
5273 *) Enable TLS extensions by default.
5276 *) Allow the CHIL engine to be loaded, whether the application is
5277 multithreaded or not. (This does not release the developer from the
5278 obligation to set up the dynamic locking callbacks.)
5279 [Sander Temme <sander@temme.net>]
5281 *) Use correct exit code if there is an error in dgst command.
5282 [Steve Henson; problem pointed out by Roland Dirlewanger]
5284 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5285 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5288 *) Add experimental JPAKE support, including demo authentication in
5289 s_client and s_server.
5292 *) Set the comparison function in v3_addr_canonize().
5293 [Rob Austein <sra@hactrn.net>]
5295 *) Add support for XMPP STARTTLS in s_client.
5296 [Philip Paeps <philip@freebsd.org>]
5298 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5299 to ensure that even with this option, only ciphersuites in the
5300 server's preference list will be accepted. (Note that the option
5301 applies only when resuming a session, so the earlier behavior was
5302 just about the algorithm choice for symmetric cryptography.)
5305 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5307 *) Fix NULL pointer dereference if a DTLS server received
5308 ChangeCipherSpec as first record (CVE-2009-1386).
5311 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5312 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5315 *) The fix in 0.9.8c that supposedly got rid of unsafe
5316 double-checked locking was incomplete for RSA blinding,
5317 addressing just one layer of what turns out to have been
5318 doubly unsafe triple-checked locking.
5320 So now fix this for real by retiring the MONT_HELPER macro
5321 in crypto/rsa/rsa_eay.c.
5323 [Bodo Moeller; problem pointed out by Marius Schilder]
5325 *) Various precautionary measures:
5327 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5329 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5330 (NB: This would require knowledge of the secret session ticket key
5331 to exploit, in which case you'd be SOL either way.)
5333 - Change bn_nist.c so that it will properly handle input BIGNUMs
5334 outside the expected range.
5336 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5339 [Neel Mehta, Bodo Moeller]
5341 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5342 the load fails. Useful for distros.
5343 [Ben Laurie and the FreeBSD team]
5345 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5348 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5351 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5353 This work was sponsored by Logica.
5356 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5357 keystores. Support for SSL/TLS client authentication too.
5358 Not compiled unless enable-capieng specified to Configure.
5360 This work was sponsored by Logica.
5363 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5364 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5365 attribute creation routines such as certificate requests and PKCS#12
5369 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5371 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5372 handshake which could lead to a client crash as found using the
5373 Codenomicon TLS test suite (CVE-2008-1672)
5374 [Steve Henson, Mark Cox]
5376 *) Fix double free in TLS server name extensions which could lead to
5377 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5380 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5382 Clear the error queue to ensure that error entries left from
5383 older function calls do not interfere with the correct operation.
5384 [Lutz Jaenicke, Erik de Castro Lopo]
5386 *) Remove root CA certificates of commercial CAs:
5388 The OpenSSL project does not recommend any specific CA and does not
5389 have any policy with respect to including or excluding any CA.
5390 Therefore it does not make any sense to ship an arbitrary selection
5391 of root CA certificates with the OpenSSL software.
5394 *) RSA OAEP patches to fix two separate invalid memory reads.
5395 The first one involves inputs when 'lzero' is greater than
5396 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5397 before the beginning of from). The second one involves inputs where
5398 the 'db' section contains nothing but zeroes (there is a one-byte
5399 invalid read after the end of 'db').
5400 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5402 *) Partial backport from 0.9.9-dev:
5404 Introduce bn_mul_mont (dedicated Montgomery multiplication
5405 procedure) as a candidate for BIGNUM assembler implementation.
5406 While 0.9.9-dev uses assembler for various architectures, only
5407 x86_64 is available by default here in the 0.9.8 branch, and
5408 32-bit x86 is available through a compile-time setting.
5410 To try the 32-bit x86 assembler implementation, use Configure
5411 option "enable-montasm" (which exists only for this backport).
5413 As "enable-montasm" for 32-bit x86 disclaims code stability
5414 anyway, in this constellation we activate additional code
5415 backported from 0.9.9-dev for further performance improvements,
5416 namely BN_from_montgomery_word. (To enable this otherwise,
5417 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5419 [Andy Polyakov (backport partially by Bodo Moeller)]
5421 *) Add TLS session ticket callback. This allows an application to set
5422 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5423 values. This is useful for key rollover for example where several key
5424 sets may exist with different names.
5427 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5428 This was broken until now in 0.9.8 releases, such that the only way
5429 a registered ENGINE could be used (assuming it initialises
5430 successfully on the host) was to explicitly set it as the default
5431 for the relevant algorithms. This is in contradiction with 0.9.7
5432 behaviour and the documentation. With this fix, when an ENGINE is
5433 registered into a given algorithm's table of implementations, the
5434 'uptodate' flag is reset so that auto-discovery will be used next
5435 time a new context for that algorithm attempts to select an
5437 [Ian Lister (tweaked by Geoff Thorpe)]
5439 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5440 implementation in the following ways:
5442 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5445 Lack of BER streaming support means one pass streaming processing is
5446 only supported if data is detached: setting the streaming flag is
5447 ignored for embedded content.
5449 CMS support is disabled by default and must be explicitly enabled
5450 with the enable-cms configuration option.
5453 *) Update the GMP engine glue to do direct copies between BIGNUM and
5454 mpz_t when openssl and GMP use the same limb size. Otherwise the
5455 existing "conversion via a text string export" trick is still used.
5456 [Paul Sheer <paulsheer@gmail.com>]
5458 *) Zlib compression BIO. This is a filter BIO which compressed and
5459 uncompresses any data passed through it.
5462 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5463 RFC3394 compatible AES key wrapping.
5466 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5467 sets string data without copying. X509_ALGOR_set0() and
5468 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5469 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5470 from an X509_ATTRIBUTE structure optionally checking it occurs only
5471 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5475 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5476 to get the expected BN_FLG_CONSTTIME behavior.
5477 [Bodo Moeller (Google)]
5481 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5482 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5483 - added some more tests to do_tests.pl
5484 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5485 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5486 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5487 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5488 - various changes to netware.pl to enable gcc-cross builds on Win32
5490 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5491 - various changes to fix missing prototype warnings
5492 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5493 - added AES, WHIRLPOOL and CPUID assembler code to build files
5494 - added missing AES assembler make rules to mk1mf.pl
5495 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5496 [Guenter Knauf <eflash@gmx.net>]
5498 *) Implement certificate status request TLS extension defined in RFC3546.
5499 A client can set the appropriate parameters and receive the encoded
5500 OCSP response via a callback. A server can query the supplied parameters
5501 and set the encoded OCSP response in the callback. Add simplified examples
5502 to s_client and s_server.
5505 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5507 *) Fix various bugs:
5508 + Binary incompatibility of ssl_ctx_st structure
5509 + DTLS interoperation with non-compliant servers
5510 + Don't call get_session_cb() without proposed session
5511 + Fix ia64 assembler code
5512 [Andy Polyakov, Steve Henson]
5514 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5516 *) DTLS Handshake overhaul. There were longstanding issues with
5517 OpenSSL DTLS implementation, which were making it impossible for
5518 RFC 4347 compliant client to communicate with OpenSSL server.
5519 Unfortunately just fixing these incompatibilities would "cut off"
5520 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5521 server keeps tolerating non RFC compliant syntax. The opposite is
5522 not true, 0.9.8f client can not communicate with earlier server.
5523 This update even addresses CVE-2007-4995.
5526 *) Changes to avoid need for function casts in OpenSSL: some compilers
5527 (gcc 4.2 and later) reject their use.
5528 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5531 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5532 RFC4507bis. The encrypted ticket format is an encrypted encoded
5533 SSL_SESSION structure, that way new session features are automatically
5536 If a client application caches session in an SSL_SESSION structure
5537 support is transparent because tickets are now stored in the encoded
5540 The SSL_CTX structure automatically generates keys for ticket
5541 protection in servers so again support should be possible
5542 with no application modification.
5544 If a client or server wishes to disable RFC4507 support then the option
5545 SSL_OP_NO_TICKET can be set.
5547 Add a TLS extension debugging callback to allow the contents of any client
5548 or server extensions to be examined.
5550 This work was sponsored by Google.
5553 *) Add initial support for TLS extensions, specifically for the server_name
5554 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5555 have new members for a host name. The SSL data structure has an
5556 additional member SSL_CTX *initial_ctx so that new sessions can be
5557 stored in that context to allow for session resumption, even after the
5558 SSL has been switched to a new SSL_CTX in reaction to a client's
5559 server_name extension.
5561 New functions (subject to change):
5563 SSL_get_servername()
5564 SSL_get_servername_type()
5567 New CTRL codes and macros (subject to change):
5569 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5570 - SSL_CTX_set_tlsext_servername_callback()
5571 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5572 - SSL_CTX_set_tlsext_servername_arg()
5573 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5575 openssl s_client has a new '-servername ...' option.
5577 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5578 '-key2 ...', '-servername_fatal' (subject to change). This allows
5579 testing the HostName extension for a specific single host name ('-cert'
5580 and '-key' remain fallbacks for handshakes without HostName
5581 negotiation). If the unrecognized_name alert has to be sent, this by
5582 default is a warning; it becomes fatal with the '-servername_fatal'
5585 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5587 *) Add AES and SSE2 assembly language support to VC++ build.
5590 *) Mitigate attack on final subtraction in Montgomery reduction.
5593 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5594 (which previously caused an internal error).
5597 *) Squeeze another 10% out of IGE mode when in != out.
5600 *) AES IGE mode speedup.
5601 [Dean Gaudet (Google)]
5603 *) Add the Korean symmetric 128-bit cipher SEED (see
5604 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5605 add SEED ciphersuites from RFC 4162:
5607 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5608 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5609 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5610 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5612 To minimize changes between patchlevels in the OpenSSL 0.9.8
5613 series, SEED remains excluded from compilation unless OpenSSL
5614 is configured with 'enable-seed'.
5615 [KISA, Bodo Moeller]
5617 *) Mitigate branch prediction attacks, which can be practical if a
5618 single processor is shared, allowing a spy process to extract
5619 information. For detailed background information, see
5620 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5621 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5622 and Necessary Software Countermeasures"). The core of the change
5623 are new versions BN_div_no_branch() and
5624 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5625 respectively, which are slower, but avoid the security-relevant
5626 conditional branches. These are automatically called by BN_div()
5627 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5628 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5629 remove a conditional branch.
5631 BN_FLG_CONSTTIME is the new name for the previous
5632 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5633 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5634 in the exponent causes BN_mod_exp_mont() to use the alternative
5635 implementation in BN_mod_exp_mont_consttime().) The old name
5636 remains as a deprecated alias.
5638 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5639 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5640 constant-time implementations for more than just exponentiation.
5641 Here too the old name is kept as a deprecated alias.
5643 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5644 the BN_BLINDING structure gets an independent copy of the
5645 modulus. This means that the previous "BIGNUM *m" argument to
5646 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5647 essentially becomes "const BIGNUM *m", although we can't actually
5648 change this in the header file before 0.9.9. It allows
5649 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5650 enable BN_FLG_CONSTTIME.
5652 [Matthew D Wood (Intel Corp)]
5654 *) In the SSL/TLS server implementation, be strict about session ID
5655 context matching (which matters if an application uses a single
5656 external cache for different purposes). Previously,
5657 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5658 set. This did ensure strict client verification, but meant that,
5659 with applications using a single external cache for quite
5660 different requirements, clients could circumvent ciphersuite
5661 restrictions for a given session ID context by starting a session
5662 in a different context.
5665 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5666 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5667 authentication-only ciphersuites.
5670 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5671 not complete and could lead to a possible single byte overflow
5672 (CVE-2007-5135) [Ben Laurie]
5674 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5676 *) Since AES128 and AES256 (and similarly Camellia128 and
5677 Camellia256) share a single mask bit in the logic of
5678 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5679 kludge to work properly if AES128 is available and AES256 isn't
5680 (or if Camellia128 is available and Camellia256 isn't).
5683 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5684 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5685 When a point or a seed is encoded in a BIT STRING, we need to
5686 prevent the removal of trailing zero bits to get the proper DER
5687 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5688 of a NamedBitList, for which trailing 0 bits need to be removed.)
5691 *) Have SSL/TLS server implementation tolerate "mismatched" record
5692 protocol version while receiving ClientHello even if the
5693 ClientHello is fragmented. (The server can't insist on the
5694 particular protocol version it has chosen before the ServerHello
5695 message has informed the client about his choice.)
5698 *) Add RFC 3779 support.
5699 [Rob Austein for ARIN, Ben Laurie]
5701 *) Load error codes if they are not already present instead of using a
5702 static variable. This allows them to be cleanly unloaded and reloaded.
5703 Improve header file function name parsing.
5706 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5707 or CAPABILITY handshake as required by RFCs.
5710 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5712 *) Introduce limits to prevent malicious keys being able to
5713 cause a denial of service. (CVE-2006-2940)
5714 [Steve Henson, Bodo Moeller]
5716 *) Fix ASN.1 parsing of certain invalid structures that can result
5717 in a denial of service. (CVE-2006-2937) [Steve Henson]
5719 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5720 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5722 *) Fix SSL client code which could crash if connecting to a
5723 malicious SSLv2 server. (CVE-2006-4343)
5724 [Tavis Ormandy and Will Drewry, Google Security Team]
5726 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5727 match only those. Before that, "AES256-SHA" would be interpreted
5728 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5729 the same strength classification in 0.9.7h) as we currently only
5730 have a single AES bit in the ciphersuite description bitmap.
5731 That change, however, also applied to ciphersuite strings such as
5732 "RC4-MD5" that intentionally matched multiple ciphersuites --
5733 namely, SSL 2.0 ciphersuites in addition to the more common ones
5734 from SSL 3.0/TLS 1.0.
5736 So we change the selection algorithm again: Naming an explicit
5737 ciphersuite selects this one ciphersuite, and any other similar
5738 ciphersuite (same bitmap) from *other* protocol versions.
5739 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5740 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5742 Since SSL 2.0 does not have any ciphersuites for which the
5743 128/256 bit distinction would be relevant, this works for now.
5744 The proper fix will be to use different bits for AES128 and
5745 AES256, which would have avoided the problems from the beginning;
5746 however, bits are scarce, so we can only do this in a new release
5747 (not just a patchlevel) when we can change the SSL_CIPHER
5748 definition to split the single 'unsigned long mask' bitmap into
5749 multiple values to extend the available space.
5753 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5755 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5756 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5758 *) Add AES IGE and biIGE modes.
5761 *) Change the Unix randomness entropy gathering to use poll() when
5762 possible instead of select(), since the latter has some
5763 undesirable limitations.
5764 [Darryl Miles via Richard Levitte and Bodo Moeller]
5766 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5767 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5768 cannot be implicitly activated as part of, e.g., the "AES" alias.
5769 However, please upgrade to OpenSSL 0.9.9[-dev] for
5770 non-experimental use of the ECC ciphersuites to get TLS extension
5771 support, which is required for curve and point format negotiation
5772 to avoid potential handshake problems.
5775 *) Disable rogue ciphersuites:
5777 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5778 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5779 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5781 The latter two were purportedly from
5782 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5785 Also deactivate the remaining ciphersuites from
5786 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5787 unofficial, and the ID has long expired.
5790 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5791 dual-core machines) and other potential thread-safety issues.
5794 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5795 versions), which is now available for royalty-free use
5796 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5797 Also, add Camellia TLS ciphersuites from RFC 4132.
5799 To minimize changes between patchlevels in the OpenSSL 0.9.8
5800 series, Camellia remains excluded from compilation unless OpenSSL
5801 is configured with 'enable-camellia'.
5804 *) Disable the padding bug check when compression is in use. The padding
5805 bug check assumes the first packet is of even length, this is not
5806 necessarily true if compression is enabled and can result in false
5807 positives causing handshake failure. The actual bug test is ancient
5808 code so it is hoped that implementations will either have fixed it by
5809 now or any which still have the bug do not support compression.
5812 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5814 *) When applying a cipher rule check to see if string match is an explicit
5815 cipher suite and only match that one cipher suite if it is.
5818 *) Link in manifests for VC++ if needed.
5819 [Austin Ziegler <halostatue@gmail.com>]
5821 *) Update support for ECC-based TLS ciphersuites according to
5822 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5823 TLS extensions, which are supported starting with the 0.9.9
5824 branch, not in the OpenSSL 0.9.8 branch).
5827 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5828 opaque EVP_CIPHER_CTX handling.
5831 *) Fixes and enhancements to zlib compression code. We now only use
5832 "zlib1.dll" and use the default __cdecl calling convention on Win32
5833 to conform with the standards mentioned here:
5834 http://www.zlib.net/DLL_FAQ.txt
5835 Static zlib linking now works on Windows and the new --with-zlib-include
5836 --with-zlib-lib options to Configure can be used to supply the location
5837 of the headers and library. Gracefully handle case where zlib library
5841 *) Several fixes and enhancements to the OID generation code. The old code
5842 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5843 handle numbers larger than ULONG_MAX, truncated printing and had a
5844 non standard OBJ_obj2txt() behaviour.
5847 *) Add support for building of engines under engine/ as shared libraries
5848 under VC++ build system.
5851 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5852 Hopefully, we will not see any false combination of paths any more.
5855 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5857 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5858 (part of SSL_OP_ALL). This option used to disable the
5859 countermeasure against man-in-the-middle protocol-version
5860 rollback in the SSL 2.0 server implementation, which is a bad
5861 idea. (CVE-2005-2969)
5863 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5864 for Information Security, National Institute of Advanced Industrial
5865 Science and Technology [AIST], Japan)]
5867 *) Add two function to clear and return the verify parameter flags.
5870 *) Keep cipherlists sorted in the source instead of sorting them at
5871 runtime, thus removing the need for a lock.
5874 *) Avoid some small subgroup attacks in Diffie-Hellman.
5875 [Nick Mathewson and Ben Laurie]
5877 *) Add functions for well-known primes.
5880 *) Extended Windows CE support.
5881 [Satoshi Nakamura and Andy Polyakov]
5883 *) Initialize SSL_METHOD structures at compile time instead of during
5884 runtime, thus removing the need for a lock.
5887 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5888 attempting to decrypt each encrypted key in turn. Add support to
5892 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5894 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5897 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5900 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5901 key into the same file any more.
5904 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5907 *) Add -utf8 command line and config file option to 'ca'.
5908 [Stefan <stf@udoma.org]
5910 *) Removed the macro des_crypt(), as it seems to conflict with some
5911 libraries. Use DES_crypt().
5914 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5915 involves renaming the source and generated shared-libs for
5916 both. The engines will accept the corrected or legacy ids
5917 ('ncipher' and '4758_cca' respectively) when binding. NB,
5918 this only applies when building 'shared'.
5919 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5921 *) Add attribute functions to EVP_PKEY structure. Modify
5922 PKCS12_create() to recognize a CSP name attribute and
5923 use it. Make -CSP option work again in pkcs12 utility.
5926 *) Add new functionality to the bn blinding code:
5927 - automatic re-creation of the BN_BLINDING parameters after
5928 a fixed number of uses (currently 32)
5929 - add new function for parameter creation
5930 - introduce flags to control the update behaviour of the
5931 BN_BLINDING parameters
5932 - hide BN_BLINDING structure
5933 Add a second BN_BLINDING slot to the RSA structure to improve
5934 performance when a single RSA object is shared among several
5938 *) Add support for DTLS.
5939 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5941 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5942 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5945 *) Remove buggy and incomplete DH cert support from
5946 ssl/ssl_rsa.c and ssl/s3_both.c
5949 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5950 the apps/openssl applications.
5953 *) Compile clean with "-Wall -Wmissing-prototypes
5954 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5955 DEBUG_SAFESTACK must also be set.
5958 *) Change ./Configure so that certain algorithms can be disabled by default.
5959 The new counterpiece to "no-xxx" is "enable-xxx".
5961 The patented RC5 and MDC2 algorithms will now be disabled unless
5962 "enable-rc5" and "enable-mdc2", respectively, are specified.
5964 (IDEA remains enabled despite being patented. This is because IDEA
5965 is frequently required for interoperability, and there is no license
5966 fee for non-commercial use. As before, "no-idea" can be used to
5967 avoid this algorithm.)
5971 *) Add processing of proxy certificates (see RFC 3820). This work was
5972 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5973 EGEE (Enabling Grids for E-science in Europe).
5976 *) RC4 performance overhaul on modern architectures/implementations, such
5977 as Intel P4, IA-64 and AMD64.
5980 *) New utility extract-section.pl. This can be used specify an alternative
5981 section number in a pod file instead of having to treat each file as
5982 a separate case in Makefile. This can be done by adding two lines to the
5985 =for comment openssl_section:XXX
5987 The blank line is mandatory.
5991 *) New arguments -certform, -keyform and -pass for s_client and s_server
5992 to allow alternative format key and certificate files and passphrase
5996 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5997 update associated structures and add various utility functions.
5999 Add new policy related verify parameters, include policy checking in
6000 standard verify code. Enhance 'smime' application with extra parameters
6001 to support policy checking and print out.
6004 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
6005 Nehemiah processors. These extensions support AES encryption in hardware
6006 as well as RNG (though RNG support is currently disabled).
6007 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
6009 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
6012 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
6013 [Andy Polyakov and a number of other people]
6015 *) Improved PowerPC platform support. Most notably BIGNUM assembler
6016 implementation contributed by IBM.
6017 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
6019 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
6020 exponent rather than 'unsigned long'. There is a corresponding change to
6021 the new 'rsa_keygen' element of the RSA_METHOD structure.
6022 [Jelte Jansen, Geoff Thorpe]
6024 *) Functionality for creating the initial serial number file is now
6025 moved from CA.pl to the 'ca' utility with a new option -create_serial.
6027 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
6028 number file to 1, which is bound to cause problems. To avoid
6029 the problems while respecting compatibility between different 0.9.7
6030 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
6031 CA.pl for serial number initialization. With the new release 0.9.8,
6032 we can fix the problem directly in the 'ca' utility.)
6035 *) Reduced header interdependencies by declaring more opaque objects in
6036 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
6037 give fewer recursive includes, which could break lazy source code - so
6038 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
6039 developers should define this symbol when building and using openssl to
6040 ensure they track the recommended behaviour, interfaces, [etc], but
6041 backwards-compatible behaviour prevails when this isn't defined.
6044 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
6047 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
6048 This will generate a random key of the appropriate length based on the
6049 cipher context. The EVP_CIPHER can provide its own random key generation
6050 routine to support keys of a specific form. This is used in the des and
6051 3des routines to generate a key of the correct parity. Update S/MIME
6052 code to use new functions and hence generate correct parity DES keys.
6053 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
6054 valid (weak or incorrect parity).
6057 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
6058 as looking them up. This is useful when the verified structure may contain
6059 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
6060 present unless the new PKCS7_NO_CRL flag is asserted.
6063 *) Extend ASN1 oid configuration module. It now additionally accepts the
6066 shortName = some long name, 1.2.3.4
6069 *) Reimplemented the BN_CTX implementation. There is now no more static
6070 limitation on the number of variables it can handle nor the depth of the
6071 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
6072 information can now expand as required, and rather than having a single
6073 static array of bignums, BN_CTX now uses a linked-list of such arrays
6074 allowing it to expand on demand whilst maintaining the usefulness of
6075 BN_CTX's "bundling".
6078 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
6079 to allow all RSA operations to function using a single BN_CTX.
6082 *) Preliminary support for certificate policy evaluation and checking. This
6083 is initially intended to pass the tests outlined in "Conformance Testing
6084 of Relying Party Client Certificate Path Processing Logic" v1.07.
6087 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6088 remained unused and not that useful. A variety of other little bignum
6089 tweaks and fixes have also been made continuing on from the audit (see
6093 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6094 associated ASN1, EVP and SSL functions and old ASN1 macros.
6097 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6098 and this should never fail. So the return value from the use of
6099 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6100 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6103 *) BN_CTX_get() should return zero-valued bignums, providing the same
6104 initialised value as BN_new().
6105 [Geoff Thorpe, suggested by Ulf Möller]
6107 *) Support for inhibitAnyPolicy certificate extension.
6110 *) An audit of the BIGNUM code is underway, for which debugging code is
6111 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6112 is considered valid when processing BIGNUMs, and causes execution to
6113 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6114 further steps are taken to deliberately pollute unused data in BIGNUM
6115 structures to try and expose faulty code further on. For now, openssl will
6116 (in its default mode of operation) continue to tolerate the inconsistent
6117 forms that it has tolerated in the past, but authors and packagers should
6118 consider trying openssl and their own applications when compiled with
6119 these debugging symbols defined. It will help highlight potential bugs in
6120 their own code, and will improve the test coverage for OpenSSL itself. At
6121 some point, these tighter rules will become openssl's default to improve
6122 maintainability, though the assert()s and other overheads will remain only
6123 in debugging configurations. See bn.h for more details.
6124 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6126 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6127 that can only be obtained through BN_CTX_new() (which implicitly
6128 initialises it). The presence of this function only made it possible
6129 to overwrite an existing structure (and cause memory leaks).
6132 *) Because of the callback-based approach for implementing LHASH as a
6133 template type, lh_insert() adds opaque objects to hash-tables and
6134 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6135 to clean up those corresponding objects before destroying the hash table
6136 (and losing the object pointers). So some over-zealous constifications in
6137 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6138 objects as "const" and the lh_doall[_arg] callback wrappers are not
6139 prototyped to have "const" restrictions on the object pointers they are
6140 given (and so aren't required to cast them away any more).
6143 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6144 (speed) prefers to use its own implementation. The two implementations
6145 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6146 its object type properly exposed (MS_TM) instead of casting to/from "char
6147 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6148 aren't necessarily the greatest nomenclatures - but this is what was used
6149 internally to the implementation so I've used that for now.
6152 *) Ensure that deprecated functions do not get compiled when
6153 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6154 the self-tests were still using deprecated key-generation functions so
6155 these have been updated also.
6158 *) Reorganise PKCS#7 code to separate the digest location functionality
6159 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6160 New function PKCS7_set_digest() to set the digest type for PKCS#7
6161 digestedData type. Add additional code to correctly generate the
6162 digestedData type and add support for this type in PKCS7 initialization
6166 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6167 structure of type "other".
6170 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6171 sure the loop does correctly stop and breaking ("division by zero")
6172 modulus operations are not performed. The (pre-generated) prime
6173 table crypto/bn/bn_prime.h was already correct, but it could not be
6174 re-generated on some platforms because of the "division by zero"
6175 situation in the script.
6176 [Ralf S. Engelschall]
6178 *) Update support for ECC-based TLS ciphersuites according to
6179 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6180 SHA-1 now is only used for "small" curves (where the
6181 representation of a field element takes up to 24 bytes); for
6182 larger curves, the field element resulting from ECDH is directly
6183 used as premaster secret.
6184 [Douglas Stebila (Sun Microsystems Laboratories)]
6186 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6187 curve secp160r1 to the tests.
6188 [Douglas Stebila (Sun Microsystems Laboratories)]
6190 *) Add the possibility to load symbols globally with DSO.
6191 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6193 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6194 control of the error stack.
6197 *) Add support for STORE in ENGINE.
6200 *) Add the STORE type. The intention is to provide a common interface
6201 to certificate and key stores, be they simple file-based stores, or
6202 HSM-type store, or LDAP stores, or...
6203 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6206 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6207 pass a list of arguments to any function as well as provide a way
6208 for a function to pass data back to the caller.
6211 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6212 works like BUF_strdup() but can be used to duplicate a portion of
6213 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6217 *) Add the function sk_find_ex() which works like sk_find(), but will
6218 return an index to an element even if an exact match couldn't be
6219 found. The index is guaranteed to point at the element where the
6220 searched-for key would be inserted to preserve sorting order.
6223 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6224 takes an extra flags argument for optional functionality. Currently,
6225 the following flags are defined:
6227 OBJ_BSEARCH_VALUE_ON_NOMATCH
6228 This one gets OBJ_bsearch_ex() to return a pointer to the first
6229 element where the comparing function returns a negative or zero
6232 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6233 This one gets OBJ_bsearch_ex() to return a pointer to the first
6234 element where the comparing function returns zero. This is useful
6235 if there are more than one element where the comparing function
6239 *) Make it possible to create self-signed certificates with 'openssl ca'
6240 in such a way that the self-signed certificate becomes part of the
6241 CA database and uses the same mechanisms for serial number generation
6242 as all other certificate signing. The new flag '-selfsign' enables
6243 this functionality. Adapt CA.sh and CA.pl.in.
6246 *) Add functionality to check the public key of a certificate request
6247 against a given private. This is useful to check that a certificate
6248 request can be signed by that key (self-signing).
6251 *) Make it possible to have multiple active certificates with the same
6252 subject in the CA index file. This is done only if the keyword
6253 'unique_subject' is set to 'no' in the main CA section (default
6254 if 'CA_default') of the configuration file. The value is saved
6255 with the database itself in a separate index attribute file,
6256 named like the index file with '.attr' appended to the name.
6259 *) Generate multi-valued AVAs using '+' notation in config files for
6263 *) Support for nameConstraints certificate extension.
6266 *) Support for policyConstraints certificate extension.
6269 *) Support for policyMappings certificate extension.
6272 *) Make sure the default DSA_METHOD implementation only uses its
6273 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6274 and change its own handlers to be NULL so as to remove unnecessary
6275 indirection. This lets alternative implementations fallback to the
6276 default implementation more easily.
6279 *) Support for directoryName in GeneralName related extensions
6283 *) Make it possible to link applications using Makefile.shared.
6284 Make that possible even when linking against static libraries!
6287 *) Support for single pass processing for S/MIME signing. This now
6288 means that S/MIME signing can be done from a pipe, in addition
6289 cleartext signing (multipart/signed type) is effectively streaming
6290 and the signed data does not need to be all held in memory.
6292 This is done with a new flag PKCS7_STREAM. When this flag is set
6293 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6294 is done after the data is output (and digests calculated) in
6295 SMIME_write_PKCS7().
6298 *) Add full support for -rpath/-R, both in shared libraries and
6299 applications, at least on the platforms where it's known how
6303 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6304 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6305 will now compute a table of multiples of the generator that
6306 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6307 faster (notably in the case of a single point multiplication,
6308 scalar * generator).
6309 [Nils Larsch, Bodo Moeller]
6311 *) IPv6 support for certificate extensions. The various extensions
6312 which use the IP:a.b.c.d can now take IPv6 addresses using the
6313 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6317 *) Added an ENGINE that implements RSA by performing private key
6318 exponentiations with the GMP library. The conversions to and from
6319 GMP's mpz_t format aren't optimised nor are any montgomery forms
6320 cached, and on x86 it appears OpenSSL's own performance has caught up.
6321 However there are likely to be other architectures where GMP could
6322 provide a boost. This ENGINE is not built in by default, but it can be
6323 specified at Configure time and should be accompanied by the necessary
6324 linker additions, eg;
6325 ./config -DOPENSSL_USE_GMP -lgmp
6328 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6329 testing availability of engines with "-t" - the old behaviour is
6330 produced by increasing the feature's verbosity with "-tt".
6333 *) ECDSA routines: under certain error conditions uninitialized BN objects
6334 could be freed. Solution: make sure initialization is performed early
6335 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6339 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6340 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6341 software implementations. For DSA and DH, parameter generation can
6342 also be overridden by providing the appropriate method callbacks.
6345 *) Change the "progress" mechanism used in key-generation and
6346 primality testing to functions that take a new BN_GENCB pointer in
6347 place of callback/argument pairs. The new API functions have "_ex"
6348 postfixes and the older functions are reimplemented as wrappers for
6349 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6350 declarations of the old functions to help (graceful) attempts to
6351 migrate to the new functions. Also, the new key-generation API
6352 functions operate on a caller-supplied key-structure and return
6353 success/failure rather than returning a key or NULL - this is to
6354 help make "keygen" another member function of RSA_METHOD etc.
6356 Example for using the new callback interface:
6358 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6362 BN_GENCB_set(&my_cb, my_callback, my_arg);
6364 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6365 /* For the meaning of a, b in calls to my_callback(), see the
6366 * documentation of the function that calls the callback.
6367 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6368 * my_callback should return 1 if it wants BN_is_prime_ex()
6369 * to continue, or 0 to stop.
6374 *) Change the ZLIB compression method to be stateful, and make it
6375 available to TLS with the number defined in
6376 draft-ietf-tls-compression-04.txt.
6379 *) Add the ASN.1 structures and functions for CertificatePair, which
6380 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6382 CertificatePair ::= SEQUENCE {
6383 forward [0] Certificate OPTIONAL,
6384 reverse [1] Certificate OPTIONAL,
6385 -- at least one of the pair shall be present -- }
6387 Also implement the PEM functions to read and write certificate
6388 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6390 This needed to be defined, mostly for the sake of the LDAP
6391 attribute crossCertificatePair, but may prove useful elsewhere as
6395 *) Make it possible to inhibit symlinking of shared libraries in
6396 Makefile.shared, for Cygwin's sake.
6399 *) Extend the BIGNUM API by creating a function
6400 void BN_set_negative(BIGNUM *a, int neg);
6401 and a macro that behave like
6402 int BN_is_negative(const BIGNUM *a);
6404 to avoid the need to access 'a->neg' directly in applications.
6407 *) Implement fast modular reduction for pseudo-Mersenne primes
6408 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6409 EC_GROUP_new_curve_GFp() will now automatically use this
6411 [Nils Larsch <nla@trustcenter.de>]
6413 *) Add new lock type (CRYPTO_LOCK_BN).
6416 *) Change the ENGINE framework to automatically load engines
6417 dynamically from specific directories unless they could be
6418 found to already be built in or loaded. Move all the
6419 current engines except for the cryptodev one to a new
6421 The engines in engines/ are built as shared libraries if
6422 the "shared" options was given to ./Configure or ./config.
6423 Otherwise, they are inserted in libcrypto.a.
6424 /usr/local/ssl/engines is the default directory for dynamic
6425 engines, but that can be overridden at configure time through
6426 the usual use of --prefix and/or --openssldir, and at run
6427 time with the environment variable OPENSSL_ENGINES.
6428 [Geoff Thorpe and Richard Levitte]
6430 *) Add Makefile.shared, a helper makefile to build shared
6431 libraries. Adapt Makefile.org.
6434 *) Add version info to Win32 DLLs.
6435 [Peter 'Luna' Runestig" <peter@runestig.com>]
6437 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6438 can be added using this API to created arbitrary PKCS#12
6439 files while avoiding the low level API.
6441 New options to PKCS12_create(), key or cert can be NULL and
6442 will then be omitted from the output file. The encryption
6443 algorithm NIDs can be set to -1 for no encryption, the mac
6444 iteration count can be set to 0 to omit the mac.
6446 Enhance pkcs12 utility by making the -nokeys and -nocerts
6447 options work when creating a PKCS#12 file. New option -nomac
6448 to omit the mac, NONE can be set for an encryption algorithm.
6449 New code is modified to use the enhanced PKCS12_create()
6450 instead of the low level API.
6453 *) Extend ASN1 encoder to support indefinite length constructed
6454 encoding. This can output sequences tags and octet strings in
6455 this form. Modify pk7_asn1.c to support indefinite length
6456 encoding. This is experimental and needs additional code to
6457 be useful, such as an ASN1 bio and some enhanced streaming
6460 Extend template encode functionality so that tagging is passed
6461 down to the template encoder.
6464 *) Let 'openssl req' fail if an argument to '-newkey' is not
6465 recognized instead of using RSA as a default.
6468 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6469 As these are not official, they are not included in "ALL";
6470 the "ECCdraft" ciphersuite group alias can be used to select them.
6471 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6473 *) Add ECDH engine support.
6474 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6476 *) Add ECDH in new directory crypto/ecdh/.
6477 [Douglas Stebila (Sun Microsystems Laboratories)]
6479 *) Let BN_rand_range() abort with an error after 100 iterations
6480 without success (which indicates a broken PRNG).
6483 *) Change BN_mod_sqrt() so that it verifies that the input value
6484 is really the square of the return value. (Previously,
6485 BN_mod_sqrt would show GIGO behaviour.)
6488 *) Add named elliptic curves over binary fields from X9.62, SECG,
6489 and WAP/WTLS; add OIDs that were still missing.
6491 [Sheueling Chang Shantz and Douglas Stebila
6492 (Sun Microsystems Laboratories)]
6494 *) Extend the EC library for elliptic curves over binary fields
6495 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6498 EC_GF2m_simple_method
6502 EC_GROUP_new_curve_GF2m
6503 EC_GROUP_set_curve_GF2m
6504 EC_GROUP_get_curve_GF2m
6505 EC_POINT_set_affine_coordinates_GF2m
6506 EC_POINT_get_affine_coordinates_GF2m
6507 EC_POINT_set_compressed_coordinates_GF2m
6509 Point compression for binary fields is disabled by default for
6510 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6513 As binary polynomials are represented as BIGNUMs, various members
6514 of the EC_GROUP and EC_POINT data structures can be shared
6515 between the implementations for prime fields and binary fields;
6516 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6517 are essentially identical to their ..._GFp counterparts.
6518 (For simplicity, the '..._GFp' prefix has been dropped from
6519 various internal method names.)
6521 An internal 'field_div' method (similar to 'field_mul' and
6522 'field_sqr') has been added; this is used only for binary fields.
6524 [Sheueling Chang Shantz and Douglas Stebila
6525 (Sun Microsystems Laboratories)]
6527 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6528 through methods ('mul', 'precompute_mult').
6530 The generic implementations (now internally called 'ec_wNAF_mul'
6531 and 'ec_wNAF_precomputed_mult') remain the default if these
6532 methods are undefined.
6534 [Sheueling Chang Shantz and Douglas Stebila
6535 (Sun Microsystems Laboratories)]
6537 *) New function EC_GROUP_get_degree, which is defined through
6538 EC_METHOD. For curves over prime fields, this returns the bit
6539 length of the modulus.
6541 [Sheueling Chang Shantz and Douglas Stebila
6542 (Sun Microsystems Laboratories)]
6544 *) New functions EC_GROUP_dup, EC_POINT_dup.
6545 (These simply call ..._new and ..._copy).
6547 [Sheueling Chang Shantz and Douglas Stebila
6548 (Sun Microsystems Laboratories)]
6550 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6551 Polynomials are represented as BIGNUMs (where the sign bit is not
6552 used) in the following functions [macros]:
6555 BN_GF2m_sub [= BN_GF2m_add]
6556 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6557 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6558 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6560 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6561 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6562 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6563 BN_GF2m_cmp [= BN_ucmp]
6565 (Note that only the 'mod' functions are actually for fields GF(2^m).
6566 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6568 For some functions, an the irreducible polynomial defining a
6569 field can be given as an 'unsigned int[]' with strictly
6570 decreasing elements giving the indices of those bits that are set;
6571 i.e., p[] represents the polynomial
6572 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6574 p[0] > p[1] > ... > p[k] = 0.
6575 This applies to the following functions:
6580 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6581 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6583 BN_GF2m_mod_sqrt_arr
6584 BN_GF2m_mod_solve_quad_arr
6588 Conversion can be performed by the following functions:
6593 bntest.c has additional tests for binary polynomial arithmetic.
6595 Two implementations for BN_GF2m_mod_div() are available.
6596 The default algorithm simply uses BN_GF2m_mod_inv() and
6597 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6598 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6599 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6601 [Sheueling Chang Shantz and Douglas Stebila
6602 (Sun Microsystems Laboratories)]
6604 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6605 functionality is disabled at compile-time.
6606 [Douglas Stebila <douglas.stebila@sun.com>]
6608 *) Change default behaviour of 'openssl asn1parse' so that more
6609 information is visible when viewing, e.g., a certificate:
6611 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6612 mode the content of non-printable OCTET STRINGs is output in a
6613 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6614 avoid the appearance of a printable string.
6615 [Nils Larsch <nla@trustcenter.de>]
6617 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6619 EC_GROUP_set_asn1_flag()
6620 EC_GROUP_get_asn1_flag()
6621 EC_GROUP_set_point_conversion_form()
6622 EC_GROUP_get_point_conversion_form()
6623 These control ASN1 encoding details:
6624 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6625 has been set to OPENSSL_EC_NAMED_CURVE.
6626 - Points are encoded in uncompressed form by default; options for
6627 asn1_for are as for point2oct, namely
6628 POINT_CONVERSION_COMPRESSED
6629 POINT_CONVERSION_UNCOMPRESSED
6630 POINT_CONVERSION_HYBRID
6632 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6635 EC_GROUP_get0_seed()
6636 EC_GROUP_get_seed_len()
6637 This is used only for ASN1 purposes (so far).
6638 [Nils Larsch <nla@trustcenter.de>]
6640 *) Add 'field_type' member to EC_METHOD, which holds the NID
6641 of the appropriate field type OID. The new function
6642 EC_METHOD_get_field_type() returns this value.
6643 [Nils Larsch <nla@trustcenter.de>]
6648 EC_POINT_point2hex()
6649 EC_POINT_hex2point()
6650 providing useful interfaces to EC_POINT_point2oct() and
6651 EC_POINT_oct2point().
6652 [Nils Larsch <nla@trustcenter.de>]
6654 *) Change internals of the EC library so that the functions
6655 EC_GROUP_set_generator()
6656 EC_GROUP_get_generator()
6657 EC_GROUP_get_order()
6658 EC_GROUP_get_cofactor()
6659 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6660 to methods, which would lead to unnecessary code duplication when
6661 adding different types of curves.
6662 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6664 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6665 arithmetic, and such that modified wNAFs are generated
6666 (which avoid length expansion in many cases).
6669 *) Add a function EC_GROUP_check_discriminant() (defined via
6670 EC_METHOD) that verifies that the curve discriminant is non-zero.
6672 Add a function EC_GROUP_check() that makes some sanity tests
6673 on a EC_GROUP, its generator and order. This includes
6674 EC_GROUP_check_discriminant().
6675 [Nils Larsch <nla@trustcenter.de>]
6677 *) Add ECDSA in new directory crypto/ecdsa/.
6679 Add applications 'openssl ecparam' and 'openssl ecdsa'
6680 (these are based on 'openssl dsaparam' and 'openssl dsa').
6682 ECDSA support is also included in various other files across the
6683 library. Most notably,
6684 - 'openssl req' now has a '-newkey ecdsa:file' option;
6685 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6686 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6687 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6688 them suitable for ECDSA where domain parameters must be
6689 extracted before the specific public key;
6690 - ECDSA engine support has been added.
6691 [Nils Larsch <nla@trustcenter.de>]
6693 *) Include some named elliptic curves, and add OIDs from X9.62,
6694 SECG, and WAP/WTLS. Each curve can be obtained from the new
6696 EC_GROUP_new_by_curve_name(),
6697 and the list of available named curves can be obtained with
6698 EC_get_builtin_curves().
6699 Also add a 'curve_name' member to EC_GROUP objects, which can be
6701 EC_GROUP_set_curve_name()
6702 EC_GROUP_get_curve_name()
6703 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6705 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6706 was actually never needed) and in BN_mul(). The removal in BN_mul()
6707 required a small change in bn_mul_part_recursive() and the addition
6708 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6709 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6710 bn_sub_words() and bn_add_words() except they take arrays with
6714 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6716 *) Cleanse PEM buffers before freeing them since they may contain
6718 [Benjamin Bennett <ben@psc.edu>]
6720 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6721 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6722 authentication-only ciphersuites.
6725 *) Since AES128 and AES256 share a single mask bit in the logic of
6726 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6727 kludge to work properly if AES128 is available and AES256 isn't.
6730 *) Expand security boundary to match 1.1.1 module.
6733 *) Remove redundant features: hash file source, editing of test vectors
6734 modify fipsld to use external fips_premain.c signature.
6737 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6738 run algorithm test programs.
6741 *) Make algorithm test programs more tolerant of whitespace.
6744 *) Have SSL/TLS server implementation tolerate "mismatched" record
6745 protocol version while receiving ClientHello even if the
6746 ClientHello is fragmented. (The server can't insist on the
6747 particular protocol version it has chosen before the ServerHello
6748 message has informed the client about his choice.)
6751 *) Load error codes if they are not already present instead of using a
6752 static variable. This allows them to be cleanly unloaded and reloaded.
6755 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6757 *) Introduce limits to prevent malicious keys being able to
6758 cause a denial of service. (CVE-2006-2940)
6759 [Steve Henson, Bodo Moeller]
6761 *) Fix ASN.1 parsing of certain invalid structures that can result
6762 in a denial of service. (CVE-2006-2937) [Steve Henson]
6764 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6765 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6767 *) Fix SSL client code which could crash if connecting to a
6768 malicious SSLv2 server. (CVE-2006-4343)
6769 [Tavis Ormandy and Will Drewry, Google Security Team]
6771 *) Change ciphersuite string processing so that an explicit
6772 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6773 will no longer include "AES128-SHA"), and any other similar
6774 ciphersuite (same bitmap) from *other* protocol versions (so that
6775 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6776 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6777 changes from 0.9.8b and 0.9.8d.
6780 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6782 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6783 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6785 *) Change the Unix randomness entropy gathering to use poll() when
6786 possible instead of select(), since the latter has some
6787 undesirable limitations.
6788 [Darryl Miles via Richard Levitte and Bodo Moeller]
6790 *) Disable rogue ciphersuites:
6792 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6793 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6794 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6796 The latter two were purportedly from
6797 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6800 Also deactivate the remaining ciphersuites from
6801 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6802 unofficial, and the ID has long expired.
6805 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6806 dual-core machines) and other potential thread-safety issues.
6809 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6811 *) Adapt fipsld and the build system to link against the validated FIPS
6812 module in FIPS mode.
6815 *) Fixes for VC++ 2005 build under Windows.
6818 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6819 from a Windows bash shell such as MSYS. It is autodetected from the
6820 "config" script when run from a VC++ environment. Modify standard VC++
6821 build to use fipscanister.o from the GNU make build.
6824 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6826 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6827 The value now differs depending on if you build for FIPS or not.
6828 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6829 safely run with a non-FIPSed libcrypto, as it may crash because of
6830 the difference induced by this change.
6833 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6835 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6836 (part of SSL_OP_ALL). This option used to disable the
6837 countermeasure against man-in-the-middle protocol-version
6838 rollback in the SSL 2.0 server implementation, which is a bad
6839 idea. (CVE-2005-2969)
6841 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6842 for Information Security, National Institute of Advanced Industrial
6843 Science and Technology [AIST], Japan)]
6845 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6846 mainly for FIPS compliance and not fully integrated at this stage.
6849 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6850 the exponentiation using a fixed-length exponent. (Otherwise,
6851 the information leaked through timing could expose the secret key
6852 after many signatures; cf. Bleichenbacher's attack on DSA with
6856 *) Make a new fixed-window mod_exp implementation the default for
6857 RSA, DSA, and DH private-key operations so that the sequence of
6858 squares and multiplies and the memory access pattern are
6859 independent of the particular secret key. This will mitigate
6860 cache-timing and potential related attacks.
6862 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6863 and this is automatically used by BN_mod_exp_mont() if the new flag
6864 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6865 will use this BN flag for private exponents unless the flag
6866 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6867 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6869 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6871 *) Change the client implementation for SSLv23_method() and
6872 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6873 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6874 (Previously, the SSL 2.0 backwards compatible Client Hello
6875 message format would be used even with SSL_OP_NO_SSLv2.)
6878 *) Add support for smime-type MIME parameter in S/MIME messages which some
6882 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6883 a threadsafe manner. Modify rsa code to use new function and add calls
6884 to dsa and dh code (which had race conditions before).
6887 *) Include the fixed error library code in the C error file definitions
6888 instead of fixing them up at runtime. This keeps the error code
6889 structures constant.
6892 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6894 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6897 *) Fixes for newer kerberos headers. NB: the casts are needed because
6898 the 'length' field is signed on one version and unsigned on another
6899 with no (?) obvious way to tell the difference, without these VC++
6900 complains. Also the "definition" of FAR (blank) is no longer included
6901 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6902 some needed definitions.
6905 *) Undo Cygwin change.
6908 *) Added support for proxy certificates according to RFC 3820.
6909 Because they may be a security thread to unaware applications,
6910 they must be explicitly allowed in run-time. See
6911 docs/HOWTO/proxy_certificates.txt for further information.
6914 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6916 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6917 server and client random values. Previously
6918 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6919 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6921 This change has negligible security impact because:
6923 1. Server and client random values still have 24 bytes of pseudo random
6926 2. Server and client random values are sent in the clear in the initial
6929 3. The master secret is derived using the premaster secret (48 bytes in
6930 size for static RSA ciphersuites) as well as client server and random
6933 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6936 [Stephen Henson, reported by UK NISCC]
6938 *) Use Windows randomness collection on Cygwin.
6941 *) Fix hang in EGD/PRNGD query when communication socket is closed
6942 prematurely by EGD/PRNGD.
6943 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6945 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6948 *) Back-port of selected performance improvements from development
6949 branch, as well as improved support for PowerPC platforms.
6952 *) Add lots of checks for memory allocation failure, error codes to indicate
6953 failure and freeing up memory if a failure occurs.
6954 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6956 *) Add new -passin argument to dgst.
6959 *) Perform some character comparisons of different types in X509_NAME_cmp:
6960 this is needed for some certificates that re-encode DNs into UTF8Strings
6961 (in violation of RFC3280) and can't or won't issue name rollover
6965 *) Make an explicit check during certificate validation to see that
6966 the CA setting in each certificate on the chain is correct. As a
6967 side effect always do the following basic checks on extensions,
6968 not just when there's an associated purpose to the check:
6970 - if there is an unhandled critical extension (unless the user
6971 has chosen to ignore this fault)
6972 - if the path length has been exceeded (if one is set at all)
6973 - that certain extensions fit the associated purpose (if one has
6977 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6979 *) Avoid a race condition when CRLs are checked in a multi threaded
6980 environment. This would happen due to the reordering of the revoked
6981 entries during signature checking and serial number lookup. Now the
6982 encoding is cached and the serial number sort performed under a lock.
6983 Add new STACK function sk_is_sorted().
6986 *) Add Delta CRL to the extension code.
6989 *) Various fixes to s3_pkt.c so alerts are sent properly.
6990 [David Holmes <d.holmes@f5.com>]
6992 *) Reduce the chances of duplicate issuer name and serial numbers (in
6993 violation of RFC3280) using the OpenSSL certificate creation utilities.
6994 This is done by creating a random 64 bit value for the initial serial
6995 number when a serial number file is created or when a self signed
6996 certificate is created using 'openssl req -x509'. The initial serial
6997 number file is created using 'openssl x509 -next_serial' in CA.pl
6998 rather than being initialized to 1.
7001 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
7003 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7004 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7005 [Joe Orton, Steve Henson]
7007 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
7009 [Joe Orton, Steve Henson]
7011 *) Make it possible to have multiple active certificates with the same
7012 subject in the CA index file. This is done only if the keyword
7013 'unique_subject' is set to 'no' in the main CA section (default
7014 if 'CA_default') of the configuration file. The value is saved
7015 with the database itself in a separate index attribute file,
7016 named like the index file with '.attr' appended to the name.
7019 *) X509 verify fixes. Disable broken certificate workarounds when
7020 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
7021 keyUsage extension present. Don't accept CRLs with unhandled critical
7022 extensions: since verify currently doesn't process CRL extensions this
7023 rejects a CRL with *any* critical extensions. Add new verify error codes
7027 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
7028 A clarification of RFC2560 will require the use of OCTET STRINGs and
7029 some implementations cannot handle the current raw format. Since OpenSSL
7030 copies and compares OCSP nonces as opaque blobs without any attempt at
7031 parsing them this should not create any compatibility issues.
7034 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
7035 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
7036 this HMAC (and other) operations are several times slower than OpenSSL
7040 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
7041 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7043 *) Use the correct content when signing type "other".
7046 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
7048 *) Fix various bugs revealed by running the NISCC test suite:
7050 Stop out of bounds reads in the ASN1 code when presented with
7051 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7053 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
7055 If verify callback ignores invalid public key errors don't try to check
7056 certificate signature with the NULL public key.
7060 *) New -ignore_err option in ocsp application to stop the server
7061 exiting on the first error in a request.
7064 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7065 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7069 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7070 extra data after the compression methods not only for TLS 1.0
7071 but also for SSL 3.0 (as required by the specification).
7072 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7074 *) Change X509_certificate_type() to mark the key as exported/exportable
7075 when it's 512 *bits* long, not 512 bytes.
7078 *) Change AES_cbc_encrypt() so it outputs exact multiple of
7079 blocks during encryption.
7082 *) Various fixes to base64 BIO and non blocking I/O. On write
7083 flushes were not handled properly if the BIO retried. On read
7084 data was not being buffered properly and had various logic bugs.
7085 This also affects blocking I/O when the data being decoded is a
7089 *) Various S/MIME bugfixes and compatibility changes:
7090 output correct application/pkcs7 MIME type if
7091 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7092 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7093 of files as .eml work). Correctly handle very long lines in MIME
7097 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7099 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7100 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7101 a protocol version number mismatch like a decryption error
7102 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7105 *) Turn on RSA blinding by default in the default implementation
7106 to avoid a timing attack. Applications that don't want it can call
7107 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7108 They would be ill-advised to do so in most cases.
7109 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7111 *) Change RSA blinding code so that it works when the PRNG is not
7112 seeded (in this case, the secret RSA exponent is abused as
7113 an unpredictable seed -- if it is not unpredictable, there
7114 is no point in blinding anyway). Make RSA blinding thread-safe
7115 by remembering the creator's thread ID in rsa->blinding and
7116 having all other threads use local one-time blinding factors
7117 (this requires more computation than sharing rsa->blinding, but
7118 avoids excessive locking; and if an RSA object is not shared
7119 between threads, blinding will still be very fast).
7122 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7123 ENGINE as defaults for all supported algorithms irrespective of
7124 the 'flags' parameter. 'flags' is now honoured, so applications
7125 should make sure they are passing it correctly.
7128 *) Target "mingw" now allows native Windows code to be generated in
7129 the Cygwin environment as well as with the MinGW compiler.
7132 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7134 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7135 via timing by performing a MAC computation even if incorrect
7136 block cipher padding has been found. This is a countermeasure
7137 against active attacks where the attacker has to distinguish
7138 between bad padding and a MAC verification error. (CVE-2003-0078)
7140 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7141 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7142 Martin Vuagnoux (EPFL, Ilion)]
7144 *) Make the no-err option work as intended. The intention with no-err
7145 is not to have the whole error stack handling routines removed from
7146 libcrypto, it's only intended to remove all the function name and
7147 reason texts, thereby removing some of the footprint that may not
7148 be interesting if those errors aren't displayed anyway.
7150 NOTE: it's still possible for any application or module to have its
7151 own set of error texts inserted. The routines are there, just not
7152 used by default when no-err is given.
7155 *) Add support for FreeBSD on IA64.
7156 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7158 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7159 Kerberos function mit_des_cbc_cksum(). Before this change,
7160 the value returned by DES_cbc_cksum() was like the one from
7161 mit_des_cbc_cksum(), except the bytes were swapped.
7162 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7164 *) Allow an application to disable the automatic SSL chain building.
7165 Before this a rather primitive chain build was always performed in
7166 ssl3_output_cert_chain(): an application had no way to send the
7167 correct chain if the automatic operation produced an incorrect result.
7169 Now the chain builder is disabled if either:
7171 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7173 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7175 The reasoning behind this is that an application would not want the
7176 auto chain building to take place if extra chain certificates are
7177 present and it might also want a means of sending no additional
7178 certificates (for example the chain has two certificates and the
7182 *) Add the possibility to build without the ENGINE framework.
7183 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7185 *) Under Win32 gmtime() can return NULL: check return value in
7186 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7189 *) DSA routines: under certain error conditions uninitialized BN objects
7190 could be freed. Solution: make sure initialization is performed early
7191 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7192 Nils Larsch <nla@trustcenter.de> via PR#459)
7195 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7196 checked on reconnect on the client side, therefore session resumption
7197 could still fail with a "ssl session id is different" error. This
7198 behaviour is masked when SSL_OP_ALL is used due to
7199 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7200 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7201 followup to PR #377.
7204 *) IA-32 assembler support enhancements: unified ELF targets, support
7205 for SCO/Caldera platforms, fix for Cygwin shared build.
7208 *) Add support for FreeBSD on sparc64. As a consequence, support for
7209 FreeBSD on non-x86 processors is separate from x86 processors on
7210 the config script, much like the NetBSD support.
7211 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7213 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7215 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7218 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7219 code (06) was taken as the first octet of the session ID and the last
7220 octet was ignored consequently. As a result SSLv2 client side session
7221 caching could not have worked due to the session ID mismatch between
7223 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7227 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7228 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7232 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7233 seems that in spite of existing for more than a year, many application
7234 author have done nothing to provide the necessary callbacks, which
7235 means that this particular engine will not work properly anywhere.
7236 This is a very unfortunate situation which forces us, in the name
7237 of usability, to give the hw_ncipher.c a static lock, which is part
7239 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7240 appear in 0.9.8 or later. We EXPECT application authors to have
7241 dealt properly with this when 0.9.8 is released (unless we actually
7242 make such changes in the libcrypto locking code that changes will
7243 have to be made anyway).
7246 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7247 octets have been read, EOF or an error occurs. Without this change
7248 some truncated ASN1 structures will not produce an error.
7251 *) Disable Heimdal support, since it hasn't been fully implemented.
7252 Still give the possibility to force the use of Heimdal, but with
7253 warnings and a request that patches get sent to openssl-dev.
7256 *) Add the VC-CE target, introduce the WINCE sysname, and add
7257 INSTALL.WCE and appropriate conditionals to make it build.
7258 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7260 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7261 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7262 edit numbers of the version.
7263 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7265 *) Introduce safe string copy and catenation functions
7266 (BUF_strlcpy() and BUF_strlcat()).
7267 [Ben Laurie (CHATS) and Richard Levitte]
7269 *) Avoid using fixed-size buffers for one-line DNs.
7270 [Ben Laurie (CHATS)]
7272 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7273 resizing buffers containing secrets, and use where appropriate.
7274 [Ben Laurie (CHATS)]
7276 *) Avoid using fixed size buffers for configuration file location.
7277 [Ben Laurie (CHATS)]
7279 *) Avoid filename truncation for various CA files.
7280 [Ben Laurie (CHATS)]
7282 *) Use sizeof in preference to magic numbers.
7283 [Ben Laurie (CHATS)]
7285 *) Avoid filename truncation in cert requests.
7286 [Ben Laurie (CHATS)]
7288 *) Add assertions to check for (supposedly impossible) buffer
7290 [Ben Laurie (CHATS)]
7292 *) Don't cache truncated DNS entries in the local cache (this could
7293 potentially lead to a spoofing attack).
7294 [Ben Laurie (CHATS)]
7296 *) Fix various buffers to be large enough for hex/decimal
7297 representations in a platform independent manner.
7298 [Ben Laurie (CHATS)]
7300 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7301 resizing buffers containing secrets, and use where appropriate.
7302 [Ben Laurie (CHATS)]
7304 *) Add BIO_indent() to avoid much slightly worrying code to do
7306 [Ben Laurie (CHATS)]
7308 *) Convert sprintf()/BIO_puts() to BIO_printf().
7309 [Ben Laurie (CHATS)]
7311 *) buffer_gets() could terminate with the buffer only half
7313 [Ben Laurie (CHATS)]
7315 *) Add assertions to prevent user-supplied crypto functions from
7316 overflowing internal buffers by having large block sizes, etc.
7317 [Ben Laurie (CHATS)]
7319 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7321 [Ben Laurie (CHATS)]
7323 *) Eliminate unused copy of key in RC4.
7324 [Ben Laurie (CHATS)]
7326 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7327 [Ben Laurie (CHATS)]
7329 *) Fix off-by-one error in EGD path.
7330 [Ben Laurie (CHATS)]
7332 *) If RANDFILE path is too long, ignore instead of truncating.
7333 [Ben Laurie (CHATS)]
7335 *) Eliminate unused and incorrectly sized X.509 structure
7337 [Ben Laurie (CHATS)]
7339 *) Eliminate unused and dangerous function knumber().
7340 [Ben Laurie (CHATS)]
7342 *) Eliminate unused and dangerous structure, KSSL_ERR.
7343 [Ben Laurie (CHATS)]
7345 *) Protect against overlong session ID context length in an encoded
7346 session object. Since these are local, this does not appear to be
7348 [Ben Laurie (CHATS)]
7350 *) Change from security patch (see 0.9.6e below) that did not affect
7351 the 0.9.6 release series:
7353 Remote buffer overflow in SSL3 protocol - an attacker could
7354 supply an oversized master key in Kerberos-enabled versions.
7356 [Ben Laurie (CHATS)]
7358 *) Change the SSL kerb5 codes to match RFC 2712.
7361 *) Make -nameopt work fully for req and add -reqopt switch.
7362 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7364 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7365 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7367 *) Make sure tests can be performed even if the corresponding algorithms
7368 have been removed entirely. This was also the last step to make
7369 OpenSSL compilable with DJGPP under all reasonable conditions.
7370 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7372 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7373 to allow version independent disabling of normally unselected ciphers,
7374 which may be activated as a side-effect of selecting a single cipher.
7376 (E.g., cipher list string "RSA" enables ciphersuites that are left
7377 out of "ALL" because they do not provide symmetric encryption.
7378 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7379 [Lutz Jaenicke, Bodo Moeller]
7381 *) Add appropriate support for separate platform-dependent build
7382 directories. The recommended way to make a platform-dependent
7383 build directory is the following (tested on Linux), maybe with
7386 # Place yourself outside of the OpenSSL source tree. In
7387 # this example, the environment variable OPENSSL_SOURCE
7388 # is assumed to contain the absolute OpenSSL source directory.
7389 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7390 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7391 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7392 mkdir -p `dirname $F`
7393 ln -s $OPENSSL_SOURCE/$F $F
7396 To be absolutely sure not to disturb the source tree, a "make clean"
7397 is a good thing. If it isn't successful, don't worry about it,
7398 it probably means the source directory is very clean.
7401 *) Make sure any ENGINE control commands make local copies of string
7402 pointers passed to them whenever necessary. Otherwise it is possible
7403 the caller may have overwritten (or deallocated) the original string
7404 data when a later ENGINE operation tries to use the stored values.
7405 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7407 *) Improve diagnostics in file reading and command-line digests.
7408 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7410 *) Add AES modes CFB and OFB to the object database. Correct an
7411 error in AES-CFB decryption.
7414 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7415 allows existing EVP_CIPHER_CTX structures to be reused after
7416 calling EVP_*Final(). This behaviour is used by encryption
7417 BIOs and some applications. This has the side effect that
7418 applications must explicitly clean up cipher contexts with
7419 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7422 *) Check the values of dna and dnb in bn_mul_recursive before calling
7423 bn_mul_comba (a non zero value means the a or b arrays do not contain
7424 n2 elements) and fallback to bn_mul_normal if either is not zero.
7427 *) Fix escaping of non-ASCII characters when using the -subj option
7428 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7431 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7432 form for "surname", serialNumber has no short form.
7433 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7434 therefore remove "mail" short name for "internet 7".
7435 The OID for unique identifiers in X509 certificates is
7436 x500UniqueIdentifier, not uniqueIdentifier.
7437 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7440 *) Add an "init" command to the ENGINE config module and auto initialize
7441 ENGINEs. Without any "init" command the ENGINE will be initialized
7442 after all ctrl commands have been executed on it. If init=1 the
7443 ENGINE is initialized at that point (ctrls before that point are run
7444 on the uninitialized ENGINE and after on the initialized one). If
7445 init=0 then the ENGINE will not be initialized at all.
7448 *) Fix the 'app_verify_callback' interface so that the user-defined
7449 argument is actually passed to the callback: In the
7450 SSL_CTX_set_cert_verify_callback() prototype, the callback
7451 declaration has been changed from
7454 int (*cb)(X509_STORE_CTX *,void *);
7455 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7456 i=s->ctx->app_verify_callback(&ctx)
7457 has been changed into
7458 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7460 To update applications using SSL_CTX_set_cert_verify_callback(),
7461 a dummy argument can be added to their callback functions.
7462 [D. K. Smetters <smetters@parc.xerox.com>]
7464 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7465 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7467 *) Add and OPENSSL_LOAD_CONF define which will cause
7468 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7469 This allows older applications to transparently support certain
7470 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7471 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7472 load the config file and OPENSSL_add_all_algorithms_conf() which will
7473 always load it have also been added.
7476 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7477 Adjust NIDs and EVP layer.
7478 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7480 *) Config modules support in openssl utility.
7482 Most commands now load modules from the config file,
7483 though in a few (such as version) this isn't done
7484 because it couldn't be used for anything.
7486 In the case of ca and req the config file used is
7487 the same as the utility itself: that is the -config
7488 command line option can be used to specify an
7492 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7493 use "openssl_conf" if filename is NULL use default openssl config file.
7496 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7497 config section name. Add a new flag to tolerate a missing config file
7498 and move code to CONF_modules_load_file().
7501 *) Support for crypto accelerator cards from Accelerated Encryption
7502 Processing, www.aep.ie. (Use engine 'aep')
7503 The support was copied from 0.9.6c [engine] and adapted/corrected
7504 to work with the new engine framework.
7505 [AEP Inc. and Richard Levitte]
7507 *) Support for SureWare crypto accelerator cards from Baltimore
7508 Technologies. (Use engine 'sureware')
7509 The support was copied from 0.9.6c [engine] and adapted
7510 to work with the new engine framework.
7513 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7514 make the newer ENGINE framework commands for the CHIL engine work.
7515 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7517 *) Make it possible to produce shared libraries on ReliantUNIX.
7518 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7520 *) Add the configuration target debug-linux-ppro.
7521 Make 'openssl rsa' use the general key loading routines
7522 implemented in apps.c, and make those routines able to
7523 handle the key format FORMAT_NETSCAPE and the variant
7525 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7527 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7528 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7530 *) Add -keyform to rsautl, and document -engine.
7531 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7533 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7534 BIO_R_NO_SUCH_FILE error code rather than the generic
7535 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7538 *) Add new functions
7540 ERR_peek_last_error_line
7541 ERR_peek_last_error_line_data.
7542 These are similar to
7545 ERR_peek_error_line_data,
7546 but report on the latest error recorded rather than the first one
7547 still in the error queue.
7548 [Ben Laurie, Bodo Moeller]
7550 *) default_algorithms option in ENGINE config module. This allows things
7552 default_algorithms = ALL
7553 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7556 *) Preliminary ENGINE config module.
7559 *) New experimental application configuration code.
7562 *) Change the AES code to follow the same name structure as all other
7563 symmetric ciphers, and behave the same way. Move everything to
7564 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7565 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7567 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7568 [Ben Laurie and Theo de Raadt]
7570 *) Add option to output public keys in req command.
7571 [Massimiliano Pala madwolf@openca.org]
7573 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7574 (up to about 10% better than before for P-192 and P-224).
7577 *) New functions/macros
7579 SSL_CTX_set_msg_callback(ctx, cb)
7580 SSL_CTX_set_msg_callback_arg(ctx, arg)
7581 SSL_set_msg_callback(ssl, cb)
7582 SSL_set_msg_callback_arg(ssl, arg)
7584 to request calling a callback function
7586 void cb(int write_p, int version, int content_type,
7587 const void *buf, size_t len, SSL *ssl, void *arg)
7589 whenever a protocol message has been completely received
7590 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7591 protocol version according to which the SSL library interprets
7592 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7593 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7594 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7595 specification (change_cipher_spec(20), alert(21), handshake(22)).
7596 'buf' and 'len' point to the actual message, 'ssl' to the
7597 SSL object, and 'arg' is the application-defined value set by
7598 SSL[_CTX]_set_msg_callback_arg().
7600 'openssl s_client' and 'openssl s_server' have new '-msg' options
7601 to enable a callback that displays all protocol messages.
7604 *) Change the shared library support so shared libraries are built as
7605 soon as the corresponding static library is finished, and thereby get
7606 openssl and the test programs linked against the shared library.
7607 This still only happens when the keyword "shard" has been given to
7608 the configuration scripts.
7610 NOTE: shared library support is still an experimental thing, and
7611 backward binary compatibility is still not guaranteed.
7612 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7614 *) Add support for Subject Information Access extension.
7615 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7617 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7618 additional bytes when new memory had to be allocated, not just
7619 when reusing an existing buffer.
7622 *) New command line and configuration option 'utf8' for the req command.
7623 This allows field values to be specified as UTF8 strings.
7626 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7627 runs for the former and machine-readable output for the latter.
7630 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7631 of the e-mail address in the DN (i.e., it will go into a certificate
7632 extension only). The new configuration file option 'email_in_dn = no'
7633 has the same effect.
7634 [Massimiliano Pala madwolf@openca.org]
7636 *) Change all functions with names starting with des_ to be starting
7637 with DES_ instead. Add wrappers that are compatible with libdes,
7638 but are named _ossl_old_des_*. Finally, add macros that map the
7639 des_* symbols to the corresponding _ossl_old_des_* if libdes
7640 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7641 desired, the des_* symbols will be mapped to DES_*, with one
7644 Since we provide two compatibility mappings, the user needs to
7645 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7646 compatibility is desired. The default (i.e., when that macro
7647 isn't defined) is OpenSSL 0.9.6c compatibility.
7649 There are also macros that enable and disable the support of old
7650 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7651 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7652 are defined, the default will apply: to support the old des routines.
7654 In either case, one must include openssl/des.h to get the correct
7655 definitions. Do not try to just include openssl/des_old.h, that
7658 NOTE: This is a major break of an old API into a new one. Software
7659 authors are encouraged to switch to the DES_ style functions. Some
7660 time in the future, des_old.h and the libdes compatibility functions
7661 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7662 default), and then completely removed.
7665 *) Test for certificates which contain unsupported critical extensions.
7666 If such a certificate is found during a verify operation it is
7667 rejected by default: this behaviour can be overridden by either
7668 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7669 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7670 X509_supported_extension() has also been added which returns 1 if a
7671 particular extension is supported.
7674 *) Modify the behaviour of EVP cipher functions in similar way to digests
7675 to retain compatibility with existing code.
7678 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7679 compatibility with existing code. In particular the 'ctx' parameter does
7680 not have to be to be initialized before the call to EVP_DigestInit() and
7681 it is tidied up after a call to EVP_DigestFinal(). New function
7682 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7683 EVP_MD_CTX_copy() changed to not require the destination to be
7684 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7685 requires the destination to be valid.
7687 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7688 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7691 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7692 so that complete 'Handshake' protocol structures are kept in memory
7693 instead of overwriting 'msg_type' and 'length' with 'body' data.
7696 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7697 [Massimo Santin via Richard Levitte]
7699 *) Major restructuring to the underlying ENGINE code. This includes
7700 reduction of linker bloat, separation of pure "ENGINE" manipulation
7701 (initialisation, etc) from functionality dealing with implementations
7702 of specific crypto interfaces. This change also introduces integrated
7703 support for symmetric ciphers and digest implementations - so ENGINEs
7704 can now accelerate these by providing EVP_CIPHER and EVP_MD
7705 implementations of their own. This is detailed in crypto/engine/README
7706 as it couldn't be adequately described here. However, there are a few
7707 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7708 were changed in the original introduction of ENGINE code have now
7709 reverted back - the hooking from this code to ENGINE is now a good
7710 deal more passive and at run-time, operations deal directly with
7711 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7712 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7713 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7714 they were not being used by the framework as there is no concept of a
7715 BIGNUM_METHOD and they could not be generalised to the new
7716 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7717 ENGINE_cpy() has been removed as it cannot be consistently defined in
7721 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7724 *) Change mkdef.pl to sort symbols that get the same entry number,
7725 and make sure the automatically generated functions ERR_load_*
7726 become part of libeay.num as well.
7729 *) New function SSL_renegotiate_pending(). This returns true once
7730 renegotiation has been requested (either SSL_renegotiate() call
7731 or HelloRequest/ClientHello received from the peer) and becomes
7732 false once a handshake has been completed.
7733 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7734 sends a HelloRequest, but does not ensure that a handshake takes
7735 place. SSL_renegotiate_pending() is useful for checking if the
7736 client has followed the request.)
7739 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7740 By default, clients may request session resumption even during
7741 renegotiation (if session ID contexts permit); with this option,
7742 session resumption is possible only in the first handshake.
7744 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7745 more bits available for options that should not be part of
7746 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7749 *) Add some demos for certificate and certificate request creation.
7752 *) Make maximum certificate chain size accepted from the peer application
7753 settable (SSL*_get/set_max_cert_list()), as proposed by
7754 "Douglas E. Engert" <deengert@anl.gov>.
7757 *) Add support for shared libraries for Unixware-7
7758 (Boyd Lynn Gerber <gerberb@zenez.com>).
7761 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7762 be done prior to destruction. Use this to unload error strings from
7763 ENGINEs that load their own error strings. NB: This adds two new API
7764 functions to "get" and "set" this destroy handler in an ENGINE.
7767 *) Alter all existing ENGINE implementations (except "openssl" and
7768 "openbsd") to dynamically instantiate their own error strings. This
7769 makes them more flexible to be built both as statically-linked ENGINEs
7770 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7771 Also, add stub code to each that makes building them as self-contained
7772 shared-libraries easier (see README.ENGINE).
7775 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7776 implementations into applications that are completely implemented in
7777 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7778 commands that can be used to configure what shared-library to load and
7779 to control aspects of the way it is handled. Also, made an update to
7780 the README.ENGINE file that brings its information up-to-date and
7781 provides some information and instructions on the "dynamic" ENGINE
7782 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7785 *) Make it possible to unload ranges of ERR strings with a new
7786 "ERR_unload_strings" function.
7789 *) Add a copy() function to EVP_MD.
7792 *) Make EVP_MD routines take a context pointer instead of just the
7793 md_data void pointer.
7796 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7797 that the digest can only process a single chunk of data
7798 (typically because it is provided by a piece of
7799 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7800 is only going to provide a single chunk of data, and hence the
7801 framework needn't accumulate the data for oneshot drivers.
7804 *) As with "ERR", make it possible to replace the underlying "ex_data"
7805 functions. This change also alters the storage and management of global
7806 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7807 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7808 index counters. The API functions that use this state have been changed
7809 to take a "class_index" rather than pointers to the class's local STACK
7810 and counter, and there is now an API function to dynamically create new
7811 classes. This centralisation allows us to (a) plug a lot of the
7812 thread-safety problems that existed, and (b) makes it possible to clean
7813 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7814 such data would previously have always leaked in application code and
7815 workarounds were in place to make the memory debugging turn a blind eye
7816 to it. Application code that doesn't use this new function will still
7817 leak as before, but their memory debugging output will announce it now
7818 rather than letting it slide.
7820 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7821 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7822 has a return value to indicate success or failure.
7825 *) Make it possible to replace the underlying "ERR" functions such that the
7826 global state (2 LHASH tables and 2 locks) is only used by the "default"
7827 implementation. This change also adds two functions to "get" and "set"
7828 the implementation prior to it being automatically set the first time
7829 any other ERR function takes place. Ie. an application can call "get",
7830 pass the return value to a module it has just loaded, and that module
7831 can call its own "set" function using that value. This means the
7832 module's "ERR" operations will use (and modify) the error state in the
7833 application and not in its own statically linked copy of OpenSSL code.
7836 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7837 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7838 the operation, and provides a more encapsulated way for external code
7839 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7840 to use these functions rather than manually incrementing the counts.
7842 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7845 *) Add EVP test program.
7848 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7851 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7852 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7853 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7854 These allow a CRL to be built without having to access X509_CRL fields
7855 directly. Modify 'ca' application to use new functions.
7858 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7859 bug workarounds. Rollback attack detection is a security feature.
7860 The problem will only arise on OpenSSL servers when TLSv1 is not
7861 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7862 Software authors not wanting to support TLSv1 will have special reasons
7863 for their choice and can explicitly enable this option.
7864 [Bodo Moeller, Lutz Jaenicke]
7866 *) Rationalise EVP so it can be extended: don't include a union of
7867 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7868 (similar to those existing for EVP_CIPHER_CTX).
7873 EVP_MD_CTX_init(&md); /* new function call */
7874 EVP_DigestInit(&md, EVP_sha1());
7875 EVP_DigestUpdate(&md, in, len);
7876 EVP_DigestFinal(&md, out, NULL);
7877 EVP_MD_CTX_cleanup(&md); /* new function call */
7881 *) Make DES key schedule conform to the usual scheme, as well as
7882 correcting its structure. This means that calls to DES functions
7883 now have to pass a pointer to a des_key_schedule instead of a
7884 plain des_key_schedule (which was actually always a pointer
7887 des_key_schedule ks;
7889 des_set_key_checked(..., &ks);
7890 des_ncbc_encrypt(..., &ks, ...);
7892 (Note that a later change renames 'des_...' into 'DES_...'.)
7895 *) Initial reduction of linker bloat: the use of some functions, such as
7896 PEM causes large amounts of unused functions to be linked in due to
7897 poor organisation. For example pem_all.c contains every PEM function
7898 which has a knock on effect of linking in large amounts of (unused)
7899 ASN1 code. Grouping together similar functions and splitting unrelated
7900 functions prevents this.
7903 *) Cleanup of EVP macros.
7906 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7907 correct _ecb suffix.
7910 *) Add initial OCSP responder support to ocsp application. The
7911 revocation information is handled using the text based index
7912 use by the ca application. The responder can either handle
7913 requests generated internally, supplied in files (for example
7914 via a CGI script) or using an internal minimal server.
7917 *) Add configuration choices to get zlib compression for TLS.
7920 *) Changes to Kerberos SSL for RFC 2712 compliance:
7921 1. Implemented real KerberosWrapper, instead of just using
7922 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7923 2. Implemented optional authenticator field of KerberosWrapper.
7925 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7926 and authenticator structs; see crypto/krb5/.
7928 Generalized Kerberos calls to support multiple Kerberos libraries.
7929 [Vern Staats <staatsvr@asc.hpc.mil>,
7930 Jeffrey Altman <jaltman@columbia.edu>
7931 via Richard Levitte]
7933 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7934 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7935 values for each of the key sizes rather than having just
7936 parameters (and 'speed' generating keys each time).
7939 *) Speed up EVP routines.
7942 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7943 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7944 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7945 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7947 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7948 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7949 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7952 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7954 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7957 *) Added the OS2-EMX target.
7958 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7960 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7961 to support NCONF routines in extension code. New function CONF_set_nconf()
7962 to allow functions which take an NCONF to also handle the old LHASH
7963 structure: this means that the old CONF compatible routines can be
7964 retained (in particular wrt extensions) without having to duplicate the
7965 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7968 *) Enhance the general user interface with mechanisms for inner control
7969 and with possibilities to have yes/no kind of prompts.
7972 *) Change all calls to low level digest routines in the library and
7973 applications to use EVP. Add missing calls to HMAC_cleanup() and
7974 don't assume HMAC_CTX can be copied using memcpy().
7975 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7977 *) Add the possibility to control engines through control names but with
7978 arbitrary arguments instead of just a string.
7979 Change the key loaders to take a UI_METHOD instead of a callback
7980 function pointer. NOTE: this breaks binary compatibility with earlier
7981 versions of OpenSSL [engine].
7982 Adapt the nCipher code for these new conditions and add a card insertion
7986 *) Enhance the general user interface with mechanisms to better support
7987 dialog box interfaces, application-defined prompts, the possibility
7988 to use defaults (for example default passwords from somewhere else)
7989 and interrupts/cancellations.
7992 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7993 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7996 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7997 tidy up some unnecessarily weird code in 'sk_new()').
7998 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
8000 *) Change the key loading routines for ENGINEs to use the same kind
8001 callback (pem_password_cb) as all other routines that need this
8005 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
8006 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
8007 than this minimum value is recommended.
8010 *) New random seeder for OpenVMS, using the system process statistics
8011 that are easily reachable.
8014 *) Windows apparently can't transparently handle global
8015 variables defined in DLLs. Initialisations such as:
8017 const ASN1_ITEM *it = &ASN1_INTEGER_it;
8019 won't compile. This is used by the any applications that need to
8020 declare their own ASN1 modules. This was fixed by adding the option
8021 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
8022 needed for static libraries under Win32.
8025 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
8026 setting of purpose and trust fields. New X509_STORE trust and
8027 purpose functions and tidy up setting in other SSL functions.
8030 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
8031 structure. These are inherited by X509_STORE_CTX when it is
8032 initialised. This allows various defaults to be set in the
8033 X509_STORE structure (such as flags for CRL checking and custom
8034 purpose or trust settings) for functions which only use X509_STORE_CTX
8035 internally such as S/MIME.
8037 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
8038 trust settings if they are not set in X509_STORE. This allows X509_STORE
8039 purposes and trust (in S/MIME for example) to override any set by default.
8041 Add command line options for CRL checking to smime, s_client and s_server
8045 *) Initial CRL based revocation checking. If the CRL checking flag(s)
8046 are set then the CRL is looked up in the X509_STORE structure and
8047 its validity and signature checked, then if the certificate is found
8048 in the CRL the verify fails with a revoked error.
8050 Various new CRL related callbacks added to X509_STORE_CTX structure.
8052 Command line options added to 'verify' application to support this.
8054 This needs some additional work, such as being able to handle multiple
8055 CRLs with different times, extension based lookup (rather than just
8056 by subject name) and ultimately more complete V2 CRL extension
8060 *) Add a general user interface API (crypto/ui/). This is designed
8061 to replace things like des_read_password and friends (backward
8062 compatibility functions using this new API are provided).
8063 The purpose is to remove prompting functions from the DES code
8064 section as well as provide for prompting through dialog boxes in
8065 a window system and the like.
8068 *) Add "ex_data" support to ENGINE so implementations can add state at a
8069 per-structure level rather than having to store it globally.
8072 *) Make it possible for ENGINE structures to be copied when retrieved by
8073 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
8074 This causes the "original" ENGINE structure to act like a template,
8075 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
8076 operational state can be localised to each ENGINE structure, despite the
8077 fact they all share the same "methods". New ENGINE structures returned in
8078 this case have no functional references and the return value is the single
8079 structural reference. This matches the single structural reference returned
8080 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8084 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8085 needs to match any other type at all we need to manually clear the
8089 *) Changes to the "openssl engine" utility to include;
8090 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8091 about an ENGINE's available control commands.
8092 - executing control commands from command line arguments using the
8093 '-pre' and '-post' switches. '-post' is only used if '-t' is
8094 specified and the ENGINE is successfully initialised. The syntax for
8095 the individual commands are colon-separated, for example;
8096 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8099 *) New dynamic control command support for ENGINEs. ENGINEs can now
8100 declare their own commands (numbers), names (strings), descriptions,
8101 and input types for run-time discovery by calling applications. A
8102 subset of these commands are implicitly classed as "executable"
8103 depending on their input type, and only these can be invoked through
8104 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8105 can be based on user input, config files, etc). The distinction is
8106 that "executable" commands cannot return anything other than a boolean
8107 result and can only support numeric or string input, whereas some
8108 discoverable commands may only be for direct use through
8109 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8110 pointers, or other custom uses. The "executable" commands are to
8111 support parameterisations of ENGINE behaviour that can be
8112 unambiguously defined by ENGINEs and used consistently across any
8113 OpenSSL-based application. Commands have been added to all the
8114 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8115 control over shared-library paths without source code alterations.
8118 *) Changed all ENGINE implementations to dynamically allocate their
8119 ENGINEs rather than declaring them statically. Apart from this being
8120 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8121 this also allows the implementations to compile without using the
8122 internal engine_int.h header.
8125 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8126 'const' value. Any code that should be able to modify a RAND_METHOD
8127 should already have non-const pointers to it (ie. they should only
8128 modify their own ones).
8131 *) Made a variety of little tweaks to the ENGINE code.
8132 - "atalla" and "ubsec" string definitions were moved from header files
8133 to C code. "nuron" string definitions were placed in variables
8134 rather than hard-coded - allowing parameterisation of these values
8135 later on via ctrl() commands.
8136 - Removed unused "#if 0"'d code.
8137 - Fixed engine list iteration code so it uses ENGINE_free() to release
8138 structural references.
8139 - Constified the RAND_METHOD element of ENGINE structures.
8140 - Constified various get/set functions as appropriate and added
8141 missing functions (including a catch-all ENGINE_cpy that duplicates
8142 all ENGINE values onto a new ENGINE except reference counts/state).
8143 - Removed NULL parameter checks in get/set functions. Setting a method
8144 or function to NULL is a way of cancelling out a previously set
8145 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8146 and doesn't justify the extra error symbols and code.
8147 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8148 flags from engine_int.h to engine.h.
8149 - Changed prototypes for ENGINE handler functions (init(), finish(),
8150 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8153 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8154 to the algorithm using long division. The binary algorithm can be
8155 used only if the modulus is odd. On 32-bit systems, it is faster
8156 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8157 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8158 up to 450 bits. In 64-bit environments, the binary algorithm
8159 appears to be advantageous for much longer moduli; here we use it
8160 for moduli up to 2048 bits.
8163 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8164 could not support the combine flag in choice fields.
8167 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8168 extensions from a certificate request to the certificate.
8171 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8172 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8173 file: this allows the display of the certificate about to be
8174 signed to be customised, to allow certain fields to be included
8175 or excluded and extension details. The old system didn't display
8176 multicharacter strings properly, omitted fields not in the policy
8177 and couldn't display additional details such as extensions.
8180 *) Function EC_POINTs_mul for multiple scalar multiplication
8181 of an arbitrary number of elliptic curve points
8182 \sum scalars[i]*points[i],
8183 optionally including the generator defined for the EC_GROUP:
8184 scalar*generator + \sum scalars[i]*points[i].
8186 EC_POINT_mul is a simple wrapper function for the typical case
8187 that the point list has just one item (besides the optional
8191 *) First EC_METHODs for curves over GF(p):
8193 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8194 operations and provides various method functions that can also
8195 operate with faster implementations of modular arithmetic.
8197 EC_GFp_mont_method() reuses most functions that are part of
8198 EC_GFp_simple_method, but uses Montgomery arithmetic.
8200 [Bodo Moeller; point addition and point doubling
8201 implementation directly derived from source code provided by
8202 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8204 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8205 crypto/ec/ec_lib.c):
8207 Curves are EC_GROUP objects (with an optional group generator)
8208 based on EC_METHODs that are built into the library.
8210 Points are EC_POINT objects based on EC_GROUP objects.
8212 Most of the framework would be able to handle curves over arbitrary
8213 finite fields, but as there are no obvious types for fields other
8214 than GF(p), some functions are limited to that for now.
8217 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8218 that the file contains a complete HTTP response.
8221 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8222 change the def and num file printf format specifier from "%-40sXXX"
8223 to "%-39s XXX". The latter will always guarantee a space after the
8224 field while the former will cause them to run together if the field
8225 is 40 of more characters long.
8228 *) Constify the cipher and digest 'method' functions and structures
8229 and modify related functions to take constant EVP_MD and EVP_CIPHER
8233 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8234 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8237 *) Modify EVP_Digest*() routines so they now return values. Although the
8238 internal software routines can never fail additional hardware versions
8242 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8244 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8245 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8248 ERR_R_NESTED_ASN1_ERROR
8250 ERR_R_MISSING_ASN1_EOS
8251 were 4 .. 9, conflicting with
8252 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8254 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8255 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8257 Add new error code 'ERR_R_INTERNAL_ERROR'.
8260 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8264 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8265 sets the subject name for a new request or supersedes the
8266 subject name in a given request. Formats that can be parsed are
8267 'CN=Some Name, OU=myOU, C=IT'
8269 'CN=Some Name/OU=myOU/C=IT'.
8271 Add options '-batch' and '-verbose' to 'openssl req'.
8272 [Massimiliano Pala <madwolf@hackmasters.net>]
8274 *) Introduce the possibility to access global variables through
8275 functions on platform were that's the best way to handle exporting
8276 global variables in shared libraries. To enable this functionality,
8277 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8278 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8279 is normally done by Configure or something similar).
8281 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8282 in the source file (foo.c) like this:
8284 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8285 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8287 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8288 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8290 OPENSSL_DECLARE_GLOBAL(int,foo);
8291 #define foo OPENSSL_GLOBAL_REF(foo)
8292 OPENSSL_DECLARE_GLOBAL(double,bar);
8293 #define bar OPENSSL_GLOBAL_REF(bar)
8295 The #defines are very important, and therefore so is including the
8296 header file everywhere where the defined globals are used.
8298 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8299 of ASN.1 items, but that structure is a bit different.
8301 The largest change is in util/mkdef.pl which has been enhanced with
8302 better and easier to understand logic to choose which symbols should
8303 go into the Windows .def files as well as a number of fixes and code
8304 cleanup (among others, algorithm keywords are now sorted
8305 lexicographically to avoid constant rewrites).
8308 *) In BN_div() keep a copy of the sign of 'num' before writing the
8309 result to 'rm' because if rm==num the value will be overwritten
8310 and produce the wrong result if 'num' is negative: this caused
8311 problems with BN_mod() and BN_nnmod().
8314 *) Function OCSP_request_verify(). This checks the signature on an
8315 OCSP request and verifies the signer certificate. The signer
8316 certificate is just checked for a generic purpose and OCSP request
8320 *) Add OCSP_check_validity() function to check the validity of OCSP
8321 responses. OCSP responses are prepared in real time and may only
8322 be a few seconds old. Simply checking that the current time lies
8323 between thisUpdate and nextUpdate max reject otherwise valid responses
8324 caused by either OCSP responder or client clock inaccuracy. Instead
8325 we allow thisUpdate and nextUpdate to fall within a certain period of
8326 the current time. The age of the response can also optionally be
8327 checked. Two new options -validity_period and -status_age added to
8331 *) If signature or public key algorithm is unrecognized print out its
8332 OID rather that just UNKNOWN.
8335 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8336 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8337 ID to be generated from the issuer certificate alone which can then be
8338 passed to OCSP_id_issuer_cmp().
8341 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8342 ASN1 modules to export functions returning ASN1_ITEM pointers
8343 instead of the ASN1_ITEM structures themselves. This adds several
8344 new macros which allow the underlying ASN1 function/structure to
8345 be accessed transparently. As a result code should not use ASN1_ITEM
8346 references directly (such as &X509_it) but instead use the relevant
8347 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8348 use of the new ASN1 code on platforms where exporting structures
8349 is problematical (for example in shared libraries) but exporting
8350 functions returning pointers to structures is not.
8353 *) Add support for overriding the generation of SSL/TLS session IDs.
8354 These callbacks can be registered either in an SSL_CTX or per SSL.
8355 The purpose of this is to allow applications to control, if they wish,
8356 the arbitrary values chosen for use as session IDs, particularly as it
8357 can be useful for session caching in multiple-server environments. A
8358 command-line switch for testing this (and any client code that wishes
8359 to use such a feature) has been added to "s_server".
8360 [Geoff Thorpe, Lutz Jaenicke]
8362 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8363 of the form '#if defined(...) || defined(...) || ...' and
8364 '#if !defined(...) && !defined(...) && ...'. This also avoids
8365 the growing number of special cases it was previously handling.
8368 *) Make all configuration macros available for application by making
8369 sure they are available in opensslconf.h, by giving them names starting
8370 with "OPENSSL_" to avoid conflicts with other packages and by making
8371 sure e_os2.h will cover all platform-specific cases together with
8373 Additionally, it is now possible to define configuration/platform-
8374 specific names (called "system identities"). In the C code, these
8375 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8376 macro with the name beginning with "OPENSSL_SYS_", which is determined
8377 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8381 *) New option -set_serial to 'req' and 'x509' this allows the serial
8382 number to use to be specified on the command line. Previously self
8383 signed certificates were hard coded with serial number 0 and the
8384 CA options of 'x509' had to use a serial number in a file which was
8388 *) New options to 'ca' utility to support V2 CRL entry extensions.
8389 Currently CRL reason, invalidity date and hold instruction are
8390 supported. Add new CRL extensions to V3 code and some new objects.
8393 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8394 disable standard block padding (aka PKCS#5 padding) in the EVP
8395 API, which was previously mandatory. This means that the data is
8396 not padded in any way and so the total length much be a multiple
8397 of the block size, otherwise an error occurs.
8400 *) Initial (incomplete) OCSP SSL support.
8403 *) New function OCSP_parse_url(). This splits up a URL into its host,
8404 port and path components: primarily to parse OCSP URLs. New -url
8405 option to ocsp utility.
8408 *) New nonce behavior. The return value of OCSP_check_nonce() now
8409 reflects the various checks performed. Applications can decide
8410 whether to tolerate certain situations such as an absent nonce
8411 in a response when one was present in a request: the ocsp application
8412 just prints out a warning. New function OCSP_add1_basic_nonce()
8413 this is to allow responders to include a nonce in a response even if
8414 the request is nonce-less.
8417 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8418 skipped when using openssl x509 multiple times on a single input file,
8419 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8422 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8423 set string type: to handle setting ASN1_TIME structures. Fix ca
8424 utility to correctly initialize revocation date of CRLs.
8427 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8428 the clients preferred ciphersuites and rather use its own preferences.
8429 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8430 Internet Explorer by ensuring unchanged hash method during stepup.
8431 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8434 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8435 to aes and add a new 'exist' option to print out symbols that don't
8439 *) Additional options to ocsp utility to allow flags to be set and
8440 additional certificates supplied.
8443 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8444 OCSP client a number of certificate to only verify the response
8448 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8449 handle the new API. Currently only ECB, CBC modes supported. Add new
8452 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8453 Encryption Standard (AES) Ciphersuites for Transport Layer
8454 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8455 not enabled by default and were not part of the "ALL" ciphersuite
8456 alias because they were not yet official; they could be
8457 explicitly requested by specifying the "AESdraft" ciphersuite
8458 group alias. In the final release of OpenSSL 0.9.7, the group
8459 alias is called "AES" and is part of "ALL".)
8460 [Ben Laurie, Steve Henson, Bodo Moeller]
8462 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8463 request to response.
8466 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8467 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8468 extract information from a certificate request. OCSP_response_create()
8469 creates a response and optionally adds a basic response structure.
8470 OCSP_basic_add1_status() adds a complete single response to a basic
8471 response and returns the OCSP_SINGLERESP structure just added (to allow
8472 extensions to be included for example). OCSP_basic_add1_cert() adds a
8473 certificate to a basic response and OCSP_basic_sign() signs a basic
8474 response with various flags. New helper functions ASN1_TIME_check()
8475 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8476 (converts ASN1_TIME to GeneralizedTime).
8479 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8480 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8481 structure from a certificate. X509_pubkey_digest() digests the public_key
8482 contents: this is used in various key identifiers.
8485 *) Make sk_sort() tolerate a NULL argument.
8486 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8488 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8489 passed by the function are trusted implicitly. If any of them signed the
8490 response then it is assumed to be valid and is not verified.
8493 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8494 to data. This was previously part of the PKCS7 ASN1 code. This
8495 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8496 [Steve Henson, reported by Kenneth R. Robinette
8497 <support@securenetterm.com>]
8499 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8500 routines: without these tracing memory leaks is very painful.
8501 Fix leaks in PKCS12 and PKCS7 routines.
8504 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8505 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8506 effectively meant GeneralizedTime would never be used. Now it
8507 is initialised to -1 but X509_time_adj() now has to check the value
8508 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8509 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8510 [Steve Henson, reported by Kenneth R. Robinette
8511 <support@securenetterm.com>]
8513 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8514 result in a zero length in the ASN1_INTEGER structure which was
8515 not consistent with the structure when d2i_ASN1_INTEGER() was used
8516 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8517 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8518 where it did not print out a minus for negative ASN1_INTEGER.
8521 *) Add summary printout to ocsp utility. The various functions which
8522 convert status values to strings have been renamed to:
8523 OCSP_response_status_str(), OCSP_cert_status_str() and
8524 OCSP_crl_reason_str() and are no longer static. New options
8525 to verify nonce values and to disable verification. OCSP response
8526 printout format cleaned up.
8529 *) Add additional OCSP certificate checks. These are those specified
8530 in RFC2560. This consists of two separate checks: the CA of the
8531 certificate being checked must either be the OCSP signer certificate
8532 or the issuer of the OCSP signer certificate. In the latter case the
8533 OCSP signer certificate must contain the OCSP signing extended key
8534 usage. This check is performed by attempting to match the OCSP
8535 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8536 in the OCSP_CERTID structures of the response.
8539 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8540 and related routines. This uses the standard OpenSSL certificate
8541 verify routines to perform initial checks (just CA validity) and
8542 to obtain the certificate chain. Then additional checks will be
8543 performed on the chain. Currently the root CA is checked to see
8544 if it is explicitly trusted for OCSP signing. This is used to set
8545 a root CA as a global signing root: that is any certificate that
8546 chains to that CA is an acceptable OCSP signing certificate.
8549 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8550 extensions from a separate configuration file.
8551 As when reading extensions from the main configuration file,
8552 the '-extensions ...' option may be used for specifying the
8554 [Massimiliano Pala <madwolf@comune.modena.it>]
8556 *) New OCSP utility. Allows OCSP requests to be generated or
8557 read. The request can be sent to a responder and the output
8558 parsed, outputed or printed in text form. Not complete yet:
8559 still needs to check the OCSP response validity.
8562 *) New subcommands for 'openssl ca':
8563 'openssl ca -status <serial>' prints the status of the cert with
8564 the given serial number (according to the index file).
8565 'openssl ca -updatedb' updates the expiry status of certificates
8567 [Massimiliano Pala <madwolf@comune.modena.it>]
8569 *) New '-newreq-nodes' command option to CA.pl. This is like
8570 '-newreq', but calls 'openssl req' with the '-nodes' option
8571 so that the resulting key is not encrypted.
8572 [Damien Miller <djm@mindrot.org>]
8574 *) New configuration for the GNU Hurd.
8575 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8577 *) Initial code to implement OCSP basic response verify. This
8578 is currently incomplete. Currently just finds the signer's
8579 certificate and verifies the signature on the response.
8582 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8583 value of OPENSSLDIR. This is available via the new '-d' option
8584 to 'openssl version', and is also included in 'openssl version -a'.
8587 *) Allowing defining memory allocation callbacks that will be given
8588 file name and line number information in additional arguments
8589 (a const char* and an int). The basic functionality remains, as
8590 well as the original possibility to just replace malloc(),
8591 realloc() and free() by functions that do not know about these
8592 additional arguments. To register and find out the current
8593 settings for extended allocation functions, the following
8594 functions are provided:
8596 CRYPTO_set_mem_ex_functions
8597 CRYPTO_set_locked_mem_ex_functions
8598 CRYPTO_get_mem_ex_functions
8599 CRYPTO_get_locked_mem_ex_functions
8601 These work the same way as CRYPTO_set_mem_functions and friends.
8602 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8603 extended allocation function is enabled.
8604 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8605 a conventional allocation function is enabled.
8606 [Richard Levitte, Bodo Moeller]
8608 *) Finish off removing the remaining LHASH function pointer casts.
8609 There should no longer be any prototype-casting required when using
8610 the LHASH abstraction, and any casts that remain are "bugs". See
8611 the callback types and macros at the head of lhash.h for details
8612 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8615 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8616 If /dev/[u]random devices are not available or do not return enough
8617 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8619 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8620 /etc/entropy will be queried once each in this sequence, querying stops
8621 when enough entropy was collected without querying more sockets.
8624 *) Change the Unix RAND_poll() variant to be able to poll several
8625 random devices, as specified by DEVRANDOM, until a sufficient amount
8626 of data has been collected. We spend at most 10 ms on each file
8627 (select timeout) and read in non-blocking mode. DEVRANDOM now
8628 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8629 (previously it was just the string "/dev/urandom"), so on typical
8630 platforms the 10 ms delay will never occur.
8631 Also separate out the Unix variant to its own file, rand_unix.c.
8632 For VMS, there's a currently-empty rand_vms.c.
8635 *) Move OCSP client related routines to ocsp_cl.c. These
8636 provide utility functions which an application needing
8637 to issue a request to an OCSP responder and analyse the
8638 response will typically need: as opposed to those which an
8639 OCSP responder itself would need which will be added later.
8641 OCSP_request_sign() signs an OCSP request with an API similar
8642 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8643 response. OCSP_response_get1_basic() extracts basic response
8644 from response. OCSP_resp_find_status(): finds and extracts status
8645 information from an OCSP_CERTID structure (which will be created
8646 when the request structure is built). These are built from lower
8647 level functions which work on OCSP_SINGLERESP structures but
8648 won't normally be used unless the application wishes to examine
8649 extensions in the OCSP response for example.
8651 Replace nonce routines with a pair of functions.
8652 OCSP_request_add1_nonce() adds a nonce value and optionally
8653 generates a random value. OCSP_check_nonce() checks the
8654 validity of the nonce in an OCSP response.
8657 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8658 This doesn't copy the supplied OCSP_CERTID and avoids the
8659 need to free up the newly created id. Change return type
8660 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8661 This can then be used to add extensions to the request.
8662 Deleted OCSP_request_new(), since most of its functionality
8663 is now in OCSP_REQUEST_new() (and the case insensitive name
8664 clash) apart from the ability to set the request name which
8665 will be added elsewhere.
8668 *) Update OCSP API. Remove obsolete extensions argument from
8669 various functions. Extensions are now handled using the new
8670 OCSP extension code. New simple OCSP HTTP function which
8671 can be used to send requests and parse the response.
8674 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8675 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8676 uses the special reorder version of SET OF to sort the attributes
8677 and reorder them to match the encoded order. This resolves a long
8678 standing problem: a verify on a PKCS7 structure just after signing
8679 it used to fail because the attribute order did not match the
8680 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8681 it uses the received order. This is necessary to tolerate some broken
8682 software that does not order SET OF. This is handled by encoding
8683 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8684 to produce the required SET OF.
8687 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8688 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8689 files to get correct declarations of the ASN.1 item variables.
8692 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8693 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8694 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8695 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8696 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8697 ASN1_ITEM and no wrapper functions.
8700 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8701 replace the old function pointer based I/O routines. Change most of
8702 the *_d2i_bio() and *_d2i_fp() functions to use these.
8705 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8706 lines, recognize more "algorithms" that can be deselected, and make
8707 it complain about algorithm deselection that isn't recognised.
8710 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8711 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8712 to use new functions. Add NO_ASN1_OLD which can be set to remove
8713 some old style ASN1 functions: this can be used to determine if old
8714 code will still work when these eventually go away.
8717 *) New extension functions for OCSP structures, these follow the
8718 same conventions as certificates and CRLs.
8721 *) New function X509V3_add1_i2d(). This automatically encodes and
8722 adds an extension. Its behaviour can be customised with various
8723 flags to append, replace or delete. Various wrappers added for
8724 certificates and CRLs.
8727 *) Fix to avoid calling the underlying ASN1 print routine when
8728 an extension cannot be parsed. Correct a typo in the
8729 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8732 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8733 entries for variables.
8736 *) Add functionality to apps/openssl.c for detecting locking
8737 problems: As the program is single-threaded, all we have
8738 to do is register a locking callback using an array for
8739 storing which locks are currently held by the program.
8742 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8743 SSL_get_ex_data_X509_STORE_idx(), which is used in
8744 ssl_verify_cert_chain() and thus can be called at any time
8745 during TLS/SSL handshakes so that thread-safety is essential.
8746 Unfortunately, the ex_data design is not at all suited
8747 for multi-threaded use, so it probably should be abolished.
8750 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8751 [Broadcom, tweaked and integrated by Geoff Thorpe]
8753 *) Move common extension printing code to new function
8754 X509V3_print_extensions(). Reorganise OCSP print routines and
8755 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8758 *) New function X509_signature_print() to remove duplication in some
8762 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8763 set (this was treated exactly the same as SET OF previously). This
8764 is used to reorder the STACK representing the structure to match the
8765 encoding. This will be used to get round a problem where a PKCS7
8766 structure which was signed could not be verified because the STACK
8767 order did not reflect the encoded order.
8770 *) Reimplement the OCSP ASN1 module using the new code.
8773 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8774 for its ASN1 operations. The old style function pointers still exist
8775 for now but they will eventually go away.
8778 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8779 completely replaces the old ASN1 functionality with a table driven
8780 encoder and decoder which interprets an ASN1_ITEM structure describing
8781 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8782 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8783 has also been converted to the new form.
8786 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8787 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8788 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8789 for negative moduli.
8792 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8793 of not touching the result's sign bit.
8796 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8800 *) Changed the LHASH code to use prototypes for callbacks, and created
8801 macros to declare and implement thin (optionally static) functions
8802 that provide type-safety and avoid function pointer casting for the
8803 type-specific callbacks.
8806 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8808 [Veers Staats <staatsvr@asc.hpc.mil>,
8809 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8811 *) Reformat the FAQ so the different questions and answers can be divided
8812 in sections depending on the subject.
8815 *) Have the zlib compression code load ZLIB.DLL dynamically under
8819 *) New function BN_mod_sqrt for computing square roots modulo a prime
8820 (using the probabilistic Tonelli-Shanks algorithm unless
8821 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8822 be handled deterministically).
8823 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8825 *) Make BN_mod_inverse faster by explicitly handling small quotients
8826 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8827 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8830 *) New function BN_kronecker.
8833 *) Fix BN_gcd so that it works on negative inputs; the result is
8834 positive unless both parameters are zero.
8835 Previously something reasonably close to an infinite loop was
8836 possible because numbers could be growing instead of shrinking
8837 in the implementation of Euclid's algorithm.
8840 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8841 sign of the number in question.
8843 Fix BN_is_word(a,w) to work correctly for w == 0.
8845 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8846 because its test if the absolute value of 'a' equals 'w'.
8847 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8848 it exists mostly for use in the implementations of BN_is_zero(),
8849 BN_is_one(), and BN_is_word().
8852 *) New function BN_swap.
8855 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8856 the exponentiation functions are more likely to produce reasonable
8857 results on negative inputs.
8860 *) Change BN_mod_mul so that the result is always non-negative.
8861 Previously, it could be negative if one of the factors was negative;
8862 I don't think anyone really wanted that behaviour.
8865 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8866 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8867 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8868 and add new functions:
8877 BN_mod_lshift1_quick
8881 These functions always generate non-negative results.
8883 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8884 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8886 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8887 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8888 be reduced modulo m.
8889 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8892 The following entry accidentally appeared in the CHANGES file
8893 distributed with OpenSSL 0.9.7. The modifications described in
8894 it do *not* apply to OpenSSL 0.9.7.
8896 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8897 was actually never needed) and in BN_mul(). The removal in BN_mul()
8898 required a small change in bn_mul_part_recursive() and the addition
8899 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8900 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8901 bn_sub_words() and bn_add_words() except they take arrays with
8906 *) In 'openssl passwd', verify passwords read from the terminal
8907 unless the '-salt' option is used (which usually means that
8908 verification would just waste user's time since the resulting
8909 hash is going to be compared with some given password hash)
8910 or the new '-noverify' option is used.
8912 This is an incompatible change, but it does not affect
8913 non-interactive use of 'openssl passwd' (passwords on the command
8914 line, '-stdin' option, '-in ...' option) and thus should not
8918 *) Remove all references to RSAref, since there's no more need for it.
8921 *) Make DSO load along a path given through an environment variable
8922 (SHLIB_PATH) with shl_load().
8925 *) Constify the ENGINE code as a result of BIGNUM constification.
8926 Also constify the RSA code and most things related to it. In a
8927 few places, most notable in the depth of the ASN.1 code, ugly
8928 casts back to non-const were required (to be solved at a later
8932 *) Make it so the openssl application has all engines loaded by default.
8935 *) Constify the BIGNUM routines a little more.
8938 *) Add the following functions:
8940 ENGINE_load_cswift()
8942 ENGINE_load_atalla()
8944 ENGINE_load_builtin_engines()
8946 That way, an application can itself choose if external engines that
8947 are built-in in OpenSSL shall ever be used or not. The benefit is
8948 that applications won't have to be linked with libdl or other dso
8949 libraries unless it's really needed.
8951 Changed 'openssl engine' to load all engines on demand.
8952 Changed the engine header files to avoid the duplication of some
8953 declarations (they differed!).
8956 *) 'openssl engine' can now list capabilities.
8959 *) Better error reporting in 'openssl engine'.
8962 *) Never call load_dh_param(NULL) in s_server.
8965 *) Add engine application. It can currently list engines by name and
8966 identity, and test if they are actually available.
8969 *) Improve RPM specification file by forcing symbolic linking and making
8970 sure the installed documentation is also owned by root.root.
8971 [Damien Miller <djm@mindrot.org>]
8973 *) Give the OpenSSL applications more possibilities to make use of
8974 keys (public as well as private) handled by engines.
8977 *) Add OCSP code that comes from CertCo.
8980 *) Add VMS support for the Rijndael code.
8983 *) Added untested support for Nuron crypto accelerator.
8986 *) Add support for external cryptographic devices. This code was
8987 previously distributed separately as the "engine" branch.
8988 [Geoff Thorpe, Richard Levitte]
8990 *) Rework the filename-translation in the DSO code. It is now possible to
8991 have far greater control over how a "name" is turned into a filename
8992 depending on the operating environment and any oddities about the
8993 different shared library filenames on each system.
8996 *) Support threads on FreeBSD-elf in Configure.
8999 *) Fix for SHA1 assembly problem with MASM: it produces
9000 warnings about corrupt line number information when assembling
9001 with debugging information. This is caused by the overlapping
9003 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
9006 NCONF_get_number() has no error checking at all. As a replacement,
9007 NCONF_get_number_e() is defined (_e for "error checking") and is
9008 promoted strongly. The old NCONF_get_number is kept around for
9009 binary backward compatibility.
9010 Make it possible for methods to load from something other than a BIO,
9011 by providing a function pointer that is given a name instead of a BIO.
9012 For example, this could be used to load configuration data from an
9016 *) Fix for non blocking accept BIOs. Added new I/O special reason
9017 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
9018 with non blocking I/O was not possible because no retry code was
9019 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
9023 *) Added the beginnings of Rijndael support.
9026 *) Fix for bug in DirectoryString mask setting. Add support for
9027 X509_NAME_print_ex() in 'req' and X509_print_ex() function
9028 to allow certificate printing to more controllable, additional
9029 'certopt' option to 'x509' to allow new printing options to be
9033 *) Clean old EAY MD5 hack from e_os.h.
9036 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
9038 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
9039 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
9040 [Joe Orton, Steve Henson]
9042 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
9044 *) Fix additional bug revealed by the NISCC test suite:
9046 Stop bug triggering large recursion when presented with
9047 certain ASN.1 tags (CVE-2003-0851)
9050 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
9052 *) Fix various bugs revealed by running the NISCC test suite:
9054 Stop out of bounds reads in the ASN1 code when presented with
9055 invalid tags (CVE-2003-0543 and CVE-2003-0544).
9057 If verify callback ignores invalid public key errors don't try to check
9058 certificate signature with the NULL public key.
9062 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
9063 if the server requested one: as stated in TLS 1.0 and SSL 3.0
9067 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
9068 extra data after the compression methods not only for TLS 1.0
9069 but also for SSL 3.0 (as required by the specification).
9070 [Bodo Moeller; problem pointed out by Matthias Loepfe]
9072 *) Change X509_certificate_type() to mark the key as exported/exportable
9073 when it's 512 *bits* long, not 512 bytes.
9076 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
9078 *) Countermeasure against the Klima-Pokorny-Rosa extension of
9079 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
9080 a protocol version number mismatch like a decryption error
9081 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9084 *) Turn on RSA blinding by default in the default implementation
9085 to avoid a timing attack. Applications that don't want it can call
9086 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9087 They would be ill-advised to do so in most cases.
9088 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9090 *) Change RSA blinding code so that it works when the PRNG is not
9091 seeded (in this case, the secret RSA exponent is abused as
9092 an unpredictable seed -- if it is not unpredictable, there
9093 is no point in blinding anyway). Make RSA blinding thread-safe
9094 by remembering the creator's thread ID in rsa->blinding and
9095 having all other threads use local one-time blinding factors
9096 (this requires more computation than sharing rsa->blinding, but
9097 avoids excessive locking; and if an RSA object is not shared
9098 between threads, blinding will still be very fast).
9101 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9103 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9104 via timing by performing a MAC computation even if incorrect
9105 block cipher padding has been found. This is a countermeasure
9106 against active attacks where the attacker has to distinguish
9107 between bad padding and a MAC verification error. (CVE-2003-0078)
9109 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9110 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9111 Martin Vuagnoux (EPFL, Ilion)]
9113 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9115 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9116 memory from its contents. This is done with a counter that will
9117 place alternating values in each byte. This can be used to solve
9118 two issues: 1) the removal of calls to memset() by highly optimizing
9119 compilers, and 2) cleansing with other values than 0, since those can
9120 be read through on certain media, for example a swap space on disk.
9123 *) Bugfix: client side session caching did not work with external caching,
9124 because the session->cipher setting was not restored when reloading
9125 from the external cache. This problem was masked, when
9126 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9127 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9130 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9131 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9132 [Zeev Lieber <zeev-l@yahoo.com>]
9134 *) Undo an undocumented change introduced in 0.9.6e which caused
9135 repeated calls to OpenSSL_add_all_ciphers() and
9136 OpenSSL_add_all_digests() to be ignored, even after calling
9140 *) Change the default configuration reader to deal with last line not
9141 being properly terminated.
9144 *) Change X509_NAME_cmp() so it applies the special rules on handling
9145 DN values that are of type PrintableString, as well as RDNs of type
9146 emailAddress where the value has the type ia5String.
9147 [stefank@valicert.com via Richard Levitte]
9149 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9150 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9151 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9152 the bitwise-OR of the two for use by the majority of applications
9153 wanting this behaviour, and update the docs. The documented
9154 behaviour and actual behaviour were inconsistent and had been
9155 changing anyway, so this is more a bug-fix than a behavioural
9157 [Geoff Thorpe, diagnosed by Nadav Har'El]
9159 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9160 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9163 *) Fix initialization code race conditions in
9164 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9165 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9166 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9167 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9168 ssl2_get_cipher_by_char(),
9169 ssl3_get_cipher_by_char().
9170 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9172 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9173 the cached sessions are flushed, as the remove_cb() might use ex_data
9174 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9175 (see [openssl.org #212]).
9176 [Geoff Thorpe, Lutz Jaenicke]
9178 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9179 length, instead of the encoding length to d2i_ASN1_OBJECT.
9182 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9184 *) [In 0.9.6g-engine release:]
9185 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9186 [Lynn Gazis <lgazis@rainbow.com>]
9188 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9190 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9191 and get fix the header length calculation.
9192 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9193 Alon Kantor <alonk@checkpoint.com> (and others),
9196 *) Use proper error handling instead of 'assertions' in buffer
9197 overflow checks added in 0.9.6e. This prevents DoS (the
9198 assertions could call abort()).
9199 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9201 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9203 *) Add various sanity checks to asn1_get_length() to reject
9204 the ASN1 length bytes if they exceed sizeof(long), will appear
9205 negative or the content length exceeds the length of the
9207 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9209 *) Fix cipher selection routines: ciphers without encryption had no flags
9210 for the cipher strength set and where therefore not handled correctly
9211 by the selection routines (PR #130).
9214 *) Fix EVP_dsa_sha macro.
9218 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9219 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9220 that was added in OpenSSL 0.9.6d.
9222 As the countermeasure turned out to be incompatible with some
9223 broken SSL implementations, the new option is part of SSL_OP_ALL.
9224 SSL_OP_ALL is usually employed when compatibility with weird SSL
9225 implementations is desired (e.g. '-bugs' option to 's_client' and
9226 's_server'), so the new option is automatically set in many
9230 *) Changes in security patch:
9232 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9233 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9234 Air Force Materiel Command, USAF, under agreement number
9237 *) Add various sanity checks to asn1_get_length() to reject
9238 the ASN1 length bytes if they exceed sizeof(long), will appear
9239 negative or the content length exceeds the length of the
9240 supplied buffer. (CVE-2002-0659)
9241 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9243 *) Assertions for various potential buffer overflows, not known to
9245 [Ben Laurie (CHATS)]
9247 *) Various temporary buffers to hold ASCII versions of integers were
9248 too small for 64 bit platforms. (CVE-2002-0655)
9249 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9251 *) Remote buffer overflow in SSL3 protocol - an attacker could
9252 supply an oversized session ID to a client. (CVE-2002-0656)
9253 [Ben Laurie (CHATS)]
9255 *) Remote buffer overflow in SSL2 protocol - an attacker could
9256 supply an oversized client master key. (CVE-2002-0656)
9257 [Ben Laurie (CHATS)]
9259 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9261 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9262 encoded as NULL) with id-dsa-with-sha1.
9263 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9265 *) Check various X509_...() return values in apps/req.c.
9266 [Nils Larsch <nla@trustcenter.de>]
9268 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9269 an end-of-file condition would erroneously be flagged, when the CRLF
9270 was just at the end of a processed block. The bug was discovered when
9271 processing data through a buffering memory BIO handing the data to a
9272 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9273 <ptsekov@syntrex.com> and Nedelcho Stanev.
9276 *) Implement a countermeasure against a vulnerability recently found
9277 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9278 before application data chunks to avoid the use of known IVs
9279 with data potentially chosen by the attacker.
9282 *) Fix length checks in ssl3_get_client_hello().
9285 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9286 to prevent ssl3_read_internal() from incorrectly assuming that
9287 ssl3_read_bytes() found application data while handshake
9288 processing was enabled when in fact s->s3->in_read_app_data was
9289 merely automatically cleared during the initial handshake.
9290 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9292 *) Fix object definitions for Private and Enterprise: they were not
9293 recognized in their shortname (=lowercase) representation. Extend
9294 obj_dat.pl to issue an error when using undefined keywords instead
9295 of silently ignoring the problem (Svenning Sorensen
9296 <sss@sss.dnsalias.net>).
9299 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9300 generators, i.e. generators other than 2 and 5. (Previously, the
9301 code did not properly initialise the 'add' and 'rem' values to
9302 BN_generate_prime().)
9304 In the new general case, we do not insist that 'generator' is
9305 actually a primitive root: This requirement is rather pointless;
9306 a generator of the order-q subgroup is just as good, if not
9310 *) Map new X509 verification errors to alerts. Discovered and submitted by
9311 Tom Wu <tom@arcot.com>.
9314 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9315 returning non-zero before the data has been completely received
9316 when using non-blocking I/O.
9317 [Bodo Moeller; problem pointed out by John Hughes]
9319 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9320 [Ben Laurie, Lutz Jaenicke]
9322 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9323 Yoram Zahavi <YoramZ@gilian.com>).
9326 *) Add information about CygWin 1.3 and on, and preserve proper
9327 configuration for the versions before that.
9328 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9330 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9331 check whether we deal with a copy of a session and do not delete from
9332 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9333 <izhar@checkpoint.com>.
9336 *) Do not store session data into the internal session cache, if it
9337 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9338 flag is set). Proposed by Aslam <aslam@funk.com>.
9341 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9345 *) [In 0.9.6d-engine release:]
9346 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9347 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9349 *) Add the configuration target linux-s390x.
9350 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9352 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9353 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9354 variable as an indication that a ClientHello message has been
9355 received. As the flag value will be lost between multiple
9356 invocations of ssl3_accept when using non-blocking I/O, the
9357 function may not be aware that a handshake has actually taken
9358 place, thus preventing a new session from being added to the
9361 To avoid this problem, we now set s->new_session to 2 instead of
9362 using a local variable.
9363 [Lutz Jaenicke, Bodo Moeller]
9365 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9366 if the SSL_R_LENGTH_MISMATCH error is detected.
9367 [Geoff Thorpe, Bodo Moeller]
9369 *) New 'shared_ldflag' column in Configure platform table.
9372 *) Fix EVP_CIPHER_mode macro.
9373 ["Dan S. Camper" <dan@bti.net>]
9375 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9376 type, we must throw them away by setting rr->length to 0.
9377 [D P Chang <dpc@qualys.com>]
9379 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9381 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9382 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9383 worked incorrectly for those cases where range = 10..._2 and
9384 3*range is two bits longer than range.)
9387 *) Only add signing time to PKCS7 structures if it is not already
9391 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9392 OBJ_ld_ce should be OBJ_id_ce.
9393 Also some ip-pda OIDs in crypto/objects/objects.txt were
9394 incorrect (cf. RFC 3039).
9395 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9397 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9398 returns early because it has nothing to do.
9399 [Andy Schneider <andy.schneider@bjss.co.uk>]
9401 *) [In 0.9.6c-engine release:]
9402 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9403 [Andy Schneider <andy.schneider@bjss.co.uk>]
9405 *) [In 0.9.6c-engine release:]
9406 Add support for Cryptographic Appliance's keyserver technology.
9407 (Use engine 'keyclient')
9408 [Cryptographic Appliances and Geoff Thorpe]
9410 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9411 is called via tools/c89.sh because arguments have to be
9412 rearranged (all '-L' options must appear before the first object
9414 [Richard Shapiro <rshapiro@abinitio.com>]
9416 *) [In 0.9.6c-engine release:]
9417 Add support for Broadcom crypto accelerator cards, backported
9419 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9421 *) [In 0.9.6c-engine release:]
9422 Add support for SureWare crypto accelerator cards from
9423 Baltimore Technologies. (Use engine 'sureware')
9424 [Baltimore Technologies and Mark Cox]
9426 *) [In 0.9.6c-engine release:]
9427 Add support for crypto accelerator cards from Accelerated
9428 Encryption Processing, www.aep.ie. (Use engine 'aep')
9429 [AEP Inc. and Mark Cox]
9431 *) Add a configuration entry for gcc on UnixWare.
9432 [Gary Benson <gbenson@redhat.com>]
9434 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9435 messages are stored in a single piece (fixed-length part and
9436 variable-length part combined) and fix various bugs found on the way.
9439 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9440 instead. BIO_gethostbyname() does not know what timeouts are
9441 appropriate, so entries would stay in cache even when they have
9443 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9445 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9446 faced with a pathologically small ClientHello fragment that does
9447 not contain client_version: Instead of aborting with an error,
9448 simply choose the highest available protocol version (i.e.,
9449 TLS 1.0 unless it is disabled). In practice, ClientHello
9450 messages are never sent like this, but this change gives us
9451 strictly correct behaviour at least for TLS.
9454 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9455 never resets s->method to s->ctx->method when called from within
9456 one of the SSL handshake functions.
9457 [Bodo Moeller; problem pointed out by Niko Baric]
9459 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9460 (sent using the client's version number) if client_version is
9461 smaller than the protocol version in use. Also change
9462 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9463 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9464 the client will at least see that alert.
9467 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9471 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9472 client receives HelloRequest while in a handshake.
9473 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9475 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9476 should end in 'break', not 'goto end' which circumvents various
9477 cleanups done in state SSL_ST_OK. But session related stuff
9478 must be disabled for SSL_ST_OK in the case that we just sent a
9481 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9482 before just sending a HelloRequest.
9483 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9485 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9486 reveal whether illegal block cipher padding was found or a MAC
9487 verification error occurred. (Neither SSLerr() codes nor alerts
9488 are directly visible to potential attackers, but the information
9489 may leak via logfiles.)
9491 Similar changes are not required for the SSL 2.0 implementation
9492 because the number of padding bytes is sent in clear for SSL 2.0,
9493 and the extra bytes are just ignored. However ssl/s2_pkt.c
9494 failed to verify that the purported number of padding bytes is in
9498 *) Add OpenUNIX-8 support including shared libraries
9499 (Boyd Lynn Gerber <gerberb@zenez.com>).
9502 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9503 'wristwatch attack' using huge encoding parameters (cf.
9504 James H. Manger's CRYPTO 2001 paper). Note that the
9505 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9506 encoding parameters and hence was not vulnerable.
9509 *) BN_sqr() bug fix.
9510 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9512 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9513 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9514 followed by modular reduction.
9515 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9517 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9518 equivalent based on BN_pseudo_rand() instead of BN_rand().
9521 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9522 This function was broken, as the check for a new client hello message
9523 to handle SGC did not allow these large messages.
9524 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9527 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9530 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9531 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9534 *) Rework the configuration and shared library support for Tru64 Unix.
9535 The configuration part makes use of modern compiler features and
9536 still retains old compiler behavior for those that run older versions
9537 of the OS. The shared library support part includes a variant that
9538 uses the RPATH feature, and is available through the special
9539 configuration target "alpha-cc-rpath", which will never be selected
9541 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9543 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9544 with the same message size as in ssl3_get_certificate_request().
9545 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9546 messages might inadvertently be reject as too long.
9547 [Petr Lampa <lampa@fee.vutbr.cz>]
9549 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9552 *) Modified SSL library such that the verify_callback that has been set
9553 specificly for an SSL object with SSL_set_verify() is actually being
9554 used. Before the change, a verify_callback set with this function was
9555 ignored and the verify_callback() set in the SSL_CTX at the time of
9556 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9557 to allow the necessary settings.
9560 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9561 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9562 done automatically (in contradiction to the requirements of the C
9563 standard). This made problems when used from OpenSSH.
9566 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9567 dh->length and always used
9569 BN_rand_range(priv_key, dh->p).
9571 BN_rand_range() is not necessary for Diffie-Hellman, and this
9572 specific range makes Diffie-Hellman unnecessarily inefficient if
9573 dh->length (recommended exponent length) is much smaller than the
9574 length of dh->p. We could use BN_rand_range() if the order of
9575 the subgroup was stored in the DH structure, but we only have
9580 BN_rand(priv_key, l, ...)
9582 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9588 RSA_eay_public_encrypt
9589 RSA_eay_private_decrypt
9590 RSA_eay_private_encrypt (signing)
9591 RSA_eay_public_decrypt (signature verification)
9593 (default implementations for RSA_public_encrypt,
9594 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9595 always reject numbers >= n.
9598 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9599 to synchronize access to 'locking_thread'. This is necessary on
9600 systems where access to 'locking_thread' (an 'unsigned long'
9601 variable) is not atomic.
9604 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9605 *before* setting the 'crypto_lock_rand' flag. The previous code had
9606 a race condition if 0 is a valid thread ID.
9607 [Travis Vitek <vitek@roguewave.com>]
9609 *) Add support for shared libraries under Irix.
9610 [Albert Chin-A-Young <china@thewrittenword.com>]
9612 *) Add configuration option to build on Linux on both big-endian and
9614 [Ralf Baechle <ralf@uni-koblenz.de>]
9616 *) Add the possibility to create shared libraries on HP-UX.
9619 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9621 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9622 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9623 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9624 PRNG state recovery was possible based on the output of
9625 one PRNG request appropriately sized to gain knowledge on
9626 'md' followed by enough consecutive 1-byte PRNG requests
9627 to traverse all of 'state'.
9629 1. When updating 'md_local' (the current thread's copy of 'md')
9630 during PRNG output generation, hash all of the previous
9631 'md_local' value, not just the half used for PRNG output.
9633 2. Make the number of bytes from 'state' included into the hash
9634 independent from the number of PRNG bytes requested.
9636 The first measure alone would be sufficient to avoid
9637 Markku-Juhani's attack. (Actually it had never occurred
9638 to me that the half of 'md_local' used for chaining was the
9639 half from which PRNG output bytes were taken -- I had always
9640 assumed that the secret half would be used.) The second
9641 measure makes sure that additional data from 'state' is never
9642 mixed into 'md_local' in small portions; this heuristically
9643 further strengthens the PRNG.
9646 *) Fix crypto/bn/asm/mips3.s.
9649 *) When only the key is given to "enc", the IV is undefined. Print out
9650 an error message in this case.
9653 *) Handle special case when X509_NAME is empty in X509 printing routines.
9656 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9657 positive and less than q.
9660 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9661 used: it isn't thread safe and the add_lock_callback should handle
9663 [Paul Rose <Paul.Rose@bridge.com>]
9665 *) Verify that incoming data obeys the block size in
9666 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9670 [Ulf Möller, Bodo Möller]
9672 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9673 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9674 when fixing the server behaviour for backwards-compatible 'client
9675 hello' messages. (Note that the attack is impractical against
9676 SSL 3.0 and TLS 1.0 anyway because length and version checking
9677 means that the probability of guessing a valid ciphertext is
9678 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9681 Before 0.9.5, the countermeasure (hide the error by generating a
9682 random 'decryption result') did not work properly because
9683 ERR_clear_error() was missing, meaning that SSL_get_error() would
9684 detect the supposedly ignored error.
9686 Both problems are now fixed.
9689 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9690 (previously it was 1024).
9693 *) Fix for compatibility mode trust settings: ignore trust settings
9694 unless some valid trust or reject settings are present.
9697 *) Fix for blowfish EVP: its a variable length cipher.
9700 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9701 parameters in DSA public key structures and return an error in the
9702 DSA routines if parameters are absent.
9705 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9706 in the current directory if neither $RANDFILE nor $HOME was set.
9707 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9708 caused some confusion to Windows users who haven't defined $HOME.
9709 Thus RAND_file_name() is changed again: e_os.h can define a
9710 DEFAULT_HOME, which will be used if $HOME is not set.
9711 For Windows, we use "C:"; on other platforms, we still require
9712 environment variables.
9714 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9715 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9716 having multiple threads call RAND_poll() concurrently.
9719 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9720 combination of a flag and a thread ID variable.
9721 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9722 flag), *other* threads can enter ssleay_add_bytes without obeying
9723 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9724 that they do not hold after the first thread unsets add_do_not_lock).
9727 *) Change bctest again: '-x' expressions are not available in all
9731 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9733 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9734 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9736 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9737 the default extension for executables, if any. Also, make the perl
9738 scripts that use symlink() to test if it really exists and use "cp"
9739 if it doesn't. All this made OpenSSL compilable and installable in
9743 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9744 If SEQUENCE is length is indefinite just set c->slen to the total
9745 amount of data available.
9746 [Steve Henson, reported by shige@FreeBSD.org]
9747 [This change does not apply to 0.9.7.]
9749 *) Change bctest to avoid here-documents inside command substitution
9750 (workaround for FreeBSD /bin/sh bug).
9751 For compatibility with Ultrix, avoid shell functions (introduced
9752 in the bctest version that searches along $PATH).
9755 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9756 with des_encrypt() defined on some operating systems, like Solaris
9760 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9761 On the Importance of Eliminating Errors in Cryptographic
9762 Computations, J. Cryptology 14 (2001) 2, 101-119,
9763 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9766 *) MIPS assembler BIGNUM division bug fix.
9769 *) Disabled incorrect Alpha assembler code.
9772 *) Fix PKCS#7 decode routines so they correctly update the length
9773 after reading an EOC for the EXPLICIT tag.
9775 [This change does not apply to 0.9.7.]
9777 *) Fix bug in PKCS#12 key generation routines. This was triggered
9778 if a 3DES key was generated with a 0 initial byte. Include
9779 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9780 (but broken) behaviour.
9783 *) Enhance bctest to search for a working bc along $PATH and print
9785 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9787 *) Fix memory leaks in err.c: free err_data string if necessary;
9788 don't write to the wrong index in ERR_set_error_data.
9791 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9795 *) Replace rdtsc with _emit statements for VC++ version 5.
9796 [Jeremy Cooper <jeremy@baymoo.org>]
9798 *) Make it possible to reuse SSLv2 sessions.
9801 *) In copy_email() check for >= 0 as a return value for
9802 X509_NAME_get_index_by_NID() since 0 is a valid index.
9803 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9805 *) Avoid coredump with unsupported or invalid public keys by checking if
9806 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9807 PKCS7_verify() fails with non detached data.
9810 *) Don't use getenv in library functions when run as setuid/setgid.
9811 New function OPENSSL_issetugid().
9814 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9815 due to incorrect handling of multi-threading:
9817 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9819 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9821 3. Count how many times MemCheck_off() has been called so that
9822 nested use can be treated correctly. This also avoids
9823 inband-signalling in the previous code (which relied on the
9824 assumption that thread ID 0 is impossible).
9827 *) Add "-rand" option also to s_client and s_server.
9830 *) Fix CPU detection on Irix 6.x.
9831 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9832 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9834 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9837 [This change does not apply to 0.9.7.]
9839 *) Use the cached encoding of an X509_NAME structure rather than
9840 copying it. This is apparently the reason for the libsafe "errors"
9841 but the code is actually correct.
9844 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9845 Bleichenbacher's DSA attack.
9846 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9847 to be set and top=0 forces the highest bit to be set; top=-1 is new
9848 and leaves the highest bit random.
9849 [Ulf Moeller, Bodo Moeller]
9851 *) In the NCONF_...-based implementations for CONF_... queries
9852 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9853 a temporary CONF structure with the data component set to NULL
9854 (which gives segmentation faults in lh_retrieve).
9855 Instead, use NULL for the CONF pointer in CONF_get_string and
9856 CONF_get_number (which may use environment variables) and directly
9857 return NULL from CONF_get_section.
9860 *) Fix potential buffer overrun for EBCDIC.
9863 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9864 keyUsage if basicConstraints absent for a CA.
9867 *) Make SMIME_write_PKCS7() write mail header values with a format that
9868 is more generally accepted (no spaces before the semicolon), since
9869 some programs can't parse those values properly otherwise. Also make
9870 sure BIO's that break lines after each write do not create invalid
9874 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9875 macros previously used would not encode an empty SEQUENCE OF
9876 and break the signature.
9878 [This change does not apply to 0.9.7.]
9880 *) Zero the premaster secret after deriving the master secret in
9884 *) Add some EVP_add_digest_alias registrations (as found in
9885 OpenSSL_add_all_digests()) to SSL_library_init()
9886 aka OpenSSL_add_ssl_algorithms(). This provides improved
9887 compatibility with peers using X.509 certificates
9888 with unconventional AlgorithmIdentifier OIDs.
9891 *) Fix for Irix with NO_ASM.
9892 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9894 *) ./config script fixes.
9895 [Ulf Moeller, Richard Levitte]
9897 *) Fix 'openssl passwd -1'.
9900 *) Change PKCS12_key_gen_asc() so it can cope with non null
9901 terminated strings whose length is passed in the passlen
9902 parameter, for example from PEM callbacks. This was done
9903 by adding an extra length parameter to asc2uni().
9904 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9906 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9907 call failed, free the DSA structure.
9910 *) Fix to uni2asc() to cope with zero length Unicode strings.
9911 These are present in some PKCS#12 files.
9914 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9915 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9916 when writing a 32767 byte record.
9917 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9919 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9920 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9922 (RSA objects have a reference count access to which is protected
9923 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9924 so they are meant to be shared between threads.)
9925 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9926 "Reddie, Steven" <Steven.Reddie@ca.com>]
9928 *) Fix a deadlock in CRYPTO_mem_leaks().
9931 *) Use better test patterns in bntest.
9934 *) rand_win.c fix for Borland C.
9937 *) BN_rshift bugfix for n == 0.
9940 *) Add a 'bctest' script that checks for some known 'bc' bugs
9941 so that 'make test' does not abort just because 'bc' is broken.
9944 *) Store verify_result within SSL_SESSION also for client side to
9945 avoid potential security hole. (Re-used sessions on the client side
9946 always resulted in verify_result==X509_V_OK, not using the original
9947 result of the server certificate verification.)
9950 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9951 SSL3_RT_APPLICATION_DATA, return 0.
9952 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9956 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9957 releases, have been re-implemented by renaming the previous
9958 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9959 and ssl3_read_internal, respectively, and adding 'peek' parameters
9960 to them. The new ssl[23]_{read,peek} functions are calls to
9961 ssl[23]_read_internal with the 'peek' flag set appropriately.
9962 A 'peek' parameter has also been added to ssl3_read_bytes, which
9963 does the actual work for ssl3_read_internal.
9966 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9967 the method-specific "init()" handler. Also clean up ex_data after
9968 calling the method-specific "finish()" handler. Previously, this was
9969 happening the other way round.
9972 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9973 The previous value, 12, was not always sufficient for BN_mod_exp().
9976 *) Make sure that shared libraries get the internal name engine with
9977 the full version number and not just 0. This should mark the
9978 shared libraries as not backward compatible. Of course, this should
9979 be changed again when we can guarantee backward binary compatibility.
9982 *) Fix typo in get_cert_by_subject() in by_dir.c
9983 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9985 *) Rework the system to generate shared libraries:
9987 - Make note of the expected extension for the shared libraries and
9988 if there is a need for symbolic links from for example libcrypto.so.0
9989 to libcrypto.so.0.9.7. There is extended info in Configure for
9992 - Make as few rebuilds of the shared libraries as possible.
9994 - Still avoid linking the OpenSSL programs with the shared libraries.
9996 - When installing, install the shared libraries separately from the
10000 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
10002 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
10003 and not in SSL_clear because the latter is also used by the
10004 accept/connect functions; previously, the settings made by
10005 SSL_set_read_ahead would be lost during the handshake.
10006 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
10008 *) Correct util/mkdef.pl to be selective about disabled algorithms.
10009 Previously, it would create entries for disabled algorithms no
10013 *) Added several new manual pages for SSL_* function.
10016 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
10018 *) In ssl23_get_client_hello, generate an error message when faced
10019 with an initial SSL 3.0/TLS record that is too small to contain the
10020 first two bytes of the ClientHello message, i.e. client_version.
10021 (Note that this is a pathologic case that probably has never happened
10022 in real life.) The previous approach was to use the version number
10023 from the record header as a substitute; but our protocol choice
10024 should not depend on that one because it is not authenticated
10025 by the Finished messages.
10028 *) More robust randomness gathering functions for Windows.
10029 [Jeffrey Altman <jaltman@columbia.edu>]
10031 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
10032 not set then we don't setup the error code for issuer check errors
10033 to avoid possibly overwriting other errors which the callback does
10034 handle. If an application does set the flag then we assume it knows
10035 what it is doing and can handle the new informational codes
10039 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
10040 a general "ANY" type, as such it should be able to decode anything
10041 including tagged types. However it didn't check the class so it would
10042 wrongly interpret tagged types in the same way as their universal
10043 counterpart and unknown types were just rejected. Changed so that the
10044 tagged and unknown types are handled in the same way as a SEQUENCE:
10045 that is the encoding is stored intact. There is also a new type
10046 "V_ASN1_OTHER" which is used when the class is not universal, in this
10047 case we have no idea what the actual type is so we just lump them all
10051 *) On VMS, stdout may very well lead to a file that is written to
10052 in a record-oriented fashion. That means that every write() will
10053 write a separate record, which will be read separately by the
10054 programs trying to read from it. This can be very confusing.
10056 The solution is to put a BIO filter in the way that will buffer
10057 text until a linefeed is reached, and then write everything a
10058 line at a time, so every record written will be an actual line,
10059 not chunks of lines and not (usually doesn't happen, but I've
10060 seen it once) several lines in one record. BIO_f_linebuffer() is
10063 Currently, it's a VMS-only method, because that's where it has
10064 been tested well enough.
10067 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
10068 it can return incorrect results.
10069 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
10070 but it was in 0.9.6-beta[12].)
10073 *) Disable the check for content being present when verifying detached
10074 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
10075 include zero length content when signing messages.
10078 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
10079 BIO_ctrl (for BIO pairs).
10082 *) Add DSO method for VMS.
10085 *) Bug fix: Montgomery multiplication could produce results with the
10089 *) Add RPM specification openssl.spec and modify it to build three
10090 packages. The default package contains applications, application
10091 documentation and run-time libraries. The devel package contains
10092 include files, static libraries and function documentation. The
10093 doc package contains the contents of the doc directory. The original
10094 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10097 *) Add a large number of documentation files for many SSL routines.
10098 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10100 *) Add a configuration entry for Sony News 4.
10101 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10103 *) Don't set the two most significant bits to one when generating a
10104 random number < q in the DSA library.
10107 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10108 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10109 the underlying transport is blocking) if a handshake took place.
10110 (The default behaviour is needed by applications such as s_client
10111 and s_server that use select() to determine when to use SSL_read;
10112 but for applications that know in advance when to expect data, it
10113 just makes things more complicated.)
10116 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10120 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10121 work better on such systems.
10122 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10124 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10125 Update PKCS12_parse() so it copies the friendlyName and the
10126 keyid to the certificates aux info.
10129 *) Fix bug in PKCS7_verify() which caused an infinite loop
10130 if there was more than one signature.
10131 [Sven Uszpelkat <su@celocom.de>]
10133 *) Major change in util/mkdef.pl to include extra information
10134 about each symbol, as well as presenting variables as well
10135 as functions. This change means that there's n more need
10136 to rebuild the .num files when some algorithms are excluded.
10139 *) Allow the verify time to be set by an application,
10140 rather than always using the current time.
10143 *) Phase 2 verify code reorganisation. The certificate
10144 verify code now looks up an issuer certificate by a
10145 number of criteria: subject name, authority key id
10146 and key usage. It also verifies self signed certificates
10147 by the same criteria. The main comparison function is
10148 X509_check_issued() which performs these checks.
10150 Lot of changes were necessary in order to support this
10151 without completely rewriting the lookup code.
10153 Authority and subject key identifier are now cached.
10155 The LHASH 'certs' is X509_STORE has now been replaced
10156 by a STACK_OF(X509_OBJECT). This is mainly because an
10157 LHASH can't store or retrieve multiple objects with
10158 the same hash value.
10160 As a result various functions (which were all internal
10161 use only) have changed to handle the new X509_STORE
10162 structure. This will break anything that messed round
10163 with X509_STORE internally.
10165 The functions X509_STORE_add_cert() now checks for an
10166 exact match, rather than just subject name.
10168 The X509_STORE API doesn't directly support the retrieval
10169 of multiple certificates matching a given criteria, however
10170 this can be worked round by performing a lookup first
10171 (which will fill the cache with candidate certificates)
10172 and then examining the cache for matches. This is probably
10173 the best we can do without throwing out X509_LOOKUP
10174 entirely (maybe later...).
10176 The X509_VERIFY_CTX structure has been enhanced considerably.
10178 All certificate lookup operations now go via a get_issuer()
10179 callback. Although this currently uses an X509_STORE it
10180 can be replaced by custom lookups. This is a simple way
10181 to bypass the X509_STORE hackery necessary to make this
10182 work and makes it possible to use more efficient techniques
10183 in future. A very simple version which uses a simple
10184 STACK for its trusted certificate store is also provided
10185 using X509_STORE_CTX_trusted_stack().
10187 The verify_cb() and verify() callbacks now have equivalents
10188 in the X509_STORE_CTX structure.
10190 X509_STORE_CTX also has a 'flags' field which can be used
10191 to customise the verify behaviour.
10194 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10195 excludes S/MIME capabilities.
10198 *) When a certificate request is read in keep a copy of the
10199 original encoding of the signed data and use it when outputting
10200 again. Signatures then use the original encoding rather than
10201 a decoded, encoded version which may cause problems if the
10202 request is improperly encoded.
10205 *) For consistency with other BIO_puts implementations, call
10206 buffer_write(b, ...) directly in buffer_puts instead of calling
10209 In BIO_puts, increment b->num_write as in BIO_write.
10210 [Peter.Sylvester@EdelWeb.fr]
10212 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10213 BN_zero, we may not return a BIGNUM with an array consisting of
10214 words set to zero.)
10217 *) Avoid calling abort() from within the library when problems are
10218 detected, except if preprocessor symbols have been defined
10219 (such as REF_CHECK, BN_DEBUG etc.).
10222 *) New openssl application 'rsautl'. This utility can be
10223 used for low level RSA operations. DER public key
10224 BIO/fp routines also added.
10227 *) New Configure entry and patches for compiling on QNX 4.
10228 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10230 *) A demo state-machine implementation was sponsored by
10231 Nuron (http://www.nuron.com/) and is now available in
10232 demos/state_machine.
10235 *) New options added to the 'dgst' utility for signature
10236 generation and verification.
10239 *) Unrecognized PKCS#7 content types are now handled via a
10240 catch all ASN1_TYPE structure. This allows unsupported
10241 types to be stored as a "blob" and an application can
10242 encode and decode it manually.
10245 *) Fix various signed/unsigned issues to make a_strex.c
10246 compile under VC++.
10247 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10249 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10250 length if passed a buffer. ASN1_INTEGER_to_BN failed
10251 if passed a NULL BN and its argument was negative.
10252 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10254 *) Modification to PKCS#7 encoding routines to output definite
10255 length encoding. Since currently the whole structures are in
10256 memory there's not real point in using indefinite length
10257 constructed encoding. However if OpenSSL is compiled with
10258 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10261 *) Added BIO_vprintf() and BIO_vsnprintf().
10264 *) Added more prefixes to parse for in the strings written
10265 through a logging bio, to cover all the levels that are available
10266 through syslog. The prefixes are now:
10268 PANIC, EMERG, EMR => LOG_EMERG
10269 ALERT, ALR => LOG_ALERT
10270 CRIT, CRI => LOG_CRIT
10271 ERROR, ERR => LOG_ERR
10272 WARNING, WARN, WAR => LOG_WARNING
10273 NOTICE, NOTE, NOT => LOG_NOTICE
10274 INFO, INF => LOG_INFO
10275 DEBUG, DBG => LOG_DEBUG
10277 and as before, if none of those prefixes are present at the
10278 beginning of the string, LOG_ERR is chosen.
10280 On Win32, the LOG_* levels are mapped according to this:
10282 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10283 LOG_WARNING => EVENTLOG_WARNING_TYPE
10284 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10288 *) Made it possible to reconfigure with just the configuration
10289 argument "reconf" or "reconfigure". The command line arguments
10290 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10291 and are retrieved from there when reconfiguring.
10294 *) MD4 implemented.
10295 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10297 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10300 *) The obj_dat.pl script was messing up the sorting of object
10301 names. The reason was that it compared the quoted version
10302 of strings as a result "OCSP" > "OCSP Signing" because
10303 " > SPACE. Changed script to store unquoted versions of
10304 names and add quotes on output. It was also omitting some
10305 names from the lookup table if they were given a default
10306 value (that is if SN is missing it is given the same
10307 value as LN and vice versa), these are now added on the
10308 grounds that if an object has a name we should be able to
10309 look it up. Finally added warning output when duplicate
10310 short or long names are found.
10313 *) Changes needed for Tandem NSK.
10314 [Scott Uroff <scott@xypro.com>]
10316 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10317 RSA_padding_check_SSLv23(), special padding was never detected
10318 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10319 version rollback attacks was not effective.
10321 In s23_clnt.c, don't use special rollback-attack detection padding
10322 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10323 client; similarly, in s23_srvr.c, don't do the rollback check if
10324 SSL 2.0 is the only protocol enabled in the server.
10327 *) Make it possible to get hexdumps of unprintable data with 'openssl
10328 asn1parse'. By implication, the functions ASN1_parse_dump() and
10329 BIO_dump_indent() are added.
10332 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10333 these print out strings and name structures based on various
10334 flags including RFC2253 support and proper handling of
10335 multibyte characters. Added options to the 'x509' utility
10336 to allow the various flags to be set.
10339 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10340 Also change the functions X509_cmp_current_time() and
10341 X509_gmtime_adj() work with an ASN1_TIME structure,
10342 this will enable certificates using GeneralizedTime in validity
10343 dates to be checked.
10346 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10347 negative public key encodings) on by default,
10348 NO_NEG_PUBKEY_BUG can be set to disable it.
10351 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10352 content octets. An i2c_ASN1_OBJECT is unnecessary because
10353 the encoding can be trivially obtained from the structure.
10356 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10357 not read locks (CRYPTO_r_[un]lock).
10360 *) A first attempt at creating official support for shared
10361 libraries through configuration. I've kept it so the
10362 default is static libraries only, and the OpenSSL programs
10363 are always statically linked for now, but there are
10364 preparations for dynamic linking in place.
10365 This has been tested on Linux and Tru64.
10368 *) Randomness polling function for Win9x, as described in:
10369 Peter Gutmann, Software Generation of Practically Strong
10373 *) Fix so PRNG is seeded in req if using an already existing
10377 *) New options to smime application. -inform and -outform
10378 allow alternative formats for the S/MIME message including
10379 PEM and DER. The -content option allows the content to be
10380 specified separately. This should allow things like Netscape
10381 form signing output easier to verify.
10384 *) Fix the ASN1 encoding of tags using the 'long form'.
10387 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10388 STRING types. These convert content octets to and from the
10389 underlying type. The actual tag and length octets are
10390 already assumed to have been read in and checked. These
10391 are needed because all other string types have virtually
10392 identical handling apart from the tag. By having versions
10393 of the ASN1 functions that just operate on content octets
10394 IMPLICIT tagging can be handled properly. It also allows
10395 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10396 and ASN1_INTEGER are identical apart from the tag.
10399 *) Change the handling of OID objects as follows:
10401 - New object identifiers are inserted in objects.txt, following
10402 the syntax given in objects.README.
10403 - objects.pl is used to process obj_mac.num and create a new
10405 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10408 This is currently kind of a hack, and the perl code in objects.pl
10409 isn't very elegant, but it works as I intended. The simplest way
10410 to check that it worked correctly is to look in obj_dat.h and
10411 check the array nid_objs and make sure the objects haven't moved
10412 around (this is important!). Additions are OK, as well as
10413 consistent name changes.
10416 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10419 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10420 The given file adds to whatever has already been seeded into the
10421 random pool through the RANDFILE configuration file option or
10422 environment variable, or the default random state file.
10425 *) mkstack.pl now sorts each macro group into lexical order.
10426 Previously the output order depended on the order the files
10427 appeared in the directory, resulting in needless rewriting
10431 *) Patches to make OpenSSL compile under Win32 again. Mostly
10432 work arounds for the VC++ problem that it treats func() as
10433 func(void). Also stripped out the parts of mkdef.pl that
10434 added extra typesafe functions: these no longer exist.
10437 *) Reorganisation of the stack code. The macros are now all
10438 collected in safestack.h . Each macro is defined in terms of
10439 a "stack macro" of the form SKM_<name>(type, a, b). The
10440 DEBUG_SAFESTACK is now handled in terms of function casts,
10441 this has the advantage of retaining type safety without the
10442 use of additional functions. If DEBUG_SAFESTACK is not defined
10443 then the non typesafe macros are used instead. Also modified the
10444 mkstack.pl script to handle the new form. Needs testing to see
10445 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10446 the default if no major problems. Similar behaviour for ASN1_SET_OF
10447 and PKCS12_STACK_OF.
10450 *) When some versions of IIS use the 'NET' form of private key the
10451 key derivation algorithm is different. Normally MD5(password) is
10452 used as a 128 bit RC4 key. In the modified case
10453 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10454 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10455 as the old Netscape_RSA functions except they have an additional
10456 'sgckey' parameter which uses the modified algorithm. Also added
10457 an -sgckey command line option to the rsa utility. Thanks to
10458 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10459 algorithm to openssl-dev.
10462 *) The evp_local.h macros were using 'c.##kname' which resulted in
10463 invalid expansion on some systems (SCO 5.0.5 for example).
10464 Corrected to 'c.kname'.
10465 [Phillip Porch <root@theporch.com>]
10467 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10468 a STACK of email addresses from a certificate or request, these look
10469 in the subject name and the subject alternative name extensions and
10470 omit any duplicate addresses.
10473 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10474 This makes DSA verification about 2 % faster.
10477 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10478 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10479 plus overhead for 1024 bit moduli).
10480 This makes exponentiations about 0.5 % faster for 1024 bit
10481 exponents (as measured by "openssl speed rsa2048").
10484 *) Rename memory handling macros to avoid conflicts with other
10486 Malloc => OPENSSL_malloc
10487 Malloc_locked => OPENSSL_malloc_locked
10488 Realloc => OPENSSL_realloc
10489 Free => OPENSSL_free
10492 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10493 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10496 *) CygWin32 support.
10497 [John Jarvie <jjarvie@newsguy.com>]
10499 *) The type-safe stack code has been rejigged. It is now only compiled
10500 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10501 by default all type-specific stack functions are "#define"d back to
10502 standard stack functions. This results in more streamlined output
10503 but retains the type-safety checking possibilities of the original
10507 *) The STACK code has been cleaned up, and certain type declarations
10508 that didn't make a lot of sense have been brought in line. This has
10509 also involved a cleanup of sorts in safestack.h to more correctly
10510 map type-safe stack functions onto their plain stack counterparts.
10511 This work has also resulted in a variety of "const"ifications of
10512 lots of the code, especially "_cmp" operations which should normally
10513 be prototyped with "const" parameters anyway.
10516 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10517 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10518 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10519 where all of 'md' is used each time the PRNG is used, but 'state'
10520 is used only indexed by a cyclic counter. As entropy may not be
10521 well distributed from the beginning, 'md' is important as a
10522 chaining variable. However, the output function chains only half
10523 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10524 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10525 in all of 'state' being rewritten, with the new values depending
10526 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10529 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10530 the handshake is continued after ssl_verify_cert_chain();
10531 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10532 can lead to 'unexplainable' connection aborts later.
10533 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10535 *) Major EVP API cipher revision.
10536 Add hooks for extra EVP features. This allows various cipher
10537 parameters to be set in the EVP interface. Support added for variable
10538 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10539 setting of RC2 and RC5 parameters.
10541 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10544 Remove lots of duplicated code from the EVP library. For example *every*
10545 cipher init() function handles the 'iv' in the same way according to the
10546 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10547 for CFB and OFB modes they zero ctx->num.
10549 New functionality allows removal of S/MIME code RC2 hack.
10551 Most of the routines have the same form and so can be declared in terms
10554 By shifting this to the top level EVP_CipherInit() it can be removed from
10555 all individual ciphers. If the cipher wants to handle IVs or keys
10556 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10559 Change lots of functions like EVP_EncryptUpdate() to now return a
10560 value: although software versions of the algorithms cannot fail
10561 any installed hardware versions can.
10564 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10565 this option is set, tolerate broken clients that send the negotiated
10566 protocol version number instead of the requested protocol version
10570 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10571 i.e. non-zero for export ciphersuites, zero otherwise.
10572 Previous versions had this flag inverted, inconsistent with
10573 rsa_tmp_cb (..._TMP_RSA_CB).
10574 [Bodo Moeller; problem reported by Amit Chopra]
10576 *) Add missing DSA library text string. Work around for some IIS
10577 key files with invalid SEQUENCE encoding.
10580 *) Add a document (doc/standards.txt) that list all kinds of standards
10581 and so on that are implemented in OpenSSL.
10584 *) Enhance c_rehash script. Old version would mishandle certificates
10585 with the same subject name hash and wouldn't handle CRLs at all.
10586 Added -fingerprint option to crl utility, to support new c_rehash
10590 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10593 *) Fix for SSL server purpose checking. Server checking was
10594 rejecting certificates which had extended key usage present
10595 but no ssl client purpose.
10596 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10598 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10599 is a little unclear about how a blank password is handled.
10600 Since the password in encoded as a BMPString with terminating
10601 double NULL a zero length password would end up as just the
10602 double NULL. However no password at all is different and is
10603 handled differently in the PKCS#12 key generation code. NS
10604 treats a blank password as zero length. MSIE treats it as no
10605 password on export: but it will try both on import. We now do
10606 the same: PKCS12_parse() tries zero length and no password if
10607 the password is set to "" or NULL (NULL is now a valid password:
10608 it wasn't before) as does the pkcs12 application.
10611 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10612 perror when PEM_read_bio_X509_REQ fails, the error message must
10613 be obtained from the error queue.
10616 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10617 it in ERR_remove_state if appropriate, and change ERR_get_state
10618 accordingly to avoid race conditions (this is necessary because
10619 thread_hash is no longer constant once set).
10622 *) Bugfix for linux-elf makefile.one.
10625 *) RSA_get_default_method() will now cause a default
10626 RSA_METHOD to be chosen if one doesn't exist already.
10627 Previously this was only set during a call to RSA_new()
10628 or RSA_new_method(NULL) meaning it was possible for
10629 RSA_get_default_method() to return NULL.
10632 *) Added native name translation to the existing DSO code
10633 that will convert (if the flag to do so is set) filenames
10634 that are sufficiently small and have no path information
10635 into a canonical native form. Eg. "blah" converted to
10636 "libblah.so" or "blah.dll" etc.
10639 *) New function ERR_error_string_n(e, buf, len) which is like
10640 ERR_error_string(e, buf), but writes at most 'len' bytes
10641 including the 0 terminator. For ERR_error_string_n, 'buf'
10643 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10645 *) CONF library reworked to become more general. A new CONF
10646 configuration file reader "class" is implemented as well as a
10647 new functions (NCONF_*, for "New CONF") to handle it. The now
10648 old CONF_* functions are still there, but are reimplemented to
10649 work in terms of the new functions. Also, a set of functions
10650 to handle the internal storage of the configuration data is
10651 provided to make it easier to write new configuration file
10652 reader "classes" (I can definitely see something reading a
10653 configuration file in XML format, for example), called _CONF_*,
10654 or "the configuration storage API"...
10656 The new configuration file reading functions are:
10658 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10659 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10661 NCONF_default, NCONF_WIN32
10663 NCONF_dump_fp, NCONF_dump_bio
10665 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10666 NCONF_new creates a new CONF object. This works in the same way
10667 as other interfaces in OpenSSL, like the BIO interface.
10668 NCONF_dump_* dump the internal storage of the configuration file,
10669 which is useful for debugging. All other functions take the same
10670 arguments as the old CONF_* functions wth the exception of the
10671 first that must be a `CONF *' instead of a `LHASH *'.
10673 To make it easer to use the new classes with the old CONF_* functions,
10674 the function CONF_set_default_method is provided.
10677 *) Add '-tls1' option to 'openssl ciphers', which was already
10678 mentioned in the documentation but had not been implemented.
10679 (This option is not yet really useful because even the additional
10680 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10683 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10684 OpenSSL-based applications) load shared libraries and bind to
10685 them in a portable way.
10686 [Geoff Thorpe, with contributions from Richard Levitte]
10688 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10690 *) Make sure _lrotl and _lrotr are only used with MSVC.
10692 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10693 (the default implementation of RAND_status).
10695 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10696 to '-clrext' (= clear extensions), as intended and documented.
10697 [Bodo Moeller; inconsistency pointed out by Michael Attili
10698 <attili@amaxo.com>]
10700 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10701 was larger than the MD block size.
10702 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10704 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10705 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10706 using the passed key: if the passed key was a private key the result
10707 of X509_print(), for example, would be to print out all the private key
10711 *) des_quad_cksum() byte order bug fix.
10712 [Ulf Möller, using the problem description in krb4-0.9.7, where
10713 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10715 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10717 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10719 *) For easily testing in shell scripts whether some command
10720 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10721 returns with exit code 0 iff no command of the given name is available.
10722 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10723 the output goes to stdout and nothing is printed to stderr.
10724 Additional arguments are always ignored.
10726 Since for each cipher there is a command of the same name,
10727 the 'no-cipher' compilation switches can be tested this way.
10729 ('openssl no-XXX' is not able to detect pseudo-commands such
10730 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10733 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10736 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10737 is set; it will be thrown away anyway because each handshake creates
10739 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10740 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10741 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10742 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10745 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10746 'Q' and 'R' lose their special meanings (quit/renegotiate).
10747 This is part of what -quiet does; unlike -quiet, -ign_eof
10748 does not suppress any output.
10751 *) Add compatibility options to the purpose and trust code. The
10752 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10753 accepts a certificate or CA, this was the previous behaviour,
10754 with all the associated security issues.
10756 X509_TRUST_COMPAT is the old trust behaviour: only and
10757 automatically trust self signed roots in certificate store. A
10758 new trust setting X509_TRUST_DEFAULT is used to specify that
10759 a purpose has no associated trust setting and it should instead
10760 use the value in the default purpose.
10763 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10764 and fix a memory leak.
10767 *) In util/mkerr.pl (which implements 'make errors'), preserve
10768 reason strings from the previous version of the .c file, as
10769 the default to have only downcase letters (and digits) in
10770 automatically generated reasons codes is not always appropriate.
10773 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10774 using strerror. Previously, ERR_reason_error_string() returned
10775 library names as reason strings for SYSerr; but SYSerr is a special
10776 case where small numbers are errno values, not library numbers.
10779 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10780 converts DSA parameters into DH parameters. (When creating parameters,
10781 DSA_generate_parameters is used.)
10784 *) Include 'length' (recommended exponent length) in C code generated
10785 by 'openssl dhparam -C'.
10788 *) The second argument to set_label in perlasm was already being used
10789 so couldn't be used as a "file scope" flag. Moved to third argument
10793 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10794 instead of RAND_bytes for encryption IVs and salts.
10797 *) Include RAND_status() into RAND_METHOD instead of implementing
10798 it only for md_rand.c Otherwise replacing the PRNG by calling
10799 RAND_set_rand_method would be impossible.
10802 *) Don't let DSA_generate_key() enter an infinite loop if the random
10803 number generation fails.
10806 *) New 'rand' application for creating pseudo-random output.
10809 *) Added configuration support for Linux/IA64
10810 [Rolf Haberrecker <rolf@suse.de>]
10812 *) Assembler module support for Mingw32.
10815 *) Shared library support for HPUX (in shlib/).
10816 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10818 *) Shared library support for Solaris gcc.
10819 [Lutz Behnke <behnke@trustcenter.de>]
10821 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10823 *) PKCS7_encrypt() was adding text MIME headers twice because they
10824 were added manually and by SMIME_crlf_copy().
10827 *) In bntest.c don't call BN_rand with zero bits argument.
10828 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10830 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10831 case was implemented. This caused BN_div_recp() to fail occasionally.
10834 *) Add an optional second argument to the set_label() in the perl
10835 assembly language builder. If this argument exists and is set
10836 to 1 it signals that the assembler should use a symbol whose
10837 scope is the entire file, not just the current function. This
10838 is needed with MASM which uses the format label:: for this scope.
10839 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10841 *) Change the ASN1 types so they are typedefs by default. Before
10842 almost all types were #define'd to ASN1_STRING which was causing
10843 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10847 *) Change names of new functions to the new get1/get0 naming
10848 convention: After 'get1', the caller owns a reference count
10849 and has to call ..._free; 'get0' returns a pointer to some
10850 data structure without incrementing reference counters.
10851 (Some of the existing 'get' functions increment a reference
10852 counter, some don't.)
10853 Similarly, 'set1' and 'add1' functions increase reference
10854 counters or duplicate objects.
10857 *) Allow for the possibility of temp RSA key generation failure:
10858 the code used to assume it always worked and crashed on failure.
10861 *) Fix potential buffer overrun problem in BIO_printf().
10862 [Ulf Möller, using public domain code by Patrick Powell; problem
10863 pointed out by David Sacerdote <das33@cornell.edu>]
10865 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10866 RAND_egd() and RAND_status(). In the command line application,
10867 the EGD socket can be specified like a seed file using RANDFILE
10871 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10872 Some CAs (e.g. Verisign) distribute certificates in this form.
10875 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10876 list to exclude them. This means that no special compilation option
10877 is needed to use anonymous DH: it just needs to be included in the
10881 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10882 EVP_MD_type. The old functionality is available in a new macro called
10883 EVP_MD_md(). Change code that uses it and update docs.
10886 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10887 where the 'void *' argument is replaced by a function pointer argument.
10888 Previously 'void *' was abused to point to functions, which works on
10889 many platforms, but is not correct. As these functions are usually
10890 called by macros defined in OpenSSL header files, most source code
10891 should work without changes.
10894 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10895 sections with information on -D... compiler switches used for
10896 compiling the library so that applications can see them. To enable
10897 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10898 must be defined. E.g.,
10899 #define OPENSSL_ALGORITHM_DEFINES
10900 #include <openssl/opensslconf.h>
10901 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10902 [Richard Levitte, Ulf and Bodo Möller]
10904 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10908 *) Change the 'other' type in certificate aux info to a STACK_OF
10909 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10910 the required ASN1 format: arbitrary types determined by an OID.
10913 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10914 argument to 'req'. This is not because the function is newer or
10915 better than others it just uses the work 'NEW' in the certificate
10916 request header lines. Some software needs this.
10919 *) Reorganise password command line arguments: now passwords can be
10920 obtained from various sources. Delete the PEM_cb function and make
10921 it the default behaviour: i.e. if the callback is NULL and the
10922 usrdata argument is not NULL interpret it as a null terminated pass
10923 phrase. If usrdata and the callback are NULL then the pass phrase
10924 is prompted for as usual.
10927 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10928 the support is automatically enabled. The resulting binaries will
10929 autodetect the card and use it if present.
10930 [Ben Laurie and Compaq Inc.]
10932 *) Work around for Netscape hang bug. This sends certificate request
10933 and server done in one record. Since this is perfectly legal in the
10934 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10935 the bugs/SSLv3 entry for more info.
10938 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10941 *) Add -rand argument to smime and pkcs12 applications and read/write
10945 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10948 *) Add command line password options to the remaining applications.
10951 *) Bug fix for BN_div_recp() for numerators with an even number of
10955 *) More tests in bntest.c, and changed test_bn output.
10958 *) ./config recognizes MacOS X now.
10961 *) Bug fix for BN_div() when the first words of num and divisor are
10962 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10965 *) Add support for various broken PKCS#8 formats, and command line
10966 options to produce them.
10969 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10970 get temporary BIGNUMs from a BN_CTX.
10973 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10977 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10978 include a #define from the old name to the new. The original intent
10979 was that statically linked binaries could for example just call
10980 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10981 link with digests. This never worked because SSLeay_add_all_digests()
10982 and SSLeay_add_all_ciphers() were in the same source file so calling
10983 one would link with the other. They are now in separate source files.
10986 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10989 *) Use a less unusual form of the Miller-Rabin primality test (it used
10990 a binary algorithm for exponentiation integrated into the Miller-Rabin
10991 loop, our standard modexp algorithms are faster).
10994 *) Support for the EBCDIC character set completed.
10995 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10997 *) Source code cleanups: use const where appropriate, eliminate casts,
10998 use void * instead of char * in lhash.
11001 *) Bugfix: ssl3_send_server_key_exchange was not restartable
11002 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
11003 this the server could overwrite ephemeral keys that the client
11007 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
11008 using 50 iterations of the Rabin-Miller test.
11010 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
11011 iterations of the Rabin-Miller test as required by the appendix
11012 to FIPS PUB 186[-1]) instead of DSA_is_prime.
11013 As BN_is_prime_fasttest includes trial division, DSA parameter
11014 generation becomes much faster.
11016 This implies a change for the callback functions in DSA_is_prime
11017 and DSA_generate_parameters: The callback function is called once
11018 for each positive witness in the Rabin-Miller test, not just
11019 occasionally in the inner loop; and the parameters to the
11020 callback function now provide an iteration count for the outer
11021 loop rather than for the current invocation of the inner loop.
11022 DSA_generate_parameters additionally can call the callback
11023 function with an 'iteration count' of -1, meaning that a
11024 candidate has passed the trial division test (when q is generated
11025 from an application-provided seed, trial division is skipped).
11028 *) New function BN_is_prime_fasttest that optionally does trial
11029 division before starting the Rabin-Miller test and has
11030 an additional BN_CTX * argument (whereas BN_is_prime always
11031 has to allocate at least one BN_CTX).
11032 'callback(1, -1, cb_arg)' is called when a number has passed the
11033 trial division stage.
11036 *) Fix for bug in CRL encoding. The validity dates weren't being handled
11040 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
11043 *) New function BN_pseudo_rand().
11046 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
11047 bignum version of BN_from_montgomery() with the working code from
11048 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
11052 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
11053 made it impossible to use the same SSL_SESSION data structure in
11054 SSL2 clients in multiple threads.
11057 *) The return value of RAND_load_file() no longer counts bytes obtained
11058 by stat(). RAND_load_file(..., -1) is new and uses the complete file
11059 to seed the PRNG (previously an explicit byte count was required).
11060 [Ulf Möller, Bodo Möller]
11062 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
11063 used (char *) instead of (void *) and had casts all over the place.
11066 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
11069 *) Retain source code compatibility for BN_prime_checks macro:
11070 BN_is_prime(..., BN_prime_checks, ...) now uses
11071 BN_prime_checks_for_size to determine the appropriate number of
11072 Rabin-Miller iterations.
11075 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
11076 DH_CHECK_P_NOT_SAFE_PRIME.
11077 (Check if this is true? OpenPGP calls them "strong".)
11080 *) Merge the functionality of "dh" and "gendh" programs into a new program
11081 "dhparam". The old programs are retained for now but will handle DH keys
11082 (instead of parameters) in future.
11085 *) Make the ciphers, s_server and s_client programs check the return values
11086 when a new cipher list is set.
11089 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11090 ciphers. Before when the 56bit ciphers were enabled the sorting was
11093 The syntax for the cipher sorting has been extended to support sorting by
11094 cipher-strength (using the strength_bits hard coded in the tables).
11095 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11097 Fix a bug in the cipher-command parser: when supplying a cipher command
11098 string with an "undefined" symbol (neither command nor alphanumeric
11099 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11100 an error is flagged.
11102 Due to the strength-sorting extension, the code of the
11103 ssl_create_cipher_list() function was completely rearranged. I hope that
11104 the readability was also increased :-)
11105 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11107 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11108 for the first serial number and places 2 in the serial number file. This
11109 avoids problems when the root CA is created with serial number zero and
11110 the first user certificate has the same issuer name and serial number
11114 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11115 the new code. Add documentation for this stuff.
11118 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11119 X509_*() to X509at_*() on the grounds that they don't handle X509
11120 structures and behave in an analogous way to the X509v3 functions:
11121 they shouldn't be called directly but wrapper functions should be used
11124 So we also now have some wrapper functions that call the X509at functions
11125 when passed certificate requests. (TO DO: similar things can be done with
11126 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11127 things. Some of these need some d2i or i2d and print functionality
11128 because they handle more complex structures.)
11131 *) Add missing #ifndefs that caused missing symbols when building libssl
11132 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11133 NO_RSA in ssl/s2*.c.
11134 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11136 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11137 has a return value which indicates the quality of the random data
11138 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11139 error queue. New function RAND_pseudo_bytes() generates output that is
11140 guaranteed to be unique but not unpredictable. RAND_add is like
11141 RAND_seed, but takes an extra argument for an entropy estimate
11142 (RAND_seed always assumes full entropy).
11145 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11146 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11147 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11148 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11149 false-positive rate of at most 2^-80 for random input.
11152 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11155 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11156 in the 0.9.5 release), this returns the chain
11157 from an X509_CTX structure with a dup of the stack and all
11158 the X509 reference counts upped: so the stack will exist
11159 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11162 Also make SSL_SESSION_print() print out the verify return
11166 *) Add manpage for the pkcs12 command. Also change the default
11167 behaviour so MAC iteration counts are used unless the new
11168 -nomaciter option is used. This improves file security and
11169 only older versions of MSIE (4.0 for example) need it.
11172 *) Honor the no-xxx Configure options when creating .DEF files.
11175 *) Add PKCS#10 attributes to field table: challengePassword,
11176 unstructuredName and unstructuredAddress. These are taken from
11177 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11178 international characters are used.
11180 More changes to X509_ATTRIBUTE code: allow the setting of types
11181 based on strings. Remove the 'loc' parameter when adding
11182 attributes because these will be a SET OF encoding which is sorted
11186 *) Initial changes to the 'req' utility to allow request generation
11187 automation. This will allow an application to just generate a template
11188 file containing all the field values and have req construct the
11191 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11192 used all over the place including certificate requests and PKCS#7
11193 structures. They are currently handled manually where necessary with
11194 some primitive wrappers for PKCS#7. The new functions behave in a
11195 manner analogous to the X509 extension functions: they allow
11196 attributes to be looked up by NID and added.
11198 Later something similar to the X509V3 code would be desirable to
11199 automatically handle the encoding, decoding and printing of the
11200 more complex types. The string types like challengePassword can
11201 be handled by the string table functions.
11203 Also modified the multi byte string table handling. Now there is
11204 a 'global mask' which masks out certain types. The table itself
11205 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11206 is useful when for example there is only one permissible type
11207 (as in countryName) and using the mask might result in no valid
11211 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11212 SSL_get_peer_finished to allow applications to obtain the latest
11213 Finished messages sent to the peer or expected from the peer,
11214 respectively. (SSL_get_peer_finished is usually the Finished message
11215 actually received from the peer, otherwise the protocol will be aborted.)
11217 As the Finished message are message digests of the complete handshake
11218 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11219 be used for external authentication procedures when the authentication
11220 provided by SSL/TLS is not desired or is not enough.
11223 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11224 the host supports BWX extension and if Compaq C is present on the
11225 $PATH. Just exploiting of the BWX extension results in 20-30%
11226 performance kick for some algorithms, e.g. DES and RC4 to mention
11227 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11231 *) Add support for MS "fast SGC". This is arguably a violation of the
11232 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11233 weak crypto and after checking the certificate is SGC a second one
11234 with strong crypto. MS SGC stops the first handshake after receiving
11235 the server certificate message and sends a second client hello. Since
11236 a server will typically do all the time consuming operations before
11237 expecting any further messages from the client (server key exchange
11238 is the most expensive) there is little difference between the two.
11240 To get OpenSSL to support MS SGC we have to permit a second client
11241 hello message after we have sent server done. In addition we have to
11242 reset the MAC if we do get this second client hello.
11245 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11246 if a DER encoded private key is RSA or DSA traditional format. Changed
11247 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11248 format DER encoded private key. Newer code should use PKCS#8 format which
11249 has the key type encoded in the ASN1 structure. Added DER private key
11250 support to pkcs8 application.
11253 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11254 ciphersuites has been selected (as required by the SSL 3/TLS 1
11255 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11256 is set, we interpret this as a request to violate the specification
11257 (the worst that can happen is a handshake failure, and 'correct'
11258 behaviour would result in a handshake failure anyway).
11261 *) In SSL_CTX_add_session, take into account that there might be multiple
11262 SSL_SESSION structures with the same session ID (e.g. when two threads
11263 concurrently obtain them from an external cache).
11264 The internal cache can handle only one SSL_SESSION with a given ID,
11265 so if there's a conflict, we now throw out the old one to achieve
11269 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11270 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11271 some routines that use cipher OIDs: some ciphers do not have OIDs
11272 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11276 *) Simplify the trust setting structure and code. Now we just have
11277 two sequences of OIDs for trusted and rejected settings. These will
11278 typically have values the same as the extended key usage extension
11279 and any application specific purposes.
11281 The trust checking code now has a default behaviour: it will just
11282 check for an object with the same NID as the passed id. Functions can
11283 be provided to override either the default behaviour or the behaviour
11284 for a given id. SSL client, server and email already have functions
11285 in place for compatibility: they check the NID and also return "trusted"
11286 if the certificate is self signed.
11289 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11290 traditional format into an EVP_PKEY structure.
11293 *) Add a password callback function PEM_cb() which either prompts for
11294 a password if usr_data is NULL or otherwise assumes it is a null
11295 terminated password. Allow passwords to be passed on command line
11296 environment or config files in a few more utilities.
11299 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11300 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11301 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11302 Update documentation.
11305 *) Support for ASN1 "NULL" type. This could be handled before by using
11306 ASN1_TYPE but there wasn't any function that would try to read a NULL
11307 and produce an error if it couldn't. For compatibility we also have
11308 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11309 don't allocate anything because they don't need to.
11312 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11314 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11316 *) Rebuild of the memory allocation routines used by OpenSSL code and
11317 possibly others as well. The purpose is to make an interface that
11318 provide hooks so anyone can build a separate set of allocation and
11319 deallocation routines to be used by OpenSSL, for example memory
11320 pool implementations, or something else, which was previously hard
11321 since Malloc(), Realloc() and Free() were defined as macros having
11322 the values malloc, realloc and free, respectively (except for Win32
11323 compilations). The same is provided for memory debugging code.
11324 OpenSSL already comes with functionality to find memory leaks, but
11325 this gives people a chance to debug other memory problems.
11327 With these changes, a new set of functions and macros have appeared:
11329 CRYPTO_set_mem_debug_functions() [F]
11330 CRYPTO_get_mem_debug_functions() [F]
11331 CRYPTO_dbg_set_options() [F]
11332 CRYPTO_dbg_get_options() [F]
11333 CRYPTO_malloc_debug_init() [M]
11335 The memory debug functions are NULL by default, unless the library
11336 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11337 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11338 gives the standard debugging functions that come with OpenSSL) or
11339 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11340 provided by the library user) must be used. When the standard
11341 debugging functions are used, CRYPTO_dbg_set_options can be used to
11342 request additional information:
11343 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11344 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11346 Also, things like CRYPTO_set_mem_functions will always give the
11347 expected result (the new set of functions is used for allocation
11348 and deallocation) at all times, regardless of platform and compiler
11351 To finish it up, some functions that were never use in any other
11352 way than through macros have a new API and new semantic:
11354 CRYPTO_dbg_malloc()
11355 CRYPTO_dbg_realloc()
11358 All macros of value have retained their old syntax.
11359 [Richard Levitte and Bodo Moeller]
11361 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11362 ordering of SMIMECapabilities wasn't in "strength order" and there
11363 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11367 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11368 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11369 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11371 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11372 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11373 functionality to handle multipart/signed properly) and a utility
11374 called 'smime' to call all this stuff. This is based on code I
11375 originally wrote for Celo who have kindly allowed it to be
11376 included in OpenSSL.
11379 *) Add variants des_set_key_checked and des_set_key_unchecked of
11380 des_set_key (aka des_key_sched). Global variable des_check_key
11381 decides which of these is called by des_set_key; this way
11382 des_check_key behaves as it always did, but applications and
11383 the library itself, which was buggy for des_check_key == 1,
11384 have a cleaner way to pick the version they need.
11387 *) New function PKCS12_newpass() which changes the password of a
11391 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11392 dynamic mix. In both cases the ids can be used as an index into the
11393 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11394 functions so they accept a list of the field values and the
11395 application doesn't need to directly manipulate the X509_TRUST
11399 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11403 *) Modify the way the V3 extension code looks up extensions. This now
11404 works in a similar way to the object code: we have some "standard"
11405 extensions in a static table which is searched with OBJ_bsearch()
11406 and the application can add dynamic ones if needed. The file
11407 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11408 updated whenever a new extension is added to the core code and kept
11409 in ext_nid order. There is a simple program 'tabtest.c' which checks
11410 this. New extensions are not added too often so this file can readily
11411 be maintained manually.
11413 There are two big advantages in doing things this way. The extensions
11414 can be looked up immediately and no longer need to be "added" using
11415 X509V3_add_standard_extensions(): this function now does nothing.
11416 [Side note: I get *lots* of email saying the extension code doesn't
11417 work because people forget to call this function]
11418 Also no dynamic allocation is done unless new extensions are added:
11419 so if we don't add custom extensions there is no need to call
11420 X509V3_EXT_cleanup().
11423 *) Modify enc utility's salting as follows: make salting the default. Add a
11424 magic header, so unsalted files fail gracefully instead of just decrypting
11425 to garbage. This is because not salting is a big security hole, so people
11426 should be discouraged from doing it.
11429 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11430 digest to be passed on the command line but it only used this
11431 parameter when signing a certificate. Modified so all relevant
11432 operations are affected by the digest parameter including the
11433 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11434 DSA key was used because it didn't fix the digest.
11437 *) Initial certificate chain verify code. Currently tests the untrusted
11438 certificates for consistency with the verify purpose (which is set
11439 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11441 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11442 this is because it will reject chains with invalid extensions whereas
11443 every previous version of OpenSSL and SSLeay made no checks at all.
11445 Trust code: checks the root CA for the relevant trust settings. Trust
11446 settings have an initial value consistent with the verify purpose: e.g.
11447 if the verify purpose is for SSL client use it expects the CA to be
11448 trusted for SSL client use. However the default value can be changed to
11449 permit custom trust settings: one example of this would be to only trust
11450 certificates from a specific "secure" set of CAs.
11452 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11453 which should be used for version portability: especially since the
11454 verify structure is likely to change more often now.
11456 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11457 to set them. If not set then assume SSL clients will verify SSL servers
11460 Two new options to the verify program: -untrusted allows a set of
11461 untrusted certificates to be passed in and -purpose which sets the
11462 intended purpose of the certificate. If a purpose is set then the
11463 new chain verify code is used to check extension consistency.
11466 *) Support for the authority information access extension.
11469 *) Modify RSA and DSA PEM read routines to transparently handle
11470 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11471 public keys in a format compatible with certificate
11472 SubjectPublicKeyInfo structures. Unfortunately there were already
11473 functions called *_PublicKey_* which used various odd formats so
11474 these are retained for compatibility: however the DSA variants were
11475 never in a public release so they have been deleted. Changed dsa/rsa
11476 utilities to handle the new format: note no releases ever handled public
11477 keys so we should be OK.
11479 The primary motivation for this change is to avoid the same fiasco
11480 that dogs private keys: there are several incompatible private key
11481 formats some of which are standard and some OpenSSL specific and
11482 require various evil hacks to allow partial transparent handling and
11483 even then it doesn't work with DER formats. Given the option anything
11484 other than PKCS#8 should be dumped: but the other formats have to
11485 stay in the name of compatibility.
11487 With public keys and the benefit of hindsight one standard format
11488 is used which works with EVP_PKEY, RSA or DSA structures: though
11489 it clearly returns an error if you try to read the wrong kind of key.
11491 Added a -pubkey option to the 'x509' utility to output the public key.
11492 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11493 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11494 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11495 that do the same as the EVP_PKEY_assign_*() except they up the
11496 reference count of the added key (they don't "swallow" the
11500 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11501 CRLs would fail if the file contained no certificates or no CRLs:
11502 added a new function to read in both types and return the number
11503 read: this means that if none are read it will be an error. The
11504 DER versions of the certificate and CRL reader would always fail
11505 because it isn't possible to mix certificates and CRLs in DER format
11506 without choking one or the other routine. Changed this to just read
11507 a certificate: this is the best we can do. Also modified the code
11508 in apps/verify.c to take notice of return codes: it was previously
11509 attempting to read in certificates from NULL pointers and ignoring
11510 any errors: this is one reason why the cert and CRL reader seemed
11511 to work. It doesn't check return codes from the default certificate
11512 routines: these may well fail if the certificates aren't installed.
11515 *) Code to support otherName option in GeneralName.
11518 *) First update to verify code. Change the verify utility
11519 so it warns if it is passed a self signed certificate:
11520 for consistency with the normal behaviour. X509_verify
11521 has been modified to it will now verify a self signed
11522 certificate if *exactly* the same certificate appears
11523 in the store: it was previously impossible to trust a
11524 single self signed certificate. This means that:
11525 openssl verify ss.pem
11526 now gives a warning about a self signed certificate but
11527 openssl verify -CAfile ss.pem ss.pem
11531 *) For servers, store verify_result in SSL_SESSION data structure
11532 (and add it to external session representation).
11533 This is needed when client certificate verifications fails,
11534 but an application-provided verification callback (set by
11535 SSL_CTX_set_cert_verify_callback) allows accepting the session
11536 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11537 but returns 1): When the session is reused, we have to set
11538 ssl->verify_result to the appropriate error code to avoid
11540 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11542 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11543 case in PKCS7_dataInit() where the signed PKCS7 structure
11544 didn't contain any existing data because it was being created.
11545 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11547 *) Add a salt to the key derivation routines in enc.c. This
11548 forms the first 8 bytes of the encrypted file. Also add a
11549 -S option to allow a salt to be input on the command line.
11552 *) New function X509_cmp(). Oddly enough there wasn't a function
11553 to compare two certificates. We do this by working out the SHA1
11554 hash and comparing that. X509_cmp() will be needed by the trust
11558 *) SSL_get1_session() is like SSL_get_session(), but increments
11559 the reference count in the SSL_SESSION returned.
11560 [Geoff Thorpe <geoff@eu.c2.net>]
11562 *) Fix for 'req': it was adding a null to request attributes.
11563 Also change the X509_LOOKUP and X509_INFO code to handle
11564 certificate auxiliary information.
11567 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11571 *) Add the possibility to add extra information to the memory leak
11572 detecting output, to form tracebacks, showing from where each
11573 allocation was originated: CRYPTO_push_info("constant string") adds
11574 the string plus current file name and line number to a per-thread
11575 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11576 is like calling CYRPTO_pop_info() until the stack is empty.
11577 Also updated memory leak detection code to be multi-thread-safe.
11580 *) Add options -text and -noout to pkcs7 utility and delete the
11581 encryption options which never did anything. Update docs.
11584 *) Add options to some of the utilities to allow the pass phrase
11585 to be included on either the command line (not recommended on
11586 OSes like Unix) or read from the environment. Update the
11587 manpages and fix a few bugs.
11590 *) Add a few manpages for some of the openssl commands.
11593 *) Fix the -revoke option in ca. It was freeing up memory twice,
11594 leaking and not finding already revoked certificates.
11597 *) Extensive changes to support certificate auxiliary information.
11598 This involves the use of X509_CERT_AUX structure and X509_AUX
11599 functions. An X509_AUX function such as PEM_read_X509_AUX()
11600 can still read in a certificate file in the usual way but it
11601 will also read in any additional "auxiliary information". By
11602 doing things this way a fair degree of compatibility can be
11603 retained: existing certificates can have this information added
11604 using the new 'x509' options.
11606 Current auxiliary information includes an "alias" and some trust
11607 settings. The trust settings will ultimately be used in enhanced
11608 certificate chain verification routines: currently a certificate
11609 can only be trusted if it is self signed and then it is trusted
11613 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11614 The problem was that one of the replacement routines had not been working
11615 since SSLeay releases. For now the offending routine has been replaced
11616 with non-optimised assembler. Even so, this now gives around 95%
11617 performance improvement for 1024 bit RSA signs.
11620 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11621 handling. Most clients have the effective key size in bits equal to
11622 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11623 A few however don't do this and instead use the size of the decrypted key
11624 to determine the RC2 key length and the AlgorithmIdentifier to determine
11625 the effective key length. In this case the effective key length can still
11626 be 40 bits but the key length can be 168 bits for example. This is fixed
11627 by manually forcing an RC2 key into the EVP_PKEY structure because the
11628 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11629 the key length and effective key length are equal.
11632 *) Add a bunch of functions that should simplify the creation of
11633 X509_NAME structures. Now you should be able to do:
11634 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11635 and have it automatically work out the correct field type and fill in
11636 the structures. The more adventurous can try:
11637 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11638 and it will (hopefully) work out the correct multibyte encoding.
11641 *) Change the 'req' utility to use the new field handling and multibyte
11642 copy routines. Before the DN field creation was handled in an ad hoc
11643 way in req, ca, and x509 which was rather broken and didn't support
11644 BMPStrings or UTF8Strings. Since some software doesn't implement
11645 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11646 using the dirstring_type option. See the new comment in the default
11647 openssl.cnf for more info.
11650 *) Make crypto/rand/md_rand.c more robust:
11651 - Assure unique random numbers after fork().
11652 - Make sure that concurrent threads access the global counter and
11653 md serializably so that we never lose entropy in them
11654 or use exactly the same state in multiple threads.
11655 Access to the large state is not always serializable because
11656 the additional locking could be a performance killer, and
11657 md should be large enough anyway.
11660 *) New file apps/app_rand.c with commonly needed functionality
11661 for handling the random seed file.
11663 Use the random seed file in some applications that previously did not:
11665 dsaparam -genkey (which also ignored its '-rand' option),
11668 x509 (when signing).
11669 Except on systems with /dev/urandom, it is crucial to have a random
11670 seed file at least for key creation, DSA signing, and for DH exchanges;
11671 for RSA signatures we could do without one.
11673 gendh and gendsa (unlike genrsa) used to read only the first byte
11674 of each file listed in the '-rand' option. The function as previously
11675 found in genrsa is now in app_rand.c and is used by all programs
11676 that support '-rand'.
11679 *) In RAND_write_file, use mode 0600 for creating files;
11680 don't just chmod when it may be too late.
11683 *) Report an error from X509_STORE_load_locations
11684 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11687 *) New function ASN1_mbstring_copy() this copies a string in either
11688 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11689 into an ASN1_STRING type. A mask of permissible types is passed
11690 and it chooses the "minimal" type to use or an error if not type
11694 *) Add function equivalents to the various macros in asn1.h. The old
11695 macros are retained with an M_ prefix. Code inside the library can
11696 use the M_ macros. External code (including the openssl utility)
11697 should *NOT* in order to be "shared library friendly".
11700 *) Add various functions that can check a certificate's extensions
11701 to see if it usable for various purposes such as SSL client,
11702 server or S/MIME and CAs of these types. This is currently
11703 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11704 verification. Also added a -purpose flag to x509 utility to
11705 print out all the purposes.
11708 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11712 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11713 for, obtain and decode and extension and obtain its critical flag.
11714 This allows all the necessary extension code to be handled in a
11715 single function call.
11718 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11719 platforms. See crypto/rc4/rc4_enc.c for further details.
11722 *) New -noout option to asn1parse. This causes no output to be produced
11723 its main use is when combined with -strparse and -out to extract data
11724 from a file (which may not be in ASN.1 format).
11727 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11728 when producing the local key id.
11729 [Richard Levitte <levitte@stacken.kth.se>]
11731 *) New option -dhparam in s_server. This allows a DH parameter file to be
11732 stated explicitly. If it is not stated then it tries the first server
11733 certificate file. The previous behaviour hard coded the filename
11737 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11738 a public key to be input or output. For example:
11739 openssl rsa -in key.pem -pubout -out pubkey.pem
11740 Also added necessary DSA public key functions to handle this.
11743 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11744 in the message. This was handled by allowing
11745 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11746 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11748 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11749 to the end of the strings whereas this didn't. This would cause problems
11750 if strings read with d2i_ASN1_bytes() were later modified.
11751 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11753 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11754 data and it contains EOF it will end up returning an error. This is
11755 caused by input 46 bytes long. The cause is due to the way base64
11756 BIOs find the start of base64 encoded data. They do this by trying a
11757 trial decode on each line until they find one that works. When they
11758 do a flag is set and it starts again knowing it can pass all the
11759 data directly through the decoder. Unfortunately it doesn't reset
11760 the context it uses. This means that if EOF is reached an attempt
11761 is made to pass two EOFs through the context and this causes the
11762 resulting error. This can also cause other problems as well. As is
11763 usual with these problems it takes *ages* to find and the fix is
11764 trivial: move one line.
11765 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11767 *) Ugly workaround to get s_client and s_server working under Windows. The
11768 old code wouldn't work because it needed to select() on sockets and the
11769 tty (for keypresses and to see if data could be written). Win32 only
11770 supports select() on sockets so we select() with a 1s timeout on the
11771 sockets and then see if any characters are waiting to be read, if none
11772 are present then we retry, we also assume we can always write data to
11773 the tty. This isn't nice because the code then blocks until we've
11774 received a complete line of data and it is effectively polling the
11775 keyboard at 1s intervals: however it's quite a bit better than not
11776 working at all :-) A dedicated Windows application might handle this
11777 with an event loop for example.
11780 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11781 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11782 will be called when RSA_sign() and RSA_verify() are used. This is useful
11783 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11784 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11785 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11786 This necessitated the support of an extra signature type NID_md5_sha1
11787 for SSL signatures and modifications to the SSL library to use it instead
11788 of calling RSA_public_decrypt() and RSA_private_encrypt().
11791 *) Add new -verify -CAfile and -CApath options to the crl program, these
11792 will lookup a CRL issuers certificate and verify the signature in a
11793 similar way to the verify program. Tidy up the crl program so it
11794 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11795 less strict. It will now permit CRL extensions even if it is not
11796 a V2 CRL: this will allow it to tolerate some broken CRLs.
11799 *) Initialize all non-automatic variables each time one of the openssl
11800 sub-programs is started (this is necessary as they may be started
11801 multiple times from the "OpenSSL>" prompt).
11802 [Lennart Bang, Bodo Moeller]
11804 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11805 removing all other RSA functionality (this is what NO_RSA does). This
11806 is so (for example) those in the US can disable those operations covered
11807 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11811 *) Non-copying interface to BIO pairs.
11812 (still largely untested)
11815 *) New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
11816 ASCII string. This was handled independently in various places before.
11819 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11820 UTF8 strings a character at a time.
11823 *) Use client_version from client hello to select the protocol
11824 (s23_srvr.c) and for RSA client key exchange verification
11825 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11828 *) Add various utility functions to handle SPKACs, these were previously
11829 handled by poking round in the structure internals. Added new function
11830 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11831 print, verify and generate SPKACs. Based on an original idea from
11832 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11835 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11838 *) Allow the config file extension section to be overwritten on the
11839 command line. Based on an original idea from Massimiliano Pala
11840 <madwolf@comune.modena.it>. The new option is called -extensions
11841 and can be applied to ca, req and x509. Also -reqexts to override
11842 the request extensions in req and -crlexts to override the crl extensions
11846 *) Add new feature to the SPKAC handling in ca. Now you can include
11847 the same field multiple times by preceding it by "XXXX." for example:
11850 this is the same syntax as used in the req config file.
11853 *) Allow certificate extensions to be added to certificate requests. These
11854 are specified in a 'req_extensions' option of the req section of the
11855 config file. They can be printed out with the -text option to req but
11856 are otherwise ignored at present.
11859 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11860 data read consists of only the final block it would not decrypted because
11861 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11862 A misplaced 'break' also meant the decrypted final block might not be
11863 copied until the next read.
11866 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11867 a few extra parameters to the DH structure: these will be useful if
11868 for example we want the value of 'q' or implement X9.42 DH.
11871 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11872 provides hooks that allow the default DSA functions or functions on a
11873 "per key" basis to be replaced. This allows hardware acceleration and
11874 hardware key storage to be handled without major modification to the
11875 library. Also added low level modexp hooks and CRYPTO_EX structure and
11876 associated functions.
11879 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11880 as "read only": it can't be written to and the buffer it points to will
11881 not be freed. Reading from a read only BIO is much more efficient than
11882 a normal memory BIO. This was added because there are several times when
11883 an area of memory needs to be read from a BIO. The previous method was
11884 to create a memory BIO and write the data to it, this results in two
11885 copies of the data and an O(n^2) reading algorithm. There is a new
11886 function BIO_new_mem_buf() which creates a read only memory BIO from
11887 an area of memory. Also modified the PKCS#7 routines to use read only
11891 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11892 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11893 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11894 but a retry condition occurred while trying to read the rest.
11897 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11898 NID_pkcs7_encrypted by default: this was wrong since this should almost
11899 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11900 the encrypted data type: this is a more sensible place to put it and it
11901 allows the PKCS#12 code to be tidied up that duplicated this
11905 *) Changed obj_dat.pl script so it takes its input and output files on
11906 the command line. This should avoid shell escape redirection problems
11910 *) Initial support for certificate extension requests, these are included
11911 in things like Xenroll certificate requests. Included functions to allow
11912 extensions to be obtained and added.
11915 *) -crlf option to s_client and s_server for sending newlines as
11916 CRLF (as required by many protocols).
11919 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11921 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11922 [Ralf S. Engelschall]
11924 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11925 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11927 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11931 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11932 DH parameters/keys (q is lost during that conversion, but the resulting
11933 DH parameters contain its length).
11935 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11936 much faster than DH_generate_parameters (which creates parameters
11937 where p = 2*q + 1), and also the smaller q makes DH computations
11938 much more efficient (160-bit exponentiation instead of 1024-bit
11939 exponentiation); so this provides a convenient way to support DHE
11940 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11941 utter importance to use
11942 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11944 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11945 when such DH parameters are used, because otherwise small subgroup
11946 attacks may become possible!
11949 *) Avoid memory leak in i2d_DHparams.
11952 *) Allow the -k option to be used more than once in the enc program:
11953 this allows the same encrypted message to be read by multiple recipients.
11956 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11957 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11958 it will always use the numerical form of the OID, even if it has a short
11962 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11963 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11964 otherwise bn_mod_exp was called. In the case of hardware keys for example
11965 no private key components need be present and it might store extra data
11966 in the RSA structure, which cannot be accessed from bn_mod_exp.
11967 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11968 private key operations.
11971 *) Added support for SPARC Linux.
11974 *) pem_password_cb function type incompatibly changed from
11975 typedef int pem_password_cb(char *buf, int size, int rwflag);
11977 ....(char *buf, int size, int rwflag, void *userdata);
11978 so that applications can pass data to their callbacks:
11979 The PEM[_ASN1]_{read,write}... functions and macros now take an
11980 additional void * argument, which is just handed through whenever
11981 the password callback is called.
11982 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11984 New function SSL_CTX_set_default_passwd_cb_userdata.
11986 Compatibility note: As many C implementations push function arguments
11987 onto the stack in reverse order, the new library version is likely to
11988 interoperate with programs that have been compiled with the old
11989 pem_password_cb definition (PEM_whatever takes some data that
11990 happens to be on the stack as its last argument, and the callback
11991 just ignores this garbage); but there is no guarantee whatsoever that
11994 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11995 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11996 problems not only on Windows, but also on some Unix platforms.
11997 To avoid problematic command lines, these definitions are now in an
11998 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11999 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
12002 *) MIPS III/IV assembler module is reimplemented.
12005 *) More DES library cleanups: remove references to srand/rand and
12006 delete an unused file.
12009 *) Add support for the free Netwide assembler (NASM) under Win32,
12010 since not many people have MASM (ml) and it can be hard to obtain.
12011 This is currently experimental but it seems to work OK and pass all
12012 the tests. Check out INSTALL.W32 for info.
12015 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
12016 without temporary keys kept an extra copy of the server key,
12017 and connections with temporary keys did not free everything in case
12021 *) New function RSA_check_key and new openssl rsa option -check
12022 for verifying the consistency of RSA keys.
12023 [Ulf Moeller, Bodo Moeller]
12025 *) Various changes to make Win32 compile work:
12026 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
12027 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
12028 comparison" warnings.
12029 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
12032 *) Add a debugging option to PKCS#5 v2 key generation function: when
12033 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
12034 derived keys are printed to stderr.
12037 *) Copy the flags in ASN1_STRING_dup().
12038 [Roman E. Pavlov <pre@mo.msk.ru>]
12040 *) The x509 application mishandled signing requests containing DSA
12041 keys when the signing key was also DSA and the parameters didn't match.
12043 It was supposed to omit the parameters when they matched the signing key:
12044 the verifying software was then supposed to automatically use the CA's
12045 parameters if they were absent from the end user certificate.
12047 Omitting parameters is no longer recommended. The test was also
12048 the wrong way round! This was probably due to unusual behaviour in
12049 EVP_cmp_parameters() which returns 1 if the parameters match.
12050 This meant that parameters were omitted when they *didn't* match and
12051 the certificate was useless. Certificates signed with 'ca' didn't have
12053 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
12055 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
12056 The interface is as follows:
12057 Applications can use
12058 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
12059 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
12060 "off" is now the default.
12061 The library internally uses
12062 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
12063 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
12064 to disable memory-checking temporarily.
12066 Some inconsistent states that previously were possible (and were
12067 even the default) are now avoided.
12069 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
12070 with each memory chunk allocated; this is occasionally more helpful
12071 than just having a counter.
12073 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
12075 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
12079 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
12080 which largely parallels "options", but is for changing API behaviour,
12081 whereas "options" are about protocol behaviour.
12082 Initial "mode" flags are:
12084 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12085 a single record has been written.
12086 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12087 retries use the same buffer location.
12088 (But all of the contents must be
12092 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12095 *) Fix problems with no-hmac etc.
12096 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12098 *) New functions RSA_get_default_method(), RSA_set_method() and
12099 RSA_get_method(). These allows replacement of RSA_METHODs without having
12100 to mess around with the internals of an RSA structure.
12103 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12104 Also really enable memory leak checks in openssl.c and in some
12106 [Chad C. Mulligan, Bodo Moeller]
12108 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12109 up the length of negative integers. This has now been simplified to just
12110 store the length when it is first determined and use it later, rather
12111 than trying to keep track of where data is copied and updating it to
12113 [Steve Henson, reported by Brien Wheeler
12114 <bwheeler@authentica-security.com>]
12116 *) Add a new function PKCS7_signatureVerify. This allows the verification
12117 of a PKCS#7 signature but with the signing certificate passed to the
12118 function itself. This contrasts with PKCS7_dataVerify which assumes the
12119 certificate is present in the PKCS#7 structure. This isn't always the
12120 case: certificates can be omitted from a PKCS#7 structure and be
12121 distributed by "out of band" means (such as a certificate database).
12124 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12125 function prototypes in pem.h, also change util/mkdef.pl to add the
12126 necessary function names.
12129 *) mk1mf.pl (used by Windows builds) did not properly read the
12130 options set by Configure in the top level Makefile, and Configure
12131 was not even able to write more than one option correctly.
12132 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12135 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12136 file to be loaded from a BIO or FILE pointer. The BIO version will
12137 for example allow memory BIOs to contain config info.
12140 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12141 Whoever hopes to achieve shared-library compatibility across versions
12142 must use this, not the compile-time macro.
12143 (Exercise 0.9.4: Which is the minimum library version required by
12145 Note: All this applies only to multi-threaded programs, others don't
12149 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12150 through a BIO pair triggered the default case, i.e.
12151 SSLerr(...,SSL_R_UNKNOWN_STATE).
12154 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12155 can use the SSL library even if none of the specific BIOs is
12159 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12160 for the encoded length.
12161 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12163 *) Add initial documentation of the X509V3 functions.
12166 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12167 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12168 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12169 secure PKCS#8 private key format with a high iteration count.
12172 *) Fix determination of Perl interpreter: A perl or perl5
12173 _directory_ in $PATH was also accepted as the interpreter.
12174 [Ralf S. Engelschall]
12176 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12177 wrong with it but it was very old and did things like calling
12178 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12179 unusual formatting.
12182 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12183 to use the new extension code.
12186 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12187 with macros. This should make it easier to change their form, add extra
12188 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12192 *) Add to configuration table a new entry that can specify an alternative
12193 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12194 according to Mark Crispin <MRC@Panda.COM>.
12198 *) DES CBC did not update the IV. Weird.
12201 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12202 Changing the behaviour of the former might break existing programs --
12203 where IV updating is needed, des_ncbc_encrypt can be used.
12206 *) When bntest is run from "make test" it drives bc to check its
12207 calculations, as well as internally checking them. If an internal check
12208 fails, it needs to cause bc to give a non-zero result or make test carries
12209 on without noticing the failure. Fixed.
12212 *) DES library cleanups.
12215 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12216 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12217 ciphers. NOTE: although the key derivation function has been verified
12218 against some published test vectors it has not been extensively tested
12219 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12223 *) Instead of "mkdir -p", which is not fully portable, use new
12224 Perl script "util/mkdir-p.pl".
12227 *) Rewrite the way password based encryption (PBE) is handled. It used to
12228 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12229 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12230 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12231 the 'parameter' field of the AlgorithmIdentifier is passed to the
12232 underlying key generation function so it must do its own ASN1 parsing.
12233 This has also changed the EVP_PBE_CipherInit() function which now has a
12234 'parameter' argument instead of literal salt and iteration count values
12235 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12238 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12239 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12240 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12241 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12242 value was just used as a "magic string" and not used directly its
12243 value doesn't matter.
12246 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12250 *) "linux-sparc64" configuration (ultrapenguin).
12251 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12252 "linux-sparc" configuration.
12253 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12255 *) config now generates no-xxx options for missing ciphers.
12258 *) Support the EBCDIC character set (work in progress).
12259 File ebcdic.c not yet included because it has a different license.
12260 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12262 *) Support BS2000/OSD-POSIX.
12263 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12265 *) Make callbacks for key generation use void * instead of char *.
12268 *) Make S/MIME samples compile (not yet tested).
12271 *) Additional typesafe stacks.
12274 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12278 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12280 *) New configuration variant "sco5-gcc".
12282 *) Updated some demos.
12283 [Sean O Riordain, Wade Scholine]
12285 *) Add missing BIO_free at exit of pkcs12 application.
12288 *) Fix memory leak in conf.c.
12291 *) Updates for Win32 to assembler version of MD5.
12294 *) Set #! path to perl in apps/der_chop to where we found it
12295 instead of using a fixed path.
12298 *) SHA library changes for irix64-mips4-cc.
12301 *) Improvements for VMS support.
12305 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12307 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12308 This also avoids the problems with SC4.2 and unpatched SC5.
12309 [Andy Polyakov <appro@fy.chalmers.se>]
12311 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12312 These are required because of the typesafe stack would otherwise break
12313 existing code. If old code used a structure member which used to be STACK
12314 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12315 sk_num or sk_value it would produce an error because the num, data members
12316 are not present in STACK_OF. Now it just produces a warning. sk_set
12317 replaces the old method of assigning a value to sk_value
12318 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12319 that does this will no longer work (and should use sk_set instead) but
12320 this could be regarded as a "questionable" behaviour anyway.
12323 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12324 correctly handle encrypted S/MIME data.
12327 *) Change type of various DES function arguments from des_cblock
12328 (which means, in function argument declarations, pointer to char)
12329 to des_cblock * (meaning pointer to array with 8 char elements),
12330 which allows the compiler to do more typechecking; it was like
12331 that back in SSLeay, but with lots of ugly casts.
12333 Introduce new type const_des_cblock.
12336 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12337 problems: find RecipientInfo structure that matches recipient certificate
12338 and initialise the ASN1 structures properly based on passed cipher.
12341 *) Belatedly make the BN tests actually check the results.
12344 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12345 to and from BNs: it was completely broken. New compilation option
12346 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12347 key elements as negative integers.
12350 *) Reorganize and speed up MD5.
12351 [Andy Polyakov <appro@fy.chalmers.se>]
12354 [Richard Levitte <richard@levitte.org>]
12356 *) New option -out to asn1parse to allow the parsed structure to be
12357 output to a file. This is most useful when combined with the -strparse
12358 option to examine the output of things like OCTET STRINGS.
12361 *) Make SSL library a little more fool-proof by not requiring any longer
12362 that SSL_set_{accept,connect}_state be called before
12363 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12364 in many applications because usually everything *appeared* to work as
12365 intended anyway -- now it really works as intended).
12368 *) Move openssl.cnf out of lib/.
12371 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12372 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12373 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12374 [Ralf S. Engelschall]
12376 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12377 handle PKCS#7 enveloped data properly.
12378 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12380 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12381 copying pointers. The cert_st handling is changed by this in
12382 various ways (and thus what used to be known as ctx->default_cert
12383 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12384 any longer when s->cert does not give us what we need).
12385 ssl_cert_instantiate becomes obsolete by this change.
12386 As soon as we've got the new code right (possibly it already is?),
12387 we have solved a couple of bugs of the earlier code where s->cert
12388 was used as if it could not have been shared with other SSL structures.
12390 Note that using the SSL API in certain dirty ways now will result
12391 in different behaviour than observed with earlier library versions:
12392 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12393 does not influence s as it used to.
12395 In order to clean up things more thoroughly, inside SSL_SESSION
12396 we don't use CERT any longer, but a new structure SESS_CERT
12397 that holds per-session data (if available); currently, this is
12398 the peer's certificate chain and, for clients, the server's certificate
12399 and temporary key. CERT holds only those values that can have
12400 meaningful defaults in an SSL_CTX.
12403 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12404 from the internal representation. Various PKCS#7 fixes: remove some
12405 evil casts and set the enc_dig_alg field properly based on the signing
12409 *) Allow PKCS#12 password to be set from the command line or the
12410 environment. Let 'ca' get its config file name from the environment
12411 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12415 *) Allow certificate policies extension to use an IA5STRING for the
12416 organization field. This is contrary to the PKIX definition but
12417 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12421 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12422 without disallowing inline assembler and the like for non-pedantic builds.
12425 *) Support Borland C++ builder.
12426 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12428 *) Support Mingw32.
12431 *) SHA-1 cleanups and performance enhancements.
12432 [Andy Polyakov <appro@fy.chalmers.se>]
12434 *) Sparc v8plus assembler for the bignum library.
12435 [Andy Polyakov <appro@fy.chalmers.se>]
12437 *) Accept any -xxx and +xxx compiler options in Configure.
12440 *) Update HPUX configuration.
12443 *) Add missing sk_<type>_unshift() function to safestack.h
12444 [Ralf S. Engelschall]
12446 *) New function SSL_CTX_use_certificate_chain_file that sets the
12447 "extra_cert"s in addition to the certificate. (This makes sense
12448 only for "PEM" format files, as chains as a whole are not
12452 *) Support verify_depth from the SSL API.
12453 x509_vfy.c had what can be considered an off-by-one-error:
12454 Its depth (which was not part of the external interface)
12455 was actually counting the number of certificates in a chain;
12456 now it really counts the depth.
12459 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12460 instead of X509err, which often resulted in confusing error
12461 messages since the error codes are not globally unique
12462 (e.g. an alleged error in ssl3_accept when a certificate
12463 didn't match the private key).
12465 *) New function SSL_CTX_set_session_id_context that allows to set a default
12466 value (so that you don't need SSL_set_session_id_context for each
12467 connection using the SSL_CTX).
12470 *) OAEP decoding bug fix.
12473 *) Support INSTALL_PREFIX for package builders, as proposed by
12477 *) New Configure options "threads" and "no-threads". For systems
12478 where the proper compiler options are known (currently Solaris
12479 and Linux), "threads" is the default.
12482 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12485 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12486 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12487 such as /usr/local/bin.
12490 *) "make linux-shared" to build shared libraries.
12491 [Niels Poppe <niels@netbox.org>]
12493 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12496 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12497 extension adding in x509 utility.
12500 *) Remove NOPROTO sections and error code comments.
12503 *) Partial rewrite of the DEF file generator to now parse the ANSI
12507 *) New Configure options --prefix=DIR and --openssldir=DIR.
12510 *) Complete rewrite of the error code script(s). It is all now handled
12511 by one script at the top level which handles error code gathering,
12512 header rewriting and C source file generation. It should be much better
12513 than the old method: it now uses a modified version of Ulf's parser to
12514 read the ANSI prototypes in all header files (thus the old K&R definitions
12515 aren't needed for error creation any more) and do a better job of
12516 translating function codes into names. The old 'ASN1 error code imbedded
12517 in a comment' is no longer necessary and it doesn't use .err files which
12518 have now been deleted. Also the error code call doesn't have to appear all
12519 on one line (which resulted in some large lines...).
12522 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12525 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12526 0 (which usually indicates a closed connection), but continue reading.
12529 *) Fix some race conditions.
12532 *) Add support for CRL distribution points extension. Add Certificate
12533 Policies and CRL distribution points documentation.
12536 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12539 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12540 8 of keying material. Merlin has also confirmed interop with this fix
12541 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12542 [Merlin Hughes <merlin@baltimore.ie>]
12544 *) Fix lots of warnings.
12545 [Richard Levitte <levitte@stacken.kth.se>]
12547 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12548 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12549 [Richard Levitte <levitte@stacken.kth.se>]
12551 *) Fix problems with sizeof(long) == 8.
12552 [Andy Polyakov <appro@fy.chalmers.se>]
12554 *) Change functions to ANSI C.
12557 *) Fix typos in error codes.
12558 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12560 *) Remove defunct assembler files from Configure.
12563 *) SPARC v8 assembler BIGNUM implementation.
12564 [Andy Polyakov <appro@fy.chalmers.se>]
12566 *) Support for Certificate Policies extension: both print and set.
12567 Various additions to support the r2i method this uses.
12570 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12571 return a const string when you are expecting an allocated buffer.
12574 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12575 types DirectoryString and DisplayText.
12578 *) Add code to allow r2i extensions to access the configuration database,
12579 add an LHASH database driver and add several ctx helper functions.
12582 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12583 fail when they extended the size of a BIGNUM.
12586 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12587 support typesafe stack.
12590 *) Fix typo in SSL_[gs]et_options().
12591 [Nils Frostberg <nils@medcom.se>]
12593 *) Delete various functions and files that belonged to the (now obsolete)
12594 old X509V3 handling code.
12597 *) New Configure option "rsaref".
12600 *) Don't auto-generate pem.h.
12603 *) Introduce type-safe ASN.1 SETs.
12606 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12607 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12609 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12610 that links with OpenSSL (well at least cause lots of warnings), but fear
12611 not: the conversion is trivial, and it eliminates loads of evil casts. A
12612 few STACKed things have been converted already. Feel free to convert more.
12613 In the fullness of time, I'll do away with the STACK type altogether.
12616 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12617 specified in <certfile> by updating the entry in the index.txt file.
12618 This way one no longer has to edit the index.txt file manually for
12619 revoking a certificate. The -revoke option does the gory details now.
12620 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12622 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12623 `-text' option at all and this way the `-noout -text' combination was
12624 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12625 [Ralf S. Engelschall]
12627 *) Make sure a corresponding plain text error message exists for the
12628 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12629 verify callback function determined that a certificate was revoked.
12630 [Ralf S. Engelschall]
12632 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12633 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12634 all available ciphers including rc5, which was forgotten until now.
12635 In order to let the testing shell script know which algorithms
12636 are available, a new (up to now undocumented) command
12637 "openssl list-cipher-commands" is used.
12640 *) Bugfix: s_client occasionally would sleep in select() when
12641 it should have checked SSL_pending() first.
12644 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12645 the raw DSA values prior to ASN.1 encoding.
12648 *) Tweaks to Configure
12649 [Niels Poppe <niels@netbox.org>]
12651 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12655 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12658 *) New config option to avoid instructions that are illegal on the 80386.
12659 The default code is faster, but requires at least a 486.
12662 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12663 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12664 same as SSL2_VERSION anyway.
12667 *) New "-showcerts" option for s_client.
12670 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12671 application. Various cleanups and fixes.
12674 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12675 modify error routines to work internally. Add error codes and PBE init
12676 to library startup routines.
12679 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12680 packing functions to asn1 and evp. Changed function names and error
12681 codes along the way.
12684 *) PKCS12 integration: and so it begins... First of several patches to
12685 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12686 objects to objects.h
12689 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12690 and display support for Thawte strong extranet extension.
12693 *) Add LinuxPPC support.
12694 [Jeff Dubrule <igor@pobox.org>]
12696 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12697 bn_div_words in alpha.s.
12698 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12700 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12701 OAEP isn't supported when OpenSSL is built with RSAref.
12702 [Ulf Moeller <ulf@fitug.de>]
12704 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12705 so they no longer are missing under -DNOPROTO.
12706 [Soren S. Jorvang <soren@t.dk>]
12709 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12711 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12712 doesn't work when the session is reused. Coming soon!
12715 *) Fix a security hole, that allows sessions to be reused in the wrong
12716 context thus bypassing client cert protection! All software that uses
12717 client certs and session caches in multiple contexts NEEDS PATCHING to
12718 allow session reuse! A fuller solution is in the works.
12719 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12721 *) Some more source tree cleanups (removed obsolete files
12722 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12723 permission on "config" script to be executable) and a fix for the INSTALL
12725 [Ulf Moeller <ulf@fitug.de>]
12727 *) Remove some legacy and erroneous uses of malloc, free instead of
12729 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12731 *) Make rsa_oaep_test return non-zero on error.
12732 [Ulf Moeller <ulf@fitug.de>]
12734 *) Add support for native Solaris shared libraries. Configure
12735 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12736 if someone would make that last step automatic.
12737 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12739 *) ctx_size was not built with the right compiler during "make links". Fixed.
12742 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12743 except NULL ciphers". This means the default cipher list will no longer
12744 enable NULL ciphers. They need to be specifically enabled e.g. with
12745 the string "DEFAULT:eNULL".
12748 *) Fix to RSA private encryption routines: if p < q then it would
12749 occasionally produce an invalid result. This will only happen with
12750 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12753 *) Be less restrictive and allow also `perl util/perlpath.pl
12754 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12755 because this way one can also use an interpreter named `perl5' (which is
12756 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12757 installed as `perl').
12758 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12760 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12761 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12763 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12764 advapi32.lib to Win32 build and change the pem test comparison
12765 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12766 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12767 and crypto/des/ede_cbcm_enc.c.
12770 *) DES quad checksum was broken on big-endian architectures. Fixed.
12773 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12774 Win32 test batch file so it (might) work again. The Win32 test batch file
12775 is horrible: I feel ill....
12778 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12779 in e_os.h. Audit of header files to check ANSI and non ANSI
12780 sections: 10 functions were absent from non ANSI section and not exported
12781 from Windows DLLs. Fixed up libeay.num for new functions.
12784 *) Make `openssl version' output lines consistent.
12785 [Ralf S. Engelschall]
12787 *) Fix Win32 symbol export lists for BIO functions: Added
12788 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12789 to ms/libeay{16,32}.def.
12790 [Ralf S. Engelschall]
12792 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12793 fine under Unix and passes some trivial tests I've now added. But the
12794 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12795 added to make sure no one expects that this stuff really works in the
12796 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12797 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12799 [Ralf S. Engelschall]
12801 *) Fix the generation of two part addresses in perl.
12802 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12804 *) Add config entry for Linux on MIPS.
12805 [John Tobey <jtobey@channel1.com>]
12807 *) Make links whenever Configure is run, unless we are on Windoze.
12810 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12811 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12815 *) Add a useful kludge to allow package maintainers to specify compiler and
12816 other platforms details on the command line without having to patch the
12817 Configure script everytime: One now can use ``perl Configure
12818 <id>:<details>'', i.e. platform ids are allowed to have details appended
12819 to them (separated by colons). This is treated as there would be a static
12820 pre-configured entry in Configure's %table under key <id> with value
12821 <details> and ``perl Configure <id>'' is called. So, when you want to
12822 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12823 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12824 now, which overrides the FreeBSD-elf entry on-the-fly.
12825 [Ralf S. Engelschall]
12827 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12830 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12831 on the `perl Configure ...' command line. This way one can compile
12832 OpenSSL libraries with Position Independent Code (PIC) which is needed
12833 for linking it into DSOs.
12834 [Ralf S. Engelschall]
12836 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12840 *) Cleaned up the LICENSE document: The official contact for any license
12841 questions now is the OpenSSL core team under openssl-core@openssl.org.
12842 And add a paragraph about the dual-license situation to make sure people
12843 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12844 to the OpenSSL toolkit.
12845 [Ralf S. Engelschall]
12847 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12848 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12849 Additionally cleaned up the `make links' target: Remove unnecessary
12850 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12851 to speed processing and no longer clutter the display with confusing
12852 stuff. Instead only the actually done links are displayed.
12853 [Ralf S. Engelschall]
12855 *) Permit null encryption ciphersuites, used for authentication only. It used
12856 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12857 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12861 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12862 signed attributes when verifying signatures (this would break them),
12863 the detached data encoding was wrong and public keys obtained using
12864 X509_get_pubkey() weren't freed.
12867 *) Add text documentation for the BUFFER functions. Also added a work around
12868 to a Win95 console bug. This was triggered by the password read stuff: the
12869 last character typed gets carried over to the next fread(). If you were
12870 generating a new cert request using 'req' for example then the last
12871 character of the passphrase would be CR which would then enter the first
12875 *) Added the new `Includes OpenSSL Cryptography Software' button as
12876 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12877 button and can be used by applications based on OpenSSL to show the
12878 relationship to the OpenSSL project.
12879 [Ralf S. Engelschall]
12881 *) Remove confusing variables in function signatures in files
12882 ssl/ssl_lib.c and ssl/ssl.h.
12883 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12885 *) Don't install bss_file.c under PREFIX/include/
12886 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12888 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12889 functions that return function pointers and has support for NT specific
12890 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12891 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12892 unsigned to signed types: this was killing the Win32 compile.
12895 *) Add new certificate file to stack functions,
12896 SSL_add_dir_cert_subjects_to_stack() and
12897 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12898 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12899 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12900 This means that Apache-SSL and similar packages don't have to mess around
12901 to add as many CAs as they want to the preferred list.
12904 *) Experiment with doxygen documentation. Currently only partially applied to
12906 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12907 openssl.doxy as the configuration file.
12910 *) Get rid of remaining C++-style comments which strict C compilers hate.
12911 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12913 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12914 compiled in by default: it has problems with large keys.
12917 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12918 DH private keys and/or callback functions which directly correspond to
12919 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12920 is needed for applications which have to configure certificates on a
12921 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12923 For the RSA certificate situation is makes no difference, but
12924 for the DSA certificate situation this fixes the "no shared cipher"
12925 problem where the OpenSSL cipher selection procedure failed because the
12926 temporary keys were not overtaken from the context and the API provided
12927 no way to reconfigure them.
12928 The new functions now let applications reconfigure the stuff and they
12929 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12930 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12931 non-public-API function ssl_cert_instantiate() is used as a helper
12932 function and also to reduce code redundancy inside ssl_rsa.c.
12933 [Ralf S. Engelschall]
12935 *) Move s_server -dcert and -dkey options out of the undocumented feature
12936 area because they are useful for the DSA situation and should be
12937 recognized by the users.
12938 [Ralf S. Engelschall]
12940 *) Fix the cipher decision scheme for export ciphers: the export bits are
12941 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12942 SSL_EXP_MASK. So, the original variable has to be used instead of the
12943 already masked variable.
12944 [Richard Levitte <levitte@stacken.kth.se>]
12946 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12947 [Richard Levitte <levitte@stacken.kth.se>]
12949 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12950 from `int' to `unsigned int' because it's a length and initialized by
12951 EVP_DigestFinal() which expects an `unsigned int *'.
12952 [Richard Levitte <levitte@stacken.kth.se>]
12954 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12955 script. Instead use the usual Shell->Perl transition trick.
12956 [Ralf S. Engelschall]
12958 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12959 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12960 -noout -modulus' as it's already the case for `openssl rsa -noout
12961 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12962 currently the public key is printed (a decision which was already done by
12963 `openssl dsa -modulus' in the past) which serves a similar purpose.
12964 Additionally the NO_RSA no longer completely removes the whole -modulus
12965 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12967 [Ralf S. Engelschall]
12969 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12970 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12971 [Arne Ansper <arne@ats.cyber.ee>]
12973 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12974 to be added. Now both 'req' and 'ca' can use new objects defined in the
12978 *) Add cool BIO that does syslog (or event log on NT).
12979 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12981 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12982 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12983 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12984 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12987 *) Add preliminary config info for new extension code.
12990 *) Make RSA_NO_PADDING really use no padding.
12991 [Ulf Moeller <ulf@fitug.de>]
12993 *) Generate errors when private/public key check is done.
12996 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12997 for some CRL extensions and new objects added.
13000 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
13001 key usage extension and fuller support for authority key id.
13004 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
13005 padding method for RSA, which is recommended for new applications in PKCS
13006 #1 v2.0 (RFC 2437, October 1998).
13007 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
13008 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
13009 against Bleichbacher's attack on RSA.
13010 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
13013 *) Updates to the new SSL compression code
13014 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13016 *) Fix so that the version number in the master secret, when passed
13017 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
13018 (because the server will not accept higher), that the version number
13019 is 0x03,0x01, not 0x03,0x00
13020 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13022 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
13023 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
13024 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
13027 *) Support for RAW extensions where an arbitrary extension can be
13028 created by including its DER encoding. See apps/openssl.cnf for
13032 *) Make sure latest Perl versions don't interpret some generated C array
13033 code as Perl array code in the crypto/err/err_genc.pl script.
13034 [Lars Weber <3weber@informatik.uni-hamburg.de>]
13036 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
13037 not many people have the assembler. Various Win32 compilation fixes and
13038 update to the INSTALL.W32 file with (hopefully) more accurate Win32
13039 build instructions.
13042 *) Modify configure script 'Configure' to automatically create crypto/date.h
13043 file under Win32 and also build pem.h from pem.org. New script
13044 util/mkfiles.pl to create the MINFO file on environments that can't do a
13045 'make files': perl util/mkfiles.pl >MINFO should work.
13048 *) Major rework of DES function declarations, in the pursuit of correctness
13049 and purity. As a result, many evil casts evaporated, and some weirdness,
13050 too. You may find this causes warnings in your code. Zapping your evil
13051 casts will probably fix them. Mostly.
13054 *) Fix for a typo in asn1.h. Bug fix to object creation script
13055 obj_dat.pl. It considered a zero in an object definition to mean
13056 "end of object": none of the objects in objects.h have any zeros
13057 so it wasn't spotted.
13058 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
13060 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
13061 Masking (CBCM). In the absence of test vectors, the best I have been able
13062 to do is check that the decrypt undoes the encrypt, so far. Send me test
13063 vectors if you have them.
13066 *) Correct calculation of key length for export ciphers (too much space was
13067 allocated for null ciphers). This has not been tested!
13070 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
13071 message is now correct (it understands "crypto" and "ssl" on its
13072 command line). There is also now an "update" option. This will update
13073 the util/ssleay.num and util/libeay.num files with any new functions.
13075 perl util/mkdef.pl crypto ssl update
13076 it will update them.
13079 *) Overhauled the Perl interface (perl/*):
13080 - ported BN stuff to OpenSSL's different BN library
13081 - made the perl/ source tree CVS-aware
13082 - renamed the package from SSLeay to OpenSSL (the files still contain
13083 their history because I've copied them in the repository)
13084 - removed obsolete files (the test scripts will be replaced
13085 by better Test::Harness variants in the future)
13086 [Ralf S. Engelschall]
13088 *) First cut for a very conservative source tree cleanup:
13089 1. merge various obsolete readme texts into doc/ssleay.txt
13090 where we collect the old documents and readme texts.
13091 2. remove the first part of files where I'm already sure that we no
13092 longer need them because of three reasons: either they are just temporary
13093 files which were left by Eric or they are preserved original files where
13094 I've verified that the diff is also available in the CVS via "cvs diff
13095 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13096 the crypto/md/ stuff).
13097 [Ralf S. Engelschall]
13099 *) More extension code. Incomplete support for subject and issuer alt
13100 name, issuer and authority key id. Change the i2v function parameters
13101 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13102 what that's for :-) Fix to ASN1 macro which messed up
13103 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13106 *) Preliminary support for ENUMERATED type. This is largely copied from the
13110 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13111 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13113 *) Make sure `make rehash' target really finds the `openssl' program.
13114 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13116 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13117 like to hear about it if this slows down other processors.
13120 *) Add CygWin32 platform information to Configure script.
13121 [Alan Batie <batie@aahz.jf.intel.com>]
13123 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13124 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13126 *) New program nseq to manipulate netscape certificate sequences
13129 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13133 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13134 but the BN code had some problems that would cause failures when
13135 doing certificate verification and some other functions.
13136 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13138 *) Add ASN1 and PEM code to support netscape certificate sequences.
13141 *) Add ASN1 and PEM code to support netscape certificate sequences.
13144 *) Add several PKIX and private extended key usage OIDs.
13147 *) Modify the 'ca' program to handle the new extension code. Modify
13148 openssl.cnf for new extension format, add comments.
13151 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13152 and add a sample to openssl.cnf so req -x509 now adds appropriate
13156 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13157 error code, add initial support to X509_print() and x509 application.
13160 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13161 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13162 stuff is currently isolated and isn't even compiled yet.
13165 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13166 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13167 Removed the versions check from X509 routines when loading extensions:
13168 this allows certain broken certificates that don't set the version
13169 properly to be processed.
13172 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13173 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13174 can still be regenerated with "make depend".
13177 *) Spelling mistake in C version of CAST-128.
13178 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13180 *) Changes to the error generation code. The perl script err-code.pl
13181 now reads in the old error codes and retains the old numbers, only
13182 adding new ones if necessary. It also only changes the .err files if new
13183 codes are added. The makefiles have been modified to only insert errors
13184 when needed (to avoid needlessly modifying header files). This is done
13185 by only inserting errors if the .err file is newer than the auto generated
13186 C file. To rebuild all the error codes from scratch (the old behaviour)
13187 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13188 or delete all the .err files.
13191 *) CAST-128 was incorrectly implemented for short keys. The C version has
13192 been fixed, but is untested. The assembler versions are also fixed, but
13193 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13194 to regenerate it if needed.
13195 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13196 Hagino <itojun@kame.net>]
13198 *) File was opened incorrectly in randfile.c.
13199 [Ulf Möller <ulf@fitug.de>]
13201 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13202 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13203 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13204 al: it's just almost always a UTCTime. Note this patch adds new error
13205 codes so do a "make errors" if there are problems.
13208 *) Correct Linux 1 recognition in config.
13209 [Ulf Möller <ulf@fitug.de>]
13211 *) Remove pointless MD5 hash when using DSA keys in ca.
13212 [Anonymous <nobody@replay.com>]
13214 *) Generate an error if given an empty string as a cert directory. Also
13215 generate an error if handed NULL (previously returned 0 to indicate an
13216 error, but didn't set one).
13217 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13219 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13222 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13223 parameters. This was causing a warning which killed off the Win32 compile.
13226 *) Remove C++ style comments from crypto/bn/bn_local.h.
13227 [Neil Costigan <neil.costigan@celocom.com>]
13229 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13230 based on a text string, looking up short and long names and finally
13231 "dot" format. The "dot" format stuff didn't work. Added new function
13232 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13233 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13234 OID is not part of the table.
13237 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13238 X509_LOOKUP_by_alias().
13241 *) Sort openssl functions by name.
13244 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13245 encryption from sample DSA keys (in case anyone is interested the password
13249 *) Make _all_ *_free functions accept a NULL pointer.
13250 [Frans Heymans <fheymans@isaserver.be>]
13252 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13254 [Anonymous <nobody@replay.com>]
13256 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13257 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13259 *) Don't blow it for numeric -newkey arguments to apps/req.
13260 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13262 *) Temp key "for export" tests were wrong in s3_srvr.c.
13263 [Anonymous <nobody@replay.com>]
13265 *) Add prototype for temp key callback functions
13266 SSL_CTX_set_tmp_{rsa,dh}_callback().
13269 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13270 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13273 *) X509_name_add_entry() freed the wrong thing after an error.
13274 [Arne Ansper <arne@ats.cyber.ee>]
13276 *) rsa_eay.c would attempt to free a NULL context.
13277 [Arne Ansper <arne@ats.cyber.ee>]
13279 *) BIO_s_socket() had a broken should_retry() on Windoze.
13280 [Arne Ansper <arne@ats.cyber.ee>]
13282 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13283 [Arne Ansper <arne@ats.cyber.ee>]
13285 *) Make sure the already existing X509_STORE->depth variable is initialized
13286 in X509_STORE_new(), but document the fact that this variable is still
13287 unused in the certificate verification process.
13288 [Ralf S. Engelschall]
13290 *) Fix the various library and apps files to free up pkeys obtained from
13291 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13294 *) Fix reference counting in X509_PUBKEY_get(). This makes
13295 demos/maurice/example2.c work, amongst others, probably.
13296 [Steve Henson and Ben Laurie]
13298 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13299 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13300 are no longer created. This way we have a single and consistent command
13301 line interface `openssl <command>', similar to `cvs <command>'.
13302 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13304 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13305 BIT STRING wrapper always have zero unused bits.
13308 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13311 *) Make the top-level INSTALL documentation easier to understand.
13314 *) Makefiles updated to exit if an error occurs in a sub-directory
13315 make (including if user presses ^C) [Paul Sutton]
13317 *) Make Montgomery context stuff explicit in RSA data structure.
13320 *) Fix build order of pem and err to allow for generated pem.h.
13323 *) Fix renumbering bug in X509_NAME_delete_entry().
13326 *) Enhanced the err-ins.pl script so it makes the error library number
13327 global and can add a library name. This is needed for external ASN1 and
13328 other error libraries.
13331 *) Fixed sk_insert which never worked properly.
13334 *) Fix ASN1 macros so they can handle indefinite length constructed
13335 EXPLICIT tags. Some non standard certificates use these: they can now
13339 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13340 into a single doc/ssleay.txt bundle. This way the information is still
13341 preserved but no longer messes up this directory. Now it's new room for
13342 the new set of documentation files.
13343 [Ralf S. Engelschall]
13345 *) SETs were incorrectly DER encoded. This was a major pain, because they
13346 shared code with SEQUENCEs, which aren't coded the same. This means that
13347 almost everything to do with SETs or SEQUENCEs has either changed name or
13348 number of arguments.
13349 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13351 *) Fix test data to work with the above.
13354 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13355 was already fixed by Eric for 0.9.1 it seems.
13356 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13358 *) Autodetect FreeBSD3.
13361 *) Fix various bugs in Configure. This affects the following platforms:
13365 unixware-2.0-pentium
13369 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13370 before they are needed.
13373 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13377 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13379 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13380 changed SSLeay to OpenSSL in version strings.
13381 [Ralf S. Engelschall]
13383 *) Some fixups to the top-level documents.
13386 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13387 because the symlink to include/ was missing.
13388 [Ralf S. Engelschall]
13390 *) Incorporated the popular no-RSA/DSA-only patches
13391 which allow to compile a RSA-free SSLeay.
13392 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13394 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13395 when "ssleay" is still not found.
13396 [Ralf S. Engelschall]
13398 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13399 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13401 *) Updated the README file.
13402 [Ralf S. Engelschall]
13404 *) Added various .cvsignore files in the CVS repository subdirs
13405 to make a "cvs update" really silent.
13406 [Ralf S. Engelschall]
13408 *) Recompiled the error-definition header files and added
13409 missing symbols to the Win32 linker tables.
13410 [Ralf S. Engelschall]
13412 *) Cleaned up the top-level documents;
13413 o new files: CHANGES and LICENSE
13414 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13415 o merged COPYRIGHT into LICENSE
13416 o removed obsolete TODO file
13417 o renamed MICROSOFT to INSTALL.W32
13418 [Ralf S. Engelschall]
13420 *) Removed dummy files from the 0.9.1b source tree:
13421 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13422 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13423 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13424 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13425 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13426 [Ralf S. Engelschall]
13428 *) Added various platform portability fixes.
13431 *) The Genesis of the OpenSSL rpject:
13432 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13433 Young and Tim J. Hudson created while they were working for C2Net until
13435 [The OpenSSL Project]
13438 Changes between 0.9.0b and 0.9.1b [not released]
13440 *) Updated a few CA certificates under certs/
13443 *) Changed some BIGNUM api stuff.
13446 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13447 DGUX x86, Linux Alpha, etc.
13450 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13451 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13455 *) Add -strparse option to asn1pars program which parses nested
13457 [Dr Stephen Henson <shenson@bigfoot.com>]
13459 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13462 *) DSA fix for "ca" program.
13465 *) Added "-genkey" option to "dsaparam" program.
13468 *) Added RIPE MD160 (rmd160) message digest.
13471 *) Added -a (all) option to "ssleay version" command.
13474 *) Added PLATFORM define which is the id given to Configure.
13477 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13480 *) Extended the ASN.1 parser routines.
13483 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13486 *) Added a BN_CTX to the BN library.
13489 *) Fixed the weak key values in DES library
13492 *) Changed API in EVP library for cipher aliases.
13495 *) Added support for RC2/64bit cipher.
13498 *) Converted the lhash library to the crypto/mem.c functions.
13501 *) Added more recognized ASN.1 object ids.
13504 *) Added more RSA padding checks for SSL/TLS.
13507 *) Added BIO proxy/filter functionality.
13510 *) Added extra_certs to SSL_CTX which can be used
13511 send extra CA certificates to the client in the CA cert chain sending
13512 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13515 *) Now Fortezza is denied in the authentication phase because
13516 this is key exchange mechanism is not supported by SSLeay at all.
13519 *) Additional PKCS1 checks.
13522 *) Support the string "TLSv1" for all TLS v1 ciphers.
13525 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13526 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13529 *) Fixed a few memory leaks.
13532 *) Fixed various code and comment typos.
13535 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13536 bytes sent in the client random.
13537 [Edward Bishop <ebishop@spyglass.com>]