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 *) Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
13 when primes for RSA keys are computed.
14 Since we previously always generated primes == 2 (mod 3) for RSA keys,
15 the 2-prime and 3-prime RSA modules were easy to distinguish, since
16 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
17 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
18 This avoids possible fingerprinting of newly generated RSA modules.
21 *) Correct the extended master secret constant on EBCDIC systems. Without this
22 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
23 negotiate EMS will fail. Unfortunately this also means that TLS connections
24 between EBCDIC systems with this fix, and EBCDIC systems without this
25 fix will fail if they negotiate EMS.
28 *) Changed the library initialisation so that the config file is now loaded
29 by default. This was already the case for libssl. It now occurs for both
30 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
31 OPENSSL_init_crypto() to suppress automatic loading of a config file.
34 *) Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
35 where the former acts as a replacement for ERR_put_error(), and the
36 latter replaces the combination ERR_put_error()+ERR_add_error_data().
37 ERR_raise_data() adds more flexibility by taking a format string and
38 an arbitrary number of arguments following it, to be processed with
42 *) Introduced a new function, OSSL_PROVIDER_available(), which can be used
43 to check if a named provider is loaded and available. When called, it
44 will also activate all fallback providers if such are still present.
47 *) Enforce a minimum DH modulus size of 512 bits.
50 *) Changed DH parameters to generate the order q subgroup instead of 2q.
51 Previously generated DH parameters are still accepted by DH_check
52 but DH_generate_key works around that by clearing bit 0 of the
53 private key for those. This avoids leaking bit 0 of the private key.
56 *) Significantly reduce secure memory usage by the randomness pools.
59 *) {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
63 *) A new type, EVP_KEYEXCH, has been introduced to represent key exchange
64 algorithms. An implementation of a key exchange algorithm can be obtained
65 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
66 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
67 the older EVP_PKEY_derive_init() function. See the man pages for the new
68 functions for further details.
71 *) The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
74 *) Removed the function names from error messages and deprecated the
77 *) Removed NextStep support and the macro OPENSSL_UNISTD
80 *) Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
81 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
82 Also removed "export var as function" capability; we do not export
83 variables, only functions.
86 *) RC5_32_set_key has been changed to return an int type, with 0 indicating
87 an error and 1 indicating success. In previous versions of OpenSSL this
88 was a void type. If a key was set longer than the maximum possible this
92 *) Support SM2 signing and verification schemes with X509 certificate.
95 *) Use SHA256 as the default digest for TS query in the ts app.
98 *) Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
99 This checks that the salt length is at least 128 bits, the derived key
100 length is at least 112 bits, and that the iteration count is at least 1000.
101 For backwards compatibility these checks are disabled by default in the
102 default provider, but are enabled by default in the fips provider.
103 To enable or disable these checks use the control
104 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
107 *) Default cipher lists/suites are now available via a function, the
108 #defines are deprecated.
111 *) Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
112 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
113 for Windows Store apps easier. Also, the "no-uplink" option has been added.
116 *) Join the directories crypto/x509 and crypto/x509v3
119 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
120 This changes the size when using the genpkey app when no size is given. It
121 fixes an omission in earlier changes that changed all RSA, DSA and DH
122 generation apps to use 2048 bits by default.
125 *) Added command 'openssl kdf' that uses the EVP_KDF API.
128 *) Added command 'openssl mac' that uses the EVP_MAC API.
131 *) Added OPENSSL_info() to get diverse built-in OpenSSL data, such
132 as default directories. Also added the command 'openssl info'
133 for scripting purposes.
136 *) The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
137 deprecated. These undocumented functions were never integrated into the EVP
138 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
139 Bi-directional IGE mode. These modes were never formally standardised and
140 usage of these functions is believed to be very small. In particular
141 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
142 is ever used. The security implications are believed to be minimal, but
143 this issue was never fixed for backwards compatibility reasons. New code
144 should not use these modes.
147 *) Add prediction resistance to the DRBG reseeding process.
150 *) Limit the number of blocks in a data unit for AES-XTS to 2^20 as
151 mandated by IEEE Std 1619-2018.
154 *) Added newline escaping functionality to a filename when using openssl dgst.
155 This output format is to replicate the output format found in the '*sum'
156 checksum programs. This aims to preserve backward compatibility.
157 [Matt Eaton, Richard Levitte, and Paul Dale]
159 *) Removed the heartbeat message in DTLS feature, as it has very
160 little usage and doesn't seem to fulfill a valuable purpose.
161 The configuration option is now deprecated.
164 *) Changed the output of 'openssl {digestname} < file' to display the
165 digest name in its output.
168 *) Added a new generic trace API which provides support for enabling
169 instrumentation through trace output. This feature is mainly intended
170 as an aid for developers and is disabled by default. To utilize it,
171 OpenSSL needs to be configured with the `enable-trace` option.
173 If the tracing API is enabled, the application can activate trace output
174 by registering BIOs as trace channels for a number of tracing and debugging
177 The 'openssl' application has been expanded to enable any of the types
178 available via environment variables defined by the user, and serves as
179 one possible example on how to use this functionality.
180 [Richard Levitte & Matthias St. Pierre]
182 *) Added build tests for C++. These are generated files that only do one
183 thing, to include one public OpenSSL head file each. This tests that
184 the public header files can be usefully included in a C++ application.
186 This test isn't enabled by default. It can be enabled with the option
187 'enable-buildtest-c++'.
190 *) Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
193 *) Add KMAC to EVP_MAC.
196 *) Added property based algorithm implementation selection framework to
200 *) Added SCA hardening for modular field inversion in EC_GROUP through
201 a new dedicated field_inv() pointer in EC_METHOD.
202 This also addresses a leakage affecting conversions from projective
203 to affine coordinates.
204 [Billy Bob Brumley, Nicola Tuveri]
206 *) Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
207 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
208 those algorithms that were already supported through the EVP_PKEY API
209 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
210 and scrypt are now wrappers that call EVP_KDF.
213 *) Build devcrypto engine as a dynamic engine.
216 *) Add keyed BLAKE2 to EVP_MAC.
219 *) Fix a bug in the computation of the endpoint-pair shared secret used
220 by DTLS over SCTP. This breaks interoperability with older versions
221 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
222 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
223 interoperability with such broken implementations. However, enabling
224 this switch breaks interoperability with correct implementations.
226 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
227 re-used X509_PUBKEY object if the second PUBKEY is malformed.
230 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
233 *) Change the license to the Apache License v2.0.
236 *) Change the possible version information given with OPENSSL_API_COMPAT.
237 It may be a pre-3.0.0 style numerical version number as it was defined
238 in 1.1.0, and it may also simply take the major version number.
240 Because of the version numbering of pre-3.0.0 releases, the values 0,
241 1 and 2 are equivalent to 0x00908000L (0.9.8), 0x10000000L (1.0.0) and
242 0x10100000L (1.1.0), respectively.
245 *) Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
247 o Major releases (indicated by incrementing the MAJOR release number)
248 may introduce incompatible API/ABI changes.
249 o Minor releases (indicated by incrementing the MINOR release number)
250 may introduce new features but retain API/ABI compatibility.
251 o Patch releases (indicated by incrementing the PATCH number)
252 are intended for bug fixes and other improvements of existing
253 features only (like improving performance or adding documentation)
254 and retain API/ABI compatibility.
257 *) Add support for RFC5297 SIV mode (siv128), including AES-SIV.
260 *) Remove the 'dist' target and add a tarball building script. The
261 'dist' target has fallen out of use, and it shouldn't be
262 necessary to configure just to create a source distribution.
265 *) Recreate the OS390-Unix config target. It no longer relies on a
266 special script like it did for OpenSSL pre-1.1.0.
269 *) Instead of having the source directories listed in Configure, add
270 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
274 *) Add GMAC to EVP_MAC.
277 *) Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
280 *) Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
281 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
282 to facilitate the continued use of MACs through raw private keys in
283 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
286 *) Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
287 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
290 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
291 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
292 are retained for backwards compatibility.
295 *) AES-XTS mode now enforces that its two keys are different to mitigate
296 the attacked described in "Efficient Instantiations of Tweakable
297 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
298 Details of this attack can be obtained from:
299 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
302 *) Rename the object files, i.e. give them other names than in previous
303 versions. Their names now include the name of the final product, as
304 well as its type mnemonic (bin, lib, shlib).
307 *) Added new option for 'openssl list', '-objects', which will display the
308 list of built in objects, i.e. OIDs with names.
311 *) Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
312 improves application performance by removing data copies and providing
313 applications with zero-copy system calls such as sendfile and splice.
316 Changes between 1.1.1a and 1.1.1b [xx XXX xxxx]
318 *) Change the info callback signals for the start and end of a post-handshake
319 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
320 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
321 confused by this and assume that a TLSv1.2 renegotiation has started. This
322 can break KeyUpdate handling. Instead we no longer signal the start and end
323 of a post handshake message exchange (although the messages themselves are
324 still signalled). This could break some applications that were expecting
325 the old signals. However without this KeyUpdate is not usable for many
329 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
331 *) Timing vulnerability in DSA signature generation
333 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
334 timing side channel attack. An attacker could use variations in the signing
335 algorithm to recover the private key.
337 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
341 *) Timing vulnerability in ECDSA signature generation
343 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
344 timing side channel attack. An attacker could use variations in the signing
345 algorithm to recover the private key.
347 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
351 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
352 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
353 of two gigabytes and the error handling improved.
355 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
356 categorized as a normal bug, not a security issue, because the DRBG reseeds
357 automatically and is fully functional even without additional randomness
358 provided by the application.
360 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
362 *) Add a new ClientHello callback. Provides a callback interface that gives
363 the application the ability to adjust the nascent SSL object at the
364 earliest stage of ClientHello processing, immediately after extensions have
365 been collected but before they have been processed. In particular, this
366 callback can adjust the supported TLS versions in response to the contents
370 *) Add SM2 base algorithm support.
373 *) s390x assembly pack: add (improved) hardware-support for the following
374 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
375 aes-cfb/cfb8, aes-ecb.
378 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
379 parameter is no longer accepted, as it leads to a corrupt table. NULL
380 pem_str is reserved for alias entries only.
383 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
384 step for prime curves. The new implementation is based on formulae from
385 differential addition-and-doubling in homogeneous projective coordinates
386 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
387 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
388 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
389 to work in projective coordinates.
390 [Billy Bob Brumley, Nicola Tuveri]
392 *) Change generating and checking of primes so that the error rate of not
393 being prime depends on the intended use based on the size of the input.
394 For larger primes this will result in more rounds of Miller-Rabin.
395 The maximal error rate for primes with more than 1080 bits is lowered
397 [Kurt Roeckx, Annie Yousar]
399 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
402 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
403 moving between systems, and to avoid confusion when a Windows build is
404 done with mingw vs with MSVC. For POSIX installs, there's still a
405 symlink or copy named 'tsget' to avoid that confusion as well.
408 *) Revert blinding in ECDSA sign and instead make problematic addition
409 length-invariant. Switch even to fixed-length Montgomery multiplication.
412 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
413 step for binary curves. The new implementation is based on formulae from
414 differential addition-and-doubling in mixed Lopez-Dahab projective
415 coordinates, modified to independently blind the operands.
416 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
418 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
419 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
420 EC_METHODs to implement their own specialized "ladder step", to take
421 advantage of more favorable coordinate systems or more efficient
422 differential addition-and-doubling algorithms.
423 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
425 *) Modified the random device based seed sources to keep the relevant
426 file descriptors open rather than reopening them on each access.
427 This allows such sources to operate in a chroot() jail without
428 the associated device nodes being available. This behaviour can be
429 controlled using RAND_keep_random_devices_open().
432 *) Numerous side-channel attack mitigations have been applied. This may have
433 performance impacts for some algorithms for the benefit of improved
434 security. Specific changes are noted in this change log by their respective
438 *) AIX shared library support overhaul. Switch to AIX "natural" way of
439 handling shared libraries, which means collecting shared objects of
440 different versions and bitnesses in one common archive. This allows to
441 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
442 doesn't affect the way 3rd party applications are linked, only how
443 multi-version installation is managed.
446 *) Make ec_group_do_inverse_ord() more robust and available to other
447 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
448 mitigations are applied to the fallback BN_mod_inverse().
449 When using this function rather than BN_mod_inverse() directly, new
450 EC cryptosystem implementations are then safer-by-default.
453 *) Add coordinate blinding for EC_POINT and implement projective
454 coordinate blinding for generic prime curves as a countermeasure to
455 chosen point SCA attacks.
456 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
458 *) Add blinding to ECDSA and DSA signatures to protect against side channel
459 attacks discovered by Keegan Ryan (NCC Group).
462 *) Enforce checking in the pkeyutl command line app to ensure that the input
463 length does not exceed the maximum supported digest length when performing
464 a sign, verify or verifyrecover operation.
467 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
468 I/O in combination with something like select() or poll() will hang. This
469 can be turned off again using SSL_CTX_clear_mode().
470 Many applications do not properly handle non-application data records, and
471 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
472 around the problems in those applications, but can also break some.
473 It's recommended to read the manpages about SSL_read(), SSL_write(),
474 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
475 SSL_CTX_set_read_ahead() again.
478 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
479 now allow empty (zero character) pass phrases.
482 *) Apply blinding to binary field modular inversion and remove patent
483 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
486 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
487 binary and prime elliptic curves.
490 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
491 constant time fixed point multiplication.
494 *) Revise elliptic curve scalar multiplication with timing attack
495 defenses: ec_wNAF_mul redirects to a constant time implementation
496 when computing fixed point and variable point multiplication (which
497 in OpenSSL are mostly used with secret scalars in keygen, sign,
498 ECDH derive operations).
499 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
502 *) Updated CONTRIBUTING
505 *) Updated DRBG / RAND to request nonce and additional low entropy
506 randomness from the system.
507 [Matthias St. Pierre]
509 *) Updated 'openssl rehash' to use OpenSSL consistent default.
512 *) Moved the load of the ssl_conf module to libcrypto, which helps
513 loading engines that libssl uses before libssl is initialised.
516 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
519 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
520 [Ingo Schwarze, Rich Salz]
522 *) Added output of accepting IP address and port for 'openssl s_server'
525 *) Added a new API for TLSv1.3 ciphersuites:
526 SSL_CTX_set_ciphersuites()
527 SSL_set_ciphersuites()
530 *) Memory allocation failures consistently add an error to the error
534 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
535 in libcrypto when run as setuid/setgid.
538 *) Load any config file by default when libssl is used.
541 *) Added new public header file <openssl/rand_drbg.h> and documentation
542 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
543 [Matthias St. Pierre]
545 *) QNX support removed (cannot find contributors to get their approval
546 for the license change).
549 *) TLSv1.3 replay protection for early data has been implemented. See the
550 SSL_read_early_data() man page for further details.
553 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
554 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
555 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
556 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
557 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
558 configuration has been separated out. See the ciphers man page or the
559 SSL_CTX_set_ciphersuites() man page for more information.
562 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
563 in responder mode now supports the new "-multi" option, which
564 spawns the specified number of child processes to handle OCSP
565 requests. The "-timeout" option now also limits the OCSP
566 responder's patience to wait to receive the full client request
567 on a newly accepted connection. Child processes are respawned
568 as needed, and the CA index file is automatically reloaded
569 when changed. This makes it possible to run the "ocsp" responder
570 as a long-running service, making the OpenSSL CA somewhat more
571 feature-complete. In this mode, most diagnostic messages logged
572 after entering the event loop are logged via syslog(3) rather than
576 *) Added support for X448 and Ed448. Heavily based on original work by
580 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
581 objects loaded. This adds the functions OSSL_STORE_expect() and
582 OSSL_STORE_find() as well as needed tools to construct searches and
583 get the search data out of them.
586 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
587 version of OpenSSL should review their configuration settings to ensure
588 that they are still appropriate for TLSv1.3. For further information see:
589 https://wiki.openssl.org/index.php/TLS1.3
592 *) Grand redesign of the OpenSSL random generator
594 The default RAND method now utilizes an AES-CTR DRBG according to
595 NIST standard SP 800-90Ar1. The new random generator is essentially
596 a port of the default random generator from the OpenSSL FIPS 2.0
597 object module. It is a hybrid deterministic random bit generator
598 using an AES-CTR bit stream and which seeds and reseeds itself
599 automatically using trusted system entropy sources.
601 Some of its new features are:
602 o Support for multiple DRBG instances with seed chaining.
603 o The default RAND method makes use of a DRBG.
604 o There is a public and private DRBG instance.
605 o The DRBG instances are fork-safe.
606 o Keep all global DRBG instances on the secure heap if it is enabled.
607 o The public and private DRBG instance are per thread for lock free
609 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
611 *) Changed Configure so it only says what it does and doesn't dump
612 so much data. Instead, ./configdata.pm should be used as a script
613 to display all sorts of configuration data.
616 *) Added processing of "make variables" to Configure.
619 *) Added SHA512/224 and SHA512/256 algorithm support.
622 *) The last traces of Netware support, first removed in 1.1.0, have
626 *) Get rid of Makefile.shared, and in the process, make the processing
627 of certain files (rc.obj, or the .def/.map/.opt files produced from
628 the ordinal files) more visible and hopefully easier to trace and
629 debug (or make silent).
632 *) Make it possible to have environment variable assignments as
633 arguments to config / Configure.
636 *) Add multi-prime RSA (RFC 8017) support.
639 *) Add SM3 implemented according to GB/T 32905-2016
640 [ Jack Lloyd <jack.lloyd@ribose.com>,
641 Ronald Tse <ronald.tse@ribose.com>,
642 Erick Borsboom <erick.borsboom@ribose.com> ]
644 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
645 as documented in RFC6066.
646 Based on a patch from Tomasz Moń
647 [Filipe Raimundo da Silva]
649 *) Add SM4 implemented according to GB/T 32907-2016.
650 [ Jack Lloyd <jack.lloyd@ribose.com>,
651 Ronald Tse <ronald.tse@ribose.com>,
652 Erick Borsboom <erick.borsboom@ribose.com> ]
654 *) Reimplement -newreq-nodes and ERR_error_string_n; the
655 original author does not agree with the license change.
658 *) Add ARIA AEAD TLS support.
661 *) Some macro definitions to support VS6 have been removed. Visual
662 Studio 6 has not worked since 1.1.0
665 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
666 without clearing the errors.
669 *) Add "atfork" functions. If building on a system that without
670 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
671 requirements. The RAND facility now uses/requires this.
677 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
678 not possible to disable entirely. However, it's still possible to
679 disable the console reading UI method, UI_OpenSSL() (use UI_null()
682 To disable, configure with 'no-ui-console'. 'no-ui' is still
683 possible to use as an alias. Check at compile time with the
684 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
685 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
688 *) Add a STORE module, which implements a uniform and URI based reader of
689 stores that can contain keys, certificates, CRLs and numerous other
690 objects. The main API is loosely based on a few stdio functions,
691 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
692 OSSL_STORE_error and OSSL_STORE_close.
693 The implementation uses backends called "loaders" to implement arbitrary
694 URI schemes. There is one built in "loader" for the 'file' scheme.
697 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
698 then adjusted to work on FreeBSD 8.4 as well.
699 Enable by configuring with 'enable-devcryptoeng'. This is done by default
700 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
703 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
704 util/mkerr.pl, which is adapted to allow those prefixes, leading to
705 error code calls like this:
707 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
709 With this change, we claim the namespaces OSSL and OPENSSL in a manner
710 that can be encoded in C. For the foreseeable future, this will only
712 [Richard Levitte and Tim Hudson]
714 *) Removed BSD cryptodev engine.
717 *) Add a build target 'build_all_generated', to build all generated files
718 and only that. This can be used to prepare everything that requires
719 things like perl for a system that lacks perl and then move everything
720 to that system and do the rest of the build there.
723 *) In the UI interface, make it possible to duplicate the user data. This
724 can be used by engines that need to retain the data for a longer time
725 than just the call where this user data is passed.
728 *) Ignore the '-named_curve auto' value for compatibility of applications
730 [Tomas Mraz <tmraz@fedoraproject.org>]
732 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
733 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
734 alerts across multiple records (some of which could be empty). In practice
735 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
736 prohibts this altogether and other libraries (BoringSSL, NSS) do not
737 support this at all. Supporting it adds significant complexity to the
738 record layer, and its removal is unlikely to cause inter-operability
742 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
743 with Z. These are meant to replace LONG and ZLONG and to be size safe.
744 The use of LONG and ZLONG is discouraged and scheduled for deprecation
748 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
749 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
750 [Richard Levitte, Andy Polyakov]
752 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
756 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
757 platform rather than 'mingw'.
760 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
761 success if they are asked to add an object which already exists
762 in the store. This change cascades to other functions which load
763 certificates and CRLs.
766 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
767 facilitate stack unwinding even from assembly subroutines.
770 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
771 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
774 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
775 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
776 which is the minimum version we support.
779 *) Certificate time validation (X509_cmp_time) enforces stricter
780 compliance with RFC 5280. Fractional seconds and timezone offsets
781 are no longer allowed.
784 *) Add support for ARIA
787 *) s_client will now send the Server Name Indication (SNI) extension by
788 default unless the new "-noservername" option is used. The server name is
789 based on the host provided to the "-connect" option unless overridden by
793 *) Add support for SipHash
796 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
797 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
798 prevent issues where no progress is being made and the peer continually
799 sends unrecognised record types, using up resources processing them.
802 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
803 using the algorithm defined in
804 https://www.akkadia.org/drepper/SHA-crypt.txt
807 *) Heartbeat support has been removed; the ABI is changed for now.
808 [Richard Levitte, Rich Salz]
810 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
813 *) The RSA "null" method, which was partially supported to avoid patent
814 issues, has been replaced to always returns NULL.
818 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
820 *) Client DoS due to large DH parameter
822 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
823 malicious server can send a very large prime value to the client. This will
824 cause the client to spend an unreasonably long period of time generating a
825 key for this prime resulting in a hang until the client has finished. This
826 could be exploited in a Denial Of Service attack.
828 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
832 *) Cache timing vulnerability in RSA Key Generation
834 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
835 a cache timing side channel attack. An attacker with sufficient access to
836 mount cache timing attacks during the RSA key generation process could
837 recover the private key.
839 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
840 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
844 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
845 parameter is no longer accepted, as it leads to a corrupt table. NULL
846 pem_str is reserved for alias entries only.
849 *) Revert blinding in ECDSA sign and instead make problematic addition
850 length-invariant. Switch even to fixed-length Montgomery multiplication.
853 *) Change generating and checking of primes so that the error rate of not
854 being prime depends on the intended use based on the size of the input.
855 For larger primes this will result in more rounds of Miller-Rabin.
856 The maximal error rate for primes with more than 1080 bits is lowered
858 [Kurt Roeckx, Annie Yousar]
860 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
863 *) Add blinding to ECDSA and DSA signatures to protect against side channel
864 attacks discovered by Keegan Ryan (NCC Group).
867 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
868 now allow empty (zero character) pass phrases.
871 *) Certificate time validation (X509_cmp_time) enforces stricter
872 compliance with RFC 5280. Fractional seconds and timezone offsets
873 are no longer allowed.
876 *) Fixed a text canonicalisation bug in CMS
878 Where a CMS detached signature is used with text content the text goes
879 through a canonicalisation process first prior to signing or verifying a
880 signature. This process strips trailing space at the end of lines, converts
881 line terminators to CRLF and removes additional trailing line terminators
882 at the end of a file. A bug in the canonicalisation process meant that
883 some characters, such as form-feed, were incorrectly treated as whitespace
884 and removed. This is contrary to the specification (RFC5485). This fix
885 could mean that detached text data signed with an earlier version of
886 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
887 signed with a fixed OpenSSL may fail to verify with an earlier version of
888 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
889 and use the "-binary" flag (for the "cms" command line application) or set
890 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
893 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
895 *) Constructed ASN.1 types with a recursive definition could exceed the stack
897 Constructed ASN.1 types with a recursive definition (such as can be found
898 in PKCS7) could eventually exceed the stack given malicious input with
899 excessive recursion. This could result in a Denial Of Service attack. There
900 are no such structures used within SSL/TLS that come from untrusted sources
901 so this is considered safe.
903 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
908 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
910 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
911 effectively reduced to only comparing the least significant bit of each
912 byte. This allows an attacker to forge messages that would be considered as
913 authenticated in an amount of tries lower than that guaranteed by the
914 security claims of the scheme. The module can only be compiled by the
915 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
917 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
922 *) Add a build target 'build_all_generated', to build all generated files
923 and only that. This can be used to prepare everything that requires
924 things like perl for a system that lacks perl and then move everything
925 to that system and do the rest of the build there.
928 *) Backport SSL_OP_NO_RENGOTIATION
930 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
931 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
932 changes this is no longer possible in 1.1.0. Therefore the new
933 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
934 1.1.0 to provide equivalent functionality.
936 Note that if an application built against 1.1.0h headers (or above) is run
937 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
938 accepted but nothing will happen, i.e. renegotiation will not be prevented.
941 *) Removed the OS390-Unix config target. It relied on a script that doesn't
945 *) rsaz_1024_mul_avx2 overflow bug on x86_64
947 There is an overflow bug in the AVX2 Montgomery multiplication procedure
948 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
949 Analysis suggests that attacks against RSA and DSA as a result of this
950 defect would be very difficult to perform and are not believed likely.
951 Attacks against DH1024 are considered just feasible, because most of the
952 work necessary to deduce information about a private key may be performed
953 offline. The amount of resources required for such an attack would be
954 significant. However, for an attack on TLS to be meaningful, the server
955 would have to share the DH1024 private key among multiple clients, which is
956 no longer an option since CVE-2016-0701.
958 This only affects processors that support the AVX2 but not ADX extensions
959 like Intel Haswell (4th generation).
961 This issue was reported to OpenSSL by David Benjamin (Google). The issue
962 was originally found via the OSS-Fuzz project.
966 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
968 *) bn_sqrx8x_internal carry bug on x86_64
970 There is a carry propagating bug in the x86_64 Montgomery squaring
971 procedure. No EC algorithms are affected. Analysis suggests that attacks
972 against RSA and DSA as a result of this defect would be very difficult to
973 perform and are not believed likely. Attacks against DH are considered just
974 feasible (although very difficult) because most of the work necessary to
975 deduce information about a private key may be performed offline. The amount
976 of resources required for such an attack would be very significant and
977 likely only accessible to a limited number of attackers. An attacker would
978 additionally need online access to an unpatched system using the target
979 private key in a scenario with persistent DH parameters and a private
980 key that is shared between multiple clients.
982 This only affects processors that support the BMI1, BMI2 and ADX extensions
983 like Intel Broadwell (5th generation) and later or AMD Ryzen.
985 This issue was reported to OpenSSL by the OSS-Fuzz project.
989 *) Malformed X.509 IPAddressFamily could cause OOB read
991 If an X.509 certificate has a malformed IPAddressFamily extension,
992 OpenSSL could do a one-byte buffer overread. The most likely result
993 would be an erroneous display of the certificate in text format.
995 This issue was reported to OpenSSL by the OSS-Fuzz project.
999 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1001 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1002 platform rather than 'mingw'.
1005 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1006 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1007 which is the minimum version we support.
1010 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1012 *) Encrypt-Then-Mac renegotiation crash
1014 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1015 negotiated where it was not in the original handshake (or vice-versa) then
1016 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1017 and servers are affected.
1019 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1023 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1025 *) Truncated packet could crash via OOB read
1027 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1028 cipher is being used, then a truncated packet can cause that host to
1029 perform an out-of-bounds read, usually resulting in a crash.
1031 This issue was reported to OpenSSL by Robert Święcki of Google.
1035 *) Bad (EC)DHE parameters cause a client crash
1037 If a malicious server supplies bad parameters for a DHE or ECDHE key
1038 exchange then this can result in the client attempting to dereference a
1039 NULL pointer leading to a client crash. This could be exploited in a Denial
1042 This issue was reported to OpenSSL by Guido Vranken.
1046 *) BN_mod_exp may produce incorrect results on x86_64
1048 There is a carry propagating bug in the x86_64 Montgomery squaring
1049 procedure. No EC algorithms are affected. Analysis suggests that attacks
1050 against RSA and DSA as a result of this defect would be very difficult to
1051 perform and are not believed likely. Attacks against DH are considered just
1052 feasible (although very difficult) because most of the work necessary to
1053 deduce information about a private key may be performed offline. The amount
1054 of resources required for such an attack would be very significant and
1055 likely only accessible to a limited number of attackers. An attacker would
1056 additionally need online access to an unpatched system using the target
1057 private key in a scenario with persistent DH parameters and a private
1058 key that is shared between multiple clients. For example this can occur by
1059 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1060 similar to CVE-2015-3193 but must be treated as a separate problem.
1062 This issue was reported to OpenSSL by the OSS-Fuzz project.
1066 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1068 *) ChaCha20/Poly1305 heap-buffer-overflow
1070 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1071 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1072 crash. This issue is not considered to be exploitable beyond a DoS.
1074 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1078 *) CMS Null dereference
1080 Applications parsing invalid CMS structures can crash with a NULL pointer
1081 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1082 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1083 structure callback if an attempt is made to free certain invalid encodings.
1084 Only CHOICE structures using a callback which do not handle NULL value are
1087 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1091 *) Montgomery multiplication may produce incorrect results
1093 There is a carry propagating bug in the Broadwell-specific Montgomery
1094 multiplication procedure that handles input lengths divisible by, but
1095 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1096 and DH private keys are impossible. This is because the subroutine in
1097 question is not used in operations with the private key itself and an input
1098 of the attacker's direct choice. Otherwise the bug can manifest itself as
1099 transient authentication and key negotiation failures or reproducible
1100 erroneous outcome of public-key operations with specially crafted input.
1101 Among EC algorithms only Brainpool P-512 curves are affected and one
1102 presumably can attack ECDH key negotiation. Impact was not analyzed in
1103 detail, because pre-requisites for attack are considered unlikely. Namely
1104 multiple clients have to choose the curve in question and the server has to
1105 share the private key among them, neither of which is default behaviour.
1106 Even then only clients that chose the curve will be affected.
1108 This issue was publicly reported as transient failures and was not
1109 initially recognized as a security issue. Thanks to Richard Morgan for
1110 providing reproducible case.
1114 *) Removed automatic addition of RPATH in shared libraries and executables,
1115 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1118 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1120 *) Fix Use After Free for large message sizes
1122 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1123 message larger than approx 16k is received then the underlying buffer to
1124 store the incoming message is reallocated and moved. Unfortunately a
1125 dangling pointer to the old location is left which results in an attempt to
1126 write to the previously freed location. This is likely to result in a
1127 crash, however it could potentially lead to execution of arbitrary code.
1129 This issue only affects OpenSSL 1.1.0a.
1131 This issue was reported to OpenSSL by Robert Święcki.
1135 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1137 *) OCSP Status Request extension unbounded memory growth
1139 A malicious client can send an excessively large OCSP Status Request
1140 extension. If that client continually requests renegotiation, sending a
1141 large OCSP Status Request extension each time, then there will be unbounded
1142 memory growth on the server. This will eventually lead to a Denial Of
1143 Service attack through memory exhaustion. Servers with a default
1144 configuration are vulnerable even if they do not support OCSP. Builds using
1145 the "no-ocsp" build time option are not affected.
1147 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1151 *) SSL_peek() hang on empty record
1153 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1154 sends an empty record. This could be exploited by a malicious peer in a
1155 Denial Of Service attack.
1157 This issue was reported to OpenSSL by Alex Gaynor.
1161 *) Excessive allocation of memory in tls_get_message_header() and
1162 dtls1_preprocess_fragment()
1164 A (D)TLS message includes 3 bytes for its length in the header for the
1165 message. This would allow for messages up to 16Mb in length. Messages of
1166 this length are excessive and OpenSSL includes a check to ensure that a
1167 peer is sending reasonably sized messages in order to avoid too much memory
1168 being consumed to service a connection. A flaw in the logic of version
1169 1.1.0 means that memory for the message is allocated too early, prior to
1170 the excessive message length check. Due to way memory is allocated in
1171 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1172 to service a connection. This could lead to a Denial of Service through
1173 memory exhaustion. However, the excessive message length check still takes
1174 place, and this would cause the connection to immediately fail. Assuming
1175 that the application calls SSL_free() on the failed connection in a timely
1176 manner then the 21Mb of allocated memory will then be immediately freed
1177 again. Therefore the excessive memory allocation will be transitory in
1178 nature. This then means that there is only a security impact if:
1180 1) The application does not call SSL_free() in a timely manner in the event
1181 that the connection fails
1183 2) The application is working in a constrained environment where there is
1184 very little free memory
1186 3) The attacker initiates multiple connection attempts such that there are
1187 multiple connections in a state where memory has been allocated for the
1188 connection; SSL_free() has not yet been called; and there is insufficient
1189 memory to service the multiple requests.
1191 Except in the instance of (1) above any Denial Of Service is likely to be
1192 transitory because as soon as the connection fails the memory is
1193 subsequently freed again in the SSL_free() call. However there is an
1194 increased risk during this period of application crashes due to the lack of
1195 memory - which would then mean a more serious Denial of Service.
1197 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1198 (CVE-2016-6307 and CVE-2016-6308)
1201 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1202 had to be removed. Primary reason is that vendor assembler can't
1203 assemble our modules with -KPIC flag. As result it, assembly
1204 support, was not even available as option. But its lack means
1205 lack of side-channel resistant code, which is incompatible with
1206 security by todays standards. Fortunately gcc is readily available
1207 prepackaged option, which we firmly point at...
1210 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1212 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1213 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1214 (to any value) allows Windows user to access PKCS#12 file generated
1215 with Windows CryptoAPI and protected with non-ASCII password, as well
1216 as files generated under UTF-8 locale on Linux also protected with
1220 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1221 have been disabled by default and removed from DEFAULT, just like RC4.
1222 See the RC4 item below to re-enable both.
1225 *) The method for finding the storage location for the Windows RAND seed file
1226 has changed. First we check %RANDFILE%. If that is not set then we check
1227 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1228 all else fails we fall back to C:\.
1231 *) The EVP_EncryptUpdate() function has had its return type changed from void
1232 to int. A return of 0 indicates and error while a return of 1 indicates
1236 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1237 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1238 off the constant time implementation for RSA, DSA and DH have been made
1239 no-ops and deprecated.
1242 *) Windows RAND implementation was simplified to only get entropy by
1243 calling CryptGenRandom(). Various other RAND-related tickets
1245 [Joseph Wylie Yandle, Rich Salz]
1247 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1248 and OPENSSL_LH_, respectively. The old names are available
1249 with API compatibility. They new names are now completely documented.
1252 *) Unify TYPE_up_ref(obj) methods signature.
1253 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1254 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1255 int (instead of void) like all others TYPE_up_ref() methods.
1256 So now these methods also check the return value of CRYPTO_atomic_add(),
1257 and the validity of object reference counter.
1258 [fdasilvayy@gmail.com]
1260 *) With Windows Visual Studio builds, the .pdb files are installed
1261 alongside the installed libraries and executables. For a static
1262 library installation, ossl_static.pdb is the associate compiler
1263 generated .pdb file to be used when linking programs.
1266 *) Remove openssl.spec. Packaging files belong with the packagers.
1269 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1270 recognise x86_64 architectures automatically. You can still decide
1271 to build for a different bitness with the environment variable
1272 KERNEL_BITS (can be 32 or 64), for example:
1274 KERNEL_BITS=32 ./config
1278 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1279 256 bit AES and HMAC with SHA256.
1282 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1285 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1288 *) To enable users to have their own config files and build file templates,
1289 Configure looks in the directory indicated by the environment variable
1290 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1291 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1292 name and is used as is.
1295 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1296 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1297 X509_CERT_FILE_CTX was removed.
1300 *) "shared" builds are now the default. To create only static libraries use
1301 the "no-shared" Configure option.
1304 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1305 All of these option have not worked for some while and are fundamental
1309 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1310 global cleanup functions are no longer required because they are handled
1311 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1312 Explicitly de-initing can cause problems (e.g. where a library that uses
1313 OpenSSL de-inits, but an application is still using it). The affected
1314 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1315 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1316 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1317 COMP_zlib_cleanup().
1320 *) --strict-warnings no longer enables runtime debugging options
1321 such as REF_DEBUG. Instead, debug options are automatically
1322 enabled with '--debug' builds.
1323 [Andy Polyakov, Emilia Käsper]
1325 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1326 have been moved out of the public header files. New functions for managing
1327 these have been added.
1330 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1331 objects have been moved out of the public header files. New
1332 functions for managing these have been added.
1335 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1336 have been moved out of the public header files. New functions for managing
1337 these have been added.
1340 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1341 moved out of the public header files. New functions for managing these
1345 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1348 *) Removed the mk1mf build scripts.
1351 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1352 it is always safe to #include a header now.
1355 *) Removed the aged BC-32 config and all its supporting scripts
1358 *) Removed support for Ultrix, Netware, and OS/2.
1361 *) Add support for HKDF.
1362 [Alessandro Ghedini]
1364 *) Add support for blake2b and blake2s
1367 *) Added support for "pipelining". Ciphers that have the
1368 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1369 encryptions/decryptions simultaneously. There are currently no built-in
1370 ciphers with this property but the expectation is that engines will be able
1371 to offer it to significantly improve throughput. Support has been extended
1372 into libssl so that multiple records for a single connection can be
1373 processed in one go (for >=TLS 1.1).
1376 *) Added the AFALG engine. This is an async capable engine which is able to
1377 offload work to the Linux kernel. In this initial version it only supports
1378 AES128-CBC. The kernel must be version 4.1.0 or greater.
1381 *) OpenSSL now uses a new threading API. It is no longer necessary to
1382 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1383 are two supported threading models: pthreads and windows threads. It is
1384 also possible to configure OpenSSL at compile time for "no-threads". The
1385 old threading API should no longer be used. The functions have been
1386 replaced with "no-op" compatibility macros.
1387 [Alessandro Ghedini, Matt Caswell]
1389 *) Modify behavior of ALPN to invoke callback after SNI/servername
1390 callback, such that updates to the SSL_CTX affect ALPN.
1393 *) Add SSL_CIPHER queries for authentication and key-exchange.
1396 *) Changes to the DEFAULT cipherlist:
1397 - Prefer (EC)DHE handshakes over plain RSA.
1398 - Prefer AEAD ciphers over legacy ciphers.
1399 - Prefer ECDSA over RSA when both certificates are available.
1400 - Prefer TLSv1.2 ciphers/PRF.
1401 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1405 *) Change the ECC default curve list to be this, in order: x25519,
1406 secp256r1, secp521r1, secp384r1.
1409 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1410 disabled by default. They can be re-enabled using the
1411 enable-weak-ssl-ciphers option to Configure.
1414 *) If the server has ALPN configured, but supports no protocols that the
1415 client advertises, send a fatal "no_application_protocol" alert.
1416 This behaviour is SHALL in RFC 7301, though it isn't universally
1417 implemented by other servers.
1420 *) Add X25519 support.
1421 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1422 for public and private key encoding using the format documented in
1423 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1424 key generation and key derivation.
1426 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1430 *) Deprecate SRP_VBASE_get_by_user.
1431 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1432 In order to fix an unavoidable memory leak (CVE-2016-0798),
1433 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1434 seed, even if the seed is configured.
1436 Users should use SRP_VBASE_get1_by_user instead. Note that in
1437 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1438 also that even though configuring the SRP seed attempts to hide
1439 invalid usernames by continuing the handshake with fake
1440 credentials, this behaviour is not constant time and no strong
1441 guarantees are made that the handshake is indistinguishable from
1442 that of a valid user.
1445 *) Configuration change; it's now possible to build dynamic engines
1446 without having to build shared libraries and vice versa. This
1447 only applies to the engines in engines/, those in crypto/engine/
1448 will always be built into libcrypto (i.e. "static").
1450 Building dynamic engines is enabled by default; to disable, use
1451 the configuration option "disable-dynamic-engine".
1453 The only requirements for building dynamic engines are the
1454 presence of the DSO module and building with position independent
1455 code, so they will also automatically be disabled if configuring
1456 with "disable-dso" or "disable-pic".
1458 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1459 are also taken away from openssl/opensslconf.h, as they are
1463 *) Configuration change; if there is a known flag to compile
1464 position independent code, it will always be applied on the
1465 libcrypto and libssl object files, and never on the application
1466 object files. This means other libraries that use routines from
1467 libcrypto / libssl can be made into shared libraries regardless
1468 of how OpenSSL was configured.
1470 If this isn't desirable, the configuration options "disable-pic"
1471 or "no-pic" can be used to disable the use of PIC. This will
1472 also disable building shared libraries and dynamic engines.
1475 *) Removed JPAKE code. It was experimental and has no wide use.
1478 *) The INSTALL_PREFIX Makefile variable has been renamed to
1479 DESTDIR. That makes for less confusion on what this variable
1480 is for. Also, the configuration option --install_prefix is
1484 *) Heartbeat for TLS has been removed and is disabled by default
1485 for DTLS; configure with enable-heartbeats. Code that uses the
1486 old #define's might need to be updated.
1487 [Emilia Käsper, Rich Salz]
1489 *) Rename REF_CHECK to REF_DEBUG.
1492 *) New "unified" build system
1494 The "unified" build system is aimed to be a common system for all
1495 platforms we support. With it comes new support for VMS.
1497 This system builds supports building in a different directory tree
1498 than the source tree. It produces one Makefile (for unix family
1499 or lookalikes), or one descrip.mms (for VMS).
1501 The source of information to make the Makefile / descrip.mms is
1502 small files called 'build.info', holding the necessary
1503 information for each directory with source to compile, and a
1504 template in Configurations, like unix-Makefile.tmpl or
1507 With this change, the library names were also renamed on Windows
1508 and on VMS. They now have names that are closer to the standard
1509 on Unix, and include the major version number, and in certain
1510 cases, the architecture they are built for. See "Notes on shared
1511 libraries" in INSTALL.
1513 We rely heavily on the perl module Text::Template.
1516 *) Added support for auto-initialisation and de-initialisation of the library.
1517 OpenSSL no longer requires explicit init or deinit routines to be called,
1518 except in certain circumstances. See the OPENSSL_init_crypto() and
1519 OPENSSL_init_ssl() man pages for further information.
1522 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1523 "peer" argument is now expected to be a BIO_ADDR object.
1525 *) Rewrite of BIO networking library. The BIO library lacked consistent
1526 support of IPv6, and adding it required some more extensive
1527 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1528 which hold all types of addresses and chains of address information.
1529 It also introduces a new API, with functions like BIO_socket,
1530 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1531 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1532 have been adapted accordingly.
1535 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1539 *) CRIME protection: disable compression by default, even if OpenSSL is
1540 compiled with zlib enabled. Applications can still enable compression
1541 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1542 using the SSL_CONF library to configure compression.
1545 *) The signature of the session callback configured with
1546 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1547 was explicitly marked as 'const unsigned char*' instead of
1551 *) Always DPURIFY. Remove the use of uninitialized memory in the
1552 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1555 *) Removed many obsolete configuration items, including
1556 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1557 MD2_CHAR, MD2_INT, MD2_LONG
1559 IDEA_SHORT, IDEA_LONG
1560 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1561 [Rich Salz, with advice from Andy Polyakov]
1563 *) Many BN internals have been moved to an internal header file.
1564 [Rich Salz with help from Andy Polyakov]
1566 *) Configuration and writing out the results from it has changed.
1567 Files such as Makefile include/openssl/opensslconf.h and are now
1568 produced through general templates, such as Makefile.in and
1569 crypto/opensslconf.h.in and some help from the perl module
1572 Also, the center of configuration information is no longer
1573 Makefile. Instead, Configure produces a perl module in
1574 configdata.pm which holds most of the config data (in the hash
1575 table %config), the target data that comes from the target
1576 configuration in one of the Configurations/*.conf files (in
1580 *) To clarify their intended purposes, the Configure options
1581 --prefix and --openssldir change their semantics, and become more
1582 straightforward and less interdependent.
1584 --prefix shall be used exclusively to give the location INSTALLTOP
1585 where programs, scripts, libraries, include files and manuals are
1586 going to be installed. The default is now /usr/local.
1588 --openssldir shall be used exclusively to give the default
1589 location OPENSSLDIR where certificates, private keys, CRLs are
1590 managed. This is also where the default openssl.cnf gets
1592 If the directory given with this option is a relative path, the
1593 values of both the --prefix value and the --openssldir value will
1594 be combined to become OPENSSLDIR.
1595 The default for --openssldir is INSTALLTOP/ssl.
1597 Anyone who uses --openssldir to specify where OpenSSL is to be
1598 installed MUST change to use --prefix instead.
1601 *) The GOST engine was out of date and therefore it has been removed. An up
1602 to date GOST engine is now being maintained in an external repository.
1603 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1604 support for GOST ciphersuites (these are only activated if a GOST engine
1608 *) EGD is no longer supported by default; use enable-egd when
1610 [Ben Kaduk and Rich Salz]
1612 *) The distribution now has Makefile.in files, which are used to
1613 create Makefile's when Configure is run. *Configure must be run
1614 before trying to build now.*
1617 *) The return value for SSL_CIPHER_description() for error conditions
1621 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1623 Obtaining and performing DNSSEC validation of TLSA records is
1624 the application's responsibility. The application provides
1625 the TLSA records of its choice to OpenSSL, and these are then
1626 used to authenticate the peer.
1628 The TLSA records need not even come from DNS. They can, for
1629 example, be used to implement local end-entity certificate or
1630 trust-anchor "pinning", where the "pin" data takes the form
1631 of TLSA records, which can augment or replace verification
1632 based on the usual WebPKI public certification authorities.
1635 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1636 continues to support deprecated interfaces in default builds.
1637 However, applications are strongly advised to compile their
1638 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1639 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1640 or the 1.1.0 releases.
1642 In environments in which all applications have been ported to
1643 not use any deprecated interfaces OpenSSL's Configure script
1644 should be used with the --api=1.1.0 option to entirely remove
1645 support for the deprecated features from the library and
1646 unconditionally disable them in the installed headers.
1647 Essentially the same effect can be achieved with the "no-deprecated"
1648 argument to Configure, except that this will always restrict
1649 the build to just the latest API, rather than a fixed API
1652 As applications are ported to future revisions of the API,
1653 they should update their compile-time OPENSSL_API_COMPAT define
1654 accordingly, but in most cases should be able to continue to
1655 compile with later releases.
1657 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1658 0x10000000L and 0x00908000L, respectively. However those
1659 versions did not support the OPENSSL_API_COMPAT feature, and
1660 so applications are not typically tested for explicit support
1661 of just the undeprecated features of either release.
1664 *) Add support for setting the minimum and maximum supported protocol.
1665 It can bet set via the SSL_set_min_proto_version() and
1666 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1667 MaxProtocol. It's recommended to use the new APIs to disable
1668 protocols instead of disabling individual protocols using
1669 SSL_set_options() or SSL_CONF's Protocol. This change also
1670 removes support for disabling TLS 1.2 in the OpenSSL TLS
1671 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1674 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1677 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1678 and integrates ECDSA and ECDH functionality into EC. Implementations can
1679 now redirect key generation and no longer need to convert to or from
1682 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1683 include the ec.h header file instead.
1686 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1687 ciphers who are no longer supported and drops support the ephemeral RSA key
1688 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1691 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1692 opaque. For HMAC_CTX, the following constructors and destructors
1695 HMAC_CTX *HMAC_CTX_new(void);
1696 void HMAC_CTX_free(HMAC_CTX *ctx);
1698 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1699 destroy such methods has been added. See EVP_MD_meth_new(3) and
1700 EVP_CIPHER_meth_new(3) for documentation.
1703 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1704 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1705 EVP_MD_CTX_reset() should be called instead to reinitialise
1706 an already created structure.
1707 2) For consistency with the majority of our object creators and
1708 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1709 EVP_MD_CTX_(new|free). The old names are retained as macros
1710 for deprecated builds.
1713 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1714 cryptographic operations to be performed asynchronously as long as an
1715 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1716 further details. Libssl has also had this capability integrated with the
1717 introduction of the new mode SSL_MODE_ASYNC and associated error
1718 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1719 pages. This work was developed in partnership with Intel Corp.
1722 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1723 always enabled now. If you want to disable the support you should
1724 exclude it using the list of supported ciphers. This also means that the
1725 "-no_ecdhe" option has been removed from s_server.
1728 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1729 SSL_{CTX_}set1_curves() which can set a list.
1732 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1733 curve you want to support using SSL_{CTX_}set1_curves().
1736 *) State machine rewrite. The state machine code has been significantly
1737 refactored in order to remove much duplication of code and solve issues
1738 with the old code (see ssl/statem/README for further details). This change
1739 does have some associated API changes. Notably the SSL_state() function
1740 has been removed and replaced by SSL_get_state which now returns an
1741 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1742 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1746 *) All instances of the string "ssleay" in the public API were replaced
1747 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1748 Some error codes related to internal RSA_eay API's were renamed.
1751 *) The demo files in crypto/threads were moved to demo/threads.
1754 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1756 [Matt Caswell, Rich Salz]
1758 *) New ASN.1 embed macro.
1760 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1761 structure is not allocated: it is part of the parent. That is instead of
1769 This reduces memory fragmentation and make it impossible to accidentally
1770 set a mandatory field to NULL.
1772 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1773 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1774 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1778 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1781 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1782 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1783 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1784 DES and RC4 ciphersuites.
1787 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1788 This changes the decoding behaviour for some invalid messages,
1789 though the change is mostly in the more lenient direction, and
1790 legacy behaviour is preserved as much as possible.
1793 *) Fix no-stdio build.
1794 [ David Woodhouse <David.Woodhouse@intel.com> and also
1795 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1797 *) New testing framework
1798 The testing framework has been largely rewritten and is now using
1799 perl and the perl modules Test::Harness and an extended variant of
1800 Test::More called OpenSSL::Test to do its work. All test scripts in
1801 test/ have been rewritten into test recipes, and all direct calls to
1802 executables in test/Makefile have become individual recipes using the
1803 simplified testing OpenSSL::Test::Simple.
1805 For documentation on our testing modules, do:
1807 perldoc test/testlib/OpenSSL/Test/Simple.pm
1808 perldoc test/testlib/OpenSSL/Test.pm
1812 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1813 are used; the latter aborts on memory leaks (usually checked on exit).
1814 Some undocumented "set malloc, etc., hooks" functions were removed
1815 and others were changed. All are now documented.
1818 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1820 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1822 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1823 from RFC4279, RFC4785, RFC5487, RFC5489.
1825 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1826 original RSA_PSK patch.
1829 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1830 era flag was never set throughout the codebase (only read). Also removed
1831 SSL3_FLAGS_POP_BUFFER which was only used if
1832 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1835 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1836 to be "oneline" instead of "compat".
1839 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1840 not aware of clients that still exhibit this bug, and the workaround
1841 hasn't been working properly for a while.
1844 *) The return type of BIO_number_read() and BIO_number_written() as well as
1845 the corresponding num_read and num_write members in the BIO structure has
1846 changed from unsigned long to uint64_t. On platforms where an unsigned
1847 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1851 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1852 OpenSSL without support for them. It also means that maintaining
1853 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1854 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1857 *) Removed support for the two export grade static DH ciphersuites
1858 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1859 were newly added (along with a number of other static DH ciphersuites) to
1860 1.0.2. However the two export ones have *never* worked since they were
1861 introduced. It seems strange in any case to be adding new export
1862 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1865 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1866 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1867 and turned into macros which simply call the new preferred function names
1868 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1869 should use the new names instead. Also as part of this change the ssl23.h
1870 header file has been removed.
1873 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1874 code and the associated standard is no longer considered fit-for-purpose.
1877 *) RT2547 was closed. When generating a private key, try to make the
1878 output file readable only by the owner. This behavior change might
1879 be noticeable when interacting with other software.
1881 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1885 *) Added HTTP GET support to the ocsp command.
1888 *) Changed default digest for the dgst and enc commands from MD5 to
1892 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1895 *) Added support for TLS extended master secret from
1896 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1897 initial patch which was a great help during development.
1900 *) All libssl internal structures have been removed from the public header
1901 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1902 now redundant). Users should not attempt to access internal structures
1903 directly. Instead they should use the provided API functions.
1906 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1907 Access to deprecated functions can be re-enabled by running config with
1908 "enable-deprecated". In addition applications wishing to use deprecated
1909 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1910 will, by default, disable some transitive includes that previously existed
1911 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1914 *) Added support for OCB mode. OpenSSL has been granted a patent license
1915 compatible with the OpenSSL license for use of OCB. Details are available
1916 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1917 for OCB can be removed by calling config with no-ocb.
1920 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1921 compatible client hello.
1924 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1925 done while fixing the error code for the key-too-small case.
1926 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1928 *) CA.sh has been removed; use CA.pl instead.
1931 *) Removed old DES API.
1934 *) Remove various unsupported platforms:
1940 Sinix/ReliantUNIX RM400
1945 16-bit platforms such as WIN16
1948 *) Clean up OPENSSL_NO_xxx #define's
1949 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1950 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1951 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1952 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1953 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1954 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1955 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1956 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1957 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1958 Remove MS_STATIC; it's a relic from platforms <32 bits.
1961 *) Cleaned up dead code
1962 Remove all but one '#ifdef undef' which is to be looked at.
1965 *) Clean up calling of xxx_free routines.
1966 Just like free(), fix most of the xxx_free routines to accept
1967 NULL. Remove the non-null checks from callers. Save much code.
1970 *) Add secure heap for storage of private keys (when possible).
1971 Add BIO_s_secmem(), CBIGNUM, etc.
1972 Contributed by Akamai Technologies under our Corporate CLA.
1975 *) Experimental support for a new, fast, unbiased prime candidate generator,
1976 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1977 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1979 *) New output format NSS in the sess_id command line tool. This allows
1980 exporting the session id and the master key in NSS keylog format.
1981 [Martin Kaiser <martin@kaiser.cx>]
1983 *) Harmonize version and its documentation. -f flag is used to display
1985 [mancha <mancha1@zoho.com>]
1987 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1988 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1989 [mancha <mancha1@zoho.com>]
1991 *) Fix some double frees. These are not thought to be exploitable.
1992 [mancha <mancha1@zoho.com>]
1994 *) A missing bounds check in the handling of the TLS heartbeat extension
1995 can be used to reveal up to 64k of memory to a connected client or
1998 Thanks for Neel Mehta of Google Security for discovering this bug and to
1999 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2000 preparing the fix (CVE-2014-0160)
2001 [Adam Langley, Bodo Moeller]
2003 *) Fix for the attack described in the paper "Recovering OpenSSL
2004 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2005 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2006 http://eprint.iacr.org/2014/140
2008 Thanks to Yuval Yarom and Naomi Benger for discovering this
2009 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2010 [Yuval Yarom and Naomi Benger]
2012 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2013 this fixes a limitation in previous versions of OpenSSL.
2016 *) Experimental encrypt-then-mac support.
2018 Experimental support for encrypt then mac from
2019 draft-gutmann-tls-encrypt-then-mac-02.txt
2021 To enable it set the appropriate extension number (0x42 for the test
2022 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2024 For non-compliant peers (i.e. just about everything) this should have no
2027 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2031 *) Add EVP support for key wrapping algorithms, to avoid problems with
2032 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2033 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2034 algorithms and include tests cases.
2037 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2041 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2042 MGF1 digest and OAEP label.
2045 *) Make openssl verify return errors.
2046 [Chris Palmer <palmer@google.com> and Ben Laurie]
2048 *) New function ASN1_TIME_diff to calculate the difference between two
2049 ASN1_TIME structures or one structure and the current time.
2052 *) Update fips_test_suite to support multiple command line options. New
2053 test to induce all self test errors in sequence and check expected
2057 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2058 sign or verify all in one operation.
2061 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2062 test programs and fips_test_suite. Includes functionality to parse
2063 the minimal script output of fipsalgest.pl directly.
2066 *) Add authorisation parameter to FIPS_module_mode_set().
2069 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2072 *) Use separate DRBG fields for internal and external flags. New function
2073 FIPS_drbg_health_check() to perform on demand health checking. Add
2074 generation tests to fips_test_suite with reduced health check interval to
2075 demonstrate periodic health checking. Add "nodh" option to
2076 fips_test_suite to skip very slow DH test.
2079 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2083 *) More extensive health check for DRBG checking many more failure modes.
2084 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2085 combination: call this in fips_test_suite.
2088 *) Add support for canonical generation of DSA parameter 'g'. See
2091 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2092 POST to handle HMAC cases.
2095 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2096 to return numerical and string versions of the FIPS module number.
2099 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2100 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2101 outside the validated module in the FIPS capable OpenSSL.
2104 *) Minor change to DRBG entropy callback semantics. In some cases
2105 there is no multiple of the block length between min_len and
2106 max_len. Allow the callback to return more than max_len bytes
2107 of entropy but discard any extra: it is the callback's responsibility
2108 to ensure that the extra data discarded does not impact the
2109 requested amount of entropy.
2112 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2113 information in FIPS186-3, SP800-57 and SP800-131A.
2116 *) CCM support via EVP. Interface is very similar to GCM case except we
2117 must supply all data in one chunk (i.e. no update, final) and the
2118 message length must be supplied if AAD is used. Add algorithm test
2122 *) Initial version of POST overhaul. Add POST callback to allow the status
2123 of POST to be monitored and/or failures induced. Modify fips_test_suite
2124 to use callback. Always run all selftests even if one fails.
2127 *) XTS support including algorithm test driver in the fips_gcmtest program.
2128 Note: this does increase the maximum key length from 32 to 64 bytes but
2129 there should be no binary compatibility issues as existing applications
2130 will never use XTS mode.
2133 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2134 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2135 performs algorithm blocking for unapproved PRNG types. Also do not
2136 set PRNG type in FIPS_mode_set(): leave this to the application.
2137 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2138 the standard OpenSSL PRNG: set additional data to a date time vector.
2141 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2142 This shouldn't present any incompatibility problems because applications
2143 shouldn't be using these directly and any that are will need to rethink
2144 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2147 *) Extensive self tests and health checking required by SP800-90 DRBG.
2148 Remove strength parameter from FIPS_drbg_instantiate and always
2149 instantiate at maximum supported strength.
2152 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2155 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2158 *) New function DH_compute_key_padded() to compute a DH key and pad with
2159 leading zeroes if needed: this complies with SP800-56A et al.
2162 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2163 anything, incomplete, subject to change and largely untested at present.
2166 *) Modify fipscanisteronly build option to only build the necessary object
2167 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2170 *) Add experimental option FIPSSYMS to give all symbols in
2171 fipscanister.o and FIPS or fips prefix. This will avoid
2172 conflicts with future versions of OpenSSL. Add perl script
2173 util/fipsas.pl to preprocess assembly language source files
2174 and rename any affected symbols.
2177 *) Add selftest checks and algorithm block of non-fips algorithms in
2178 FIPS mode. Remove DES2 from selftests.
2181 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2182 return internal method without any ENGINE dependencies. Add new
2183 tiny fips sign and verify functions.
2186 *) New build option no-ec2m to disable characteristic 2 code.
2189 *) New build option "fipscanisteronly". This only builds fipscanister.o
2190 and (currently) associated fips utilities. Uses the file Makefile.fips
2191 instead of Makefile.org as the prototype.
2194 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2195 Update fips_gcmtest to use IV generator.
2198 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2199 setting output buffer to NULL. The *Final function must be
2200 called although it will not retrieve any additional data. The tag
2201 can be set or retrieved with a ctrl. The IV length is by default 12
2202 bytes (96 bits) but can be set to an alternative value. If the IV
2203 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2207 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2208 underlying do_cipher function handles all cipher semantics itself
2209 including padding and finalisation. This is useful if (for example)
2210 an ENGINE cipher handles block padding itself. The behaviour of
2211 do_cipher is subtly changed if this flag is set: the return value
2212 is the number of characters written to the output buffer (zero is
2213 no longer an error code) or a negative error code. Also if the
2214 input buffer is NULL and length 0 finalisation should be performed.
2217 *) If a candidate issuer certificate is already part of the constructed
2218 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2221 *) Improve forward-security support: add functions
2223 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2224 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2226 for use by SSL/TLS servers; the callback function will be called whenever a
2227 new session is created, and gets to decide whether the session may be
2228 cached to make it resumable (return 0) or not (return 1). (As by the
2229 SSL/TLS protocol specifications, the session_id sent by the server will be
2230 empty to indicate that the session is not resumable; also, the server will
2231 not generate RFC 4507 (RFC 5077) session tickets.)
2233 A simple reasonable callback implementation is to return is_forward_secure.
2234 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2235 by the SSL/TLS server library, indicating whether it can provide forward
2237 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2239 *) New -verify_name option in command line utilities to set verification
2243 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2244 Add CMAC pkey methods.
2247 *) Experimental renegotiation in s_server -www mode. If the client
2248 browses /reneg connection is renegotiated. If /renegcert it is
2249 renegotiated requesting a certificate.
2252 *) Add an "external" session cache for debugging purposes to s_server. This
2253 should help trace issues which normally are only apparent in deployed
2254 multi-process servers.
2257 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2258 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2259 BIO_set_cipher() and some obscure PEM functions were changed so they
2260 can now return an error. The RAND changes required a change to the
2261 RAND_METHOD structure.
2264 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2265 a gcc attribute to warn if the result of a function is ignored. This
2266 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2267 whose return value is often ignored.
2270 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2271 These allow SCTs (signed certificate timestamps) to be requested and
2272 validated when establishing a connection.
2273 [Rob Percival <robpercival@google.com>]
2275 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2277 *) Prevent padding oracle in AES-NI CBC MAC check
2279 A MITM attacker can use a padding oracle attack to decrypt traffic
2280 when the connection uses an AES CBC cipher and the server support
2283 This issue was introduced as part of the fix for Lucky 13 padding
2284 attack (CVE-2013-0169). The padding check was rewritten to be in
2285 constant time by making sure that always the same bytes are read and
2286 compared against either the MAC or padding bytes. But it no longer
2287 checked that there was enough data to have both the MAC and padding
2290 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2294 *) Fix EVP_EncodeUpdate overflow
2296 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2297 Base64 encoding of binary data. If an attacker is able to supply very large
2298 amounts of input data then a length check can overflow resulting in a heap
2301 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2302 the PEM_write_bio* family of functions. These are mainly used within the
2303 OpenSSL command line applications, so any application which processes data
2304 from an untrusted source and outputs it as a PEM file should be considered
2305 vulnerable to this issue. User applications that call these APIs directly
2306 with large amounts of untrusted data may also be vulnerable.
2308 This issue was reported by Guido Vranken.
2312 *) Fix EVP_EncryptUpdate overflow
2314 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2315 is able to supply very large amounts of input data after a previous call to
2316 EVP_EncryptUpdate() with a partial block then a length check can overflow
2317 resulting in a heap corruption. Following an analysis of all OpenSSL
2318 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2319 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2320 the first called function after an EVP_EncryptInit(), and therefore that
2321 specific call must be safe. The second form is where the length passed to
2322 EVP_EncryptUpdate() can be seen from the code to be some small value and
2323 therefore there is no possibility of an overflow. Since all instances are
2324 one of these two forms, it is believed that there can be no overflows in
2325 internal code due to this problem. It should be noted that
2326 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2327 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2328 of these calls have also been analysed too and it is believed there are no
2329 instances in internal usage where an overflow could occur.
2331 This issue was reported by Guido Vranken.
2335 *) Prevent ASN.1 BIO excessive memory allocation
2337 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2338 a short invalid encoding can cause allocation of large amounts of memory
2339 potentially consuming excessive resources or exhausting memory.
2341 Any application parsing untrusted data through d2i BIO functions is
2342 affected. The memory based functions such as d2i_X509() are *not* affected.
2343 Since the memory based functions are used by the TLS library, TLS
2344 applications are not affected.
2346 This issue was reported by Brian Carpenter.
2352 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2353 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2354 in arbitrary stack data being returned in the buffer.
2356 This issue was reported by Guido Vranken.
2360 *) Modify behavior of ALPN to invoke callback after SNI/servername
2361 callback, such that updates to the SSL_CTX affect ALPN.
2364 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2368 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2369 methods are enabled and ssl2 is disabled the methods return NULL.
2372 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2374 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2375 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2376 provide any "EXPORT" or "LOW" strength ciphers.
2379 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2380 is by default disabled at build-time. Builds that are not configured with
2381 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2382 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2383 will need to explicitly call either of:
2385 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2387 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2389 as appropriate. Even if either of those is used, or the application
2390 explicitly uses the version-specific SSLv2_method() or its client and
2391 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2392 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2393 ciphers, and SSLv2 56-bit DES are no longer available.
2397 *) Fix a double-free in DSA code
2399 A double free bug was discovered when OpenSSL parses malformed DSA private
2400 keys and could lead to a DoS attack or memory corruption for applications
2401 that receive DSA private keys from untrusted sources. This scenario is
2404 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2409 *) Disable SRP fake user seed to address a server memory leak.
2411 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2413 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2414 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2415 was changed to ignore the "fake user" SRP seed, even if the seed
2418 Users should use SRP_VBASE_get1_by_user instead. Note that in
2419 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2420 also that even though configuring the SRP seed attempts to hide
2421 invalid usernames by continuing the handshake with fake
2422 credentials, this behaviour is not constant time and no strong
2423 guarantees are made that the handshake is indistinguishable from
2424 that of a valid user.
2428 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2430 In the BN_hex2bn function the number of hex digits is calculated using an
2431 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2432 large values of |i| this can result in |bn_expand| not allocating any
2433 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2434 field as NULL leading to a subsequent NULL ptr deref. For very large values
2435 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2436 In this case memory is allocated to the internal BIGNUM data field, but it
2437 is insufficiently sized leading to heap corruption. A similar issue exists
2438 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2439 is ever called by user applications with very large untrusted hex/dec data.
2440 This is anticipated to be a rare occurrence.
2442 All OpenSSL internal usage of these functions use data that is not expected
2443 to be untrusted, e.g. config file data or application command line
2444 arguments. If user developed applications generate config file data based
2445 on untrusted data then it is possible that this could also lead to security
2446 consequences. This is also anticipated to be rare.
2448 This issue was reported to OpenSSL by Guido Vranken.
2452 *) Fix memory issues in BIO_*printf functions
2454 The internal |fmtstr| function used in processing a "%s" format string in
2455 the BIO_*printf functions could overflow while calculating the length of a
2456 string and cause an OOB read when printing very long strings.
2458 Additionally the internal |doapr_outch| function can attempt to write to an
2459 OOB memory location (at an offset from the NULL pointer) in the event of a
2460 memory allocation failure. In 1.0.2 and below this could be caused where
2461 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2462 could be in processing a very long "%s" format string. Memory leaks can
2465 The first issue may mask the second issue dependent on compiler behaviour.
2466 These problems could enable attacks where large amounts of untrusted data
2467 is passed to the BIO_*printf functions. If applications use these functions
2468 in this way then they could be vulnerable. OpenSSL itself uses these
2469 functions when printing out human-readable dumps of ASN.1 data. Therefore
2470 applications that print this data could be vulnerable if the data is from
2471 untrusted sources. OpenSSL command line applications could also be
2472 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2473 as command line arguments.
2475 Libssl is not considered directly vulnerable. Additionally certificates etc
2476 received via remote connections via libssl are also unlikely to be able to
2477 trigger these issues because of message size limits enforced within libssl.
2479 This issue was reported to OpenSSL Guido Vranken.
2483 *) Side channel attack on modular exponentiation
2485 A side-channel attack was found which makes use of cache-bank conflicts on
2486 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2487 of RSA keys. The ability to exploit this issue is limited as it relies on
2488 an attacker who has control of code in a thread running on the same
2489 hyper-threaded core as the victim thread which is performing decryptions.
2491 This issue was reported to OpenSSL by Yuval Yarom, The University of
2492 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2493 Nadia Heninger, University of Pennsylvania with more information at
2494 http://cachebleed.info.
2498 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2499 if no keysize is specified with default_bits. This fixes an
2500 omission in an earlier change that changed all RSA/DSA key generation
2501 apps to use 2048 bits by default.
2504 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2505 *) DH small subgroups
2507 Historically OpenSSL only ever generated DH parameters based on "safe"
2508 primes. More recently (in version 1.0.2) support was provided for
2509 generating X9.42 style parameter files such as those required for RFC 5114
2510 support. The primes used in such files may not be "safe". Where an
2511 application is using DH configured with parameters based on primes that are
2512 not "safe" then an attacker could use this fact to find a peer's private
2513 DH exponent. This attack requires that the attacker complete multiple
2514 handshakes in which the peer uses the same private DH exponent. For example
2515 this could be used to discover a TLS server's private DH exponent if it's
2516 reusing the private DH exponent or it's using a static DH ciphersuite.
2518 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2519 TLS. It is not on by default. If the option is not set then the server
2520 reuses the same private DH exponent for the life of the server process and
2521 would be vulnerable to this attack. It is believed that many popular
2522 applications do set this option and would therefore not be at risk.
2524 The fix for this issue adds an additional check where a "q" parameter is
2525 available (as is the case in X9.42 based parameters). This detects the
2526 only known attack, and is the only possible defense for static DH
2527 ciphersuites. This could have some performance impact.
2529 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2530 default and cannot be disabled. This could have some performance impact.
2532 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2536 *) SSLv2 doesn't block disabled ciphers
2538 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2539 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2540 been disabled, provided that the SSLv2 protocol was not also disabled via
2543 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2544 and Sebastian Schinzel.
2548 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2550 *) BN_mod_exp may produce incorrect results on x86_64
2552 There is a carry propagating bug in the x86_64 Montgomery squaring
2553 procedure. No EC algorithms are affected. Analysis suggests that attacks
2554 against RSA and DSA as a result of this defect would be very difficult to
2555 perform and are not believed likely. Attacks against DH are considered just
2556 feasible (although very difficult) because most of the work necessary to
2557 deduce information about a private key may be performed offline. The amount
2558 of resources required for such an attack would be very significant and
2559 likely only accessible to a limited number of attackers. An attacker would
2560 additionally need online access to an unpatched system using the target
2561 private key in a scenario with persistent DH parameters and a private
2562 key that is shared between multiple clients. For example this can occur by
2563 default in OpenSSL DHE based SSL/TLS ciphersuites.
2565 This issue was reported to OpenSSL by Hanno Böck.
2569 *) Certificate verify crash with missing PSS parameter
2571 The signature verification routines will crash with a NULL pointer
2572 dereference if presented with an ASN.1 signature using the RSA PSS
2573 algorithm and absent mask generation function parameter. Since these
2574 routines are used to verify certificate signature algorithms this can be
2575 used to crash any certificate verification operation and exploited in a
2576 DoS attack. Any application which performs certificate verification is
2577 vulnerable including OpenSSL clients and servers which enable client
2580 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2584 *) X509_ATTRIBUTE memory leak
2586 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2587 memory. This structure is used by the PKCS#7 and CMS routines so any
2588 application which reads PKCS#7 or CMS data from untrusted sources is
2589 affected. SSL/TLS is not affected.
2591 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2596 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2597 This changes the decoding behaviour for some invalid messages,
2598 though the change is mostly in the more lenient direction, and
2599 legacy behaviour is preserved as much as possible.
2602 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2604 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2606 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2608 *) Alternate chains certificate forgery
2610 During certificate verification, OpenSSL will attempt to find an
2611 alternative certificate chain if the first attempt to build such a chain
2612 fails. An error in the implementation of this logic can mean that an
2613 attacker could cause certain checks on untrusted certificates to be
2614 bypassed, such as the CA flag, enabling them to use a valid leaf
2615 certificate to act as a CA and "issue" an invalid certificate.
2617 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2621 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2623 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2624 incompatibility in the handling of HMAC. The previous ABI has now been
2628 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2630 *) Malformed ECParameters causes infinite loop
2632 When processing an ECParameters structure OpenSSL enters an infinite loop
2633 if the curve specified is over a specially malformed binary polynomial
2636 This can be used to perform denial of service against any
2637 system which processes public keys, certificate requests or
2638 certificates. This includes TLS clients and TLS servers with
2639 client authentication enabled.
2641 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2645 *) Exploitable out-of-bounds read in X509_cmp_time
2647 X509_cmp_time does not properly check the length of the ASN1_TIME
2648 string and can read a few bytes out of bounds. In addition,
2649 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2652 An attacker can use this to craft malformed certificates and CRLs of
2653 various sizes and potentially cause a segmentation fault, resulting in
2654 a DoS on applications that verify certificates or CRLs. TLS clients
2655 that verify CRLs are affected. TLS clients and servers with client
2656 authentication enabled may be affected if they use custom verification
2659 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2660 independently by Hanno Böck.
2664 *) PKCS7 crash with missing EnvelopedContent
2666 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2667 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2668 with missing content and trigger a NULL pointer dereference on parsing.
2670 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2671 structures from untrusted sources are affected. OpenSSL clients and
2672 servers are not affected.
2674 This issue was reported to OpenSSL by Michal Zalewski (Google).
2678 *) CMS verify infinite loop with unknown hash function
2680 When verifying a signedData message the CMS code can enter an infinite loop
2681 if presented with an unknown hash function OID. This can be used to perform
2682 denial of service against any system which verifies signedData messages using
2684 This issue was reported to OpenSSL by Johannes Bauer.
2688 *) Race condition handling NewSessionTicket
2690 If a NewSessionTicket is received by a multi-threaded client when attempting to
2691 reuse a previous ticket then a race condition can occur potentially leading to
2692 a double free of the ticket data.
2696 *) Only support 256-bit or stronger elliptic curves with the
2697 'ecdh_auto' setting (server) or by default (client). Of supported
2698 curves, prefer P-256 (both).
2701 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2703 *) ClientHello sigalgs DoS fix
2705 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2706 invalid signature algorithms extension a NULL pointer dereference will
2707 occur. This can be exploited in a DoS attack against the server.
2709 This issue was was reported to OpenSSL by David Ramos of Stanford
2712 [Stephen Henson and Matt Caswell]
2714 *) Multiblock corrupted pointer fix
2716 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2717 feature only applies on 64 bit x86 architecture platforms that support AES
2718 NI instructions. A defect in the implementation of "multiblock" can cause
2719 OpenSSL's internal write buffer to become incorrectly set to NULL when
2720 using non-blocking IO. Typically, when the user application is using a
2721 socket BIO for writing, this will only result in a failed connection.
2722 However if some other BIO is used then it is likely that a segmentation
2723 fault will be triggered, thus enabling a potential DoS attack.
2725 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2729 *) Segmentation fault in DTLSv1_listen fix
2731 The DTLSv1_listen function is intended to be stateless and processes the
2732 initial ClientHello from many peers. It is common for user code to loop
2733 over the call to DTLSv1_listen until a valid ClientHello is received with
2734 an associated cookie. A defect in the implementation of DTLSv1_listen means
2735 that state is preserved in the SSL object from one invocation to the next
2736 that can lead to a segmentation fault. Errors processing the initial
2737 ClientHello can trigger this scenario. An example of such an error could be
2738 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2741 This issue was reported to OpenSSL by Per Allansson.
2745 *) Segmentation fault in ASN1_TYPE_cmp fix
2747 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2748 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2749 certificate signature algorithm consistency this can be used to crash any
2750 certificate verification operation and exploited in a DoS attack. Any
2751 application which performs certificate verification is vulnerable including
2752 OpenSSL clients and servers which enable client authentication.
2756 *) Segmentation fault for invalid PSS parameters fix
2758 The signature verification routines will crash with a NULL pointer
2759 dereference if presented with an ASN.1 signature using the RSA PSS
2760 algorithm and invalid parameters. Since these routines are used to verify
2761 certificate signature algorithms this can be used to crash any
2762 certificate verification operation and exploited in a DoS attack. Any
2763 application which performs certificate verification is vulnerable including
2764 OpenSSL clients and servers which enable client authentication.
2766 This issue was was reported to OpenSSL by Brian Carpenter.
2770 *) ASN.1 structure reuse memory corruption fix
2772 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2773 memory corruption via an invalid write. Such reuse is and has been
2774 strongly discouraged and is believed to be rare.
2776 Applications that parse structures containing CHOICE or ANY DEFINED BY
2777 components may be affected. Certificate parsing (d2i_X509 and related
2778 functions) are however not affected. OpenSSL clients and servers are
2783 *) PKCS7 NULL pointer dereferences fix
2785 The PKCS#7 parsing code does not handle missing outer ContentInfo
2786 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2787 missing content and trigger a NULL pointer dereference on parsing.
2789 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2790 otherwise parse PKCS#7 structures from untrusted sources are
2791 affected. OpenSSL clients and servers are not affected.
2793 This issue was reported to OpenSSL by Michal Zalewski (Google).
2797 *) DoS via reachable assert in SSLv2 servers fix
2799 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2800 servers that both support SSLv2 and enable export cipher suites by sending
2801 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2803 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2804 (OpenSSL development team).
2808 *) Empty CKE with client auth and DHE fix
2810 If client auth is used then a server can seg fault in the event of a DHE
2811 ciphersuite being selected and a zero length ClientKeyExchange message
2812 being sent by the client. This could be exploited in a DoS attack.
2816 *) Handshake with unseeded PRNG fix
2818 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2819 with an unseeded PRNG. The conditions are:
2820 - The client is on a platform where the PRNG has not been seeded
2821 automatically, and the user has not seeded manually
2822 - A protocol specific client method version has been used (i.e. not
2823 SSL_client_methodv23)
2824 - A ciphersuite is used that does not require additional random data from
2825 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2827 If the handshake succeeds then the client random that has been used will
2828 have been generated from a PRNG with insufficient entropy and therefore the
2829 output may be predictable.
2831 For example using the following command with an unseeded openssl will
2832 succeed on an unpatched platform:
2834 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2838 *) Use After Free following d2i_ECPrivatekey error fix
2840 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2841 could cause a use after free condition. This, in turn, could cause a double
2842 free in several private key parsing functions (such as d2i_PrivateKey
2843 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2844 for applications that receive EC private keys from untrusted
2845 sources. This scenario is considered rare.
2847 This issue was discovered by the BoringSSL project and fixed in their
2852 *) X509_to_X509_REQ NULL pointer deref fix
2854 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2855 the certificate key is invalid. This function is rarely used in practice.
2857 This issue was discovered by Brian Carpenter.
2861 *) Removed the export ciphers from the DEFAULT ciphers
2864 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2866 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2867 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2868 So far those who have to target multiple platforms would compromise
2869 and argue that binary targeting say ARMv5 would still execute on
2870 ARMv8. "Universal" build resolves this compromise by providing
2871 near-optimal performance even on newer platforms.
2874 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2875 (other platforms pending).
2876 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2878 *) Add support for the SignedCertificateTimestampList certificate and
2879 OCSP response extensions from RFC6962.
2882 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2883 for corner cases. (Certain input points at infinity could lead to
2884 bogus results, with non-infinity inputs mapped to infinity too.)
2887 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2888 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2889 common cases are optimized and there still is room for further
2890 improvements. Vector Permutation AES for Altivec is also added.
2893 *) Add support for little-endian ppc64 Linux target.
2894 [Marcelo Cerri (IBM)]
2896 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2897 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2898 are optimized and there still is room for further improvements.
2899 Both 32- and 64-bit modes are supported.
2900 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2902 *) Improved ARMv7 NEON support.
2905 *) Support for SPARC Architecture 2011 crypto extensions, first
2906 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2907 SHA256/512, MD5, GHASH and modular exponentiation.
2908 [Andy Polyakov, David Miller]
2910 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2912 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2914 *) Support for new and upcoming Intel processors, including AVX2,
2915 BMI and SHA ISA extensions. This includes additional "stitched"
2916 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2919 This work was sponsored by Intel Corp.
2922 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2923 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2924 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2927 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2928 this fixes a limitation in previous versions of OpenSSL.
2931 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2932 MGF1 digest and OAEP label.
2935 *) Add EVP support for key wrapping algorithms, to avoid problems with
2936 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2937 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2938 algorithms and include tests cases.
2941 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2943 [Douglas E. Engert, Steve Henson]
2945 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2946 difference in days and seconds between two tm or ASN1_TIME structures.
2949 *) Add -rev test option to s_server to just reverse order of characters
2950 received by client and send back to server. Also prints an abbreviated
2951 summary of the connection parameters.
2954 *) New option -brief for s_client and s_server to print out a brief summary
2955 of connection parameters.
2958 *) Add callbacks for arbitrary TLS extensions.
2959 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2961 *) New option -crl_download in several openssl utilities to download CRLs
2962 from CRLDP extension in certificates.
2965 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2968 *) New function X509_CRL_diff to generate a delta CRL from the difference
2969 of two full CRLs. Add support to "crl" utility.
2972 *) New functions to set lookup_crls function and to retrieve
2973 X509_STORE from X509_STORE_CTX.
2976 *) Print out deprecated issuer and subject unique ID fields in
2980 *) Extend OCSP I/O functions so they can be used for simple general purpose
2981 HTTP as well as OCSP. New wrapper function which can be used to download
2982 CRLs using the OCSP API.
2985 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2988 *) SSL_CONF* functions. These provide a common framework for application
2989 configuration using configuration files or command lines.
2992 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2993 message callback and prints the results. Needs compile time option
2994 "enable-ssl-trace". New options to s_client and s_server to enable
2998 *) New ctrl and macro to retrieve supported points extensions.
2999 Print out extension in s_server and s_client.
3002 *) New functions to retrieve certificate signature and signature
3006 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3010 *) New Suite B modes for TLS code. These use and enforce the requirements
3011 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3012 only use Suite B curves. The Suite B modes can be set by using the
3013 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3016 *) New chain verification flags for Suite B levels of security. Check
3017 algorithms are acceptable when flags are set in X509_verify_cert.
3020 *) Make tls1_check_chain return a set of flags indicating checks passed
3021 by a certificate chain. Add additional tests to handle client
3022 certificates: checks for matching certificate type and issuer name
3026 *) If an attempt is made to use a signature algorithm not in the peer
3027 preference list abort the handshake. If client has no suitable
3028 signature algorithms in response to a certificate request do not
3029 use the certificate.
3032 *) If server EC tmp key is not in client preference list abort handshake.
3035 *) Add support for certificate stores in CERT structure. This makes it
3036 possible to have different stores per SSL structure or one store in
3037 the parent SSL_CTX. Include distinct stores for certificate chain
3038 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3039 to build and store a certificate chain in CERT structure: returning
3040 an error if the chain cannot be built: this will allow applications
3041 to test if a chain is correctly configured.
3043 Note: if the CERT based stores are not set then the parent SSL_CTX
3044 store is used to retain compatibility with existing behaviour.
3048 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3049 mask based on the current session, check mask when sending client
3050 hello and checking the requested ciphersuite.
3053 *) New ctrls to retrieve and set certificate types in a certificate
3054 request message. Print out received values in s_client. If certificate
3055 types is not set with custom values set sensible values based on
3056 supported signature algorithms.
3059 *) Support for distinct client and server supported signature algorithms.
3062 *) Add certificate callback. If set this is called whenever a certificate
3063 is required by client or server. An application can decide which
3064 certificate chain to present based on arbitrary criteria: for example
3065 supported signature algorithms. Add very simple example to s_server.
3066 This fixes many of the problems and restrictions of the existing client
3067 certificate callback: for example you can now clear an existing
3068 certificate and specify the whole chain.
3071 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3072 the certificate can be used for (if anything). Set valid_flags field
3073 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3074 to have similar checks in it.
3076 Add new "cert_flags" field to CERT structure and include a "strict mode".
3077 This enforces some TLS certificate requirements (such as only permitting
3078 certificate signature algorithms contained in the supported algorithms
3079 extension) which some implementations ignore: this option should be used
3080 with caution as it could cause interoperability issues.
3083 *) Update and tidy signature algorithm extension processing. Work out
3084 shared signature algorithms based on preferences and peer algorithms
3085 and print them out in s_client and s_server. Abort handshake if no
3086 shared signature algorithms.
3089 *) Add new functions to allow customised supported signature algorithms
3090 for SSL and SSL_CTX structures. Add options to s_client and s_server
3094 *) New function SSL_certs_clear() to delete all references to certificates
3095 from an SSL structure. Before this once a certificate had been added
3096 it couldn't be removed.
3099 *) Integrate hostname, email address and IP address checking with certificate
3100 verification. New verify options supporting checking in openssl utility.
3103 *) Fixes and wildcard matching support to hostname and email checking
3104 functions. Add manual page.
3105 [Florian Weimer (Red Hat Product Security Team)]
3107 *) New functions to check a hostname email or IP address against a
3108 certificate. Add options x509 utility to print results of checks against
3112 *) Fix OCSP checking.
3113 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3115 *) Initial experimental support for explicitly trusted non-root CAs.
3116 OpenSSL still tries to build a complete chain to a root but if an
3117 intermediate CA has a trust setting included that is used. The first
3118 setting is used: whether to trust (e.g., -addtrust option to the x509
3122 *) Add -trusted_first option which attempts to find certificates in the
3123 trusted store even if an untrusted chain is also supplied.
3126 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3127 platform support for Linux and Android.
3130 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3133 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3134 When in FIPS mode the approved implementations are used as normal,
3135 when not in FIPS mode the internal unapproved versions are used instead.
3136 This means that the FIPS capable OpenSSL isn't forced to use the
3137 (often lower performance) FIPS implementations outside FIPS mode.
3140 *) Transparently support X9.42 DH parameters when calling
3141 PEM_read_bio_DHparameters. This means existing applications can handle
3142 the new parameter format automatically.
3145 *) Initial experimental support for X9.42 DH parameter format: mainly
3146 to support use of 'q' parameter for RFC5114 parameters.
3149 *) Add DH parameters from RFC5114 including test data to dhtest.
3152 *) Support for automatic EC temporary key parameter selection. If enabled
3153 the most preferred EC parameters are automatically used instead of
3154 hardcoded fixed parameters. Now a server just has to call:
3155 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3156 support ECDH and use the most appropriate parameters.
3159 *) Enhance and tidy EC curve and point format TLS extension code. Use
3160 static structures instead of allocation if default values are used.
3161 New ctrls to set curves we wish to support and to retrieve shared curves.
3162 Print out shared curves in s_server. New options to s_server and s_client
3163 to set list of supported curves.
3166 *) New ctrls to retrieve supported signature algorithms and
3167 supported curve values as an array of NIDs. Extend openssl utility
3168 to print out received values.
3171 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3172 between NIDs and the more common NIST names such as "P-256". Enhance
3173 ecparam utility and ECC method to recognise the NIST names for curves.
3176 *) Enhance SSL/TLS certificate chain handling to support different
3177 chains for each certificate instead of one chain in the parent SSL_CTX.
3180 *) Support for fixed DH ciphersuite client authentication: where both
3181 server and client use DH certificates with common parameters.
3184 *) Support for fixed DH ciphersuites: those requiring DH server
3188 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3190 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3191 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3192 X509_CINF_get_signature were reverted post internal team review.
3194 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3196 *) Build fixes for the Windows and OpenVMS platforms
3197 [Matt Caswell and Richard Levitte]
3199 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3201 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3202 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3203 dereference. This could lead to a Denial Of Service attack. Thanks to
3204 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3208 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3209 dtls1_buffer_record function under certain conditions. In particular this
3210 could occur if an attacker sent repeated DTLS records with the same
3211 sequence number but for the next epoch. The memory leak could be exploited
3212 by an attacker in a Denial of Service attack through memory exhaustion.
3213 Thanks to Chris Mueller for reporting this issue.
3217 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3218 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3219 method would be set to NULL which could later result in a NULL pointer
3220 dereference. Thanks to Frank Schmirler for reporting this issue.
3224 *) Abort handshake if server key exchange message is omitted for ephemeral
3227 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3228 reporting this issue.
3232 *) Remove non-export ephemeral RSA code on client and server. This code
3233 violated the TLS standard by allowing the use of temporary RSA keys in
3234 non-export ciphersuites and could be used by a server to effectively
3235 downgrade the RSA key length used to a value smaller than the server
3236 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3237 INRIA or reporting this issue.
3241 *) Fixed issue where DH client certificates are accepted without verification.
3242 An OpenSSL server will accept a DH certificate for client authentication
3243 without the certificate verify message. This effectively allows a client to
3244 authenticate without the use of a private key. This only affects servers
3245 which trust a client certificate authority which issues certificates
3246 containing DH keys: these are extremely rare and hardly ever encountered.
3247 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3252 *) Ensure that the session ID context of an SSL is updated when its
3253 SSL_CTX is updated via SSL_set_SSL_CTX.
3255 The session ID context is typically set from the parent SSL_CTX,
3256 and can vary with the CTX.
3259 *) Fix various certificate fingerprint issues.
3261 By using non-DER or invalid encodings outside the signed portion of a
3262 certificate the fingerprint can be changed without breaking the signature.
3263 Although no details of the signed portion of the certificate can be changed
3264 this can cause problems with some applications: e.g. those using the
3265 certificate fingerprint for blacklists.
3267 1. Reject signatures with non zero unused bits.
3269 If the BIT STRING containing the signature has non zero unused bits reject
3270 the signature. All current signature algorithms require zero unused bits.
3272 2. Check certificate algorithm consistency.
3274 Check the AlgorithmIdentifier inside TBS matches the one in the
3275 certificate signature. NB: this will result in signature failure
3276 errors for some broken certificates.
3278 Thanks to Konrad Kraszewski from Google for reporting this issue.
3280 3. Check DSA/ECDSA signatures use DER.
3282 Re-encode DSA/ECDSA signatures and compare with the original received
3283 signature. Return an error if there is a mismatch.
3285 This will reject various cases including garbage after signature
3286 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3287 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3288 (negative or with leading zeroes).
3290 Further analysis was conducted and fixes were developed by Stephen Henson
3291 of the OpenSSL core team.
3296 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3297 results on some platforms, including x86_64. This bug occurs at random
3298 with a very low probability, and is not known to be exploitable in any
3299 way, though its exact impact is difficult to determine. Thanks to Pieter
3300 Wuille (Blockstream) who reported this issue and also suggested an initial
3301 fix. Further analysis was conducted by the OpenSSL development team and
3302 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3303 the OpenSSL core team.
3307 *) Do not resume sessions on the server if the negotiated protocol
3308 version does not match the session's version. Resuming with a different
3309 version, while not strictly forbidden by the RFC, is of questionable
3310 sanity and breaks all known clients.
3311 [David Benjamin, Emilia Käsper]
3313 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3314 early CCS messages during renegotiation. (Note that because
3315 renegotiation is encrypted, this early CCS was not exploitable.)
3318 *) Tighten client-side session ticket handling during renegotiation:
3319 ensure that the client only accepts a session ticket if the server sends
3320 the extension anew in the ServerHello. Previously, a TLS client would
3321 reuse the old extension state and thus accept a session ticket if one was
3322 announced in the initial ServerHello.
3324 Similarly, ensure that the client requires a session ticket if one
3325 was advertised in the ServerHello. Previously, a TLS client would
3326 ignore a missing NewSessionTicket message.
3329 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3331 *) SRTP Memory Leak.
3333 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3334 sends a carefully crafted handshake message, to cause OpenSSL to fail
3335 to free up to 64k of memory causing a memory leak. This could be
3336 exploited in a Denial Of Service attack. This issue affects OpenSSL
3337 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3338 whether SRTP is used or configured. Implementations of OpenSSL that
3339 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3341 The fix was developed by the OpenSSL team.
3345 *) Session Ticket Memory Leak.
3347 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3348 integrity of that ticket is first verified. In the event of a session
3349 ticket integrity check failing, OpenSSL will fail to free memory
3350 causing a memory leak. By sending a large number of invalid session
3351 tickets an attacker could exploit this issue in a Denial Of Service
3356 *) Build option no-ssl3 is incomplete.
3358 When OpenSSL is configured with "no-ssl3" as a build option, servers
3359 could accept and complete a SSL 3.0 handshake, and clients could be
3360 configured to send them.
3362 [Akamai and the OpenSSL team]
3364 *) Add support for TLS_FALLBACK_SCSV.
3365 Client applications doing fallback retries should call
3366 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3368 [Adam Langley, Bodo Moeller]
3370 *) Add additional DigestInfo checks.
3372 Re-encode DigestInto in DER and check against the original when
3373 verifying RSA signature: this will reject any improperly encoded
3374 DigestInfo structures.
3376 Note: this is a precautionary measure and no attacks are currently known.
3380 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3382 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3383 SRP code can be overrun an internal buffer. Add sanity check that
3384 g, A, B < N to SRP code.
3386 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3387 Group for discovering this issue.
3391 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3392 TLS 1.0 instead of higher protocol versions when the ClientHello message
3393 is badly fragmented. This allows a man-in-the-middle attacker to force a
3394 downgrade to TLS 1.0 even if both the server and the client support a
3395 higher protocol version, by modifying the client's TLS records.
3397 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3398 researching this issue.
3402 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3403 to a denial of service attack. A malicious server can crash the client
3404 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3405 ciphersuite and sending carefully crafted handshake messages.
3407 Thanks to Felix Gröbert (Google) for discovering and researching this
3412 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3413 to leak memory. This can be exploited through a Denial of Service attack.
3414 Thanks to Adam Langley for discovering and researching this issue.
3418 *) An attacker can force openssl to consume large amounts of memory whilst
3419 processing DTLS handshake messages. This can be exploited through a
3420 Denial of Service attack.
3421 Thanks to Adam Langley for discovering and researching this issue.
3425 *) An attacker can force an error condition which causes openssl to crash
3426 whilst processing DTLS packets due to memory being freed twice. This
3427 can be exploited through a Denial of Service attack.
3428 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3433 *) If a multithreaded client connects to a malicious server using a resumed
3434 session and the server sends an ec point format extension it could write
3435 up to 255 bytes to freed memory.
3437 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3442 *) A malicious server can crash an OpenSSL client with a null pointer
3443 dereference (read) by specifying an SRP ciphersuite even though it was not
3444 properly negotiated with the client. This can be exploited through a
3445 Denial of Service attack.
3447 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3448 discovering and researching this issue.
3452 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3453 X509_name_oneline, X509_name_print_ex et al. to leak some information
3454 from the stack. Applications may be affected if they echo pretty printing
3455 output to the attacker.
3457 Thanks to Ivan Fratric (Google) for discovering this issue.
3459 [Emilia Käsper, and Steve Henson]
3461 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3462 for corner cases. (Certain input points at infinity could lead to
3463 bogus results, with non-infinity inputs mapped to infinity too.)
3466 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3468 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3469 handshake can force the use of weak keying material in OpenSSL
3470 SSL/TLS clients and servers.
3472 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3473 researching this issue. (CVE-2014-0224)
3474 [KIKUCHI Masashi, Steve Henson]
3476 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3477 OpenSSL DTLS client the code can be made to recurse eventually crashing
3480 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3482 [Imre Rad, Steve Henson]
3484 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3485 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3486 client or server. This is potentially exploitable to run arbitrary
3487 code on a vulnerable client or server.
3489 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3490 [Jüri Aedla, Steve Henson]
3492 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3493 are subject to a denial of service attack.
3495 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3496 this issue. (CVE-2014-3470)
3497 [Felix Gröbert, Ivan Fratric, Steve Henson]
3499 *) Harmonize version and its documentation. -f flag is used to display
3501 [mancha <mancha1@zoho.com>]
3503 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3504 in i2d_ECPrivateKey.
3505 [mancha <mancha1@zoho.com>]
3507 *) Fix some double frees. These are not thought to be exploitable.
3508 [mancha <mancha1@zoho.com>]
3510 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3512 *) A missing bounds check in the handling of the TLS heartbeat extension
3513 can be used to reveal up to 64k of memory to a connected client or
3516 Thanks for Neel Mehta of Google Security for discovering this bug and to
3517 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3518 preparing the fix (CVE-2014-0160)
3519 [Adam Langley, Bodo Moeller]
3521 *) Fix for the attack described in the paper "Recovering OpenSSL
3522 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3523 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3524 http://eprint.iacr.org/2014/140
3526 Thanks to Yuval Yarom and Naomi Benger for discovering this
3527 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3528 [Yuval Yarom and Naomi Benger]
3530 *) TLS pad extension: draft-agl-tls-padding-03
3532 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3533 TLS client Hello record length value would otherwise be > 255 and
3534 less that 512 pad with a dummy extension containing zeroes so it
3535 is at least 512 bytes long.
3537 [Adam Langley, Steve Henson]
3539 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3541 *) Fix for TLS record tampering bug. A carefully crafted invalid
3542 handshake could crash OpenSSL with a NULL pointer exception.
3543 Thanks to Anton Johansson for reporting this issues.
3546 *) Keep original DTLS digest and encryption contexts in retransmission
3547 structures so we can use the previous session parameters if they need
3548 to be resent. (CVE-2013-6450)
3551 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3552 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3553 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3554 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3555 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3556 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3557 [Rob Stradling, Adam Langley]
3559 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3561 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3562 supporting platforms or when small records were transferred.
3563 [Andy Polyakov, Steve Henson]
3565 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3567 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3569 This addresses the flaw in CBC record processing discovered by
3570 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3571 at: http://www.isg.rhul.ac.uk/tls/
3573 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3574 Security Group at Royal Holloway, University of London
3575 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3576 Emilia Käsper for the initial patch.
3578 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3580 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3581 ciphersuites which can be exploited in a denial of service attack.
3582 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3583 and detecting this bug and to Wolfgang Ettlinger
3584 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3588 *) Return an error when checking OCSP signatures when key is NULL.
3589 This fixes a DoS attack. (CVE-2013-0166)
3592 *) Make openssl verify return errors.
3593 [Chris Palmer <palmer@google.com> and Ben Laurie]
3595 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3596 the right response is stapled. Also change SSL_get_certificate()
3597 so it returns the certificate actually sent.
3598 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3599 [Rob Stradling <rob.stradling@comodo.com>]
3601 *) Fix possible deadlock when decoding public keys.
3604 *) Don't use TLS 1.0 record version number in initial client hello
3608 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3610 *) Sanity check record length before skipping explicit IV in TLS
3611 1.2, 1.1 and DTLS to fix DoS attack.
3613 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3614 fuzzing as a service testing platform.
3618 *) Initialise tkeylen properly when encrypting CMS messages.
3619 Thanks to Solar Designer of Openwall for reporting this issue.
3622 *) In FIPS mode don't try to use composite ciphers as they are not
3626 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3628 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3629 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3630 mean any application compiled against OpenSSL 1.0.0 headers setting
3631 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3632 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3633 0x10000000L Any application which was previously compiled against
3634 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3635 will need to be recompiled as a result. Letting be results in
3636 inability to disable specifically TLS 1.1 and in client context,
3637 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3640 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3641 disable just protocol X, but all protocols above X *if* there are
3642 protocols *below* X still enabled. In more practical terms it means
3643 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3644 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3645 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3649 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3651 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3652 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3653 in CRYPTO_realloc_clean.
3655 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3656 issue and to Adam Langley <agl@chromium.org> for fixing it.
3658 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3660 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3663 *) Workarounds for some broken servers that "hang" if a client hello
3664 record length exceeds 255 bytes.
3666 1. Do not use record version number > TLS 1.0 in initial client
3667 hello: some (but not all) hanging servers will now work.
3668 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3669 the number of ciphers sent in the client hello. This should be
3670 set to an even number, such as 50, for example by passing:
3671 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3672 Most broken servers should now work.
3673 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3674 TLS 1.2 client support entirely.
3677 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3680 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3682 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3683 STRING form instead of a DigestInfo.
3686 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3687 and the RSA_sign/RSA_verify functions. This was made more apparent when
3688 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3689 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3690 the correct format in RSA_verify so both forms transparently work.
3693 *) Some servers which support TLS 1.0 can choke if we initially indicate
3694 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3695 encrypted premaster secret. As a workaround use the maximum permitted
3696 client version in client hello, this should keep such servers happy
3697 and still work with previous versions of OpenSSL.
3700 *) Add support for TLS/DTLS heartbeats.
3701 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3703 *) Add support for SCTP.
3704 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3706 *) Improved PRNG seeding for VOS.
3707 [Paul Green <Paul.Green@stratus.com>]
3709 *) Extensive assembler packs updates, most notably:
3711 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3712 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3713 - x86_64: bit-sliced AES implementation;
3714 - ARM: NEON support, contemporary platforms optimizations;
3715 - s390x: z196 support;
3716 - *: GHASH and GF(2^m) multiplication implementations;
3720 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3721 (removal of unnecessary code)
3722 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3724 *) Add TLS key material exporter from RFC 5705.
3727 *) Add DTLS-SRTP negotiation from RFC 5764.
3730 *) Add Next Protocol Negotiation,
3731 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3732 disabled with a no-npn flag to config or Configure. Code donated
3734 [Adam Langley <agl@google.com> and Ben Laurie]
3736 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3737 NIST-P256, NIST-P521, with constant-time single point multiplication on
3738 typical inputs. Compiler support for the nonstandard type __uint128_t is
3739 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3740 Code made available under Apache License version 2.0.
3742 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3743 line to include this in your build of OpenSSL, and run "make depend" (or
3744 "make update"). This enables the following EC_METHODs:
3746 EC_GFp_nistp224_method()
3747 EC_GFp_nistp256_method()
3748 EC_GFp_nistp521_method()
3750 EC_GROUP_new_by_curve_name() will automatically use these (while
3751 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3753 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3755 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3756 all platforms. Move ssize_t definition from e_os.h to the public
3757 header file e_os2.h as it now appears in public header file cms.h
3760 *) New -sigopt option to the ca, req and x509 utilities. Additional
3761 signature parameters can be passed using this option and in
3765 *) Add RSA PSS signing function. This will generate and set the
3766 appropriate AlgorithmIdentifiers for PSS based on those in the
3767 corresponding EVP_MD_CTX structure. No application support yet.
3770 *) Support for companion algorithm specific ASN1 signing routines.
3771 New function ASN1_item_sign_ctx() signs a pre-initialised
3772 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3773 the appropriate parameters.
3776 *) Add new algorithm specific ASN1 verification initialisation function
3777 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3778 handling will be the same no matter what EVP_PKEY_METHOD is used.
3779 Add a PSS handler to support verification of PSS signatures: checked
3780 against a number of sample certificates.
3783 *) Add signature printing for PSS. Add PSS OIDs.
3784 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3786 *) Add algorithm specific signature printing. An individual ASN1 method
3787 can now print out signatures instead of the standard hex dump.
3789 More complex signatures (e.g. PSS) can print out more meaningful
3790 information. Include DSA version that prints out the signature
3794 *) Password based recipient info support for CMS library: implementing
3798 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3799 neatly separates the code into cipher and PBE sections and is required
3800 for some algorithms that split PBES2 into separate pieces (such as
3801 password based CMS).
3804 *) Session-handling fixes:
3805 - Fix handling of connections that are resuming with a session ID,
3806 but also support Session Tickets.
3807 - Fix a bug that suppressed issuing of a new ticket if the client
3808 presented a ticket with an expired session.
3809 - Try to set the ticket lifetime hint to something reasonable.
3810 - Make tickets shorter by excluding irrelevant information.
3811 - On the client side, don't ignore renewed tickets.
3812 [Adam Langley, Bodo Moeller (Google)]
3814 *) Fix PSK session representation.
3817 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3819 This work was sponsored by Intel.
3822 *) Add GCM support to TLS library. Some custom code is needed to split
3823 the IV between the fixed (from PRF) and explicit (from TLS record)
3824 portions. This adds all GCM ciphersuites supported by RFC5288 and
3825 RFC5289. Generalise some AES* cipherstrings to include GCM and
3826 add a special AESGCM string for GCM only.
3829 *) Expand range of ctrls for AES GCM. Permit setting invocation
3830 field on decrypt and retrieval of invocation field only on encrypt.
3833 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3834 As required by RFC5289 these ciphersuites cannot be used if for
3835 versions of TLS earlier than 1.2.
3838 *) For FIPS capable OpenSSL interpret a NULL default public key method
3839 as unset and return the appropriate default but do *not* set the default.
3840 This means we can return the appropriate method in applications that
3841 switch between FIPS and non-FIPS modes.
3844 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3845 ENGINE is used then we cannot handle that in the FIPS module so we
3846 keep original code iff non-FIPS operations are allowed.
3849 *) Add -attime option to openssl utilities.
3850 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3852 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3855 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3856 FIPS EC methods unconditionally for now.
3859 *) New build option no-ec2m to disable characteristic 2 code.
3862 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3863 all cases can be covered as some introduce binary incompatibilities.
3866 *) Redirect RSA operations to FIPS module including keygen,
3867 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3870 *) Add similar low level API blocking to ciphers.
3873 *) Low level digest APIs are not approved in FIPS mode: any attempt
3874 to use these will cause a fatal error. Applications that *really* want
3875 to use them can use the private_* version instead.
3878 *) Redirect cipher operations to FIPS module for FIPS builds.
3881 *) Redirect digest operations to FIPS module for FIPS builds.
3884 *) Update build system to add "fips" flag which will link in fipscanister.o
3885 for static and shared library builds embedding a signature if needed.
3888 *) Output TLS supported curves in preference order instead of numerical
3889 order. This is currently hardcoded for the highest order curves first.
3890 This should be configurable so applications can judge speed vs strength.
3893 *) Add TLS v1.2 server support for client authentication.
3896 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3900 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3901 FIPS modules versions.
3904 *) Add TLS v1.2 client side support for client authentication. Keep cache
3905 of handshake records longer as we don't know the hash algorithm to use
3906 until after the certificate request message is received.
3909 *) Initial TLS v1.2 client support. Add a default signature algorithms
3910 extension including all the algorithms we support. Parse new signature
3911 format in client key exchange. Relax some ECC signing restrictions for
3912 TLS v1.2 as indicated in RFC5246.
3915 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3916 to new signature format when needed using client digest preference.
3917 All server ciphersuites should now work correctly in TLS v1.2. No client
3918 support yet and no support for client certificates.
3921 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3922 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3923 ciphersuites. At present only RSA key exchange ciphersuites work with
3924 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3925 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3926 and version checking.
3929 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3930 with this defined it will not be affected by any changes to ssl internal
3931 structures. Add several utility functions to allow openssl application
3932 to work with OPENSSL_NO_SSL_INTERN defined.
3935 *) A long standing patch to add support for SRP from EdelWeb (Peter
3936 Sylvester and Christophe Renou) was integrated.
3937 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3938 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3941 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3944 *) Permit abbreviated handshakes when renegotiating using the function
3945 SSL_renegotiate_abbreviated().
3946 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3948 *) Add call to ENGINE_register_all_complete() to
3949 ENGINE_load_builtin_engines(), so some implementations get used
3950 automatically instead of needing explicit application support.
3953 *) Add support for TLS key exporter as described in RFC5705.
3954 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3956 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3957 a few changes are required:
3959 Add SSL_OP_NO_TLSv1_1 flag.
3960 Add TLSv1_1 methods.
3961 Update version checking logic to handle version 1.1.
3962 Add explicit IV handling (ported from DTLS code).
3963 Add command line options to s_client/s_server.
3966 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3968 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3969 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3970 content decryption and always return the same error. Note: this attack
3971 needs on average 2^20 messages so it only affects automated senders. The
3972 old behaviour can be re-enabled in the CMS code by setting the
3973 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3974 an MMA defence is not necessary.
3975 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3976 this issue. (CVE-2012-0884)
3979 *) Fix CVE-2011-4619: make sure we really are receiving a
3980 client hello before rejecting multiple SGC restarts. Thanks to
3981 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3984 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3986 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3987 Thanks to Antonio Martin, Enterprise Secure Access Research and
3988 Development, Cisco Systems, Inc. for discovering this bug and
3989 preparing a fix. (CVE-2012-0050)
3992 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3994 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3995 of the Vaudenay padding oracle attack on CBC mode encryption
3996 which enables an efficient plaintext recovery attack against
3997 the OpenSSL implementation of DTLS. Their attack exploits timing
3998 differences arising during decryption processing. A research
3999 paper describing this attack can be found at:
4000 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4001 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4002 Security Group at Royal Holloway, University of London
4003 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4004 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4005 for preparing the fix. (CVE-2011-4108)
4006 [Robin Seggelmann, Michael Tuexen]
4008 *) Clear bytes used for block padding of SSL 3.0 records.
4010 [Adam Langley (Google)]
4012 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4013 Kadianakis <desnacked@gmail.com> for discovering this issue and
4014 Adam Langley for preparing the fix. (CVE-2011-4619)
4015 [Adam Langley (Google)]
4017 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4018 [Andrey Kulikov <amdeich@gmail.com>]
4020 *) Prevent malformed RFC3779 data triggering an assertion failure.
4021 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4022 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4023 [Rob Austein <sra@hactrn.net>]
4025 *) Improved PRNG seeding for VOS.
4026 [Paul Green <Paul.Green@stratus.com>]
4028 *) Fix ssl_ciph.c set-up race.
4029 [Adam Langley (Google)]
4031 *) Fix spurious failures in ecdsatest.c.
4032 [Emilia Käsper (Google)]
4034 *) Fix the BIO_f_buffer() implementation (which was mixing different
4035 interpretations of the '..._len' fields).
4036 [Adam Langley (Google)]
4038 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4039 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4040 threads won't reuse the same blinding coefficients.
4042 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4043 lock to call BN_BLINDING_invert_ex, and avoids one use of
4044 BN_BLINDING_update for each BN_BLINDING structure (previously,
4045 the last update always remained unused).
4046 [Emilia Käsper (Google)]
4048 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4049 [Bob Buckholz (Google)]
4051 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4053 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4054 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4055 [Kaspar Brand <ossl@velox.ch>]
4057 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4058 for multi-threaded use of ECDH. (CVE-2011-3210)
4059 [Adam Langley (Google)]
4061 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4064 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4065 signature public key algorithm by using OID xref utilities instead.
4066 Before this you could only use some ECC ciphersuites with SHA1 only.
4069 *) Add protection against ECDSA timing attacks as mentioned in the paper
4070 by Billy Bob Brumley and Nicola Tuveri, see:
4072 http://eprint.iacr.org/2011/232.pdf
4074 [Billy Bob Brumley and Nicola Tuveri]
4076 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4078 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4079 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4081 *) Fix bug in string printing code: if *any* escaping is enabled we must
4082 escape the escape character (backslash) or the resulting string is
4086 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4088 *) Disable code workaround for ancient and obsolete Netscape browsers
4089 and servers: an attacker can use it in a ciphersuite downgrade attack.
4090 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4093 *) Fixed J-PAKE implementation error, originally discovered by
4094 Sebastien Martini, further info and confirmation from Stefan
4095 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4098 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4100 *) Fix extension code to avoid race conditions which can result in a buffer
4101 overrun vulnerability: resumed sessions must not be modified as they can
4102 be shared by multiple threads. CVE-2010-3864
4105 *) Fix WIN32 build system to correctly link an ENGINE directory into
4109 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4111 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4113 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4115 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4117 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4118 context. The operation can be customised via the ctrl mechanism in
4119 case ENGINEs want to include additional functionality.
4122 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4125 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4126 output hashes compatible with older versions of OpenSSL.
4127 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4129 *) Fix compression algorithm handling: if resuming a session use the
4130 compression algorithm of the resumed session instead of determining
4131 it from client hello again. Don't allow server to change algorithm.
4134 *) Add load_crls() function to apps tidying load_certs() too. Add option
4135 to verify utility to allow additional CRLs to be included.
4138 *) Update OCSP request code to permit adding custom headers to the request:
4139 some responders need this.
4142 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4144 [Julia Lawall <julia@diku.dk>]
4146 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4147 needlessly dereferenced structures, used obsolete functions and
4148 didn't handle all updated verify codes correctly.
4151 *) Disable MD2 in the default configuration.
4154 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4155 indicate the initial BIO being pushed or popped. This makes it possible
4156 to determine whether the BIO is the one explicitly called or as a result
4157 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4158 it handles reference counts correctly and doesn't zero out the I/O bio
4159 when it is not being explicitly popped. WARNING: applications which
4160 included workarounds for the old buggy behaviour will need to be modified
4161 or they could free up already freed BIOs.
4164 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4165 renaming to all platforms (within the 0.9.8 branch, this was
4166 done conditionally on Netware platforms to avoid a name clash).
4167 [Guenter <lists@gknw.net>]
4169 *) Add ECDHE and PSK support to DTLS.
4170 [Michael Tuexen <tuexen@fh-muenster.de>]
4172 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4176 *) Add "missing" function EVP_MD_flags() (without this the only way to
4177 retrieve a digest flags is by accessing the structure directly. Update
4178 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4179 or cipher is registered as in the "from" argument. Print out all
4180 registered digests in the dgst usage message instead of manually
4181 attempting to work them out.
4184 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4185 this allows the use of compression and extensions. Change default cipher
4186 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4187 by default unless an application cipher string requests it.
4190 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4191 key ids to find matching certificates and keys but some PKCS#12 files
4192 don't follow the (somewhat unwritten) rules and this strategy fails.
4193 Now just gather all certificates together and the first private key
4194 then look for the first certificate that matches the key.
4197 *) Support use of registered digest and cipher names for dgst and cipher
4198 commands instead of having to add each one as a special case. So now
4205 openssl dgst -sha256 foo
4207 and this works for ENGINE based algorithms too.
4211 *) Update Gost ENGINE to support parameter files.
4212 [Victor B. Wagner <vitus@cryptocom.ru>]
4214 *) Support GeneralizedTime in ca utility.
4215 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4217 *) Enhance the hash format used for certificate directory links. The new
4218 form uses the canonical encoding (meaning equivalent names will work
4219 even if they aren't identical) and uses SHA1 instead of MD5. This form
4220 is incompatible with the older format and as a result c_rehash should
4221 be used to rebuild symbolic links.
4224 *) Make PKCS#8 the default write format for private keys, replacing the
4225 traditional format. This form is standardised, more secure and doesn't
4226 include an implicit MD5 dependency.
4229 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4230 committed to OpenSSL should pass this lot as a minimum.
4233 *) Add session ticket override functionality for use by EAP-FAST.
4234 [Jouni Malinen <j@w1.fi>]
4236 *) Modify HMAC functions to return a value. Since these can be implemented
4237 in an ENGINE errors can occur.
4240 *) Type-checked OBJ_bsearch_ex.
4243 *) Type-checked OBJ_bsearch. Also some constification necessitated
4244 by type-checking. Still to come: TXT_DB, bsearch(?),
4245 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4249 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4250 seconds to a tm structure directly, instead of going through OS
4251 specific date routines. This avoids any issues with OS routines such
4252 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4253 and X509_time_adj_ex() to cover the extended range. The existing
4254 X509_time_adj() is still usable and will no longer have any date issues.
4257 *) Delta CRL support. New use deltas option which will attempt to locate
4258 and search any appropriate delta CRLs available.
4260 This work was sponsored by Google.
4263 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4264 code and add additional score elements. Validate alternate CRL paths
4265 as part of the CRL checking and indicate a new error "CRL path validation
4266 error" in this case. Applications wanting additional details can use
4267 the verify callback and check the new "parent" field. If this is not
4268 NULL CRL path validation is taking place. Existing applications won't
4269 see this because it requires extended CRL support which is off by
4272 This work was sponsored by Google.
4275 *) Support for freshest CRL extension.
4277 This work was sponsored by Google.
4280 *) Initial indirect CRL support. Currently only supported in the CRLs
4281 passed directly and not via lookup. Process certificate issuer
4282 CRL entry extension and lookup CRL entries by bother issuer name
4283 and serial number. Check and process CRL issuer entry in IDP extension.
4285 This work was sponsored by Google.
4288 *) Add support for distinct certificate and CRL paths. The CRL issuer
4289 certificate is validated separately in this case. Only enabled if
4290 an extended CRL support flag is set: this flag will enable additional
4291 CRL functionality in future.
4293 This work was sponsored by Google.
4296 *) Add support for policy mappings extension.
4298 This work was sponsored by Google.
4301 *) Fixes to pathlength constraint, self issued certificate handling,
4302 policy processing to align with RFC3280 and PKITS tests.
4304 This work was sponsored by Google.
4307 *) Support for name constraints certificate extension. DN, email, DNS
4308 and URI types are currently supported.
4310 This work was sponsored by Google.
4313 *) To cater for systems that provide a pointer-based thread ID rather
4314 than numeric, deprecate the current numeric thread ID mechanism and
4315 replace it with a structure and associated callback type. This
4316 mechanism allows a numeric "hash" to be extracted from a thread ID in
4317 either case, and on platforms where pointers are larger than 'long',
4318 mixing is done to help ensure the numeric 'hash' is usable even if it
4319 can't be guaranteed unique. The default mechanism is to use "&errno"
4320 as a pointer-based thread ID to distinguish between threads.
4322 Applications that want to provide their own thread IDs should now use
4323 CRYPTO_THREADID_set_callback() to register a callback that will call
4324 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4326 Note that ERR_remove_state() is now deprecated, because it is tied
4327 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4328 to free the current thread's error state should be replaced by
4329 ERR_remove_thread_state(NULL).
4331 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4332 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4333 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4334 application was previously providing a numeric thread callback that
4335 was inappropriate for distinguishing threads, then uniqueness might
4336 have been obtained with &errno that happened immediately in the
4337 intermediate development versions of OpenSSL; this is no longer the
4338 case, the numeric thread callback will now override the automatic use
4340 [Geoff Thorpe, with help from Bodo Moeller]
4342 *) Initial support for different CRL issuing certificates. This covers a
4343 simple case where the self issued certificates in the chain exist and
4344 the real CRL issuer is higher in the existing chain.
4346 This work was sponsored by Google.
4349 *) Removed effectively defunct crypto/store from the build.
4352 *) Revamp of STACK to provide stronger type-checking. Still to come:
4353 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4354 ASN1_STRING, CONF_VALUE.
4357 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4358 RAM on SSL connections. This option can save about 34k per idle SSL.
4361 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4362 STACK, TXT_DB, bsearch, qsort.
4365 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4366 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4367 support for data, signedData, compressedData, digestedData and
4368 encryptedData, envelopedData types included. Scripts to check against
4369 RFC4134 examples draft and interop and consistency checks of many
4370 content types and variants.
4373 *) Add options to enc utility to support use of zlib compression BIO.
4376 *) Extend mk1mf to support importing of options and assembly language
4377 files from Configure script, currently only included in VC-WIN32.
4378 The assembly language rules can now optionally generate the source
4379 files from the associated perl scripts.
4382 *) Implement remaining functionality needed to support GOST ciphersuites.
4383 Interop testing has been performed using CryptoPro implementations.
4384 [Victor B. Wagner <vitus@cryptocom.ru>]
4386 *) s390x assembler pack.
4389 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4393 *) Implement Opaque PRF Input TLS extension as specified in
4394 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4395 official specification yet and no extension type assignment by
4396 IANA exists, this extension (for now) will have to be explicitly
4397 enabled when building OpenSSL by providing the extension number
4398 to use. For example, specify an option
4400 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4402 to the "config" or "Configure" script to enable the extension,
4403 assuming extension number 0x9527 (which is a completely arbitrary
4404 and unofficial assignment based on the MD5 hash of the Internet
4405 Draft). Note that by doing so, you potentially lose
4406 interoperability with other TLS implementations since these might
4407 be using the same extension number for other purposes.
4409 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4410 opaque PRF input value to use in the handshake. This will create
4411 an internal copy of the length-'len' string at 'src', and will
4412 return non-zero for success.
4414 To get more control and flexibility, provide a callback function
4417 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4418 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4422 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4425 Callback function 'cb' will be called in handshakes, and is
4426 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4427 Argument 'arg' is for application purposes (the value as given to
4428 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4429 be provided to the callback function). The callback function
4430 has to return non-zero to report success: usually 1 to use opaque
4431 PRF input just if possible, or 2 to enforce use of the opaque PRF
4432 input. In the latter case, the library will abort the handshake
4433 if opaque PRF input is not successfully negotiated.
4435 Arguments 'peerinput' and 'len' given to the callback function
4436 will always be NULL and 0 in the case of a client. A server will
4437 see the client's opaque PRF input through these variables if
4438 available (NULL and 0 otherwise). Note that if the server
4439 provides an opaque PRF input, the length must be the same as the
4440 length of the client's opaque PRF input.
4442 Note that the callback function will only be called when creating
4443 a new session (session resumption can resume whatever was
4444 previously negotiated), and will not be called in SSL 2.0
4445 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4446 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4447 for applications that need to enforce opaque PRF input.
4451 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4454 [Victor B. Wagner <vitus@cryptocom.ru>]
4456 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4457 RFC4507bis. The encrypted ticket format is an encrypted encoded
4458 SSL_SESSION structure, that way new session features are automatically
4461 If a client application caches session in an SSL_SESSION structure
4462 support is transparent because tickets are now stored in the encoded
4465 The SSL_CTX structure automatically generates keys for ticket
4466 protection in servers so again support should be possible
4467 with no application modification.
4469 If a client or server wishes to disable RFC4507 support then the option
4470 SSL_OP_NO_TICKET can be set.
4472 Add a TLS extension debugging callback to allow the contents of any client
4473 or server extensions to be examined.
4475 This work was sponsored by Google.
4478 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4479 OpenSSL should now compile cleanly on gcc 4.2
4480 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4482 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4483 support including streaming MAC support: this is required for GOST
4484 ciphersuite support.
4485 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4487 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4488 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4489 to output in BER and PEM format.
4492 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4493 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4494 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4495 ENGINE support for HMAC keys which are unextractable. New -mac and
4496 -macopt options to dgst utility.
4499 *) New option -sigopt to dgst utility. Update dgst to use
4500 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4501 alternative signing parameters such as X9.31 or PSS in the dgst
4505 *) Change ssl_cipher_apply_rule(), the internal function that does
4506 the work each time a ciphersuite string requests enabling
4507 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4508 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4509 the order of disabled ciphersuites such that those ciphersuites
4510 that most recently went from enabled to disabled not only stay
4511 in order with respect to each other, but also have higher priority
4512 than other disabled ciphersuites the next time ciphersuites are
4515 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4516 the same ciphersuites as with "HIGH" alone, but in a specific
4517 order where the PSK ciphersuites come first (since they are the
4518 most recently disabled ciphersuites when "HIGH" is parsed).
4520 Also, change ssl_create_cipher_list() (using this new
4521 functionality) such that between otherwise identical
4522 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4526 *) Change ssl_create_cipher_list() so that it automatically
4527 arranges the ciphersuites in reasonable order before starting
4528 to process the rule string. Thus, the definition for "DEFAULT"
4529 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4530 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4531 This makes it much easier to arrive at a reasonable default order
4532 in applications for which anonymous ciphers are OK (meaning
4533 that you can't actually use DEFAULT).
4534 [Bodo Moeller; suggested by Victor Duchovni]
4536 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4537 processing) into multiple integers instead of setting
4538 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4539 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4540 (These masks as well as the individual bit definitions are hidden
4541 away into the non-exported interface ssl/ssl_locl.h, so this
4542 change to the definition of the SSL_CIPHER structure shouldn't
4543 affect applications.) This give us more bits for each of these
4544 categories, so there is no longer a need to coagulate AES128 and
4545 AES256 into a single algorithm bit, and to coagulate Camellia128
4546 and Camellia256 into a single algorithm bit, which has led to all
4549 Thus, among other things, the kludge introduced in 0.9.7m and
4550 0.9.8e for masking out AES256 independently of AES128 or masking
4551 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4553 With the change, we also introduce new ciphersuite aliases that
4554 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4558 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4559 Use the leftmost N bytes of the signature input if the input is
4560 larger than the prime q (with N being the size in bytes of q).
4563 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4564 it yet and it is largely untested.
4567 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4570 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4571 some compilers (gcc 4.2 and later) reject their use. Safestack is
4572 reimplemented. Update ASN1 to avoid use of legacy functions.
4575 *) Win32/64 targets are linked with Winsock2.
4578 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4579 to external functions. This can be used to increase CRL handling
4580 efficiency especially when CRLs are very large by (for example) storing
4581 the CRL revoked certificates in a database.
4584 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4585 new CRLs added to a directory can be used. New command line option
4586 -verify_return_error to s_client and s_server. This causes real errors
4587 to be returned by the verify callback instead of carrying on no matter
4588 what. This reflects the way a "real world" verify callback would behave.
4591 *) GOST engine, supporting several GOST algorithms and public key formats.
4592 Kindly donated by Cryptocom.
4595 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4596 partitioned by DP are handled but no indirect CRL or reason partitioning
4597 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4598 selected via a scoring technique which handles IDP and AKID in CRLs.
4601 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4602 will ultimately be used for all verify operations: this will remove the
4603 X509_STORE dependency on certificate verification and allow alternative
4604 lookup methods. X509_STORE based implementations of these two callbacks.
4607 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4608 Modify get_crl() to find a valid (unexpired) CRL if possible.
4611 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4612 this would be called X509_CRL_cmp() but that name is already used by
4613 a function that just compares CRL issuer names. Cache several CRL
4614 extensions in X509_CRL structure and cache CRLDP in X509.
4617 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4618 this maps equivalent X509_NAME structures into a consistent structure.
4619 Name comparison can then be performed rapidly using memcmp().
4622 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4626 *) Allow digests to supply their own micalg string for S/MIME type using
4627 the ctrl EVP_MD_CTRL_MICALG.
4630 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4631 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4632 ctrl. It can then customise the structure before and/or after signing
4636 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4637 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4638 to free up any added signature OIDs.
4641 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4642 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4643 digest and cipher tables. New options added to openssl utility:
4644 list-message-digest-algorithms and list-cipher-algorithms.
4647 *) Change the array representation of binary polynomials: the list
4648 of degrees of non-zero coefficients is now terminated with -1.
4649 Previously it was terminated with 0, which was also part of the
4650 value; thus, the array representation was not applicable to
4651 polynomials where t^0 has coefficient zero. This change makes
4652 the array representation useful in a more general context.
4655 *) Various modifications and fixes to SSL/TLS cipher string
4656 handling. For ECC, the code now distinguishes between fixed ECDH
4657 with RSA certificates on the one hand and with ECDSA certificates
4658 on the other hand, since these are separate ciphersuites. The
4659 unused code for Fortezza ciphersuites has been removed.
4661 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4662 (not "ECDHE"). For consistency with the code for DH
4663 certificates, use of ECDH certificates is now considered ECDH
4664 authentication, not RSA or ECDSA authentication (the latter is
4665 merely the CA's signing algorithm and not actively used in the
4668 The temporary ciphersuite alias "ECCdraft" is no longer
4669 available, and ECC ciphersuites are no longer excluded from "ALL"
4670 and "DEFAULT". The following aliases now exist for RFC 4492
4671 ciphersuites, most of these by analogy with the DH case:
4673 kECDHr - ECDH cert, signed with RSA
4674 kECDHe - ECDH cert, signed with ECDSA
4675 kECDH - ECDH cert (signed with either RSA or ECDSA)
4676 kEECDH - ephemeral ECDH
4677 ECDH - ECDH cert or ephemeral ECDH
4683 AECDH - anonymous ECDH
4684 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4688 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4689 Use correct micalg parameters depending on digest(s) in signed message.
4692 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4693 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4696 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4697 an engine to register a method. Add ENGINE lookups for methods and
4698 functional reference processing.
4701 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4702 EVP_{Sign,Verify}* which allow an application to customise the signature
4706 *) New -resign option to smime utility. This adds one or more signers
4707 to an existing PKCS#7 signedData structure. Also -md option to use an
4708 alternative message digest algorithm for signing.
4711 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4712 create PKCS7 structures containing multiple signers. Update smime
4713 application to support multiple signers.
4716 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4720 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4721 Reorganize PBE internals to lookup from a static table using NIDs,
4722 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4723 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4724 PRF which will be automatically used with PBES2.
4727 *) Replace the algorithm specific calls to generate keys in "req" with the
4731 *) Update PKCS#7 enveloped data routines to use new API. This is now
4732 supported by any public key method supporting the encrypt operation. A
4733 ctrl is added to allow the public key algorithm to examine or modify
4734 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4738 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4739 a default digest type to use. In most cases this will be SHA1 but some
4740 algorithms (such as GOST) need to specify an alternative digest. The
4741 return value indicates how strong the preference is 1 means optional and
4742 2 is mandatory (that is it is the only supported type). Modify
4743 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4744 use the default md. Update openssl utilities to use the default digest
4745 type for signing if it is not explicitly indicated.
4748 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4749 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4750 signing method from the key type. This effectively removes the link
4751 between digests and public key types.
4754 *) Add an OID cross reference table and utility functions. Its purpose is to
4755 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4756 rsaEncryption. This will allow some of the algorithm specific hackery
4757 needed to use the correct OID to be removed.
4760 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4761 structures for PKCS7_sign(). They are now set up by the relevant public
4765 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4768 *) Add support for key derivation (agreement) in the API, DH method and
4772 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4773 public and private key formats. As a side effect these add additional
4774 command line functionality not previously available: DSA signatures can be
4775 generated and verified using pkeyutl and DH key support and generation in
4780 [Oliver Tappe <zooey@hirschkaefer.de>]
4782 *) New make target "install_html_docs" installs HTML renditions of the
4784 [Oliver Tappe <zooey@hirschkaefer.de>]
4786 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4787 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4788 support key and parameter generation and add initial key generation
4789 functionality for RSA.
4792 *) Add functions for main EVP_PKEY_method operations. The undocumented
4793 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4794 EVP_PKEY_{encrypt,decrypt}_old.
4797 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4798 key API, doesn't do much yet.
4801 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4802 public key algorithms. New option to openssl utility:
4803 "list-public-key-algorithms" to print out info.
4806 *) Implement the Supported Elliptic Curves Extension for
4807 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4810 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4811 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4814 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4815 utilities such as rsa, dsa, dsaparam etc except they process any key
4819 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4820 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4821 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4825 *) Initial support for pluggable public key ASN1.
4826 De-spaghettify the public key ASN1 handling. Move public and private
4827 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4828 algorithm specific handling to a single module within the relevant
4829 algorithm directory. Add functions to allow (near) opaque processing
4830 of public and private key structures.
4833 *) Implement the Supported Point Formats Extension for
4834 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4837 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4838 for the psk identity [hint] and the psk callback functions to the
4839 SSL_SESSION, SSL and SSL_CTX structure.
4842 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4846 SSL_CTX_use_psk_identity_hint
4847 SSL_get_psk_identity_hint
4848 SSL_get_psk_identity
4849 SSL_use_psk_identity_hint
4851 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4853 *) Add RFC 3161 compliant time stamp request creation, response generation
4854 and response verification functionality.
4855 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4857 *) Add initial support for TLS extensions, specifically for the server_name
4858 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4859 have new members for a host name. The SSL data structure has an
4860 additional member SSL_CTX *initial_ctx so that new sessions can be
4861 stored in that context to allow for session resumption, even after the
4862 SSL has been switched to a new SSL_CTX in reaction to a client's
4863 server_name extension.
4865 New functions (subject to change):
4867 SSL_get_servername()
4868 SSL_get_servername_type()
4871 New CTRL codes and macros (subject to change):
4873 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4874 - SSL_CTX_set_tlsext_servername_callback()
4875 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4876 - SSL_CTX_set_tlsext_servername_arg()
4877 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4879 openssl s_client has a new '-servername ...' option.
4881 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4882 '-key2 ...', '-servername_fatal' (subject to change). This allows
4883 testing the HostName extension for a specific single host name ('-cert'
4884 and '-key' remain fallbacks for handshakes without HostName
4885 negotiation). If the unrecognized_name alert has to be sent, this by
4886 default is a warning; it becomes fatal with the '-servername_fatal'
4889 [Peter Sylvester, Remy Allais, Christophe Renou]
4891 *) Whirlpool hash implementation is added.
4894 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4895 bn(64,32). Because of instruction set limitations it doesn't have
4896 any negative impact on performance. This was done mostly in order
4897 to make it possible to share assembler modules, such as bn_mul_mont
4898 implementations, between 32- and 64-bit builds without hassle.
4901 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4902 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4906 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4907 dedicated Montgomery multiplication procedure, is introduced.
4908 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4909 "64-bit" performance on certain 32-bit targets.
4912 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4913 in SSL structures. New SSL ctrl to set maximum send fragment size.
4914 Save memory by setting the I/O buffer sizes dynamically instead of
4915 using the maximum available value.
4918 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4919 in addition to the text details.
4922 *) Very, very preliminary EXPERIMENTAL support for printing of general
4923 ASN1 structures. This currently produces rather ugly output and doesn't
4924 handle several customised structures at all.
4927 *) Integrated support for PVK file format and some related formats such
4928 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4929 these in the 'rsa' and 'dsa' utilities.
4932 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4935 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4936 place for the (very old) "NETSCAPE" format certificates which are now
4937 handled using new ASN1 code equivalents.
4940 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4941 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4942 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4945 *) Modify CRL distribution points extension code to print out previously
4946 unsupported fields. Enhance extension setting code to allow setting of
4950 *) Add print and set support for Issuing Distribution Point CRL extension.
4953 *) Change 'Configure' script to enable Camellia by default.
4956 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4958 *) When rejecting SSL/TLS records due to an incorrect version number, never
4959 update s->server with a new major version number. As of
4960 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4961 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4962 the previous behavior could result in a read attempt at NULL when
4963 receiving specific incorrect SSL/TLS records once record payload
4964 protection is active. (CVE-2010-0740)
4965 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4967 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4968 could be crashed if the relevant tables were not present (e.g. chrooted).
4969 [Tomas Hoger <thoger@redhat.com>]
4971 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4973 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4974 [Martin Olsson, Neel Mehta]
4976 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4977 accommodate for stack sorting, always a write lock!).
4980 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4981 excessive delays in the RAND_poll(): over a minute. As a workaround
4982 include a time check in the inner Heap32Next loop too.
4985 *) The code that handled flushing of data in SSL/TLS originally used the
4986 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4987 the problem outlined in PR#1949. The fix suggested there however can
4988 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4989 of Apache). So instead simplify the code to flush unconditionally.
4990 This should be fine since flushing with no data to flush is a no op.
4993 *) Handle TLS versions 2.0 and later properly and correctly use the
4994 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4995 off ancient servers have a habit of sticking around for a while...
4998 *) Modify compression code so it frees up structures without using the
4999 ex_data callbacks. This works around a problem where some applications
5000 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5001 restarting) then use compression (e.g. SSL with compression) later.
5002 This results in significant per-connection memory leaks and
5003 has caused some security issues including CVE-2008-1678 and
5007 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5008 change when encrypting or decrypting.
5011 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5012 connect and renegotiate with servers which do not support RI.
5013 Until RI is more widely deployed this option is enabled by default.
5016 *) Add "missing" ssl ctrls to clear options and mode.
5019 *) If client attempts to renegotiate and doesn't support RI respond with
5020 a no_renegotiation alert as required by RFC5746. Some renegotiating
5021 TLS clients will continue a connection gracefully when they receive
5022 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5023 waiting for a server hello which it will never receive. Now we treat a
5024 received no_renegotiation alert as a fatal error. This is because
5025 applications requesting a renegotiation might well expect it to succeed
5026 and would have no code in place to handle the server denying it so the
5027 only safe thing to do is to terminate the connection.
5030 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5031 peer supports secure renegotiation and 0 otherwise. Print out peer
5032 renegotiation support in s_client/s_server.
5035 *) Replace the highly broken and deprecated SPKAC certification method with
5036 the updated NID creation version. This should correctly handle UTF8.
5039 *) Implement RFC5746. Re-enable renegotiation but require the extension
5040 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5041 turns out to be a bad idea. It has been replaced by
5042 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5043 SSL_CTX_set_options(). This is really not recommended unless you
5044 know what you are doing.
5045 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5047 *) Fixes to stateless session resumption handling. Use initial_ctx when
5048 issuing and attempting to decrypt tickets in case it has changed during
5049 servername handling. Use a non-zero length session ID when attempting
5050 stateless session resumption: this makes it possible to determine if
5051 a resumption has occurred immediately after receiving server hello
5052 (several places in OpenSSL subtly assume this) instead of later in
5056 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5057 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5058 fixes for a few places where the return code is not checked
5060 [Julia Lawall <julia@diku.dk>]
5062 *) Add --strict-warnings option to Configure script to include devteam
5063 warnings in other configurations.
5066 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5067 makes it possible to install openssl libraries in locations which
5068 have names other than "lib", for example "/usr/lib64" which some
5070 [Steve Henson, based on patch from Jeremy Utley]
5072 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5073 X690 8.9.12 and can produce some misleading textual output of OIDs.
5074 [Steve Henson, reported by Dan Kaminsky]
5076 *) Delete MD2 from algorithm tables. This follows the recommendation in
5077 several standards that it is not used in new applications due to
5078 several cryptographic weaknesses. For binary compatibility reasons
5079 the MD2 API is still compiled in by default.
5082 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5086 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5087 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5089 [Guenter <lists@gknw.net>]
5091 *) Fix the server certificate chain building code to use X509_verify_cert(),
5092 it used to have an ad-hoc builder which was unable to cope with anything
5093 other than a simple chain.
5094 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5096 *) Don't check self signed certificate signatures in X509_verify_cert()
5097 by default (a flag can override this): it just wastes time without
5098 adding any security. As a useful side effect self signed root CAs
5099 with non-FIPS digests are now usable in FIPS mode.
5102 *) In dtls1_process_out_of_seq_message() the check if the current message
5103 is already buffered was missing. For every new message was memory
5104 allocated, allowing an attacker to perform an denial of service attack
5105 with sending out of seq handshake messages until there is no memory
5106 left. Additionally every future message was buffered, even if the
5107 sequence number made no sense and would be part of another handshake.
5108 So only messages with sequence numbers less than 10 in advance will be
5109 buffered. (CVE-2009-1378)
5110 [Robin Seggelmann, discovered by Daniel Mentz]
5112 *) Records are buffered if they arrive with a future epoch to be
5113 processed after finishing the corresponding handshake. There is
5114 currently no limitation to this buffer allowing an attacker to perform
5115 a DOS attack with sending records with future epochs until there is no
5116 memory left. This patch adds the pqueue_size() function to determine
5117 the size of a buffer and limits the record buffer to 100 entries.
5119 [Robin Seggelmann, discovered by Daniel Mentz]
5121 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5122 parent structure is freed. (CVE-2009-1379)
5125 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5126 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5129 [Ilya O. <vrghost@gmail.com>]
5131 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5133 *) Disable renegotiation completely - this fixes a severe security
5134 problem (CVE-2009-3555) at the cost of breaking all
5135 renegotiation. Renegotiation can be re-enabled by setting
5136 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5137 run-time. This is really not recommended unless you know what
5141 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5143 *) Don't set val to NULL when freeing up structures, it is freed up by
5144 underlying code. If sizeof(void *) > sizeof(long) this can result in
5145 zeroing past the valid field. (CVE-2009-0789)
5146 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5148 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5149 checked correctly. This would allow some invalid signed attributes to
5150 appear to verify correctly. (CVE-2009-0591)
5151 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5153 *) Reject UniversalString and BMPString types with invalid lengths. This
5154 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5155 a legal length. (CVE-2009-0590)
5158 *) Set S/MIME signing as the default purpose rather than setting it
5159 unconditionally. This allows applications to override it at the store
5163 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5164 to handle some structures.
5167 *) Improve efficiency of mem_gets: don't search whole buffer each time
5169 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5171 *) New -hex option for openssl rand.
5174 *) Print out UTF8String and NumericString when parsing ASN1.
5177 *) Support NumericString type for name components.
5180 *) Allow CC in the environment to override the automatically chosen
5181 compiler. Note that nothing is done to ensure flags work with the
5185 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5187 *) Properly check EVP_VerifyFinal() and similar return values
5189 [Ben Laurie, Bodo Moeller, Google Security Team]
5191 *) Enable TLS extensions by default.
5194 *) Allow the CHIL engine to be loaded, whether the application is
5195 multithreaded or not. (This does not release the developer from the
5196 obligation to set up the dynamic locking callbacks.)
5197 [Sander Temme <sander@temme.net>]
5199 *) Use correct exit code if there is an error in dgst command.
5200 [Steve Henson; problem pointed out by Roland Dirlewanger]
5202 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5203 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5206 *) Add experimental JPAKE support, including demo authentication in
5207 s_client and s_server.
5210 *) Set the comparison function in v3_addr_canonize().
5211 [Rob Austein <sra@hactrn.net>]
5213 *) Add support for XMPP STARTTLS in s_client.
5214 [Philip Paeps <philip@freebsd.org>]
5216 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5217 to ensure that even with this option, only ciphersuites in the
5218 server's preference list will be accepted. (Note that the option
5219 applies only when resuming a session, so the earlier behavior was
5220 just about the algorithm choice for symmetric cryptography.)
5223 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5225 *) Fix NULL pointer dereference if a DTLS server received
5226 ChangeCipherSpec as first record (CVE-2009-1386).
5229 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5230 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5233 *) The fix in 0.9.8c that supposedly got rid of unsafe
5234 double-checked locking was incomplete for RSA blinding,
5235 addressing just one layer of what turns out to have been
5236 doubly unsafe triple-checked locking.
5238 So now fix this for real by retiring the MONT_HELPER macro
5239 in crypto/rsa/rsa_eay.c.
5241 [Bodo Moeller; problem pointed out by Marius Schilder]
5243 *) Various precautionary measures:
5245 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5247 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5248 (NB: This would require knowledge of the secret session ticket key
5249 to exploit, in which case you'd be SOL either way.)
5251 - Change bn_nist.c so that it will properly handle input BIGNUMs
5252 outside the expected range.
5254 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5257 [Neel Mehta, Bodo Moeller]
5259 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5260 the load fails. Useful for distros.
5261 [Ben Laurie and the FreeBSD team]
5263 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5266 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5269 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5271 This work was sponsored by Logica.
5274 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5275 keystores. Support for SSL/TLS client authentication too.
5276 Not compiled unless enable-capieng specified to Configure.
5278 This work was sponsored by Logica.
5281 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5282 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5283 attribute creation routines such as certificate requests and PKCS#12
5287 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5289 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5290 handshake which could lead to a client crash as found using the
5291 Codenomicon TLS test suite (CVE-2008-1672)
5292 [Steve Henson, Mark Cox]
5294 *) Fix double free in TLS server name extensions which could lead to
5295 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5298 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5300 Clear the error queue to ensure that error entries left from
5301 older function calls do not interfere with the correct operation.
5302 [Lutz Jaenicke, Erik de Castro Lopo]
5304 *) Remove root CA certificates of commercial CAs:
5306 The OpenSSL project does not recommend any specific CA and does not
5307 have any policy with respect to including or excluding any CA.
5308 Therefore it does not make any sense to ship an arbitrary selection
5309 of root CA certificates with the OpenSSL software.
5312 *) RSA OAEP patches to fix two separate invalid memory reads.
5313 The first one involves inputs when 'lzero' is greater than
5314 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5315 before the beginning of from). The second one involves inputs where
5316 the 'db' section contains nothing but zeroes (there is a one-byte
5317 invalid read after the end of 'db').
5318 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5320 *) Partial backport from 0.9.9-dev:
5322 Introduce bn_mul_mont (dedicated Montgomery multiplication
5323 procedure) as a candidate for BIGNUM assembler implementation.
5324 While 0.9.9-dev uses assembler for various architectures, only
5325 x86_64 is available by default here in the 0.9.8 branch, and
5326 32-bit x86 is available through a compile-time setting.
5328 To try the 32-bit x86 assembler implementation, use Configure
5329 option "enable-montasm" (which exists only for this backport).
5331 As "enable-montasm" for 32-bit x86 disclaims code stability
5332 anyway, in this constellation we activate additional code
5333 backported from 0.9.9-dev for further performance improvements,
5334 namely BN_from_montgomery_word. (To enable this otherwise,
5335 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5337 [Andy Polyakov (backport partially by Bodo Moeller)]
5339 *) Add TLS session ticket callback. This allows an application to set
5340 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5341 values. This is useful for key rollover for example where several key
5342 sets may exist with different names.
5345 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5346 This was broken until now in 0.9.8 releases, such that the only way
5347 a registered ENGINE could be used (assuming it initialises
5348 successfully on the host) was to explicitly set it as the default
5349 for the relevant algorithms. This is in contradiction with 0.9.7
5350 behaviour and the documentation. With this fix, when an ENGINE is
5351 registered into a given algorithm's table of implementations, the
5352 'uptodate' flag is reset so that auto-discovery will be used next
5353 time a new context for that algorithm attempts to select an
5355 [Ian Lister (tweaked by Geoff Thorpe)]
5357 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5358 implementation in the following ways:
5360 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5363 Lack of BER streaming support means one pass streaming processing is
5364 only supported if data is detached: setting the streaming flag is
5365 ignored for embedded content.
5367 CMS support is disabled by default and must be explicitly enabled
5368 with the enable-cms configuration option.
5371 *) Update the GMP engine glue to do direct copies between BIGNUM and
5372 mpz_t when openssl and GMP use the same limb size. Otherwise the
5373 existing "conversion via a text string export" trick is still used.
5374 [Paul Sheer <paulsheer@gmail.com>]
5376 *) Zlib compression BIO. This is a filter BIO which compressed and
5377 uncompresses any data passed through it.
5380 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5381 RFC3394 compatible AES key wrapping.
5384 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5385 sets string data without copying. X509_ALGOR_set0() and
5386 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5387 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5388 from an X509_ATTRIBUTE structure optionally checking it occurs only
5389 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5393 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5394 to get the expected BN_FLG_CONSTTIME behavior.
5395 [Bodo Moeller (Google)]
5399 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5400 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5401 - added some more tests to do_tests.pl
5402 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5403 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5404 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5405 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5406 - various changes to netware.pl to enable gcc-cross builds on Win32
5408 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5409 - various changes to fix missing prototype warnings
5410 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5411 - added AES, WHIRLPOOL and CPUID assembler code to build files
5412 - added missing AES assembler make rules to mk1mf.pl
5413 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5414 [Guenter Knauf <eflash@gmx.net>]
5416 *) Implement certificate status request TLS extension defined in RFC3546.
5417 A client can set the appropriate parameters and receive the encoded
5418 OCSP response via a callback. A server can query the supplied parameters
5419 and set the encoded OCSP response in the callback. Add simplified examples
5420 to s_client and s_server.
5423 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5425 *) Fix various bugs:
5426 + Binary incompatibility of ssl_ctx_st structure
5427 + DTLS interoperation with non-compliant servers
5428 + Don't call get_session_cb() without proposed session
5429 + Fix ia64 assembler code
5430 [Andy Polyakov, Steve Henson]
5432 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5434 *) DTLS Handshake overhaul. There were longstanding issues with
5435 OpenSSL DTLS implementation, which were making it impossible for
5436 RFC 4347 compliant client to communicate with OpenSSL server.
5437 Unfortunately just fixing these incompatibilities would "cut off"
5438 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5439 server keeps tolerating non RFC compliant syntax. The opposite is
5440 not true, 0.9.8f client can not communicate with earlier server.
5441 This update even addresses CVE-2007-4995.
5444 *) Changes to avoid need for function casts in OpenSSL: some compilers
5445 (gcc 4.2 and later) reject their use.
5446 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5449 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5450 RFC4507bis. The encrypted ticket format is an encrypted encoded
5451 SSL_SESSION structure, that way new session features are automatically
5454 If a client application caches session in an SSL_SESSION structure
5455 support is transparent because tickets are now stored in the encoded
5458 The SSL_CTX structure automatically generates keys for ticket
5459 protection in servers so again support should be possible
5460 with no application modification.
5462 If a client or server wishes to disable RFC4507 support then the option
5463 SSL_OP_NO_TICKET can be set.
5465 Add a TLS extension debugging callback to allow the contents of any client
5466 or server extensions to be examined.
5468 This work was sponsored by Google.
5471 *) Add initial support for TLS extensions, specifically for the server_name
5472 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5473 have new members for a host name. The SSL data structure has an
5474 additional member SSL_CTX *initial_ctx so that new sessions can be
5475 stored in that context to allow for session resumption, even after the
5476 SSL has been switched to a new SSL_CTX in reaction to a client's
5477 server_name extension.
5479 New functions (subject to change):
5481 SSL_get_servername()
5482 SSL_get_servername_type()
5485 New CTRL codes and macros (subject to change):
5487 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5488 - SSL_CTX_set_tlsext_servername_callback()
5489 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5490 - SSL_CTX_set_tlsext_servername_arg()
5491 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5493 openssl s_client has a new '-servername ...' option.
5495 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5496 '-key2 ...', '-servername_fatal' (subject to change). This allows
5497 testing the HostName extension for a specific single host name ('-cert'
5498 and '-key' remain fallbacks for handshakes without HostName
5499 negotiation). If the unrecognized_name alert has to be sent, this by
5500 default is a warning; it becomes fatal with the '-servername_fatal'
5503 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5505 *) Add AES and SSE2 assembly language support to VC++ build.
5508 *) Mitigate attack on final subtraction in Montgomery reduction.
5511 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5512 (which previously caused an internal error).
5515 *) Squeeze another 10% out of IGE mode when in != out.
5518 *) AES IGE mode speedup.
5519 [Dean Gaudet (Google)]
5521 *) Add the Korean symmetric 128-bit cipher SEED (see
5522 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5523 add SEED ciphersuites from RFC 4162:
5525 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5526 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5527 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5528 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5530 To minimize changes between patchlevels in the OpenSSL 0.9.8
5531 series, SEED remains excluded from compilation unless OpenSSL
5532 is configured with 'enable-seed'.
5533 [KISA, Bodo Moeller]
5535 *) Mitigate branch prediction attacks, which can be practical if a
5536 single processor is shared, allowing a spy process to extract
5537 information. For detailed background information, see
5538 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5539 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5540 and Necessary Software Countermeasures"). The core of the change
5541 are new versions BN_div_no_branch() and
5542 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5543 respectively, which are slower, but avoid the security-relevant
5544 conditional branches. These are automatically called by BN_div()
5545 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5546 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5547 remove a conditional branch.
5549 BN_FLG_CONSTTIME is the new name for the previous
5550 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5551 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5552 in the exponent causes BN_mod_exp_mont() to use the alternative
5553 implementation in BN_mod_exp_mont_consttime().) The old name
5554 remains as a deprecated alias.
5556 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5557 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5558 constant-time implementations for more than just exponentiation.
5559 Here too the old name is kept as a deprecated alias.
5561 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5562 the BN_BLINDING structure gets an independent copy of the
5563 modulus. This means that the previous "BIGNUM *m" argument to
5564 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5565 essentially becomes "const BIGNUM *m", although we can't actually
5566 change this in the header file before 0.9.9. It allows
5567 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5568 enable BN_FLG_CONSTTIME.
5570 [Matthew D Wood (Intel Corp)]
5572 *) In the SSL/TLS server implementation, be strict about session ID
5573 context matching (which matters if an application uses a single
5574 external cache for different purposes). Previously,
5575 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5576 set. This did ensure strict client verification, but meant that,
5577 with applications using a single external cache for quite
5578 different requirements, clients could circumvent ciphersuite
5579 restrictions for a given session ID context by starting a session
5580 in a different context.
5583 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5584 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5585 authentication-only ciphersuites.
5588 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5589 not complete and could lead to a possible single byte overflow
5590 (CVE-2007-5135) [Ben Laurie]
5592 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5594 *) Since AES128 and AES256 (and similarly Camellia128 and
5595 Camellia256) share a single mask bit in the logic of
5596 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5597 kludge to work properly if AES128 is available and AES256 isn't
5598 (or if Camellia128 is available and Camellia256 isn't).
5601 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5602 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5603 When a point or a seed is encoded in a BIT STRING, we need to
5604 prevent the removal of trailing zero bits to get the proper DER
5605 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5606 of a NamedBitList, for which trailing 0 bits need to be removed.)
5609 *) Have SSL/TLS server implementation tolerate "mismatched" record
5610 protocol version while receiving ClientHello even if the
5611 ClientHello is fragmented. (The server can't insist on the
5612 particular protocol version it has chosen before the ServerHello
5613 message has informed the client about his choice.)
5616 *) Add RFC 3779 support.
5617 [Rob Austein for ARIN, Ben Laurie]
5619 *) Load error codes if they are not already present instead of using a
5620 static variable. This allows them to be cleanly unloaded and reloaded.
5621 Improve header file function name parsing.
5624 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5625 or CAPABILITY handshake as required by RFCs.
5628 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5630 *) Introduce limits to prevent malicious keys being able to
5631 cause a denial of service. (CVE-2006-2940)
5632 [Steve Henson, Bodo Moeller]
5634 *) Fix ASN.1 parsing of certain invalid structures that can result
5635 in a denial of service. (CVE-2006-2937) [Steve Henson]
5637 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5638 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5640 *) Fix SSL client code which could crash if connecting to a
5641 malicious SSLv2 server. (CVE-2006-4343)
5642 [Tavis Ormandy and Will Drewry, Google Security Team]
5644 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5645 match only those. Before that, "AES256-SHA" would be interpreted
5646 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5647 the same strength classification in 0.9.7h) as we currently only
5648 have a single AES bit in the ciphersuite description bitmap.
5649 That change, however, also applied to ciphersuite strings such as
5650 "RC4-MD5" that intentionally matched multiple ciphersuites --
5651 namely, SSL 2.0 ciphersuites in addition to the more common ones
5652 from SSL 3.0/TLS 1.0.
5654 So we change the selection algorithm again: Naming an explicit
5655 ciphersuite selects this one ciphersuite, and any other similar
5656 ciphersuite (same bitmap) from *other* protocol versions.
5657 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5658 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5660 Since SSL 2.0 does not have any ciphersuites for which the
5661 128/256 bit distinction would be relevant, this works for now.
5662 The proper fix will be to use different bits for AES128 and
5663 AES256, which would have avoided the problems from the beginning;
5664 however, bits are scarce, so we can only do this in a new release
5665 (not just a patchlevel) when we can change the SSL_CIPHER
5666 definition to split the single 'unsigned long mask' bitmap into
5667 multiple values to extend the available space.
5671 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5673 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5674 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5676 *) Add AES IGE and biIGE modes.
5679 *) Change the Unix randomness entropy gathering to use poll() when
5680 possible instead of select(), since the latter has some
5681 undesirable limitations.
5682 [Darryl Miles via Richard Levitte and Bodo Moeller]
5684 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5685 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5686 cannot be implicitly activated as part of, e.g., the "AES" alias.
5687 However, please upgrade to OpenSSL 0.9.9[-dev] for
5688 non-experimental use of the ECC ciphersuites to get TLS extension
5689 support, which is required for curve and point format negotiation
5690 to avoid potential handshake problems.
5693 *) Disable rogue ciphersuites:
5695 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5696 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5697 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5699 The latter two were purportedly from
5700 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5703 Also deactivate the remaining ciphersuites from
5704 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5705 unofficial, and the ID has long expired.
5708 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5709 dual-core machines) and other potential thread-safety issues.
5712 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5713 versions), which is now available for royalty-free use
5714 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5715 Also, add Camellia TLS ciphersuites from RFC 4132.
5717 To minimize changes between patchlevels in the OpenSSL 0.9.8
5718 series, Camellia remains excluded from compilation unless OpenSSL
5719 is configured with 'enable-camellia'.
5722 *) Disable the padding bug check when compression is in use. The padding
5723 bug check assumes the first packet is of even length, this is not
5724 necessarily true if compression is enabled and can result in false
5725 positives causing handshake failure. The actual bug test is ancient
5726 code so it is hoped that implementations will either have fixed it by
5727 now or any which still have the bug do not support compression.
5730 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5732 *) When applying a cipher rule check to see if string match is an explicit
5733 cipher suite and only match that one cipher suite if it is.
5736 *) Link in manifests for VC++ if needed.
5737 [Austin Ziegler <halostatue@gmail.com>]
5739 *) Update support for ECC-based TLS ciphersuites according to
5740 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5741 TLS extensions, which are supported starting with the 0.9.9
5742 branch, not in the OpenSSL 0.9.8 branch).
5745 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5746 opaque EVP_CIPHER_CTX handling.
5749 *) Fixes and enhancements to zlib compression code. We now only use
5750 "zlib1.dll" and use the default __cdecl calling convention on Win32
5751 to conform with the standards mentioned here:
5752 http://www.zlib.net/DLL_FAQ.txt
5753 Static zlib linking now works on Windows and the new --with-zlib-include
5754 --with-zlib-lib options to Configure can be used to supply the location
5755 of the headers and library. Gracefully handle case where zlib library
5759 *) Several fixes and enhancements to the OID generation code. The old code
5760 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5761 handle numbers larger than ULONG_MAX, truncated printing and had a
5762 non standard OBJ_obj2txt() behaviour.
5765 *) Add support for building of engines under engine/ as shared libraries
5766 under VC++ build system.
5769 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5770 Hopefully, we will not see any false combination of paths any more.
5773 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5775 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5776 (part of SSL_OP_ALL). This option used to disable the
5777 countermeasure against man-in-the-middle protocol-version
5778 rollback in the SSL 2.0 server implementation, which is a bad
5779 idea. (CVE-2005-2969)
5781 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5782 for Information Security, National Institute of Advanced Industrial
5783 Science and Technology [AIST], Japan)]
5785 *) Add two function to clear and return the verify parameter flags.
5788 *) Keep cipherlists sorted in the source instead of sorting them at
5789 runtime, thus removing the need for a lock.
5792 *) Avoid some small subgroup attacks in Diffie-Hellman.
5793 [Nick Mathewson and Ben Laurie]
5795 *) Add functions for well-known primes.
5798 *) Extended Windows CE support.
5799 [Satoshi Nakamura and Andy Polyakov]
5801 *) Initialize SSL_METHOD structures at compile time instead of during
5802 runtime, thus removing the need for a lock.
5805 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5806 attempting to decrypt each encrypted key in turn. Add support to
5810 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5812 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5815 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5818 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5819 key into the same file any more.
5822 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5825 *) Add -utf8 command line and config file option to 'ca'.
5826 [Stefan <stf@udoma.org]
5828 *) Removed the macro des_crypt(), as it seems to conflict with some
5829 libraries. Use DES_crypt().
5832 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5833 involves renaming the source and generated shared-libs for
5834 both. The engines will accept the corrected or legacy ids
5835 ('ncipher' and '4758_cca' respectively) when binding. NB,
5836 this only applies when building 'shared'.
5837 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5839 *) Add attribute functions to EVP_PKEY structure. Modify
5840 PKCS12_create() to recognize a CSP name attribute and
5841 use it. Make -CSP option work again in pkcs12 utility.
5844 *) Add new functionality to the bn blinding code:
5845 - automatic re-creation of the BN_BLINDING parameters after
5846 a fixed number of uses (currently 32)
5847 - add new function for parameter creation
5848 - introduce flags to control the update behaviour of the
5849 BN_BLINDING parameters
5850 - hide BN_BLINDING structure
5851 Add a second BN_BLINDING slot to the RSA structure to improve
5852 performance when a single RSA object is shared among several
5856 *) Add support for DTLS.
5857 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5859 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5860 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5863 *) Remove buggy and incomplete DH cert support from
5864 ssl/ssl_rsa.c and ssl/s3_both.c
5867 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5868 the apps/openssl applications.
5871 *) Compile clean with "-Wall -Wmissing-prototypes
5872 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5873 DEBUG_SAFESTACK must also be set.
5876 *) Change ./Configure so that certain algorithms can be disabled by default.
5877 The new counterpiece to "no-xxx" is "enable-xxx".
5879 The patented RC5 and MDC2 algorithms will now be disabled unless
5880 "enable-rc5" and "enable-mdc2", respectively, are specified.
5882 (IDEA remains enabled despite being patented. This is because IDEA
5883 is frequently required for interoperability, and there is no license
5884 fee for non-commercial use. As before, "no-idea" can be used to
5885 avoid this algorithm.)
5889 *) Add processing of proxy certificates (see RFC 3820). This work was
5890 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5891 EGEE (Enabling Grids for E-science in Europe).
5894 *) RC4 performance overhaul on modern architectures/implementations, such
5895 as Intel P4, IA-64 and AMD64.
5898 *) New utility extract-section.pl. This can be used specify an alternative
5899 section number in a pod file instead of having to treat each file as
5900 a separate case in Makefile. This can be done by adding two lines to the
5903 =for comment openssl_section:XXX
5905 The blank line is mandatory.
5909 *) New arguments -certform, -keyform and -pass for s_client and s_server
5910 to allow alternative format key and certificate files and passphrase
5914 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5915 update associated structures and add various utility functions.
5917 Add new policy related verify parameters, include policy checking in
5918 standard verify code. Enhance 'smime' application with extra parameters
5919 to support policy checking and print out.
5922 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5923 Nehemiah processors. These extensions support AES encryption in hardware
5924 as well as RNG (though RNG support is currently disabled).
5925 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5927 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5930 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5931 [Andy Polyakov and a number of other people]
5933 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5934 implementation contributed by IBM.
5935 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5937 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5938 exponent rather than 'unsigned long'. There is a corresponding change to
5939 the new 'rsa_keygen' element of the RSA_METHOD structure.
5940 [Jelte Jansen, Geoff Thorpe]
5942 *) Functionality for creating the initial serial number file is now
5943 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5945 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5946 number file to 1, which is bound to cause problems. To avoid
5947 the problems while respecting compatibility between different 0.9.7
5948 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5949 CA.pl for serial number initialization. With the new release 0.9.8,
5950 we can fix the problem directly in the 'ca' utility.)
5953 *) Reduced header interdependencies by declaring more opaque objects in
5954 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5955 give fewer recursive includes, which could break lazy source code - so
5956 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5957 developers should define this symbol when building and using openssl to
5958 ensure they track the recommended behaviour, interfaces, [etc], but
5959 backwards-compatible behaviour prevails when this isn't defined.
5962 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5965 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5966 This will generate a random key of the appropriate length based on the
5967 cipher context. The EVP_CIPHER can provide its own random key generation
5968 routine to support keys of a specific form. This is used in the des and
5969 3des routines to generate a key of the correct parity. Update S/MIME
5970 code to use new functions and hence generate correct parity DES keys.
5971 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5972 valid (weak or incorrect parity).
5975 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5976 as looking them up. This is useful when the verified structure may contain
5977 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5978 present unless the new PKCS7_NO_CRL flag is asserted.
5981 *) Extend ASN1 oid configuration module. It now additionally accepts the
5984 shortName = some long name, 1.2.3.4
5987 *) Reimplemented the BN_CTX implementation. There is now no more static
5988 limitation on the number of variables it can handle nor the depth of the
5989 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5990 information can now expand as required, and rather than having a single
5991 static array of bignums, BN_CTX now uses a linked-list of such arrays
5992 allowing it to expand on demand whilst maintaining the usefulness of
5993 BN_CTX's "bundling".
5996 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5997 to allow all RSA operations to function using a single BN_CTX.
6000 *) Preliminary support for certificate policy evaluation and checking. This
6001 is initially intended to pass the tests outlined in "Conformance Testing
6002 of Relying Party Client Certificate Path Processing Logic" v1.07.
6005 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6006 remained unused and not that useful. A variety of other little bignum
6007 tweaks and fixes have also been made continuing on from the audit (see
6011 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6012 associated ASN1, EVP and SSL functions and old ASN1 macros.
6015 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6016 and this should never fail. So the return value from the use of
6017 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6018 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6021 *) BN_CTX_get() should return zero-valued bignums, providing the same
6022 initialised value as BN_new().
6023 [Geoff Thorpe, suggested by Ulf Möller]
6025 *) Support for inhibitAnyPolicy certificate extension.
6028 *) An audit of the BIGNUM code is underway, for which debugging code is
6029 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6030 is considered valid when processing BIGNUMs, and causes execution to
6031 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6032 further steps are taken to deliberately pollute unused data in BIGNUM
6033 structures to try and expose faulty code further on. For now, openssl will
6034 (in its default mode of operation) continue to tolerate the inconsistent
6035 forms that it has tolerated in the past, but authors and packagers should
6036 consider trying openssl and their own applications when compiled with
6037 these debugging symbols defined. It will help highlight potential bugs in
6038 their own code, and will improve the test coverage for OpenSSL itself. At
6039 some point, these tighter rules will become openssl's default to improve
6040 maintainability, though the assert()s and other overheads will remain only
6041 in debugging configurations. See bn.h for more details.
6042 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6044 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6045 that can only be obtained through BN_CTX_new() (which implicitly
6046 initialises it). The presence of this function only made it possible
6047 to overwrite an existing structure (and cause memory leaks).
6050 *) Because of the callback-based approach for implementing LHASH as a
6051 template type, lh_insert() adds opaque objects to hash-tables and
6052 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6053 to clean up those corresponding objects before destroying the hash table
6054 (and losing the object pointers). So some over-zealous constifications in
6055 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6056 objects as "const" and the lh_doall[_arg] callback wrappers are not
6057 prototyped to have "const" restrictions on the object pointers they are
6058 given (and so aren't required to cast them away any more).
6061 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6062 (speed) prefers to use its own implementation. The two implementations
6063 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6064 its object type properly exposed (MS_TM) instead of casting to/from "char
6065 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6066 aren't necessarily the greatest nomenclatures - but this is what was used
6067 internally to the implementation so I've used that for now.
6070 *) Ensure that deprecated functions do not get compiled when
6071 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6072 the self-tests were still using deprecated key-generation functions so
6073 these have been updated also.
6076 *) Reorganise PKCS#7 code to separate the digest location functionality
6077 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6078 New function PKCS7_set_digest() to set the digest type for PKCS#7
6079 digestedData type. Add additional code to correctly generate the
6080 digestedData type and add support for this type in PKCS7 initialization
6084 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6085 structure of type "other".
6088 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6089 sure the loop does correctly stop and breaking ("division by zero")
6090 modulus operations are not performed. The (pre-generated) prime
6091 table crypto/bn/bn_prime.h was already correct, but it could not be
6092 re-generated on some platforms because of the "division by zero"
6093 situation in the script.
6094 [Ralf S. Engelschall]
6096 *) Update support for ECC-based TLS ciphersuites according to
6097 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6098 SHA-1 now is only used for "small" curves (where the
6099 representation of a field element takes up to 24 bytes); for
6100 larger curves, the field element resulting from ECDH is directly
6101 used as premaster secret.
6102 [Douglas Stebila (Sun Microsystems Laboratories)]
6104 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6105 curve secp160r1 to the tests.
6106 [Douglas Stebila (Sun Microsystems Laboratories)]
6108 *) Add the possibility to load symbols globally with DSO.
6109 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6111 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6112 control of the error stack.
6115 *) Add support for STORE in ENGINE.
6118 *) Add the STORE type. The intention is to provide a common interface
6119 to certificate and key stores, be they simple file-based stores, or
6120 HSM-type store, or LDAP stores, or...
6121 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6124 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6125 pass a list of arguments to any function as well as provide a way
6126 for a function to pass data back to the caller.
6129 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6130 works like BUF_strdup() but can be used to duplicate a portion of
6131 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6135 *) Add the function sk_find_ex() which works like sk_find(), but will
6136 return an index to an element even if an exact match couldn't be
6137 found. The index is guaranteed to point at the element where the
6138 searched-for key would be inserted to preserve sorting order.
6141 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6142 takes an extra flags argument for optional functionality. Currently,
6143 the following flags are defined:
6145 OBJ_BSEARCH_VALUE_ON_NOMATCH
6146 This one gets OBJ_bsearch_ex() to return a pointer to the first
6147 element where the comparing function returns a negative or zero
6150 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6151 This one gets OBJ_bsearch_ex() to return a pointer to the first
6152 element where the comparing function returns zero. This is useful
6153 if there are more than one element where the comparing function
6157 *) Make it possible to create self-signed certificates with 'openssl ca'
6158 in such a way that the self-signed certificate becomes part of the
6159 CA database and uses the same mechanisms for serial number generation
6160 as all other certificate signing. The new flag '-selfsign' enables
6161 this functionality. Adapt CA.sh and CA.pl.in.
6164 *) Add functionality to check the public key of a certificate request
6165 against a given private. This is useful to check that a certificate
6166 request can be signed by that key (self-signing).
6169 *) Make it possible to have multiple active certificates with the same
6170 subject in the CA index file. This is done only if the keyword
6171 'unique_subject' is set to 'no' in the main CA section (default
6172 if 'CA_default') of the configuration file. The value is saved
6173 with the database itself in a separate index attribute file,
6174 named like the index file with '.attr' appended to the name.
6177 *) Generate multi-valued AVAs using '+' notation in config files for
6181 *) Support for nameConstraints certificate extension.
6184 *) Support for policyConstraints certificate extension.
6187 *) Support for policyMappings certificate extension.
6190 *) Make sure the default DSA_METHOD implementation only uses its
6191 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6192 and change its own handlers to be NULL so as to remove unnecessary
6193 indirection. This lets alternative implementations fallback to the
6194 default implementation more easily.
6197 *) Support for directoryName in GeneralName related extensions
6201 *) Make it possible to link applications using Makefile.shared.
6202 Make that possible even when linking against static libraries!
6205 *) Support for single pass processing for S/MIME signing. This now
6206 means that S/MIME signing can be done from a pipe, in addition
6207 cleartext signing (multipart/signed type) is effectively streaming
6208 and the signed data does not need to be all held in memory.
6210 This is done with a new flag PKCS7_STREAM. When this flag is set
6211 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6212 is done after the data is output (and digests calculated) in
6213 SMIME_write_PKCS7().
6216 *) Add full support for -rpath/-R, both in shared libraries and
6217 applications, at least on the platforms where it's known how
6221 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6222 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6223 will now compute a table of multiples of the generator that
6224 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6225 faster (notably in the case of a single point multiplication,
6226 scalar * generator).
6227 [Nils Larsch, Bodo Moeller]
6229 *) IPv6 support for certificate extensions. The various extensions
6230 which use the IP:a.b.c.d can now take IPv6 addresses using the
6231 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6235 *) Added an ENGINE that implements RSA by performing private key
6236 exponentiations with the GMP library. The conversions to and from
6237 GMP's mpz_t format aren't optimised nor are any montgomery forms
6238 cached, and on x86 it appears OpenSSL's own performance has caught up.
6239 However there are likely to be other architectures where GMP could
6240 provide a boost. This ENGINE is not built in by default, but it can be
6241 specified at Configure time and should be accompanied by the necessary
6242 linker additions, eg;
6243 ./config -DOPENSSL_USE_GMP -lgmp
6246 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6247 testing availability of engines with "-t" - the old behaviour is
6248 produced by increasing the feature's verbosity with "-tt".
6251 *) ECDSA routines: under certain error conditions uninitialized BN objects
6252 could be freed. Solution: make sure initialization is performed early
6253 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6257 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6258 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6259 software implementations. For DSA and DH, parameter generation can
6260 also be overridden by providing the appropriate method callbacks.
6263 *) Change the "progress" mechanism used in key-generation and
6264 primality testing to functions that take a new BN_GENCB pointer in
6265 place of callback/argument pairs. The new API functions have "_ex"
6266 postfixes and the older functions are reimplemented as wrappers for
6267 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6268 declarations of the old functions to help (graceful) attempts to
6269 migrate to the new functions. Also, the new key-generation API
6270 functions operate on a caller-supplied key-structure and return
6271 success/failure rather than returning a key or NULL - this is to
6272 help make "keygen" another member function of RSA_METHOD etc.
6274 Example for using the new callback interface:
6276 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6280 BN_GENCB_set(&my_cb, my_callback, my_arg);
6282 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6283 /* For the meaning of a, b in calls to my_callback(), see the
6284 * documentation of the function that calls the callback.
6285 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6286 * my_callback should return 1 if it wants BN_is_prime_ex()
6287 * to continue, or 0 to stop.
6292 *) Change the ZLIB compression method to be stateful, and make it
6293 available to TLS with the number defined in
6294 draft-ietf-tls-compression-04.txt.
6297 *) Add the ASN.1 structures and functions for CertificatePair, which
6298 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6300 CertificatePair ::= SEQUENCE {
6301 forward [0] Certificate OPTIONAL,
6302 reverse [1] Certificate OPTIONAL,
6303 -- at least one of the pair shall be present -- }
6305 Also implement the PEM functions to read and write certificate
6306 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6308 This needed to be defined, mostly for the sake of the LDAP
6309 attribute crossCertificatePair, but may prove useful elsewhere as
6313 *) Make it possible to inhibit symlinking of shared libraries in
6314 Makefile.shared, for Cygwin's sake.
6317 *) Extend the BIGNUM API by creating a function
6318 void BN_set_negative(BIGNUM *a, int neg);
6319 and a macro that behave like
6320 int BN_is_negative(const BIGNUM *a);
6322 to avoid the need to access 'a->neg' directly in applications.
6325 *) Implement fast modular reduction for pseudo-Mersenne primes
6326 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6327 EC_GROUP_new_curve_GFp() will now automatically use this
6329 [Nils Larsch <nla@trustcenter.de>]
6331 *) Add new lock type (CRYPTO_LOCK_BN).
6334 *) Change the ENGINE framework to automatically load engines
6335 dynamically from specific directories unless they could be
6336 found to already be built in or loaded. Move all the
6337 current engines except for the cryptodev one to a new
6339 The engines in engines/ are built as shared libraries if
6340 the "shared" options was given to ./Configure or ./config.
6341 Otherwise, they are inserted in libcrypto.a.
6342 /usr/local/ssl/engines is the default directory for dynamic
6343 engines, but that can be overridden at configure time through
6344 the usual use of --prefix and/or --openssldir, and at run
6345 time with the environment variable OPENSSL_ENGINES.
6346 [Geoff Thorpe and Richard Levitte]
6348 *) Add Makefile.shared, a helper makefile to build shared
6349 libraries. Adapt Makefile.org.
6352 *) Add version info to Win32 DLLs.
6353 [Peter 'Luna' Runestig" <peter@runestig.com>]
6355 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6356 can be added using this API to created arbitrary PKCS#12
6357 files while avoiding the low level API.
6359 New options to PKCS12_create(), key or cert can be NULL and
6360 will then be omitted from the output file. The encryption
6361 algorithm NIDs can be set to -1 for no encryption, the mac
6362 iteration count can be set to 0 to omit the mac.
6364 Enhance pkcs12 utility by making the -nokeys and -nocerts
6365 options work when creating a PKCS#12 file. New option -nomac
6366 to omit the mac, NONE can be set for an encryption algorithm.
6367 New code is modified to use the enhanced PKCS12_create()
6368 instead of the low level API.
6371 *) Extend ASN1 encoder to support indefinite length constructed
6372 encoding. This can output sequences tags and octet strings in
6373 this form. Modify pk7_asn1.c to support indefinite length
6374 encoding. This is experimental and needs additional code to
6375 be useful, such as an ASN1 bio and some enhanced streaming
6378 Extend template encode functionality so that tagging is passed
6379 down to the template encoder.
6382 *) Let 'openssl req' fail if an argument to '-newkey' is not
6383 recognized instead of using RSA as a default.
6386 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6387 As these are not official, they are not included in "ALL";
6388 the "ECCdraft" ciphersuite group alias can be used to select them.
6389 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6391 *) Add ECDH engine support.
6392 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6394 *) Add ECDH in new directory crypto/ecdh/.
6395 [Douglas Stebila (Sun Microsystems Laboratories)]
6397 *) Let BN_rand_range() abort with an error after 100 iterations
6398 without success (which indicates a broken PRNG).
6401 *) Change BN_mod_sqrt() so that it verifies that the input value
6402 is really the square of the return value. (Previously,
6403 BN_mod_sqrt would show GIGO behaviour.)
6406 *) Add named elliptic curves over binary fields from X9.62, SECG,
6407 and WAP/WTLS; add OIDs that were still missing.
6409 [Sheueling Chang Shantz and Douglas Stebila
6410 (Sun Microsystems Laboratories)]
6412 *) Extend the EC library for elliptic curves over binary fields
6413 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6416 EC_GF2m_simple_method
6420 EC_GROUP_new_curve_GF2m
6421 EC_GROUP_set_curve_GF2m
6422 EC_GROUP_get_curve_GF2m
6423 EC_POINT_set_affine_coordinates_GF2m
6424 EC_POINT_get_affine_coordinates_GF2m
6425 EC_POINT_set_compressed_coordinates_GF2m
6427 Point compression for binary fields is disabled by default for
6428 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6431 As binary polynomials are represented as BIGNUMs, various members
6432 of the EC_GROUP and EC_POINT data structures can be shared
6433 between the implementations for prime fields and binary fields;
6434 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6435 are essentially identical to their ..._GFp counterparts.
6436 (For simplicity, the '..._GFp' prefix has been dropped from
6437 various internal method names.)
6439 An internal 'field_div' method (similar to 'field_mul' and
6440 'field_sqr') has been added; this is used only for binary fields.
6442 [Sheueling Chang Shantz and Douglas Stebila
6443 (Sun Microsystems Laboratories)]
6445 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6446 through methods ('mul', 'precompute_mult').
6448 The generic implementations (now internally called 'ec_wNAF_mul'
6449 and 'ec_wNAF_precomputed_mult') remain the default if these
6450 methods are undefined.
6452 [Sheueling Chang Shantz and Douglas Stebila
6453 (Sun Microsystems Laboratories)]
6455 *) New function EC_GROUP_get_degree, which is defined through
6456 EC_METHOD. For curves over prime fields, this returns the bit
6457 length of the modulus.
6459 [Sheueling Chang Shantz and Douglas Stebila
6460 (Sun Microsystems Laboratories)]
6462 *) New functions EC_GROUP_dup, EC_POINT_dup.
6463 (These simply call ..._new and ..._copy).
6465 [Sheueling Chang Shantz and Douglas Stebila
6466 (Sun Microsystems Laboratories)]
6468 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6469 Polynomials are represented as BIGNUMs (where the sign bit is not
6470 used) in the following functions [macros]:
6473 BN_GF2m_sub [= BN_GF2m_add]
6474 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6475 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6476 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6478 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6479 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6480 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6481 BN_GF2m_cmp [= BN_ucmp]
6483 (Note that only the 'mod' functions are actually for fields GF(2^m).
6484 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6486 For some functions, an the irreducible polynomial defining a
6487 field can be given as an 'unsigned int[]' with strictly
6488 decreasing elements giving the indices of those bits that are set;
6489 i.e., p[] represents the polynomial
6490 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6492 p[0] > p[1] > ... > p[k] = 0.
6493 This applies to the following functions:
6498 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6499 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6501 BN_GF2m_mod_sqrt_arr
6502 BN_GF2m_mod_solve_quad_arr
6506 Conversion can be performed by the following functions:
6511 bntest.c has additional tests for binary polynomial arithmetic.
6513 Two implementations for BN_GF2m_mod_div() are available.
6514 The default algorithm simply uses BN_GF2m_mod_inv() and
6515 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6516 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6517 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6519 [Sheueling Chang Shantz and Douglas Stebila
6520 (Sun Microsystems Laboratories)]
6522 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6523 functionality is disabled at compile-time.
6524 [Douglas Stebila <douglas.stebila@sun.com>]
6526 *) Change default behaviour of 'openssl asn1parse' so that more
6527 information is visible when viewing, e.g., a certificate:
6529 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6530 mode the content of non-printable OCTET STRINGs is output in a
6531 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6532 avoid the appearance of a printable string.
6533 [Nils Larsch <nla@trustcenter.de>]
6535 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6537 EC_GROUP_set_asn1_flag()
6538 EC_GROUP_get_asn1_flag()
6539 EC_GROUP_set_point_conversion_form()
6540 EC_GROUP_get_point_conversion_form()
6541 These control ASN1 encoding details:
6542 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6543 has been set to OPENSSL_EC_NAMED_CURVE.
6544 - Points are encoded in uncompressed form by default; options for
6545 asn1_for are as for point2oct, namely
6546 POINT_CONVERSION_COMPRESSED
6547 POINT_CONVERSION_UNCOMPRESSED
6548 POINT_CONVERSION_HYBRID
6550 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6553 EC_GROUP_get0_seed()
6554 EC_GROUP_get_seed_len()
6555 This is used only for ASN1 purposes (so far).
6556 [Nils Larsch <nla@trustcenter.de>]
6558 *) Add 'field_type' member to EC_METHOD, which holds the NID
6559 of the appropriate field type OID. The new function
6560 EC_METHOD_get_field_type() returns this value.
6561 [Nils Larsch <nla@trustcenter.de>]
6566 EC_POINT_point2hex()
6567 EC_POINT_hex2point()
6568 providing useful interfaces to EC_POINT_point2oct() and
6569 EC_POINT_oct2point().
6570 [Nils Larsch <nla@trustcenter.de>]
6572 *) Change internals of the EC library so that the functions
6573 EC_GROUP_set_generator()
6574 EC_GROUP_get_generator()
6575 EC_GROUP_get_order()
6576 EC_GROUP_get_cofactor()
6577 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6578 to methods, which would lead to unnecessary code duplication when
6579 adding different types of curves.
6580 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6582 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6583 arithmetic, and such that modified wNAFs are generated
6584 (which avoid length expansion in many cases).
6587 *) Add a function EC_GROUP_check_discriminant() (defined via
6588 EC_METHOD) that verifies that the curve discriminant is non-zero.
6590 Add a function EC_GROUP_check() that makes some sanity tests
6591 on a EC_GROUP, its generator and order. This includes
6592 EC_GROUP_check_discriminant().
6593 [Nils Larsch <nla@trustcenter.de>]
6595 *) Add ECDSA in new directory crypto/ecdsa/.
6597 Add applications 'openssl ecparam' and 'openssl ecdsa'
6598 (these are based on 'openssl dsaparam' and 'openssl dsa').
6600 ECDSA support is also included in various other files across the
6601 library. Most notably,
6602 - 'openssl req' now has a '-newkey ecdsa:file' option;
6603 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6604 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6605 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6606 them suitable for ECDSA where domain parameters must be
6607 extracted before the specific public key;
6608 - ECDSA engine support has been added.
6609 [Nils Larsch <nla@trustcenter.de>]
6611 *) Include some named elliptic curves, and add OIDs from X9.62,
6612 SECG, and WAP/WTLS. Each curve can be obtained from the new
6614 EC_GROUP_new_by_curve_name(),
6615 and the list of available named curves can be obtained with
6616 EC_get_builtin_curves().
6617 Also add a 'curve_name' member to EC_GROUP objects, which can be
6619 EC_GROUP_set_curve_name()
6620 EC_GROUP_get_curve_name()
6621 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6623 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6624 was actually never needed) and in BN_mul(). The removal in BN_mul()
6625 required a small change in bn_mul_part_recursive() and the addition
6626 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6627 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6628 bn_sub_words() and bn_add_words() except they take arrays with
6632 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6634 *) Cleanse PEM buffers before freeing them since they may contain
6636 [Benjamin Bennett <ben@psc.edu>]
6638 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6639 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6640 authentication-only ciphersuites.
6643 *) Since AES128 and AES256 share a single mask bit in the logic of
6644 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6645 kludge to work properly if AES128 is available and AES256 isn't.
6648 *) Expand security boundary to match 1.1.1 module.
6651 *) Remove redundant features: hash file source, editing of test vectors
6652 modify fipsld to use external fips_premain.c signature.
6655 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6656 run algorithm test programs.
6659 *) Make algorithm test programs more tolerant of whitespace.
6662 *) Have SSL/TLS server implementation tolerate "mismatched" record
6663 protocol version while receiving ClientHello even if the
6664 ClientHello is fragmented. (The server can't insist on the
6665 particular protocol version it has chosen before the ServerHello
6666 message has informed the client about his choice.)
6669 *) Load error codes if they are not already present instead of using a
6670 static variable. This allows them to be cleanly unloaded and reloaded.
6673 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6675 *) Introduce limits to prevent malicious keys being able to
6676 cause a denial of service. (CVE-2006-2940)
6677 [Steve Henson, Bodo Moeller]
6679 *) Fix ASN.1 parsing of certain invalid structures that can result
6680 in a denial of service. (CVE-2006-2937) [Steve Henson]
6682 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6683 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6685 *) Fix SSL client code which could crash if connecting to a
6686 malicious SSLv2 server. (CVE-2006-4343)
6687 [Tavis Ormandy and Will Drewry, Google Security Team]
6689 *) Change ciphersuite string processing so that an explicit
6690 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6691 will no longer include "AES128-SHA"), and any other similar
6692 ciphersuite (same bitmap) from *other* protocol versions (so that
6693 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6694 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6695 changes from 0.9.8b and 0.9.8d.
6698 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6700 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6701 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6703 *) Change the Unix randomness entropy gathering to use poll() when
6704 possible instead of select(), since the latter has some
6705 undesirable limitations.
6706 [Darryl Miles via Richard Levitte and Bodo Moeller]
6708 *) Disable rogue ciphersuites:
6710 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6711 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6712 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6714 The latter two were purportedly from
6715 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6718 Also deactivate the remaining ciphersuites from
6719 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6720 unofficial, and the ID has long expired.
6723 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6724 dual-core machines) and other potential thread-safety issues.
6727 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6729 *) Adapt fipsld and the build system to link against the validated FIPS
6730 module in FIPS mode.
6733 *) Fixes for VC++ 2005 build under Windows.
6736 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6737 from a Windows bash shell such as MSYS. It is autodetected from the
6738 "config" script when run from a VC++ environment. Modify standard VC++
6739 build to use fipscanister.o from the GNU make build.
6742 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6744 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6745 The value now differs depending on if you build for FIPS or not.
6746 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6747 safely run with a non-FIPSed libcrypto, as it may crash because of
6748 the difference induced by this change.
6751 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6753 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6754 (part of SSL_OP_ALL). This option used to disable the
6755 countermeasure against man-in-the-middle protocol-version
6756 rollback in the SSL 2.0 server implementation, which is a bad
6757 idea. (CVE-2005-2969)
6759 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6760 for Information Security, National Institute of Advanced Industrial
6761 Science and Technology [AIST], Japan)]
6763 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6764 mainly for FIPS compliance and not fully integrated at this stage.
6767 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6768 the exponentiation using a fixed-length exponent. (Otherwise,
6769 the information leaked through timing could expose the secret key
6770 after many signatures; cf. Bleichenbacher's attack on DSA with
6774 *) Make a new fixed-window mod_exp implementation the default for
6775 RSA, DSA, and DH private-key operations so that the sequence of
6776 squares and multiplies and the memory access pattern are
6777 independent of the particular secret key. This will mitigate
6778 cache-timing and potential related attacks.
6780 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6781 and this is automatically used by BN_mod_exp_mont() if the new flag
6782 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6783 will use this BN flag for private exponents unless the flag
6784 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6785 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6787 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6789 *) Change the client implementation for SSLv23_method() and
6790 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6791 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6792 (Previously, the SSL 2.0 backwards compatible Client Hello
6793 message format would be used even with SSL_OP_NO_SSLv2.)
6796 *) Add support for smime-type MIME parameter in S/MIME messages which some
6800 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6801 a threadsafe manner. Modify rsa code to use new function and add calls
6802 to dsa and dh code (which had race conditions before).
6805 *) Include the fixed error library code in the C error file definitions
6806 instead of fixing them up at runtime. This keeps the error code
6807 structures constant.
6810 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6812 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6815 *) Fixes for newer kerberos headers. NB: the casts are needed because
6816 the 'length' field is signed on one version and unsigned on another
6817 with no (?) obvious way to tell the difference, without these VC++
6818 complains. Also the "definition" of FAR (blank) is no longer included
6819 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6820 some needed definitions.
6823 *) Undo Cygwin change.
6826 *) Added support for proxy certificates according to RFC 3820.
6827 Because they may be a security thread to unaware applications,
6828 they must be explicitly allowed in run-time. See
6829 docs/HOWTO/proxy_certificates.txt for further information.
6832 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6834 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6835 server and client random values. Previously
6836 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6837 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6839 This change has negligible security impact because:
6841 1. Server and client random values still have 24 bytes of pseudo random
6844 2. Server and client random values are sent in the clear in the initial
6847 3. The master secret is derived using the premaster secret (48 bytes in
6848 size for static RSA ciphersuites) as well as client server and random
6851 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6854 [Stephen Henson, reported by UK NISCC]
6856 *) Use Windows randomness collection on Cygwin.
6859 *) Fix hang in EGD/PRNGD query when communication socket is closed
6860 prematurely by EGD/PRNGD.
6861 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6863 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6866 *) Back-port of selected performance improvements from development
6867 branch, as well as improved support for PowerPC platforms.
6870 *) Add lots of checks for memory allocation failure, error codes to indicate
6871 failure and freeing up memory if a failure occurs.
6872 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6874 *) Add new -passin argument to dgst.
6877 *) Perform some character comparisons of different types in X509_NAME_cmp:
6878 this is needed for some certificates that re-encode DNs into UTF8Strings
6879 (in violation of RFC3280) and can't or won't issue name rollover
6883 *) Make an explicit check during certificate validation to see that
6884 the CA setting in each certificate on the chain is correct. As a
6885 side effect always do the following basic checks on extensions,
6886 not just when there's an associated purpose to the check:
6888 - if there is an unhandled critical extension (unless the user
6889 has chosen to ignore this fault)
6890 - if the path length has been exceeded (if one is set at all)
6891 - that certain extensions fit the associated purpose (if one has
6895 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6897 *) Avoid a race condition when CRLs are checked in a multi threaded
6898 environment. This would happen due to the reordering of the revoked
6899 entries during signature checking and serial number lookup. Now the
6900 encoding is cached and the serial number sort performed under a lock.
6901 Add new STACK function sk_is_sorted().
6904 *) Add Delta CRL to the extension code.
6907 *) Various fixes to s3_pkt.c so alerts are sent properly.
6908 [David Holmes <d.holmes@f5.com>]
6910 *) Reduce the chances of duplicate issuer name and serial numbers (in
6911 violation of RFC3280) using the OpenSSL certificate creation utilities.
6912 This is done by creating a random 64 bit value for the initial serial
6913 number when a serial number file is created or when a self signed
6914 certificate is created using 'openssl req -x509'. The initial serial
6915 number file is created using 'openssl x509 -next_serial' in CA.pl
6916 rather than being initialized to 1.
6919 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6921 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6922 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6923 [Joe Orton, Steve Henson]
6925 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6927 [Joe Orton, Steve Henson]
6929 *) Make it possible to have multiple active certificates with the same
6930 subject in the CA index file. This is done only if the keyword
6931 'unique_subject' is set to 'no' in the main CA section (default
6932 if 'CA_default') of the configuration file. The value is saved
6933 with the database itself in a separate index attribute file,
6934 named like the index file with '.attr' appended to the name.
6937 *) X509 verify fixes. Disable broken certificate workarounds when
6938 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6939 keyUsage extension present. Don't accept CRLs with unhandled critical
6940 extensions: since verify currently doesn't process CRL extensions this
6941 rejects a CRL with *any* critical extensions. Add new verify error codes
6945 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6946 A clarification of RFC2560 will require the use of OCTET STRINGs and
6947 some implementations cannot handle the current raw format. Since OpenSSL
6948 copies and compares OCSP nonces as opaque blobs without any attempt at
6949 parsing them this should not create any compatibility issues.
6952 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6953 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6954 this HMAC (and other) operations are several times slower than OpenSSL
6958 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6959 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6961 *) Use the correct content when signing type "other".
6964 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6966 *) Fix various bugs revealed by running the NISCC test suite:
6968 Stop out of bounds reads in the ASN1 code when presented with
6969 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6971 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6973 If verify callback ignores invalid public key errors don't try to check
6974 certificate signature with the NULL public key.
6978 *) New -ignore_err option in ocsp application to stop the server
6979 exiting on the first error in a request.
6982 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6983 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6987 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6988 extra data after the compression methods not only for TLS 1.0
6989 but also for SSL 3.0 (as required by the specification).
6990 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6992 *) Change X509_certificate_type() to mark the key as exported/exportable
6993 when it's 512 *bits* long, not 512 bytes.
6996 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6997 blocks during encryption.
7000 *) Various fixes to base64 BIO and non blocking I/O. On write
7001 flushes were not handled properly if the BIO retried. On read
7002 data was not being buffered properly and had various logic bugs.
7003 This also affects blocking I/O when the data being decoded is a
7007 *) Various S/MIME bugfixes and compatibility changes:
7008 output correct application/pkcs7 MIME type if
7009 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7010 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7011 of files as .eml work). Correctly handle very long lines in MIME
7015 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7017 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7018 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7019 a protocol version number mismatch like a decryption error
7020 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7023 *) Turn on RSA blinding by default in the default implementation
7024 to avoid a timing attack. Applications that don't want it can call
7025 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7026 They would be ill-advised to do so in most cases.
7027 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7029 *) Change RSA blinding code so that it works when the PRNG is not
7030 seeded (in this case, the secret RSA exponent is abused as
7031 an unpredictable seed -- if it is not unpredictable, there
7032 is no point in blinding anyway). Make RSA blinding thread-safe
7033 by remembering the creator's thread ID in rsa->blinding and
7034 having all other threads use local one-time blinding factors
7035 (this requires more computation than sharing rsa->blinding, but
7036 avoids excessive locking; and if an RSA object is not shared
7037 between threads, blinding will still be very fast).
7040 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7041 ENGINE as defaults for all supported algorithms irrespective of
7042 the 'flags' parameter. 'flags' is now honoured, so applications
7043 should make sure they are passing it correctly.
7046 *) Target "mingw" now allows native Windows code to be generated in
7047 the Cygwin environment as well as with the MinGW compiler.
7050 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7052 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7053 via timing by performing a MAC computation even if incorrect
7054 block cipher padding has been found. This is a countermeasure
7055 against active attacks where the attacker has to distinguish
7056 between bad padding and a MAC verification error. (CVE-2003-0078)
7058 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7059 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7060 Martin Vuagnoux (EPFL, Ilion)]
7062 *) Make the no-err option work as intended. The intention with no-err
7063 is not to have the whole error stack handling routines removed from
7064 libcrypto, it's only intended to remove all the function name and
7065 reason texts, thereby removing some of the footprint that may not
7066 be interesting if those errors aren't displayed anyway.
7068 NOTE: it's still possible for any application or module to have its
7069 own set of error texts inserted. The routines are there, just not
7070 used by default when no-err is given.
7073 *) Add support for FreeBSD on IA64.
7074 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7076 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7077 Kerberos function mit_des_cbc_cksum(). Before this change,
7078 the value returned by DES_cbc_cksum() was like the one from
7079 mit_des_cbc_cksum(), except the bytes were swapped.
7080 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7082 *) Allow an application to disable the automatic SSL chain building.
7083 Before this a rather primitive chain build was always performed in
7084 ssl3_output_cert_chain(): an application had no way to send the
7085 correct chain if the automatic operation produced an incorrect result.
7087 Now the chain builder is disabled if either:
7089 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7091 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7093 The reasoning behind this is that an application would not want the
7094 auto chain building to take place if extra chain certificates are
7095 present and it might also want a means of sending no additional
7096 certificates (for example the chain has two certificates and the
7100 *) Add the possibility to build without the ENGINE framework.
7101 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7103 *) Under Win32 gmtime() can return NULL: check return value in
7104 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7107 *) DSA routines: under certain error conditions uninitialized BN objects
7108 could be freed. Solution: make sure initialization is performed early
7109 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7110 Nils Larsch <nla@trustcenter.de> via PR#459)
7113 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7114 checked on reconnect on the client side, therefore session resumption
7115 could still fail with a "ssl session id is different" error. This
7116 behaviour is masked when SSL_OP_ALL is used due to
7117 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7118 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7119 followup to PR #377.
7122 *) IA-32 assembler support enhancements: unified ELF targets, support
7123 for SCO/Caldera platforms, fix for Cygwin shared build.
7126 *) Add support for FreeBSD on sparc64. As a consequence, support for
7127 FreeBSD on non-x86 processors is separate from x86 processors on
7128 the config script, much like the NetBSD support.
7129 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7131 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7133 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7136 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7137 code (06) was taken as the first octet of the session ID and the last
7138 octet was ignored consequently. As a result SSLv2 client side session
7139 caching could not have worked due to the session ID mismatch between
7141 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7145 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7146 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7150 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7151 seems that in spite of existing for more than a year, many application
7152 author have done nothing to provide the necessary callbacks, which
7153 means that this particular engine will not work properly anywhere.
7154 This is a very unfortunate situation which forces us, in the name
7155 of usability, to give the hw_ncipher.c a static lock, which is part
7157 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7158 appear in 0.9.8 or later. We EXPECT application authors to have
7159 dealt properly with this when 0.9.8 is released (unless we actually
7160 make such changes in the libcrypto locking code that changes will
7161 have to be made anyway).
7164 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7165 octets have been read, EOF or an error occurs. Without this change
7166 some truncated ASN1 structures will not produce an error.
7169 *) Disable Heimdal support, since it hasn't been fully implemented.
7170 Still give the possibility to force the use of Heimdal, but with
7171 warnings and a request that patches get sent to openssl-dev.
7174 *) Add the VC-CE target, introduce the WINCE sysname, and add
7175 INSTALL.WCE and appropriate conditionals to make it build.
7176 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7178 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7179 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7180 edit numbers of the version.
7181 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7183 *) Introduce safe string copy and catenation functions
7184 (BUF_strlcpy() and BUF_strlcat()).
7185 [Ben Laurie (CHATS) and Richard Levitte]
7187 *) Avoid using fixed-size buffers for one-line DNs.
7188 [Ben Laurie (CHATS)]
7190 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7191 resizing buffers containing secrets, and use where appropriate.
7192 [Ben Laurie (CHATS)]
7194 *) Avoid using fixed size buffers for configuration file location.
7195 [Ben Laurie (CHATS)]
7197 *) Avoid filename truncation for various CA files.
7198 [Ben Laurie (CHATS)]
7200 *) Use sizeof in preference to magic numbers.
7201 [Ben Laurie (CHATS)]
7203 *) Avoid filename truncation in cert requests.
7204 [Ben Laurie (CHATS)]
7206 *) Add assertions to check for (supposedly impossible) buffer
7208 [Ben Laurie (CHATS)]
7210 *) Don't cache truncated DNS entries in the local cache (this could
7211 potentially lead to a spoofing attack).
7212 [Ben Laurie (CHATS)]
7214 *) Fix various buffers to be large enough for hex/decimal
7215 representations in a platform independent manner.
7216 [Ben Laurie (CHATS)]
7218 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7219 resizing buffers containing secrets, and use where appropriate.
7220 [Ben Laurie (CHATS)]
7222 *) Add BIO_indent() to avoid much slightly worrying code to do
7224 [Ben Laurie (CHATS)]
7226 *) Convert sprintf()/BIO_puts() to BIO_printf().
7227 [Ben Laurie (CHATS)]
7229 *) buffer_gets() could terminate with the buffer only half
7231 [Ben Laurie (CHATS)]
7233 *) Add assertions to prevent user-supplied crypto functions from
7234 overflowing internal buffers by having large block sizes, etc.
7235 [Ben Laurie (CHATS)]
7237 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7239 [Ben Laurie (CHATS)]
7241 *) Eliminate unused copy of key in RC4.
7242 [Ben Laurie (CHATS)]
7244 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7245 [Ben Laurie (CHATS)]
7247 *) Fix off-by-one error in EGD path.
7248 [Ben Laurie (CHATS)]
7250 *) If RANDFILE path is too long, ignore instead of truncating.
7251 [Ben Laurie (CHATS)]
7253 *) Eliminate unused and incorrectly sized X.509 structure
7255 [Ben Laurie (CHATS)]
7257 *) Eliminate unused and dangerous function knumber().
7258 [Ben Laurie (CHATS)]
7260 *) Eliminate unused and dangerous structure, KSSL_ERR.
7261 [Ben Laurie (CHATS)]
7263 *) Protect against overlong session ID context length in an encoded
7264 session object. Since these are local, this does not appear to be
7266 [Ben Laurie (CHATS)]
7268 *) Change from security patch (see 0.9.6e below) that did not affect
7269 the 0.9.6 release series:
7271 Remote buffer overflow in SSL3 protocol - an attacker could
7272 supply an oversized master key in Kerberos-enabled versions.
7274 [Ben Laurie (CHATS)]
7276 *) Change the SSL kerb5 codes to match RFC 2712.
7279 *) Make -nameopt work fully for req and add -reqopt switch.
7280 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7282 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7283 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7285 *) Make sure tests can be performed even if the corresponding algorithms
7286 have been removed entirely. This was also the last step to make
7287 OpenSSL compilable with DJGPP under all reasonable conditions.
7288 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7290 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7291 to allow version independent disabling of normally unselected ciphers,
7292 which may be activated as a side-effect of selecting a single cipher.
7294 (E.g., cipher list string "RSA" enables ciphersuites that are left
7295 out of "ALL" because they do not provide symmetric encryption.
7296 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7297 [Lutz Jaenicke, Bodo Moeller]
7299 *) Add appropriate support for separate platform-dependent build
7300 directories. The recommended way to make a platform-dependent
7301 build directory is the following (tested on Linux), maybe with
7304 # Place yourself outside of the OpenSSL source tree. In
7305 # this example, the environment variable OPENSSL_SOURCE
7306 # is assumed to contain the absolute OpenSSL source directory.
7307 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7308 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7309 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7310 mkdir -p `dirname $F`
7311 ln -s $OPENSSL_SOURCE/$F $F
7314 To be absolutely sure not to disturb the source tree, a "make clean"
7315 is a good thing. If it isn't successful, don't worry about it,
7316 it probably means the source directory is very clean.
7319 *) Make sure any ENGINE control commands make local copies of string
7320 pointers passed to them whenever necessary. Otherwise it is possible
7321 the caller may have overwritten (or deallocated) the original string
7322 data when a later ENGINE operation tries to use the stored values.
7323 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7325 *) Improve diagnostics in file reading and command-line digests.
7326 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7328 *) Add AES modes CFB and OFB to the object database. Correct an
7329 error in AES-CFB decryption.
7332 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7333 allows existing EVP_CIPHER_CTX structures to be reused after
7334 calling EVP_*Final(). This behaviour is used by encryption
7335 BIOs and some applications. This has the side effect that
7336 applications must explicitly clean up cipher contexts with
7337 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7340 *) Check the values of dna and dnb in bn_mul_recursive before calling
7341 bn_mul_comba (a non zero value means the a or b arrays do not contain
7342 n2 elements) and fallback to bn_mul_normal if either is not zero.
7345 *) Fix escaping of non-ASCII characters when using the -subj option
7346 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7349 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7350 form for "surname", serialNumber has no short form.
7351 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7352 therefore remove "mail" short name for "internet 7".
7353 The OID for unique identifiers in X509 certificates is
7354 x500UniqueIdentifier, not uniqueIdentifier.
7355 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7358 *) Add an "init" command to the ENGINE config module and auto initialize
7359 ENGINEs. Without any "init" command the ENGINE will be initialized
7360 after all ctrl commands have been executed on it. If init=1 the
7361 ENGINE is initialized at that point (ctrls before that point are run
7362 on the uninitialized ENGINE and after on the initialized one). If
7363 init=0 then the ENGINE will not be initialized at all.
7366 *) Fix the 'app_verify_callback' interface so that the user-defined
7367 argument is actually passed to the callback: In the
7368 SSL_CTX_set_cert_verify_callback() prototype, the callback
7369 declaration has been changed from
7372 int (*cb)(X509_STORE_CTX *,void *);
7373 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7374 i=s->ctx->app_verify_callback(&ctx)
7375 has been changed into
7376 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7378 To update applications using SSL_CTX_set_cert_verify_callback(),
7379 a dummy argument can be added to their callback functions.
7380 [D. K. Smetters <smetters@parc.xerox.com>]
7382 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7383 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7385 *) Add and OPENSSL_LOAD_CONF define which will cause
7386 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7387 This allows older applications to transparently support certain
7388 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7389 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7390 load the config file and OPENSSL_add_all_algorithms_conf() which will
7391 always load it have also been added.
7394 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7395 Adjust NIDs and EVP layer.
7396 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7398 *) Config modules support in openssl utility.
7400 Most commands now load modules from the config file,
7401 though in a few (such as version) this isn't done
7402 because it couldn't be used for anything.
7404 In the case of ca and req the config file used is
7405 the same as the utility itself: that is the -config
7406 command line option can be used to specify an
7410 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7411 use "openssl_conf" if filename is NULL use default openssl config file.
7414 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7415 config section name. Add a new flag to tolerate a missing config file
7416 and move code to CONF_modules_load_file().
7419 *) Support for crypto accelerator cards from Accelerated Encryption
7420 Processing, www.aep.ie. (Use engine 'aep')
7421 The support was copied from 0.9.6c [engine] and adapted/corrected
7422 to work with the new engine framework.
7423 [AEP Inc. and Richard Levitte]
7425 *) Support for SureWare crypto accelerator cards from Baltimore
7426 Technologies. (Use engine 'sureware')
7427 The support was copied from 0.9.6c [engine] and adapted
7428 to work with the new engine framework.
7431 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7432 make the newer ENGINE framework commands for the CHIL engine work.
7433 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7435 *) Make it possible to produce shared libraries on ReliantUNIX.
7436 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7438 *) Add the configuration target debug-linux-ppro.
7439 Make 'openssl rsa' use the general key loading routines
7440 implemented in apps.c, and make those routines able to
7441 handle the key format FORMAT_NETSCAPE and the variant
7443 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7445 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7446 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7448 *) Add -keyform to rsautl, and document -engine.
7449 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7451 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7452 BIO_R_NO_SUCH_FILE error code rather than the generic
7453 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7456 *) Add new functions
7458 ERR_peek_last_error_line
7459 ERR_peek_last_error_line_data.
7460 These are similar to
7463 ERR_peek_error_line_data,
7464 but report on the latest error recorded rather than the first one
7465 still in the error queue.
7466 [Ben Laurie, Bodo Moeller]
7468 *) default_algorithms option in ENGINE config module. This allows things
7470 default_algorithms = ALL
7471 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7474 *) Preliminary ENGINE config module.
7477 *) New experimental application configuration code.
7480 *) Change the AES code to follow the same name structure as all other
7481 symmetric ciphers, and behave the same way. Move everything to
7482 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7483 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7485 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7486 [Ben Laurie and Theo de Raadt]
7488 *) Add option to output public keys in req command.
7489 [Massimiliano Pala madwolf@openca.org]
7491 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7492 (up to about 10% better than before for P-192 and P-224).
7495 *) New functions/macros
7497 SSL_CTX_set_msg_callback(ctx, cb)
7498 SSL_CTX_set_msg_callback_arg(ctx, arg)
7499 SSL_set_msg_callback(ssl, cb)
7500 SSL_set_msg_callback_arg(ssl, arg)
7502 to request calling a callback function
7504 void cb(int write_p, int version, int content_type,
7505 const void *buf, size_t len, SSL *ssl, void *arg)
7507 whenever a protocol message has been completely received
7508 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7509 protocol version according to which the SSL library interprets
7510 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7511 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7512 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7513 specification (change_cipher_spec(20), alert(21), handshake(22)).
7514 'buf' and 'len' point to the actual message, 'ssl' to the
7515 SSL object, and 'arg' is the application-defined value set by
7516 SSL[_CTX]_set_msg_callback_arg().
7518 'openssl s_client' and 'openssl s_server' have new '-msg' options
7519 to enable a callback that displays all protocol messages.
7522 *) Change the shared library support so shared libraries are built as
7523 soon as the corresponding static library is finished, and thereby get
7524 openssl and the test programs linked against the shared library.
7525 This still only happens when the keyword "shard" has been given to
7526 the configuration scripts.
7528 NOTE: shared library support is still an experimental thing, and
7529 backward binary compatibility is still not guaranteed.
7530 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7532 *) Add support for Subject Information Access extension.
7533 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7535 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7536 additional bytes when new memory had to be allocated, not just
7537 when reusing an existing buffer.
7540 *) New command line and configuration option 'utf8' for the req command.
7541 This allows field values to be specified as UTF8 strings.
7544 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7545 runs for the former and machine-readable output for the latter.
7548 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7549 of the e-mail address in the DN (i.e., it will go into a certificate
7550 extension only). The new configuration file option 'email_in_dn = no'
7551 has the same effect.
7552 [Massimiliano Pala madwolf@openca.org]
7554 *) Change all functions with names starting with des_ to be starting
7555 with DES_ instead. Add wrappers that are compatible with libdes,
7556 but are named _ossl_old_des_*. Finally, add macros that map the
7557 des_* symbols to the corresponding _ossl_old_des_* if libdes
7558 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7559 desired, the des_* symbols will be mapped to DES_*, with one
7562 Since we provide two compatibility mappings, the user needs to
7563 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7564 compatibility is desired. The default (i.e., when that macro
7565 isn't defined) is OpenSSL 0.9.6c compatibility.
7567 There are also macros that enable and disable the support of old
7568 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7569 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7570 are defined, the default will apply: to support the old des routines.
7572 In either case, one must include openssl/des.h to get the correct
7573 definitions. Do not try to just include openssl/des_old.h, that
7576 NOTE: This is a major break of an old API into a new one. Software
7577 authors are encouraged to switch to the DES_ style functions. Some
7578 time in the future, des_old.h and the libdes compatibility functions
7579 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7580 default), and then completely removed.
7583 *) Test for certificates which contain unsupported critical extensions.
7584 If such a certificate is found during a verify operation it is
7585 rejected by default: this behaviour can be overridden by either
7586 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7587 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7588 X509_supported_extension() has also been added which returns 1 if a
7589 particular extension is supported.
7592 *) Modify the behaviour of EVP cipher functions in similar way to digests
7593 to retain compatibility with existing code.
7596 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7597 compatibility with existing code. In particular the 'ctx' parameter does
7598 not have to be to be initialized before the call to EVP_DigestInit() and
7599 it is tidied up after a call to EVP_DigestFinal(). New function
7600 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7601 EVP_MD_CTX_copy() changed to not require the destination to be
7602 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7603 requires the destination to be valid.
7605 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7606 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7609 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7610 so that complete 'Handshake' protocol structures are kept in memory
7611 instead of overwriting 'msg_type' and 'length' with 'body' data.
7614 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7615 [Massimo Santin via Richard Levitte]
7617 *) Major restructuring to the underlying ENGINE code. This includes
7618 reduction of linker bloat, separation of pure "ENGINE" manipulation
7619 (initialisation, etc) from functionality dealing with implementations
7620 of specific crypto interfaces. This change also introduces integrated
7621 support for symmetric ciphers and digest implementations - so ENGINEs
7622 can now accelerate these by providing EVP_CIPHER and EVP_MD
7623 implementations of their own. This is detailed in crypto/engine/README
7624 as it couldn't be adequately described here. However, there are a few
7625 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7626 were changed in the original introduction of ENGINE code have now
7627 reverted back - the hooking from this code to ENGINE is now a good
7628 deal more passive and at run-time, operations deal directly with
7629 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7630 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7631 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7632 they were not being used by the framework as there is no concept of a
7633 BIGNUM_METHOD and they could not be generalised to the new
7634 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7635 ENGINE_cpy() has been removed as it cannot be consistently defined in
7639 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7642 *) Change mkdef.pl to sort symbols that get the same entry number,
7643 and make sure the automatically generated functions ERR_load_*
7644 become part of libeay.num as well.
7647 *) New function SSL_renegotiate_pending(). This returns true once
7648 renegotiation has been requested (either SSL_renegotiate() call
7649 or HelloRequest/ClientHello received from the peer) and becomes
7650 false once a handshake has been completed.
7651 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7652 sends a HelloRequest, but does not ensure that a handshake takes
7653 place. SSL_renegotiate_pending() is useful for checking if the
7654 client has followed the request.)
7657 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7658 By default, clients may request session resumption even during
7659 renegotiation (if session ID contexts permit); with this option,
7660 session resumption is possible only in the first handshake.
7662 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7663 more bits available for options that should not be part of
7664 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7667 *) Add some demos for certificate and certificate request creation.
7670 *) Make maximum certificate chain size accepted from the peer application
7671 settable (SSL*_get/set_max_cert_list()), as proposed by
7672 "Douglas E. Engert" <deengert@anl.gov>.
7675 *) Add support for shared libraries for Unixware-7
7676 (Boyd Lynn Gerber <gerberb@zenez.com>).
7679 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7680 be done prior to destruction. Use this to unload error strings from
7681 ENGINEs that load their own error strings. NB: This adds two new API
7682 functions to "get" and "set" this destroy handler in an ENGINE.
7685 *) Alter all existing ENGINE implementations (except "openssl" and
7686 "openbsd") to dynamically instantiate their own error strings. This
7687 makes them more flexible to be built both as statically-linked ENGINEs
7688 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7689 Also, add stub code to each that makes building them as self-contained
7690 shared-libraries easier (see README.ENGINE).
7693 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7694 implementations into applications that are completely implemented in
7695 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7696 commands that can be used to configure what shared-library to load and
7697 to control aspects of the way it is handled. Also, made an update to
7698 the README.ENGINE file that brings its information up-to-date and
7699 provides some information and instructions on the "dynamic" ENGINE
7700 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7703 *) Make it possible to unload ranges of ERR strings with a new
7704 "ERR_unload_strings" function.
7707 *) Add a copy() function to EVP_MD.
7710 *) Make EVP_MD routines take a context pointer instead of just the
7711 md_data void pointer.
7714 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7715 that the digest can only process a single chunk of data
7716 (typically because it is provided by a piece of
7717 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7718 is only going to provide a single chunk of data, and hence the
7719 framework needn't accumulate the data for oneshot drivers.
7722 *) As with "ERR", make it possible to replace the underlying "ex_data"
7723 functions. This change also alters the storage and management of global
7724 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7725 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7726 index counters. The API functions that use this state have been changed
7727 to take a "class_index" rather than pointers to the class's local STACK
7728 and counter, and there is now an API function to dynamically create new
7729 classes. This centralisation allows us to (a) plug a lot of the
7730 thread-safety problems that existed, and (b) makes it possible to clean
7731 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7732 such data would previously have always leaked in application code and
7733 workarounds were in place to make the memory debugging turn a blind eye
7734 to it. Application code that doesn't use this new function will still
7735 leak as before, but their memory debugging output will announce it now
7736 rather than letting it slide.
7738 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7739 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7740 has a return value to indicate success or failure.
7743 *) Make it possible to replace the underlying "ERR" functions such that the
7744 global state (2 LHASH tables and 2 locks) is only used by the "default"
7745 implementation. This change also adds two functions to "get" and "set"
7746 the implementation prior to it being automatically set the first time
7747 any other ERR function takes place. Ie. an application can call "get",
7748 pass the return value to a module it has just loaded, and that module
7749 can call its own "set" function using that value. This means the
7750 module's "ERR" operations will use (and modify) the error state in the
7751 application and not in its own statically linked copy of OpenSSL code.
7754 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7755 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7756 the operation, and provides a more encapsulated way for external code
7757 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7758 to use these functions rather than manually incrementing the counts.
7760 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7763 *) Add EVP test program.
7766 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7769 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7770 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7771 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7772 These allow a CRL to be built without having to access X509_CRL fields
7773 directly. Modify 'ca' application to use new functions.
7776 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7777 bug workarounds. Rollback attack detection is a security feature.
7778 The problem will only arise on OpenSSL servers when TLSv1 is not
7779 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7780 Software authors not wanting to support TLSv1 will have special reasons
7781 for their choice and can explicitly enable this option.
7782 [Bodo Moeller, Lutz Jaenicke]
7784 *) Rationalise EVP so it can be extended: don't include a union of
7785 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7786 (similar to those existing for EVP_CIPHER_CTX).
7791 EVP_MD_CTX_init(&md); /* new function call */
7792 EVP_DigestInit(&md, EVP_sha1());
7793 EVP_DigestUpdate(&md, in, len);
7794 EVP_DigestFinal(&md, out, NULL);
7795 EVP_MD_CTX_cleanup(&md); /* new function call */
7799 *) Make DES key schedule conform to the usual scheme, as well as
7800 correcting its structure. This means that calls to DES functions
7801 now have to pass a pointer to a des_key_schedule instead of a
7802 plain des_key_schedule (which was actually always a pointer
7805 des_key_schedule ks;
7807 des_set_key_checked(..., &ks);
7808 des_ncbc_encrypt(..., &ks, ...);
7810 (Note that a later change renames 'des_...' into 'DES_...'.)
7813 *) Initial reduction of linker bloat: the use of some functions, such as
7814 PEM causes large amounts of unused functions to be linked in due to
7815 poor organisation. For example pem_all.c contains every PEM function
7816 which has a knock on effect of linking in large amounts of (unused)
7817 ASN1 code. Grouping together similar functions and splitting unrelated
7818 functions prevents this.
7821 *) Cleanup of EVP macros.
7824 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7825 correct _ecb suffix.
7828 *) Add initial OCSP responder support to ocsp application. The
7829 revocation information is handled using the text based index
7830 use by the ca application. The responder can either handle
7831 requests generated internally, supplied in files (for example
7832 via a CGI script) or using an internal minimal server.
7835 *) Add configuration choices to get zlib compression for TLS.
7838 *) Changes to Kerberos SSL for RFC 2712 compliance:
7839 1. Implemented real KerberosWrapper, instead of just using
7840 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7841 2. Implemented optional authenticator field of KerberosWrapper.
7843 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7844 and authenticator structs; see crypto/krb5/.
7846 Generalized Kerberos calls to support multiple Kerberos libraries.
7847 [Vern Staats <staatsvr@asc.hpc.mil>,
7848 Jeffrey Altman <jaltman@columbia.edu>
7849 via Richard Levitte]
7851 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7852 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7853 values for each of the key sizes rather than having just
7854 parameters (and 'speed' generating keys each time).
7857 *) Speed up EVP routines.
7860 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7861 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7862 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7863 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7865 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7866 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7867 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7870 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7872 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7875 *) Added the OS2-EMX target.
7876 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7878 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7879 to support NCONF routines in extension code. New function CONF_set_nconf()
7880 to allow functions which take an NCONF to also handle the old LHASH
7881 structure: this means that the old CONF compatible routines can be
7882 retained (in particular wrt extensions) without having to duplicate the
7883 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7886 *) Enhance the general user interface with mechanisms for inner control
7887 and with possibilities to have yes/no kind of prompts.
7890 *) Change all calls to low level digest routines in the library and
7891 applications to use EVP. Add missing calls to HMAC_cleanup() and
7892 don't assume HMAC_CTX can be copied using memcpy().
7893 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7895 *) Add the possibility to control engines through control names but with
7896 arbitrary arguments instead of just a string.
7897 Change the key loaders to take a UI_METHOD instead of a callback
7898 function pointer. NOTE: this breaks binary compatibility with earlier
7899 versions of OpenSSL [engine].
7900 Adapt the nCipher code for these new conditions and add a card insertion
7904 *) Enhance the general user interface with mechanisms to better support
7905 dialog box interfaces, application-defined prompts, the possibility
7906 to use defaults (for example default passwords from somewhere else)
7907 and interrupts/cancellations.
7910 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7911 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7914 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7915 tidy up some unnecessarily weird code in 'sk_new()').
7916 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7918 *) Change the key loading routines for ENGINEs to use the same kind
7919 callback (pem_password_cb) as all other routines that need this
7923 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7924 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7925 than this minimum value is recommended.
7928 *) New random seeder for OpenVMS, using the system process statistics
7929 that are easily reachable.
7932 *) Windows apparently can't transparently handle global
7933 variables defined in DLLs. Initialisations such as:
7935 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7937 won't compile. This is used by the any applications that need to
7938 declare their own ASN1 modules. This was fixed by adding the option
7939 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7940 needed for static libraries under Win32.
7943 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7944 setting of purpose and trust fields. New X509_STORE trust and
7945 purpose functions and tidy up setting in other SSL functions.
7948 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7949 structure. These are inherited by X509_STORE_CTX when it is
7950 initialised. This allows various defaults to be set in the
7951 X509_STORE structure (such as flags for CRL checking and custom
7952 purpose or trust settings) for functions which only use X509_STORE_CTX
7953 internally such as S/MIME.
7955 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7956 trust settings if they are not set in X509_STORE. This allows X509_STORE
7957 purposes and trust (in S/MIME for example) to override any set by default.
7959 Add command line options for CRL checking to smime, s_client and s_server
7963 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7964 are set then the CRL is looked up in the X509_STORE structure and
7965 its validity and signature checked, then if the certificate is found
7966 in the CRL the verify fails with a revoked error.
7968 Various new CRL related callbacks added to X509_STORE_CTX structure.
7970 Command line options added to 'verify' application to support this.
7972 This needs some additional work, such as being able to handle multiple
7973 CRLs with different times, extension based lookup (rather than just
7974 by subject name) and ultimately more complete V2 CRL extension
7978 *) Add a general user interface API (crypto/ui/). This is designed
7979 to replace things like des_read_password and friends (backward
7980 compatibility functions using this new API are provided).
7981 The purpose is to remove prompting functions from the DES code
7982 section as well as provide for prompting through dialog boxes in
7983 a window system and the like.
7986 *) Add "ex_data" support to ENGINE so implementations can add state at a
7987 per-structure level rather than having to store it globally.
7990 *) Make it possible for ENGINE structures to be copied when retrieved by
7991 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7992 This causes the "original" ENGINE structure to act like a template,
7993 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7994 operational state can be localised to each ENGINE structure, despite the
7995 fact they all share the same "methods". New ENGINE structures returned in
7996 this case have no functional references and the return value is the single
7997 structural reference. This matches the single structural reference returned
7998 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8002 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8003 needs to match any other type at all we need to manually clear the
8007 *) Changes to the "openssl engine" utility to include;
8008 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8009 about an ENGINE's available control commands.
8010 - executing control commands from command line arguments using the
8011 '-pre' and '-post' switches. '-post' is only used if '-t' is
8012 specified and the ENGINE is successfully initialised. The syntax for
8013 the individual commands are colon-separated, for example;
8014 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8017 *) New dynamic control command support for ENGINEs. ENGINEs can now
8018 declare their own commands (numbers), names (strings), descriptions,
8019 and input types for run-time discovery by calling applications. A
8020 subset of these commands are implicitly classed as "executable"
8021 depending on their input type, and only these can be invoked through
8022 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8023 can be based on user input, config files, etc). The distinction is
8024 that "executable" commands cannot return anything other than a boolean
8025 result and can only support numeric or string input, whereas some
8026 discoverable commands may only be for direct use through
8027 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8028 pointers, or other custom uses. The "executable" commands are to
8029 support parameterisations of ENGINE behaviour that can be
8030 unambiguously defined by ENGINEs and used consistently across any
8031 OpenSSL-based application. Commands have been added to all the
8032 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8033 control over shared-library paths without source code alterations.
8036 *) Changed all ENGINE implementations to dynamically allocate their
8037 ENGINEs rather than declaring them statically. Apart from this being
8038 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8039 this also allows the implementations to compile without using the
8040 internal engine_int.h header.
8043 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8044 'const' value. Any code that should be able to modify a RAND_METHOD
8045 should already have non-const pointers to it (ie. they should only
8046 modify their own ones).
8049 *) Made a variety of little tweaks to the ENGINE code.
8050 - "atalla" and "ubsec" string definitions were moved from header files
8051 to C code. "nuron" string definitions were placed in variables
8052 rather than hard-coded - allowing parameterisation of these values
8053 later on via ctrl() commands.
8054 - Removed unused "#if 0"'d code.
8055 - Fixed engine list iteration code so it uses ENGINE_free() to release
8056 structural references.
8057 - Constified the RAND_METHOD element of ENGINE structures.
8058 - Constified various get/set functions as appropriate and added
8059 missing functions (including a catch-all ENGINE_cpy that duplicates
8060 all ENGINE values onto a new ENGINE except reference counts/state).
8061 - Removed NULL parameter checks in get/set functions. Setting a method
8062 or function to NULL is a way of cancelling out a previously set
8063 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8064 and doesn't justify the extra error symbols and code.
8065 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8066 flags from engine_int.h to engine.h.
8067 - Changed prototypes for ENGINE handler functions (init(), finish(),
8068 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8071 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8072 to the algorithm using long division. The binary algorithm can be
8073 used only if the modulus is odd. On 32-bit systems, it is faster
8074 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8075 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8076 up to 450 bits. In 64-bit environments, the binary algorithm
8077 appears to be advantageous for much longer moduli; here we use it
8078 for moduli up to 2048 bits.
8081 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8082 could not support the combine flag in choice fields.
8085 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8086 extensions from a certificate request to the certificate.
8089 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8090 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8091 file: this allows the display of the certificate about to be
8092 signed to be customised, to allow certain fields to be included
8093 or excluded and extension details. The old system didn't display
8094 multicharacter strings properly, omitted fields not in the policy
8095 and couldn't display additional details such as extensions.
8098 *) Function EC_POINTs_mul for multiple scalar multiplication
8099 of an arbitrary number of elliptic curve points
8100 \sum scalars[i]*points[i],
8101 optionally including the generator defined for the EC_GROUP:
8102 scalar*generator + \sum scalars[i]*points[i].
8104 EC_POINT_mul is a simple wrapper function for the typical case
8105 that the point list has just one item (besides the optional
8109 *) First EC_METHODs for curves over GF(p):
8111 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8112 operations and provides various method functions that can also
8113 operate with faster implementations of modular arithmetic.
8115 EC_GFp_mont_method() reuses most functions that are part of
8116 EC_GFp_simple_method, but uses Montgomery arithmetic.
8118 [Bodo Moeller; point addition and point doubling
8119 implementation directly derived from source code provided by
8120 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8122 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8123 crypto/ec/ec_lib.c):
8125 Curves are EC_GROUP objects (with an optional group generator)
8126 based on EC_METHODs that are built into the library.
8128 Points are EC_POINT objects based on EC_GROUP objects.
8130 Most of the framework would be able to handle curves over arbitrary
8131 finite fields, but as there are no obvious types for fields other
8132 than GF(p), some functions are limited to that for now.
8135 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8136 that the file contains a complete HTTP response.
8139 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8140 change the def and num file printf format specifier from "%-40sXXX"
8141 to "%-39s XXX". The latter will always guarantee a space after the
8142 field while the former will cause them to run together if the field
8143 is 40 of more characters long.
8146 *) Constify the cipher and digest 'method' functions and structures
8147 and modify related functions to take constant EVP_MD and EVP_CIPHER
8151 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8152 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8155 *) Modify EVP_Digest*() routines so they now return values. Although the
8156 internal software routines can never fail additional hardware versions
8160 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8162 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8163 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8166 ERR_R_NESTED_ASN1_ERROR
8168 ERR_R_MISSING_ASN1_EOS
8169 were 4 .. 9, conflicting with
8170 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8172 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8173 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8175 Add new error code 'ERR_R_INTERNAL_ERROR'.
8178 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8182 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8183 sets the subject name for a new request or supersedes the
8184 subject name in a given request. Formats that can be parsed are
8185 'CN=Some Name, OU=myOU, C=IT'
8187 'CN=Some Name/OU=myOU/C=IT'.
8189 Add options '-batch' and '-verbose' to 'openssl req'.
8190 [Massimiliano Pala <madwolf@hackmasters.net>]
8192 *) Introduce the possibility to access global variables through
8193 functions on platform were that's the best way to handle exporting
8194 global variables in shared libraries. To enable this functionality,
8195 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8196 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8197 is normally done by Configure or something similar).
8199 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8200 in the source file (foo.c) like this:
8202 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8203 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8205 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8206 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8208 OPENSSL_DECLARE_GLOBAL(int,foo);
8209 #define foo OPENSSL_GLOBAL_REF(foo)
8210 OPENSSL_DECLARE_GLOBAL(double,bar);
8211 #define bar OPENSSL_GLOBAL_REF(bar)
8213 The #defines are very important, and therefore so is including the
8214 header file everywhere where the defined globals are used.
8216 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8217 of ASN.1 items, but that structure is a bit different.
8219 The largest change is in util/mkdef.pl which has been enhanced with
8220 better and easier to understand logic to choose which symbols should
8221 go into the Windows .def files as well as a number of fixes and code
8222 cleanup (among others, algorithm keywords are now sorted
8223 lexicographically to avoid constant rewrites).
8226 *) In BN_div() keep a copy of the sign of 'num' before writing the
8227 result to 'rm' because if rm==num the value will be overwritten
8228 and produce the wrong result if 'num' is negative: this caused
8229 problems with BN_mod() and BN_nnmod().
8232 *) Function OCSP_request_verify(). This checks the signature on an
8233 OCSP request and verifies the signer certificate. The signer
8234 certificate is just checked for a generic purpose and OCSP request
8238 *) Add OCSP_check_validity() function to check the validity of OCSP
8239 responses. OCSP responses are prepared in real time and may only
8240 be a few seconds old. Simply checking that the current time lies
8241 between thisUpdate and nextUpdate max reject otherwise valid responses
8242 caused by either OCSP responder or client clock inaccuracy. Instead
8243 we allow thisUpdate and nextUpdate to fall within a certain period of
8244 the current time. The age of the response can also optionally be
8245 checked. Two new options -validity_period and -status_age added to
8249 *) If signature or public key algorithm is unrecognized print out its
8250 OID rather that just UNKNOWN.
8253 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8254 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8255 ID to be generated from the issuer certificate alone which can then be
8256 passed to OCSP_id_issuer_cmp().
8259 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8260 ASN1 modules to export functions returning ASN1_ITEM pointers
8261 instead of the ASN1_ITEM structures themselves. This adds several
8262 new macros which allow the underlying ASN1 function/structure to
8263 be accessed transparently. As a result code should not use ASN1_ITEM
8264 references directly (such as &X509_it) but instead use the relevant
8265 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8266 use of the new ASN1 code on platforms where exporting structures
8267 is problematical (for example in shared libraries) but exporting
8268 functions returning pointers to structures is not.
8271 *) Add support for overriding the generation of SSL/TLS session IDs.
8272 These callbacks can be registered either in an SSL_CTX or per SSL.
8273 The purpose of this is to allow applications to control, if they wish,
8274 the arbitrary values chosen for use as session IDs, particularly as it
8275 can be useful for session caching in multiple-server environments. A
8276 command-line switch for testing this (and any client code that wishes
8277 to use such a feature) has been added to "s_server".
8278 [Geoff Thorpe, Lutz Jaenicke]
8280 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8281 of the form '#if defined(...) || defined(...) || ...' and
8282 '#if !defined(...) && !defined(...) && ...'. This also avoids
8283 the growing number of special cases it was previously handling.
8286 *) Make all configuration macros available for application by making
8287 sure they are available in opensslconf.h, by giving them names starting
8288 with "OPENSSL_" to avoid conflicts with other packages and by making
8289 sure e_os2.h will cover all platform-specific cases together with
8291 Additionally, it is now possible to define configuration/platform-
8292 specific names (called "system identities"). In the C code, these
8293 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8294 macro with the name beginning with "OPENSSL_SYS_", which is determined
8295 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8299 *) New option -set_serial to 'req' and 'x509' this allows the serial
8300 number to use to be specified on the command line. Previously self
8301 signed certificates were hard coded with serial number 0 and the
8302 CA options of 'x509' had to use a serial number in a file which was
8306 *) New options to 'ca' utility to support V2 CRL entry extensions.
8307 Currently CRL reason, invalidity date and hold instruction are
8308 supported. Add new CRL extensions to V3 code and some new objects.
8311 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8312 disable standard block padding (aka PKCS#5 padding) in the EVP
8313 API, which was previously mandatory. This means that the data is
8314 not padded in any way and so the total length much be a multiple
8315 of the block size, otherwise an error occurs.
8318 *) Initial (incomplete) OCSP SSL support.
8321 *) New function OCSP_parse_url(). This splits up a URL into its host,
8322 port and path components: primarily to parse OCSP URLs. New -url
8323 option to ocsp utility.
8326 *) New nonce behavior. The return value of OCSP_check_nonce() now
8327 reflects the various checks performed. Applications can decide
8328 whether to tolerate certain situations such as an absent nonce
8329 in a response when one was present in a request: the ocsp application
8330 just prints out a warning. New function OCSP_add1_basic_nonce()
8331 this is to allow responders to include a nonce in a response even if
8332 the request is nonce-less.
8335 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8336 skipped when using openssl x509 multiple times on a single input file,
8337 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8340 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8341 set string type: to handle setting ASN1_TIME structures. Fix ca
8342 utility to correctly initialize revocation date of CRLs.
8345 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8346 the clients preferred ciphersuites and rather use its own preferences.
8347 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8348 Internet Explorer by ensuring unchanged hash method during stepup.
8349 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8352 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8353 to aes and add a new 'exist' option to print out symbols that don't
8357 *) Additional options to ocsp utility to allow flags to be set and
8358 additional certificates supplied.
8361 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8362 OCSP client a number of certificate to only verify the response
8366 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8367 handle the new API. Currently only ECB, CBC modes supported. Add new
8370 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8371 Encryption Standard (AES) Ciphersuites for Transport Layer
8372 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8373 not enabled by default and were not part of the "ALL" ciphersuite
8374 alias because they were not yet official; they could be
8375 explicitly requested by specifying the "AESdraft" ciphersuite
8376 group alias. In the final release of OpenSSL 0.9.7, the group
8377 alias is called "AES" and is part of "ALL".)
8378 [Ben Laurie, Steve Henson, Bodo Moeller]
8380 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8381 request to response.
8384 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8385 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8386 extract information from a certificate request. OCSP_response_create()
8387 creates a response and optionally adds a basic response structure.
8388 OCSP_basic_add1_status() adds a complete single response to a basic
8389 response and returns the OCSP_SINGLERESP structure just added (to allow
8390 extensions to be included for example). OCSP_basic_add1_cert() adds a
8391 certificate to a basic response and OCSP_basic_sign() signs a basic
8392 response with various flags. New helper functions ASN1_TIME_check()
8393 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8394 (converts ASN1_TIME to GeneralizedTime).
8397 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8398 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8399 structure from a certificate. X509_pubkey_digest() digests the public_key
8400 contents: this is used in various key identifiers.
8403 *) Make sk_sort() tolerate a NULL argument.
8404 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8406 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8407 passed by the function are trusted implicitly. If any of them signed the
8408 response then it is assumed to be valid and is not verified.
8411 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8412 to data. This was previously part of the PKCS7 ASN1 code. This
8413 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8414 [Steve Henson, reported by Kenneth R. Robinette
8415 <support@securenetterm.com>]
8417 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8418 routines: without these tracing memory leaks is very painful.
8419 Fix leaks in PKCS12 and PKCS7 routines.
8422 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8423 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8424 effectively meant GeneralizedTime would never be used. Now it
8425 is initialised to -1 but X509_time_adj() now has to check the value
8426 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8427 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8428 [Steve Henson, reported by Kenneth R. Robinette
8429 <support@securenetterm.com>]
8431 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8432 result in a zero length in the ASN1_INTEGER structure which was
8433 not consistent with the structure when d2i_ASN1_INTEGER() was used
8434 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8435 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8436 where it did not print out a minus for negative ASN1_INTEGER.
8439 *) Add summary printout to ocsp utility. The various functions which
8440 convert status values to strings have been renamed to:
8441 OCSP_response_status_str(), OCSP_cert_status_str() and
8442 OCSP_crl_reason_str() and are no longer static. New options
8443 to verify nonce values and to disable verification. OCSP response
8444 printout format cleaned up.
8447 *) Add additional OCSP certificate checks. These are those specified
8448 in RFC2560. This consists of two separate checks: the CA of the
8449 certificate being checked must either be the OCSP signer certificate
8450 or the issuer of the OCSP signer certificate. In the latter case the
8451 OCSP signer certificate must contain the OCSP signing extended key
8452 usage. This check is performed by attempting to match the OCSP
8453 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8454 in the OCSP_CERTID structures of the response.
8457 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8458 and related routines. This uses the standard OpenSSL certificate
8459 verify routines to perform initial checks (just CA validity) and
8460 to obtain the certificate chain. Then additional checks will be
8461 performed on the chain. Currently the root CA is checked to see
8462 if it is explicitly trusted for OCSP signing. This is used to set
8463 a root CA as a global signing root: that is any certificate that
8464 chains to that CA is an acceptable OCSP signing certificate.
8467 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8468 extensions from a separate configuration file.
8469 As when reading extensions from the main configuration file,
8470 the '-extensions ...' option may be used for specifying the
8472 [Massimiliano Pala <madwolf@comune.modena.it>]
8474 *) New OCSP utility. Allows OCSP requests to be generated or
8475 read. The request can be sent to a responder and the output
8476 parsed, outputed or printed in text form. Not complete yet:
8477 still needs to check the OCSP response validity.
8480 *) New subcommands for 'openssl ca':
8481 'openssl ca -status <serial>' prints the status of the cert with
8482 the given serial number (according to the index file).
8483 'openssl ca -updatedb' updates the expiry status of certificates
8485 [Massimiliano Pala <madwolf@comune.modena.it>]
8487 *) New '-newreq-nodes' command option to CA.pl. This is like
8488 '-newreq', but calls 'openssl req' with the '-nodes' option
8489 so that the resulting key is not encrypted.
8490 [Damien Miller <djm@mindrot.org>]
8492 *) New configuration for the GNU Hurd.
8493 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8495 *) Initial code to implement OCSP basic response verify. This
8496 is currently incomplete. Currently just finds the signer's
8497 certificate and verifies the signature on the response.
8500 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8501 value of OPENSSLDIR. This is available via the new '-d' option
8502 to 'openssl version', and is also included in 'openssl version -a'.
8505 *) Allowing defining memory allocation callbacks that will be given
8506 file name and line number information in additional arguments
8507 (a const char* and an int). The basic functionality remains, as
8508 well as the original possibility to just replace malloc(),
8509 realloc() and free() by functions that do not know about these
8510 additional arguments. To register and find out the current
8511 settings for extended allocation functions, the following
8512 functions are provided:
8514 CRYPTO_set_mem_ex_functions
8515 CRYPTO_set_locked_mem_ex_functions
8516 CRYPTO_get_mem_ex_functions
8517 CRYPTO_get_locked_mem_ex_functions
8519 These work the same way as CRYPTO_set_mem_functions and friends.
8520 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8521 extended allocation function is enabled.
8522 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8523 a conventional allocation function is enabled.
8524 [Richard Levitte, Bodo Moeller]
8526 *) Finish off removing the remaining LHASH function pointer casts.
8527 There should no longer be any prototype-casting required when using
8528 the LHASH abstraction, and any casts that remain are "bugs". See
8529 the callback types and macros at the head of lhash.h for details
8530 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8533 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8534 If /dev/[u]random devices are not available or do not return enough
8535 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8537 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8538 /etc/entropy will be queried once each in this sequence, querying stops
8539 when enough entropy was collected without querying more sockets.
8542 *) Change the Unix RAND_poll() variant to be able to poll several
8543 random devices, as specified by DEVRANDOM, until a sufficient amount
8544 of data has been collected. We spend at most 10 ms on each file
8545 (select timeout) and read in non-blocking mode. DEVRANDOM now
8546 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8547 (previously it was just the string "/dev/urandom"), so on typical
8548 platforms the 10 ms delay will never occur.
8549 Also separate out the Unix variant to its own file, rand_unix.c.
8550 For VMS, there's a currently-empty rand_vms.c.
8553 *) Move OCSP client related routines to ocsp_cl.c. These
8554 provide utility functions which an application needing
8555 to issue a request to an OCSP responder and analyse the
8556 response will typically need: as opposed to those which an
8557 OCSP responder itself would need which will be added later.
8559 OCSP_request_sign() signs an OCSP request with an API similar
8560 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8561 response. OCSP_response_get1_basic() extracts basic response
8562 from response. OCSP_resp_find_status(): finds and extracts status
8563 information from an OCSP_CERTID structure (which will be created
8564 when the request structure is built). These are built from lower
8565 level functions which work on OCSP_SINGLERESP structures but
8566 won't normally be used unless the application wishes to examine
8567 extensions in the OCSP response for example.
8569 Replace nonce routines with a pair of functions.
8570 OCSP_request_add1_nonce() adds a nonce value and optionally
8571 generates a random value. OCSP_check_nonce() checks the
8572 validity of the nonce in an OCSP response.
8575 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8576 This doesn't copy the supplied OCSP_CERTID and avoids the
8577 need to free up the newly created id. Change return type
8578 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8579 This can then be used to add extensions to the request.
8580 Deleted OCSP_request_new(), since most of its functionality
8581 is now in OCSP_REQUEST_new() (and the case insensitive name
8582 clash) apart from the ability to set the request name which
8583 will be added elsewhere.
8586 *) Update OCSP API. Remove obsolete extensions argument from
8587 various functions. Extensions are now handled using the new
8588 OCSP extension code. New simple OCSP HTTP function which
8589 can be used to send requests and parse the response.
8592 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8593 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8594 uses the special reorder version of SET OF to sort the attributes
8595 and reorder them to match the encoded order. This resolves a long
8596 standing problem: a verify on a PKCS7 structure just after signing
8597 it used to fail because the attribute order did not match the
8598 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8599 it uses the received order. This is necessary to tolerate some broken
8600 software that does not order SET OF. This is handled by encoding
8601 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8602 to produce the required SET OF.
8605 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8606 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8607 files to get correct declarations of the ASN.1 item variables.
8610 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8611 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8612 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8613 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8614 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8615 ASN1_ITEM and no wrapper functions.
8618 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8619 replace the old function pointer based I/O routines. Change most of
8620 the *_d2i_bio() and *_d2i_fp() functions to use these.
8623 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8624 lines, recognize more "algorithms" that can be deselected, and make
8625 it complain about algorithm deselection that isn't recognised.
8628 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8629 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8630 to use new functions. Add NO_ASN1_OLD which can be set to remove
8631 some old style ASN1 functions: this can be used to determine if old
8632 code will still work when these eventually go away.
8635 *) New extension functions for OCSP structures, these follow the
8636 same conventions as certificates and CRLs.
8639 *) New function X509V3_add1_i2d(). This automatically encodes and
8640 adds an extension. Its behaviour can be customised with various
8641 flags to append, replace or delete. Various wrappers added for
8642 certificates and CRLs.
8645 *) Fix to avoid calling the underlying ASN1 print routine when
8646 an extension cannot be parsed. Correct a typo in the
8647 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8650 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8651 entries for variables.
8654 *) Add functionality to apps/openssl.c for detecting locking
8655 problems: As the program is single-threaded, all we have
8656 to do is register a locking callback using an array for
8657 storing which locks are currently held by the program.
8660 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8661 SSL_get_ex_data_X509_STORE_idx(), which is used in
8662 ssl_verify_cert_chain() and thus can be called at any time
8663 during TLS/SSL handshakes so that thread-safety is essential.
8664 Unfortunately, the ex_data design is not at all suited
8665 for multi-threaded use, so it probably should be abolished.
8668 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8669 [Broadcom, tweaked and integrated by Geoff Thorpe]
8671 *) Move common extension printing code to new function
8672 X509V3_print_extensions(). Reorganise OCSP print routines and
8673 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8676 *) New function X509_signature_print() to remove duplication in some
8680 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8681 set (this was treated exactly the same as SET OF previously). This
8682 is used to reorder the STACK representing the structure to match the
8683 encoding. This will be used to get round a problem where a PKCS7
8684 structure which was signed could not be verified because the STACK
8685 order did not reflect the encoded order.
8688 *) Reimplement the OCSP ASN1 module using the new code.
8691 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8692 for its ASN1 operations. The old style function pointers still exist
8693 for now but they will eventually go away.
8696 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8697 completely replaces the old ASN1 functionality with a table driven
8698 encoder and decoder which interprets an ASN1_ITEM structure describing
8699 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8700 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8701 has also been converted to the new form.
8704 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8705 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8706 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8707 for negative moduli.
8710 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8711 of not touching the result's sign bit.
8714 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8718 *) Changed the LHASH code to use prototypes for callbacks, and created
8719 macros to declare and implement thin (optionally static) functions
8720 that provide type-safety and avoid function pointer casting for the
8721 type-specific callbacks.
8724 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8726 [Veers Staats <staatsvr@asc.hpc.mil>,
8727 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8729 *) Reformat the FAQ so the different questions and answers can be divided
8730 in sections depending on the subject.
8733 *) Have the zlib compression code load ZLIB.DLL dynamically under
8737 *) New function BN_mod_sqrt for computing square roots modulo a prime
8738 (using the probabilistic Tonelli-Shanks algorithm unless
8739 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8740 be handled deterministically).
8741 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8743 *) Make BN_mod_inverse faster by explicitly handling small quotients
8744 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8745 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8748 *) New function BN_kronecker.
8751 *) Fix BN_gcd so that it works on negative inputs; the result is
8752 positive unless both parameters are zero.
8753 Previously something reasonably close to an infinite loop was
8754 possible because numbers could be growing instead of shrinking
8755 in the implementation of Euclid's algorithm.
8758 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8759 sign of the number in question.
8761 Fix BN_is_word(a,w) to work correctly for w == 0.
8763 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8764 because its test if the absolute value of 'a' equals 'w'.
8765 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8766 it exists mostly for use in the implementations of BN_is_zero(),
8767 BN_is_one(), and BN_is_word().
8770 *) New function BN_swap.
8773 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8774 the exponentiation functions are more likely to produce reasonable
8775 results on negative inputs.
8778 *) Change BN_mod_mul so that the result is always non-negative.
8779 Previously, it could be negative if one of the factors was negative;
8780 I don't think anyone really wanted that behaviour.
8783 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8784 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8785 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8786 and add new functions:
8795 BN_mod_lshift1_quick
8799 These functions always generate non-negative results.
8801 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8802 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8804 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8805 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8806 be reduced modulo m.
8807 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8810 The following entry accidentally appeared in the CHANGES file
8811 distributed with OpenSSL 0.9.7. The modifications described in
8812 it do *not* apply to OpenSSL 0.9.7.
8814 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8815 was actually never needed) and in BN_mul(). The removal in BN_mul()
8816 required a small change in bn_mul_part_recursive() and the addition
8817 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8818 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8819 bn_sub_words() and bn_add_words() except they take arrays with
8824 *) In 'openssl passwd', verify passwords read from the terminal
8825 unless the '-salt' option is used (which usually means that
8826 verification would just waste user's time since the resulting
8827 hash is going to be compared with some given password hash)
8828 or the new '-noverify' option is used.
8830 This is an incompatible change, but it does not affect
8831 non-interactive use of 'openssl passwd' (passwords on the command
8832 line, '-stdin' option, '-in ...' option) and thus should not
8836 *) Remove all references to RSAref, since there's no more need for it.
8839 *) Make DSO load along a path given through an environment variable
8840 (SHLIB_PATH) with shl_load().
8843 *) Constify the ENGINE code as a result of BIGNUM constification.
8844 Also constify the RSA code and most things related to it. In a
8845 few places, most notable in the depth of the ASN.1 code, ugly
8846 casts back to non-const were required (to be solved at a later
8850 *) Make it so the openssl application has all engines loaded by default.
8853 *) Constify the BIGNUM routines a little more.
8856 *) Add the following functions:
8858 ENGINE_load_cswift()
8860 ENGINE_load_atalla()
8862 ENGINE_load_builtin_engines()
8864 That way, an application can itself choose if external engines that
8865 are built-in in OpenSSL shall ever be used or not. The benefit is
8866 that applications won't have to be linked with libdl or other dso
8867 libraries unless it's really needed.
8869 Changed 'openssl engine' to load all engines on demand.
8870 Changed the engine header files to avoid the duplication of some
8871 declarations (they differed!).
8874 *) 'openssl engine' can now list capabilities.
8877 *) Better error reporting in 'openssl engine'.
8880 *) Never call load_dh_param(NULL) in s_server.
8883 *) Add engine application. It can currently list engines by name and
8884 identity, and test if they are actually available.
8887 *) Improve RPM specification file by forcing symbolic linking and making
8888 sure the installed documentation is also owned by root.root.
8889 [Damien Miller <djm@mindrot.org>]
8891 *) Give the OpenSSL applications more possibilities to make use of
8892 keys (public as well as private) handled by engines.
8895 *) Add OCSP code that comes from CertCo.
8898 *) Add VMS support for the Rijndael code.
8901 *) Added untested support for Nuron crypto accelerator.
8904 *) Add support for external cryptographic devices. This code was
8905 previously distributed separately as the "engine" branch.
8906 [Geoff Thorpe, Richard Levitte]
8908 *) Rework the filename-translation in the DSO code. It is now possible to
8909 have far greater control over how a "name" is turned into a filename
8910 depending on the operating environment and any oddities about the
8911 different shared library filenames on each system.
8914 *) Support threads on FreeBSD-elf in Configure.
8917 *) Fix for SHA1 assembly problem with MASM: it produces
8918 warnings about corrupt line number information when assembling
8919 with debugging information. This is caused by the overlapping
8921 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8924 NCONF_get_number() has no error checking at all. As a replacement,
8925 NCONF_get_number_e() is defined (_e for "error checking") and is
8926 promoted strongly. The old NCONF_get_number is kept around for
8927 binary backward compatibility.
8928 Make it possible for methods to load from something other than a BIO,
8929 by providing a function pointer that is given a name instead of a BIO.
8930 For example, this could be used to load configuration data from an
8934 *) Fix for non blocking accept BIOs. Added new I/O special reason
8935 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8936 with non blocking I/O was not possible because no retry code was
8937 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8941 *) Added the beginnings of Rijndael support.
8944 *) Fix for bug in DirectoryString mask setting. Add support for
8945 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8946 to allow certificate printing to more controllable, additional
8947 'certopt' option to 'x509' to allow new printing options to be
8951 *) Clean old EAY MD5 hack from e_os.h.
8954 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8956 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8957 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8958 [Joe Orton, Steve Henson]
8960 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8962 *) Fix additional bug revealed by the NISCC test suite:
8964 Stop bug triggering large recursion when presented with
8965 certain ASN.1 tags (CVE-2003-0851)
8968 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8970 *) Fix various bugs revealed by running the NISCC test suite:
8972 Stop out of bounds reads in the ASN1 code when presented with
8973 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8975 If verify callback ignores invalid public key errors don't try to check
8976 certificate signature with the NULL public key.
8980 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8981 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8985 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8986 extra data after the compression methods not only for TLS 1.0
8987 but also for SSL 3.0 (as required by the specification).
8988 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8990 *) Change X509_certificate_type() to mark the key as exported/exportable
8991 when it's 512 *bits* long, not 512 bytes.
8994 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8996 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8997 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8998 a protocol version number mismatch like a decryption error
8999 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9002 *) Turn on RSA blinding by default in the default implementation
9003 to avoid a timing attack. Applications that don't want it can call
9004 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9005 They would be ill-advised to do so in most cases.
9006 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9008 *) Change RSA blinding code so that it works when the PRNG is not
9009 seeded (in this case, the secret RSA exponent is abused as
9010 an unpredictable seed -- if it is not unpredictable, there
9011 is no point in blinding anyway). Make RSA blinding thread-safe
9012 by remembering the creator's thread ID in rsa->blinding and
9013 having all other threads use local one-time blinding factors
9014 (this requires more computation than sharing rsa->blinding, but
9015 avoids excessive locking; and if an RSA object is not shared
9016 between threads, blinding will still be very fast).
9019 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9021 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9022 via timing by performing a MAC computation even if incorrect
9023 block cipher padding has been found. This is a countermeasure
9024 against active attacks where the attacker has to distinguish
9025 between bad padding and a MAC verification error. (CVE-2003-0078)
9027 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9028 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9029 Martin Vuagnoux (EPFL, Ilion)]
9031 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9033 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9034 memory from its contents. This is done with a counter that will
9035 place alternating values in each byte. This can be used to solve
9036 two issues: 1) the removal of calls to memset() by highly optimizing
9037 compilers, and 2) cleansing with other values than 0, since those can
9038 be read through on certain media, for example a swap space on disk.
9041 *) Bugfix: client side session caching did not work with external caching,
9042 because the session->cipher setting was not restored when reloading
9043 from the external cache. This problem was masked, when
9044 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9045 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9048 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9049 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9050 [Zeev Lieber <zeev-l@yahoo.com>]
9052 *) Undo an undocumented change introduced in 0.9.6e which caused
9053 repeated calls to OpenSSL_add_all_ciphers() and
9054 OpenSSL_add_all_digests() to be ignored, even after calling
9058 *) Change the default configuration reader to deal with last line not
9059 being properly terminated.
9062 *) Change X509_NAME_cmp() so it applies the special rules on handling
9063 DN values that are of type PrintableString, as well as RDNs of type
9064 emailAddress where the value has the type ia5String.
9065 [stefank@valicert.com via Richard Levitte]
9067 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9068 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9069 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9070 the bitwise-OR of the two for use by the majority of applications
9071 wanting this behaviour, and update the docs. The documented
9072 behaviour and actual behaviour were inconsistent and had been
9073 changing anyway, so this is more a bug-fix than a behavioural
9075 [Geoff Thorpe, diagnosed by Nadav Har'El]
9077 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9078 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9081 *) Fix initialization code race conditions in
9082 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9083 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9084 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9085 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9086 ssl2_get_cipher_by_char(),
9087 ssl3_get_cipher_by_char().
9088 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9090 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9091 the cached sessions are flushed, as the remove_cb() might use ex_data
9092 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9093 (see [openssl.org #212]).
9094 [Geoff Thorpe, Lutz Jaenicke]
9096 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9097 length, instead of the encoding length to d2i_ASN1_OBJECT.
9100 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9102 *) [In 0.9.6g-engine release:]
9103 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9104 [Lynn Gazis <lgazis@rainbow.com>]
9106 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9108 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9109 and get fix the header length calculation.
9110 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9111 Alon Kantor <alonk@checkpoint.com> (and others),
9114 *) Use proper error handling instead of 'assertions' in buffer
9115 overflow checks added in 0.9.6e. This prevents DoS (the
9116 assertions could call abort()).
9117 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9119 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9121 *) Add various sanity checks to asn1_get_length() to reject
9122 the ASN1 length bytes if they exceed sizeof(long), will appear
9123 negative or the content length exceeds the length of the
9125 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9127 *) Fix cipher selection routines: ciphers without encryption had no flags
9128 for the cipher strength set and where therefore not handled correctly
9129 by the selection routines (PR #130).
9132 *) Fix EVP_dsa_sha macro.
9136 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9137 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9138 that was added in OpenSSL 0.9.6d.
9140 As the countermeasure turned out to be incompatible with some
9141 broken SSL implementations, the new option is part of SSL_OP_ALL.
9142 SSL_OP_ALL is usually employed when compatibility with weird SSL
9143 implementations is desired (e.g. '-bugs' option to 's_client' and
9144 's_server'), so the new option is automatically set in many
9148 *) Changes in security patch:
9150 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9151 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9152 Air Force Materiel Command, USAF, under agreement number
9155 *) Add various sanity checks to asn1_get_length() to reject
9156 the ASN1 length bytes if they exceed sizeof(long), will appear
9157 negative or the content length exceeds the length of the
9158 supplied buffer. (CVE-2002-0659)
9159 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9161 *) Assertions for various potential buffer overflows, not known to
9163 [Ben Laurie (CHATS)]
9165 *) Various temporary buffers to hold ASCII versions of integers were
9166 too small for 64 bit platforms. (CVE-2002-0655)
9167 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9169 *) Remote buffer overflow in SSL3 protocol - an attacker could
9170 supply an oversized session ID to a client. (CVE-2002-0656)
9171 [Ben Laurie (CHATS)]
9173 *) Remote buffer overflow in SSL2 protocol - an attacker could
9174 supply an oversized client master key. (CVE-2002-0656)
9175 [Ben Laurie (CHATS)]
9177 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9179 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9180 encoded as NULL) with id-dsa-with-sha1.
9181 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9183 *) Check various X509_...() return values in apps/req.c.
9184 [Nils Larsch <nla@trustcenter.de>]
9186 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9187 an end-of-file condition would erroneously be flagged, when the CRLF
9188 was just at the end of a processed block. The bug was discovered when
9189 processing data through a buffering memory BIO handing the data to a
9190 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9191 <ptsekov@syntrex.com> and Nedelcho Stanev.
9194 *) Implement a countermeasure against a vulnerability recently found
9195 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9196 before application data chunks to avoid the use of known IVs
9197 with data potentially chosen by the attacker.
9200 *) Fix length checks in ssl3_get_client_hello().
9203 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9204 to prevent ssl3_read_internal() from incorrectly assuming that
9205 ssl3_read_bytes() found application data while handshake
9206 processing was enabled when in fact s->s3->in_read_app_data was
9207 merely automatically cleared during the initial handshake.
9208 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9210 *) Fix object definitions for Private and Enterprise: they were not
9211 recognized in their shortname (=lowercase) representation. Extend
9212 obj_dat.pl to issue an error when using undefined keywords instead
9213 of silently ignoring the problem (Svenning Sorensen
9214 <sss@sss.dnsalias.net>).
9217 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9218 generators, i.e. generators other than 2 and 5. (Previously, the
9219 code did not properly initialise the 'add' and 'rem' values to
9220 BN_generate_prime().)
9222 In the new general case, we do not insist that 'generator' is
9223 actually a primitive root: This requirement is rather pointless;
9224 a generator of the order-q subgroup is just as good, if not
9228 *) Map new X509 verification errors to alerts. Discovered and submitted by
9229 Tom Wu <tom@arcot.com>.
9232 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9233 returning non-zero before the data has been completely received
9234 when using non-blocking I/O.
9235 [Bodo Moeller; problem pointed out by John Hughes]
9237 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9238 [Ben Laurie, Lutz Jaenicke]
9240 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9241 Yoram Zahavi <YoramZ@gilian.com>).
9244 *) Add information about CygWin 1.3 and on, and preserve proper
9245 configuration for the versions before that.
9246 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9248 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9249 check whether we deal with a copy of a session and do not delete from
9250 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9251 <izhar@checkpoint.com>.
9254 *) Do not store session data into the internal session cache, if it
9255 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9256 flag is set). Proposed by Aslam <aslam@funk.com>.
9259 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9263 *) [In 0.9.6d-engine release:]
9264 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9265 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9267 *) Add the configuration target linux-s390x.
9268 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9270 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9271 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9272 variable as an indication that a ClientHello message has been
9273 received. As the flag value will be lost between multiple
9274 invocations of ssl3_accept when using non-blocking I/O, the
9275 function may not be aware that a handshake has actually taken
9276 place, thus preventing a new session from being added to the
9279 To avoid this problem, we now set s->new_session to 2 instead of
9280 using a local variable.
9281 [Lutz Jaenicke, Bodo Moeller]
9283 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9284 if the SSL_R_LENGTH_MISMATCH error is detected.
9285 [Geoff Thorpe, Bodo Moeller]
9287 *) New 'shared_ldflag' column in Configure platform table.
9290 *) Fix EVP_CIPHER_mode macro.
9291 ["Dan S. Camper" <dan@bti.net>]
9293 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9294 type, we must throw them away by setting rr->length to 0.
9295 [D P Chang <dpc@qualys.com>]
9297 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9299 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9300 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9301 worked incorrectly for those cases where range = 10..._2 and
9302 3*range is two bits longer than range.)
9305 *) Only add signing time to PKCS7 structures if it is not already
9309 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9310 OBJ_ld_ce should be OBJ_id_ce.
9311 Also some ip-pda OIDs in crypto/objects/objects.txt were
9312 incorrect (cf. RFC 3039).
9313 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9315 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9316 returns early because it has nothing to do.
9317 [Andy Schneider <andy.schneider@bjss.co.uk>]
9319 *) [In 0.9.6c-engine release:]
9320 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9321 [Andy Schneider <andy.schneider@bjss.co.uk>]
9323 *) [In 0.9.6c-engine release:]
9324 Add support for Cryptographic Appliance's keyserver technology.
9325 (Use engine 'keyclient')
9326 [Cryptographic Appliances and Geoff Thorpe]
9328 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9329 is called via tools/c89.sh because arguments have to be
9330 rearranged (all '-L' options must appear before the first object
9332 [Richard Shapiro <rshapiro@abinitio.com>]
9334 *) [In 0.9.6c-engine release:]
9335 Add support for Broadcom crypto accelerator cards, backported
9337 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9339 *) [In 0.9.6c-engine release:]
9340 Add support for SureWare crypto accelerator cards from
9341 Baltimore Technologies. (Use engine 'sureware')
9342 [Baltimore Technologies and Mark Cox]
9344 *) [In 0.9.6c-engine release:]
9345 Add support for crypto accelerator cards from Accelerated
9346 Encryption Processing, www.aep.ie. (Use engine 'aep')
9347 [AEP Inc. and Mark Cox]
9349 *) Add a configuration entry for gcc on UnixWare.
9350 [Gary Benson <gbenson@redhat.com>]
9352 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9353 messages are stored in a single piece (fixed-length part and
9354 variable-length part combined) and fix various bugs found on the way.
9357 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9358 instead. BIO_gethostbyname() does not know what timeouts are
9359 appropriate, so entries would stay in cache even when they have
9361 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9363 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9364 faced with a pathologically small ClientHello fragment that does
9365 not contain client_version: Instead of aborting with an error,
9366 simply choose the highest available protocol version (i.e.,
9367 TLS 1.0 unless it is disabled). In practice, ClientHello
9368 messages are never sent like this, but this change gives us
9369 strictly correct behaviour at least for TLS.
9372 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9373 never resets s->method to s->ctx->method when called from within
9374 one of the SSL handshake functions.
9375 [Bodo Moeller; problem pointed out by Niko Baric]
9377 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9378 (sent using the client's version number) if client_version is
9379 smaller than the protocol version in use. Also change
9380 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9381 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9382 the client will at least see that alert.
9385 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9389 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9390 client receives HelloRequest while in a handshake.
9391 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9393 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9394 should end in 'break', not 'goto end' which circumvents various
9395 cleanups done in state SSL_ST_OK. But session related stuff
9396 must be disabled for SSL_ST_OK in the case that we just sent a
9399 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9400 before just sending a HelloRequest.
9401 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9403 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9404 reveal whether illegal block cipher padding was found or a MAC
9405 verification error occurred. (Neither SSLerr() codes nor alerts
9406 are directly visible to potential attackers, but the information
9407 may leak via logfiles.)
9409 Similar changes are not required for the SSL 2.0 implementation
9410 because the number of padding bytes is sent in clear for SSL 2.0,
9411 and the extra bytes are just ignored. However ssl/s2_pkt.c
9412 failed to verify that the purported number of padding bytes is in
9416 *) Add OpenUNIX-8 support including shared libraries
9417 (Boyd Lynn Gerber <gerberb@zenez.com>).
9420 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9421 'wristwatch attack' using huge encoding parameters (cf.
9422 James H. Manger's CRYPTO 2001 paper). Note that the
9423 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9424 encoding parameters and hence was not vulnerable.
9427 *) BN_sqr() bug fix.
9428 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9430 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9431 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9432 followed by modular reduction.
9433 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9435 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9436 equivalent based on BN_pseudo_rand() instead of BN_rand().
9439 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9440 This function was broken, as the check for a new client hello message
9441 to handle SGC did not allow these large messages.
9442 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9445 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9448 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9449 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9452 *) Rework the configuration and shared library support for Tru64 Unix.
9453 The configuration part makes use of modern compiler features and
9454 still retains old compiler behavior for those that run older versions
9455 of the OS. The shared library support part includes a variant that
9456 uses the RPATH feature, and is available through the special
9457 configuration target "alpha-cc-rpath", which will never be selected
9459 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9461 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9462 with the same message size as in ssl3_get_certificate_request().
9463 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9464 messages might inadvertently be reject as too long.
9465 [Petr Lampa <lampa@fee.vutbr.cz>]
9467 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9470 *) Modified SSL library such that the verify_callback that has been set
9471 specificly for an SSL object with SSL_set_verify() is actually being
9472 used. Before the change, a verify_callback set with this function was
9473 ignored and the verify_callback() set in the SSL_CTX at the time of
9474 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9475 to allow the necessary settings.
9478 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9479 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9480 done automatically (in contradiction to the requirements of the C
9481 standard). This made problems when used from OpenSSH.
9484 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9485 dh->length and always used
9487 BN_rand_range(priv_key, dh->p).
9489 BN_rand_range() is not necessary for Diffie-Hellman, and this
9490 specific range makes Diffie-Hellman unnecessarily inefficient if
9491 dh->length (recommended exponent length) is much smaller than the
9492 length of dh->p. We could use BN_rand_range() if the order of
9493 the subgroup was stored in the DH structure, but we only have
9498 BN_rand(priv_key, l, ...)
9500 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9506 RSA_eay_public_encrypt
9507 RSA_eay_private_decrypt
9508 RSA_eay_private_encrypt (signing)
9509 RSA_eay_public_decrypt (signature verification)
9511 (default implementations for RSA_public_encrypt,
9512 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9513 always reject numbers >= n.
9516 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9517 to synchronize access to 'locking_thread'. This is necessary on
9518 systems where access to 'locking_thread' (an 'unsigned long'
9519 variable) is not atomic.
9522 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9523 *before* setting the 'crypto_lock_rand' flag. The previous code had
9524 a race condition if 0 is a valid thread ID.
9525 [Travis Vitek <vitek@roguewave.com>]
9527 *) Add support for shared libraries under Irix.
9528 [Albert Chin-A-Young <china@thewrittenword.com>]
9530 *) Add configuration option to build on Linux on both big-endian and
9532 [Ralf Baechle <ralf@uni-koblenz.de>]
9534 *) Add the possibility to create shared libraries on HP-UX.
9537 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9539 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9540 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9541 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9542 PRNG state recovery was possible based on the output of
9543 one PRNG request appropriately sized to gain knowledge on
9544 'md' followed by enough consecutive 1-byte PRNG requests
9545 to traverse all of 'state'.
9547 1. When updating 'md_local' (the current thread's copy of 'md')
9548 during PRNG output generation, hash all of the previous
9549 'md_local' value, not just the half used for PRNG output.
9551 2. Make the number of bytes from 'state' included into the hash
9552 independent from the number of PRNG bytes requested.
9554 The first measure alone would be sufficient to avoid
9555 Markku-Juhani's attack. (Actually it had never occurred
9556 to me that the half of 'md_local' used for chaining was the
9557 half from which PRNG output bytes were taken -- I had always
9558 assumed that the secret half would be used.) The second
9559 measure makes sure that additional data from 'state' is never
9560 mixed into 'md_local' in small portions; this heuristically
9561 further strengthens the PRNG.
9564 *) Fix crypto/bn/asm/mips3.s.
9567 *) When only the key is given to "enc", the IV is undefined. Print out
9568 an error message in this case.
9571 *) Handle special case when X509_NAME is empty in X509 printing routines.
9574 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9575 positive and less than q.
9578 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9579 used: it isn't thread safe and the add_lock_callback should handle
9581 [Paul Rose <Paul.Rose@bridge.com>]
9583 *) Verify that incoming data obeys the block size in
9584 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9588 [Ulf Möller, Bodo Möller]
9590 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9591 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9592 when fixing the server behaviour for backwards-compatible 'client
9593 hello' messages. (Note that the attack is impractical against
9594 SSL 3.0 and TLS 1.0 anyway because length and version checking
9595 means that the probability of guessing a valid ciphertext is
9596 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9599 Before 0.9.5, the countermeasure (hide the error by generating a
9600 random 'decryption result') did not work properly because
9601 ERR_clear_error() was missing, meaning that SSL_get_error() would
9602 detect the supposedly ignored error.
9604 Both problems are now fixed.
9607 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9608 (previously it was 1024).
9611 *) Fix for compatibility mode trust settings: ignore trust settings
9612 unless some valid trust or reject settings are present.
9615 *) Fix for blowfish EVP: its a variable length cipher.
9618 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9619 parameters in DSA public key structures and return an error in the
9620 DSA routines if parameters are absent.
9623 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9624 in the current directory if neither $RANDFILE nor $HOME was set.
9625 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9626 caused some confusion to Windows users who haven't defined $HOME.
9627 Thus RAND_file_name() is changed again: e_os.h can define a
9628 DEFAULT_HOME, which will be used if $HOME is not set.
9629 For Windows, we use "C:"; on other platforms, we still require
9630 environment variables.
9632 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9633 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9634 having multiple threads call RAND_poll() concurrently.
9637 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9638 combination of a flag and a thread ID variable.
9639 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9640 flag), *other* threads can enter ssleay_add_bytes without obeying
9641 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9642 that they do not hold after the first thread unsets add_do_not_lock).
9645 *) Change bctest again: '-x' expressions are not available in all
9649 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9651 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9652 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9654 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9655 the default extension for executables, if any. Also, make the perl
9656 scripts that use symlink() to test if it really exists and use "cp"
9657 if it doesn't. All this made OpenSSL compilable and installable in
9661 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9662 If SEQUENCE is length is indefinite just set c->slen to the total
9663 amount of data available.
9664 [Steve Henson, reported by shige@FreeBSD.org]
9665 [This change does not apply to 0.9.7.]
9667 *) Change bctest to avoid here-documents inside command substitution
9668 (workaround for FreeBSD /bin/sh bug).
9669 For compatibility with Ultrix, avoid shell functions (introduced
9670 in the bctest version that searches along $PATH).
9673 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9674 with des_encrypt() defined on some operating systems, like Solaris
9678 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9679 On the Importance of Eliminating Errors in Cryptographic
9680 Computations, J. Cryptology 14 (2001) 2, 101-119,
9681 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9684 *) MIPS assembler BIGNUM division bug fix.
9687 *) Disabled incorrect Alpha assembler code.
9690 *) Fix PKCS#7 decode routines so they correctly update the length
9691 after reading an EOC for the EXPLICIT tag.
9693 [This change does not apply to 0.9.7.]
9695 *) Fix bug in PKCS#12 key generation routines. This was triggered
9696 if a 3DES key was generated with a 0 initial byte. Include
9697 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9698 (but broken) behaviour.
9701 *) Enhance bctest to search for a working bc along $PATH and print
9703 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9705 *) Fix memory leaks in err.c: free err_data string if necessary;
9706 don't write to the wrong index in ERR_set_error_data.
9709 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9713 *) Replace rdtsc with _emit statements for VC++ version 5.
9714 [Jeremy Cooper <jeremy@baymoo.org>]
9716 *) Make it possible to reuse SSLv2 sessions.
9719 *) In copy_email() check for >= 0 as a return value for
9720 X509_NAME_get_index_by_NID() since 0 is a valid index.
9721 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9723 *) Avoid coredump with unsupported or invalid public keys by checking if
9724 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9725 PKCS7_verify() fails with non detached data.
9728 *) Don't use getenv in library functions when run as setuid/setgid.
9729 New function OPENSSL_issetugid().
9732 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9733 due to incorrect handling of multi-threading:
9735 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9737 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9739 3. Count how many times MemCheck_off() has been called so that
9740 nested use can be treated correctly. This also avoids
9741 inband-signalling in the previous code (which relied on the
9742 assumption that thread ID 0 is impossible).
9745 *) Add "-rand" option also to s_client and s_server.
9748 *) Fix CPU detection on Irix 6.x.
9749 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9750 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9752 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9755 [This change does not apply to 0.9.7.]
9757 *) Use the cached encoding of an X509_NAME structure rather than
9758 copying it. This is apparently the reason for the libsafe "errors"
9759 but the code is actually correct.
9762 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9763 Bleichenbacher's DSA attack.
9764 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9765 to be set and top=0 forces the highest bit to be set; top=-1 is new
9766 and leaves the highest bit random.
9767 [Ulf Moeller, Bodo Moeller]
9769 *) In the NCONF_...-based implementations for CONF_... queries
9770 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9771 a temporary CONF structure with the data component set to NULL
9772 (which gives segmentation faults in lh_retrieve).
9773 Instead, use NULL for the CONF pointer in CONF_get_string and
9774 CONF_get_number (which may use environment variables) and directly
9775 return NULL from CONF_get_section.
9778 *) Fix potential buffer overrun for EBCDIC.
9781 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9782 keyUsage if basicConstraints absent for a CA.
9785 *) Make SMIME_write_PKCS7() write mail header values with a format that
9786 is more generally accepted (no spaces before the semicolon), since
9787 some programs can't parse those values properly otherwise. Also make
9788 sure BIO's that break lines after each write do not create invalid
9792 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9793 macros previously used would not encode an empty SEQUENCE OF
9794 and break the signature.
9796 [This change does not apply to 0.9.7.]
9798 *) Zero the premaster secret after deriving the master secret in
9802 *) Add some EVP_add_digest_alias registrations (as found in
9803 OpenSSL_add_all_digests()) to SSL_library_init()
9804 aka OpenSSL_add_ssl_algorithms(). This provides improved
9805 compatibility with peers using X.509 certificates
9806 with unconventional AlgorithmIdentifier OIDs.
9809 *) Fix for Irix with NO_ASM.
9810 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9812 *) ./config script fixes.
9813 [Ulf Moeller, Richard Levitte]
9815 *) Fix 'openssl passwd -1'.
9818 *) Change PKCS12_key_gen_asc() so it can cope with non null
9819 terminated strings whose length is passed in the passlen
9820 parameter, for example from PEM callbacks. This was done
9821 by adding an extra length parameter to asc2uni().
9822 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9824 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9825 call failed, free the DSA structure.
9828 *) Fix to uni2asc() to cope with zero length Unicode strings.
9829 These are present in some PKCS#12 files.
9832 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9833 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9834 when writing a 32767 byte record.
9835 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9837 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9838 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9840 (RSA objects have a reference count access to which is protected
9841 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9842 so they are meant to be shared between threads.)
9843 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9844 "Reddie, Steven" <Steven.Reddie@ca.com>]
9846 *) Fix a deadlock in CRYPTO_mem_leaks().
9849 *) Use better test patterns in bntest.
9852 *) rand_win.c fix for Borland C.
9855 *) BN_rshift bugfix for n == 0.
9858 *) Add a 'bctest' script that checks for some known 'bc' bugs
9859 so that 'make test' does not abort just because 'bc' is broken.
9862 *) Store verify_result within SSL_SESSION also for client side to
9863 avoid potential security hole. (Re-used sessions on the client side
9864 always resulted in verify_result==X509_V_OK, not using the original
9865 result of the server certificate verification.)
9868 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9869 SSL3_RT_APPLICATION_DATA, return 0.
9870 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9874 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9875 releases, have been re-implemented by renaming the previous
9876 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9877 and ssl3_read_internal, respectively, and adding 'peek' parameters
9878 to them. The new ssl[23]_{read,peek} functions are calls to
9879 ssl[23]_read_internal with the 'peek' flag set appropriately.
9880 A 'peek' parameter has also been added to ssl3_read_bytes, which
9881 does the actual work for ssl3_read_internal.
9884 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9885 the method-specific "init()" handler. Also clean up ex_data after
9886 calling the method-specific "finish()" handler. Previously, this was
9887 happening the other way round.
9890 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9891 The previous value, 12, was not always sufficient for BN_mod_exp().
9894 *) Make sure that shared libraries get the internal name engine with
9895 the full version number and not just 0. This should mark the
9896 shared libraries as not backward compatible. Of course, this should
9897 be changed again when we can guarantee backward binary compatibility.
9900 *) Fix typo in get_cert_by_subject() in by_dir.c
9901 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9903 *) Rework the system to generate shared libraries:
9905 - Make note of the expected extension for the shared libraries and
9906 if there is a need for symbolic links from for example libcrypto.so.0
9907 to libcrypto.so.0.9.7. There is extended info in Configure for
9910 - Make as few rebuilds of the shared libraries as possible.
9912 - Still avoid linking the OpenSSL programs with the shared libraries.
9914 - When installing, install the shared libraries separately from the
9918 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9920 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9921 and not in SSL_clear because the latter is also used by the
9922 accept/connect functions; previously, the settings made by
9923 SSL_set_read_ahead would be lost during the handshake.
9924 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9926 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9927 Previously, it would create entries for disabled algorithms no
9931 *) Added several new manual pages for SSL_* function.
9934 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9936 *) In ssl23_get_client_hello, generate an error message when faced
9937 with an initial SSL 3.0/TLS record that is too small to contain the
9938 first two bytes of the ClientHello message, i.e. client_version.
9939 (Note that this is a pathologic case that probably has never happened
9940 in real life.) The previous approach was to use the version number
9941 from the record header as a substitute; but our protocol choice
9942 should not depend on that one because it is not authenticated
9943 by the Finished messages.
9946 *) More robust randomness gathering functions for Windows.
9947 [Jeffrey Altman <jaltman@columbia.edu>]
9949 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9950 not set then we don't setup the error code for issuer check errors
9951 to avoid possibly overwriting other errors which the callback does
9952 handle. If an application does set the flag then we assume it knows
9953 what it is doing and can handle the new informational codes
9957 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9958 a general "ANY" type, as such it should be able to decode anything
9959 including tagged types. However it didn't check the class so it would
9960 wrongly interpret tagged types in the same way as their universal
9961 counterpart and unknown types were just rejected. Changed so that the
9962 tagged and unknown types are handled in the same way as a SEQUENCE:
9963 that is the encoding is stored intact. There is also a new type
9964 "V_ASN1_OTHER" which is used when the class is not universal, in this
9965 case we have no idea what the actual type is so we just lump them all
9969 *) On VMS, stdout may very well lead to a file that is written to
9970 in a record-oriented fashion. That means that every write() will
9971 write a separate record, which will be read separately by the
9972 programs trying to read from it. This can be very confusing.
9974 The solution is to put a BIO filter in the way that will buffer
9975 text until a linefeed is reached, and then write everything a
9976 line at a time, so every record written will be an actual line,
9977 not chunks of lines and not (usually doesn't happen, but I've
9978 seen it once) several lines in one record. BIO_f_linebuffer() is
9981 Currently, it's a VMS-only method, because that's where it has
9982 been tested well enough.
9985 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9986 it can return incorrect results.
9987 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9988 but it was in 0.9.6-beta[12].)
9991 *) Disable the check for content being present when verifying detached
9992 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9993 include zero length content when signing messages.
9996 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9997 BIO_ctrl (for BIO pairs).
10000 *) Add DSO method for VMS.
10003 *) Bug fix: Montgomery multiplication could produce results with the
10007 *) Add RPM specification openssl.spec and modify it to build three
10008 packages. The default package contains applications, application
10009 documentation and run-time libraries. The devel package contains
10010 include files, static libraries and function documentation. The
10011 doc package contains the contents of the doc directory. The original
10012 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10015 *) Add a large number of documentation files for many SSL routines.
10016 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10018 *) Add a configuration entry for Sony News 4.
10019 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10021 *) Don't set the two most significant bits to one when generating a
10022 random number < q in the DSA library.
10025 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10026 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10027 the underlying transport is blocking) if a handshake took place.
10028 (The default behaviour is needed by applications such as s_client
10029 and s_server that use select() to determine when to use SSL_read;
10030 but for applications that know in advance when to expect data, it
10031 just makes things more complicated.)
10034 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10038 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10039 work better on such systems.
10040 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10042 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10043 Update PKCS12_parse() so it copies the friendlyName and the
10044 keyid to the certificates aux info.
10047 *) Fix bug in PKCS7_verify() which caused an infinite loop
10048 if there was more than one signature.
10049 [Sven Uszpelkat <su@celocom.de>]
10051 *) Major change in util/mkdef.pl to include extra information
10052 about each symbol, as well as presenting variables as well
10053 as functions. This change means that there's n more need
10054 to rebuild the .num files when some algorithms are excluded.
10057 *) Allow the verify time to be set by an application,
10058 rather than always using the current time.
10061 *) Phase 2 verify code reorganisation. The certificate
10062 verify code now looks up an issuer certificate by a
10063 number of criteria: subject name, authority key id
10064 and key usage. It also verifies self signed certificates
10065 by the same criteria. The main comparison function is
10066 X509_check_issued() which performs these checks.
10068 Lot of changes were necessary in order to support this
10069 without completely rewriting the lookup code.
10071 Authority and subject key identifier are now cached.
10073 The LHASH 'certs' is X509_STORE has now been replaced
10074 by a STACK_OF(X509_OBJECT). This is mainly because an
10075 LHASH can't store or retrieve multiple objects with
10076 the same hash value.
10078 As a result various functions (which were all internal
10079 use only) have changed to handle the new X509_STORE
10080 structure. This will break anything that messed round
10081 with X509_STORE internally.
10083 The functions X509_STORE_add_cert() now checks for an
10084 exact match, rather than just subject name.
10086 The X509_STORE API doesn't directly support the retrieval
10087 of multiple certificates matching a given criteria, however
10088 this can be worked round by performing a lookup first
10089 (which will fill the cache with candidate certificates)
10090 and then examining the cache for matches. This is probably
10091 the best we can do without throwing out X509_LOOKUP
10092 entirely (maybe later...).
10094 The X509_VERIFY_CTX structure has been enhanced considerably.
10096 All certificate lookup operations now go via a get_issuer()
10097 callback. Although this currently uses an X509_STORE it
10098 can be replaced by custom lookups. This is a simple way
10099 to bypass the X509_STORE hackery necessary to make this
10100 work and makes it possible to use more efficient techniques
10101 in future. A very simple version which uses a simple
10102 STACK for its trusted certificate store is also provided
10103 using X509_STORE_CTX_trusted_stack().
10105 The verify_cb() and verify() callbacks now have equivalents
10106 in the X509_STORE_CTX structure.
10108 X509_STORE_CTX also has a 'flags' field which can be used
10109 to customise the verify behaviour.
10112 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10113 excludes S/MIME capabilities.
10116 *) When a certificate request is read in keep a copy of the
10117 original encoding of the signed data and use it when outputting
10118 again. Signatures then use the original encoding rather than
10119 a decoded, encoded version which may cause problems if the
10120 request is improperly encoded.
10123 *) For consistency with other BIO_puts implementations, call
10124 buffer_write(b, ...) directly in buffer_puts instead of calling
10127 In BIO_puts, increment b->num_write as in BIO_write.
10128 [Peter.Sylvester@EdelWeb.fr]
10130 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10131 BN_zero, we may not return a BIGNUM with an array consisting of
10132 words set to zero.)
10135 *) Avoid calling abort() from within the library when problems are
10136 detected, except if preprocessor symbols have been defined
10137 (such as REF_CHECK, BN_DEBUG etc.).
10140 *) New openssl application 'rsautl'. This utility can be
10141 used for low level RSA operations. DER public key
10142 BIO/fp routines also added.
10145 *) New Configure entry and patches for compiling on QNX 4.
10146 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10148 *) A demo state-machine implementation was sponsored by
10149 Nuron (http://www.nuron.com/) and is now available in
10150 demos/state_machine.
10153 *) New options added to the 'dgst' utility for signature
10154 generation and verification.
10157 *) Unrecognized PKCS#7 content types are now handled via a
10158 catch all ASN1_TYPE structure. This allows unsupported
10159 types to be stored as a "blob" and an application can
10160 encode and decode it manually.
10163 *) Fix various signed/unsigned issues to make a_strex.c
10164 compile under VC++.
10165 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10167 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10168 length if passed a buffer. ASN1_INTEGER_to_BN failed
10169 if passed a NULL BN and its argument was negative.
10170 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10172 *) Modification to PKCS#7 encoding routines to output definite
10173 length encoding. Since currently the whole structures are in
10174 memory there's not real point in using indefinite length
10175 constructed encoding. However if OpenSSL is compiled with
10176 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10179 *) Added BIO_vprintf() and BIO_vsnprintf().
10182 *) Added more prefixes to parse for in the strings written
10183 through a logging bio, to cover all the levels that are available
10184 through syslog. The prefixes are now:
10186 PANIC, EMERG, EMR => LOG_EMERG
10187 ALERT, ALR => LOG_ALERT
10188 CRIT, CRI => LOG_CRIT
10189 ERROR, ERR => LOG_ERR
10190 WARNING, WARN, WAR => LOG_WARNING
10191 NOTICE, NOTE, NOT => LOG_NOTICE
10192 INFO, INF => LOG_INFO
10193 DEBUG, DBG => LOG_DEBUG
10195 and as before, if none of those prefixes are present at the
10196 beginning of the string, LOG_ERR is chosen.
10198 On Win32, the LOG_* levels are mapped according to this:
10200 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10201 LOG_WARNING => EVENTLOG_WARNING_TYPE
10202 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10206 *) Made it possible to reconfigure with just the configuration
10207 argument "reconf" or "reconfigure". The command line arguments
10208 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10209 and are retrieved from there when reconfiguring.
10212 *) MD4 implemented.
10213 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10215 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10218 *) The obj_dat.pl script was messing up the sorting of object
10219 names. The reason was that it compared the quoted version
10220 of strings as a result "OCSP" > "OCSP Signing" because
10221 " > SPACE. Changed script to store unquoted versions of
10222 names and add quotes on output. It was also omitting some
10223 names from the lookup table if they were given a default
10224 value (that is if SN is missing it is given the same
10225 value as LN and vice versa), these are now added on the
10226 grounds that if an object has a name we should be able to
10227 look it up. Finally added warning output when duplicate
10228 short or long names are found.
10231 *) Changes needed for Tandem NSK.
10232 [Scott Uroff <scott@xypro.com>]
10234 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10235 RSA_padding_check_SSLv23(), special padding was never detected
10236 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10237 version rollback attacks was not effective.
10239 In s23_clnt.c, don't use special rollback-attack detection padding
10240 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10241 client; similarly, in s23_srvr.c, don't do the rollback check if
10242 SSL 2.0 is the only protocol enabled in the server.
10245 *) Make it possible to get hexdumps of unprintable data with 'openssl
10246 asn1parse'. By implication, the functions ASN1_parse_dump() and
10247 BIO_dump_indent() are added.
10250 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10251 these print out strings and name structures based on various
10252 flags including RFC2253 support and proper handling of
10253 multibyte characters. Added options to the 'x509' utility
10254 to allow the various flags to be set.
10257 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10258 Also change the functions X509_cmp_current_time() and
10259 X509_gmtime_adj() work with an ASN1_TIME structure,
10260 this will enable certificates using GeneralizedTime in validity
10261 dates to be checked.
10264 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10265 negative public key encodings) on by default,
10266 NO_NEG_PUBKEY_BUG can be set to disable it.
10269 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10270 content octets. An i2c_ASN1_OBJECT is unnecessary because
10271 the encoding can be trivially obtained from the structure.
10274 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10275 not read locks (CRYPTO_r_[un]lock).
10278 *) A first attempt at creating official support for shared
10279 libraries through configuration. I've kept it so the
10280 default is static libraries only, and the OpenSSL programs
10281 are always statically linked for now, but there are
10282 preparations for dynamic linking in place.
10283 This has been tested on Linux and Tru64.
10286 *) Randomness polling function for Win9x, as described in:
10287 Peter Gutmann, Software Generation of Practically Strong
10291 *) Fix so PRNG is seeded in req if using an already existing
10295 *) New options to smime application. -inform and -outform
10296 allow alternative formats for the S/MIME message including
10297 PEM and DER. The -content option allows the content to be
10298 specified separately. This should allow things like Netscape
10299 form signing output easier to verify.
10302 *) Fix the ASN1 encoding of tags using the 'long form'.
10305 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10306 STRING types. These convert content octets to and from the
10307 underlying type. The actual tag and length octets are
10308 already assumed to have been read in and checked. These
10309 are needed because all other string types have virtually
10310 identical handling apart from the tag. By having versions
10311 of the ASN1 functions that just operate on content octets
10312 IMPLICIT tagging can be handled properly. It also allows
10313 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10314 and ASN1_INTEGER are identical apart from the tag.
10317 *) Change the handling of OID objects as follows:
10319 - New object identifiers are inserted in objects.txt, following
10320 the syntax given in objects.README.
10321 - objects.pl is used to process obj_mac.num and create a new
10323 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10326 This is currently kind of a hack, and the perl code in objects.pl
10327 isn't very elegant, but it works as I intended. The simplest way
10328 to check that it worked correctly is to look in obj_dat.h and
10329 check the array nid_objs and make sure the objects haven't moved
10330 around (this is important!). Additions are OK, as well as
10331 consistent name changes.
10334 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10337 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10338 The given file adds to whatever has already been seeded into the
10339 random pool through the RANDFILE configuration file option or
10340 environment variable, or the default random state file.
10343 *) mkstack.pl now sorts each macro group into lexical order.
10344 Previously the output order depended on the order the files
10345 appeared in the directory, resulting in needless rewriting
10349 *) Patches to make OpenSSL compile under Win32 again. Mostly
10350 work arounds for the VC++ problem that it treats func() as
10351 func(void). Also stripped out the parts of mkdef.pl that
10352 added extra typesafe functions: these no longer exist.
10355 *) Reorganisation of the stack code. The macros are now all
10356 collected in safestack.h . Each macro is defined in terms of
10357 a "stack macro" of the form SKM_<name>(type, a, b). The
10358 DEBUG_SAFESTACK is now handled in terms of function casts,
10359 this has the advantage of retaining type safety without the
10360 use of additional functions. If DEBUG_SAFESTACK is not defined
10361 then the non typesafe macros are used instead. Also modified the
10362 mkstack.pl script to handle the new form. Needs testing to see
10363 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10364 the default if no major problems. Similar behaviour for ASN1_SET_OF
10365 and PKCS12_STACK_OF.
10368 *) When some versions of IIS use the 'NET' form of private key the
10369 key derivation algorithm is different. Normally MD5(password) is
10370 used as a 128 bit RC4 key. In the modified case
10371 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10372 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10373 as the old Netscape_RSA functions except they have an additional
10374 'sgckey' parameter which uses the modified algorithm. Also added
10375 an -sgckey command line option to the rsa utility. Thanks to
10376 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10377 algorithm to openssl-dev.
10380 *) The evp_local.h macros were using 'c.##kname' which resulted in
10381 invalid expansion on some systems (SCO 5.0.5 for example).
10382 Corrected to 'c.kname'.
10383 [Phillip Porch <root@theporch.com>]
10385 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10386 a STACK of email addresses from a certificate or request, these look
10387 in the subject name and the subject alternative name extensions and
10388 omit any duplicate addresses.
10391 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10392 This makes DSA verification about 2 % faster.
10395 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10396 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10397 plus overhead for 1024 bit moduli).
10398 This makes exponentiations about 0.5 % faster for 1024 bit
10399 exponents (as measured by "openssl speed rsa2048").
10402 *) Rename memory handling macros to avoid conflicts with other
10404 Malloc => OPENSSL_malloc
10405 Malloc_locked => OPENSSL_malloc_locked
10406 Realloc => OPENSSL_realloc
10407 Free => OPENSSL_free
10410 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10411 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10414 *) CygWin32 support.
10415 [John Jarvie <jjarvie@newsguy.com>]
10417 *) The type-safe stack code has been rejigged. It is now only compiled
10418 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10419 by default all type-specific stack functions are "#define"d back to
10420 standard stack functions. This results in more streamlined output
10421 but retains the type-safety checking possibilities of the original
10425 *) The STACK code has been cleaned up, and certain type declarations
10426 that didn't make a lot of sense have been brought in line. This has
10427 also involved a cleanup of sorts in safestack.h to more correctly
10428 map type-safe stack functions onto their plain stack counterparts.
10429 This work has also resulted in a variety of "const"ifications of
10430 lots of the code, especially "_cmp" operations which should normally
10431 be prototyped with "const" parameters anyway.
10434 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10435 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10436 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10437 where all of 'md' is used each time the PRNG is used, but 'state'
10438 is used only indexed by a cyclic counter. As entropy may not be
10439 well distributed from the beginning, 'md' is important as a
10440 chaining variable. However, the output function chains only half
10441 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10442 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10443 in all of 'state' being rewritten, with the new values depending
10444 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10447 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10448 the handshake is continued after ssl_verify_cert_chain();
10449 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10450 can lead to 'unexplainable' connection aborts later.
10451 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10453 *) Major EVP API cipher revision.
10454 Add hooks for extra EVP features. This allows various cipher
10455 parameters to be set in the EVP interface. Support added for variable
10456 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10457 setting of RC2 and RC5 parameters.
10459 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10462 Remove lots of duplicated code from the EVP library. For example *every*
10463 cipher init() function handles the 'iv' in the same way according to the
10464 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10465 for CFB and OFB modes they zero ctx->num.
10467 New functionality allows removal of S/MIME code RC2 hack.
10469 Most of the routines have the same form and so can be declared in terms
10472 By shifting this to the top level EVP_CipherInit() it can be removed from
10473 all individual ciphers. If the cipher wants to handle IVs or keys
10474 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10477 Change lots of functions like EVP_EncryptUpdate() to now return a
10478 value: although software versions of the algorithms cannot fail
10479 any installed hardware versions can.
10482 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10483 this option is set, tolerate broken clients that send the negotiated
10484 protocol version number instead of the requested protocol version
10488 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10489 i.e. non-zero for export ciphersuites, zero otherwise.
10490 Previous versions had this flag inverted, inconsistent with
10491 rsa_tmp_cb (..._TMP_RSA_CB).
10492 [Bodo Moeller; problem reported by Amit Chopra]
10494 *) Add missing DSA library text string. Work around for some IIS
10495 key files with invalid SEQUENCE encoding.
10498 *) Add a document (doc/standards.txt) that list all kinds of standards
10499 and so on that are implemented in OpenSSL.
10502 *) Enhance c_rehash script. Old version would mishandle certificates
10503 with the same subject name hash and wouldn't handle CRLs at all.
10504 Added -fingerprint option to crl utility, to support new c_rehash
10508 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10511 *) Fix for SSL server purpose checking. Server checking was
10512 rejecting certificates which had extended key usage present
10513 but no ssl client purpose.
10514 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10516 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10517 is a little unclear about how a blank password is handled.
10518 Since the password in encoded as a BMPString with terminating
10519 double NULL a zero length password would end up as just the
10520 double NULL. However no password at all is different and is
10521 handled differently in the PKCS#12 key generation code. NS
10522 treats a blank password as zero length. MSIE treats it as no
10523 password on export: but it will try both on import. We now do
10524 the same: PKCS12_parse() tries zero length and no password if
10525 the password is set to "" or NULL (NULL is now a valid password:
10526 it wasn't before) as does the pkcs12 application.
10529 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10530 perror when PEM_read_bio_X509_REQ fails, the error message must
10531 be obtained from the error queue.
10534 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10535 it in ERR_remove_state if appropriate, and change ERR_get_state
10536 accordingly to avoid race conditions (this is necessary because
10537 thread_hash is no longer constant once set).
10540 *) Bugfix for linux-elf makefile.one.
10543 *) RSA_get_default_method() will now cause a default
10544 RSA_METHOD to be chosen if one doesn't exist already.
10545 Previously this was only set during a call to RSA_new()
10546 or RSA_new_method(NULL) meaning it was possible for
10547 RSA_get_default_method() to return NULL.
10550 *) Added native name translation to the existing DSO code
10551 that will convert (if the flag to do so is set) filenames
10552 that are sufficiently small and have no path information
10553 into a canonical native form. Eg. "blah" converted to
10554 "libblah.so" or "blah.dll" etc.
10557 *) New function ERR_error_string_n(e, buf, len) which is like
10558 ERR_error_string(e, buf), but writes at most 'len' bytes
10559 including the 0 terminator. For ERR_error_string_n, 'buf'
10561 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10563 *) CONF library reworked to become more general. A new CONF
10564 configuration file reader "class" is implemented as well as a
10565 new functions (NCONF_*, for "New CONF") to handle it. The now
10566 old CONF_* functions are still there, but are reimplemented to
10567 work in terms of the new functions. Also, a set of functions
10568 to handle the internal storage of the configuration data is
10569 provided to make it easier to write new configuration file
10570 reader "classes" (I can definitely see something reading a
10571 configuration file in XML format, for example), called _CONF_*,
10572 or "the configuration storage API"...
10574 The new configuration file reading functions are:
10576 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10577 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10579 NCONF_default, NCONF_WIN32
10581 NCONF_dump_fp, NCONF_dump_bio
10583 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10584 NCONF_new creates a new CONF object. This works in the same way
10585 as other interfaces in OpenSSL, like the BIO interface.
10586 NCONF_dump_* dump the internal storage of the configuration file,
10587 which is useful for debugging. All other functions take the same
10588 arguments as the old CONF_* functions wth the exception of the
10589 first that must be a `CONF *' instead of a `LHASH *'.
10591 To make it easer to use the new classes with the old CONF_* functions,
10592 the function CONF_set_default_method is provided.
10595 *) Add '-tls1' option to 'openssl ciphers', which was already
10596 mentioned in the documentation but had not been implemented.
10597 (This option is not yet really useful because even the additional
10598 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10601 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10602 OpenSSL-based applications) load shared libraries and bind to
10603 them in a portable way.
10604 [Geoff Thorpe, with contributions from Richard Levitte]
10606 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10608 *) Make sure _lrotl and _lrotr are only used with MSVC.
10610 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10611 (the default implementation of RAND_status).
10613 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10614 to '-clrext' (= clear extensions), as intended and documented.
10615 [Bodo Moeller; inconsistency pointed out by Michael Attili
10616 <attili@amaxo.com>]
10618 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10619 was larger than the MD block size.
10620 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10622 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10623 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10624 using the passed key: if the passed key was a private key the result
10625 of X509_print(), for example, would be to print out all the private key
10629 *) des_quad_cksum() byte order bug fix.
10630 [Ulf Möller, using the problem description in krb4-0.9.7, where
10631 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10633 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10635 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10637 *) For easily testing in shell scripts whether some command
10638 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10639 returns with exit code 0 iff no command of the given name is available.
10640 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10641 the output goes to stdout and nothing is printed to stderr.
10642 Additional arguments are always ignored.
10644 Since for each cipher there is a command of the same name,
10645 the 'no-cipher' compilation switches can be tested this way.
10647 ('openssl no-XXX' is not able to detect pseudo-commands such
10648 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10651 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10654 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10655 is set; it will be thrown away anyway because each handshake creates
10657 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10658 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10659 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10660 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10663 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10664 'Q' and 'R' lose their special meanings (quit/renegotiate).
10665 This is part of what -quiet does; unlike -quiet, -ign_eof
10666 does not suppress any output.
10669 *) Add compatibility options to the purpose and trust code. The
10670 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10671 accepts a certificate or CA, this was the previous behaviour,
10672 with all the associated security issues.
10674 X509_TRUST_COMPAT is the old trust behaviour: only and
10675 automatically trust self signed roots in certificate store. A
10676 new trust setting X509_TRUST_DEFAULT is used to specify that
10677 a purpose has no associated trust setting and it should instead
10678 use the value in the default purpose.
10681 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10682 and fix a memory leak.
10685 *) In util/mkerr.pl (which implements 'make errors'), preserve
10686 reason strings from the previous version of the .c file, as
10687 the default to have only downcase letters (and digits) in
10688 automatically generated reasons codes is not always appropriate.
10691 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10692 using strerror. Previously, ERR_reason_error_string() returned
10693 library names as reason strings for SYSerr; but SYSerr is a special
10694 case where small numbers are errno values, not library numbers.
10697 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10698 converts DSA parameters into DH parameters. (When creating parameters,
10699 DSA_generate_parameters is used.)
10702 *) Include 'length' (recommended exponent length) in C code generated
10703 by 'openssl dhparam -C'.
10706 *) The second argument to set_label in perlasm was already being used
10707 so couldn't be used as a "file scope" flag. Moved to third argument
10711 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10712 instead of RAND_bytes for encryption IVs and salts.
10715 *) Include RAND_status() into RAND_METHOD instead of implementing
10716 it only for md_rand.c Otherwise replacing the PRNG by calling
10717 RAND_set_rand_method would be impossible.
10720 *) Don't let DSA_generate_key() enter an infinite loop if the random
10721 number generation fails.
10724 *) New 'rand' application for creating pseudo-random output.
10727 *) Added configuration support for Linux/IA64
10728 [Rolf Haberrecker <rolf@suse.de>]
10730 *) Assembler module support for Mingw32.
10733 *) Shared library support for HPUX (in shlib/).
10734 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10736 *) Shared library support for Solaris gcc.
10737 [Lutz Behnke <behnke@trustcenter.de>]
10739 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10741 *) PKCS7_encrypt() was adding text MIME headers twice because they
10742 were added manually and by SMIME_crlf_copy().
10745 *) In bntest.c don't call BN_rand with zero bits argument.
10746 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10748 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10749 case was implemented. This caused BN_div_recp() to fail occasionally.
10752 *) Add an optional second argument to the set_label() in the perl
10753 assembly language builder. If this argument exists and is set
10754 to 1 it signals that the assembler should use a symbol whose
10755 scope is the entire file, not just the current function. This
10756 is needed with MASM which uses the format label:: for this scope.
10757 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10759 *) Change the ASN1 types so they are typedefs by default. Before
10760 almost all types were #define'd to ASN1_STRING which was causing
10761 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10765 *) Change names of new functions to the new get1/get0 naming
10766 convention: After 'get1', the caller owns a reference count
10767 and has to call ..._free; 'get0' returns a pointer to some
10768 data structure without incrementing reference counters.
10769 (Some of the existing 'get' functions increment a reference
10770 counter, some don't.)
10771 Similarly, 'set1' and 'add1' functions increase reference
10772 counters or duplicate objects.
10775 *) Allow for the possibility of temp RSA key generation failure:
10776 the code used to assume it always worked and crashed on failure.
10779 *) Fix potential buffer overrun problem in BIO_printf().
10780 [Ulf Möller, using public domain code by Patrick Powell; problem
10781 pointed out by David Sacerdote <das33@cornell.edu>]
10783 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10784 RAND_egd() and RAND_status(). In the command line application,
10785 the EGD socket can be specified like a seed file using RANDFILE
10789 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10790 Some CAs (e.g. Verisign) distribute certificates in this form.
10793 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10794 list to exclude them. This means that no special compilation option
10795 is needed to use anonymous DH: it just needs to be included in the
10799 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10800 EVP_MD_type. The old functionality is available in a new macro called
10801 EVP_MD_md(). Change code that uses it and update docs.
10804 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10805 where the 'void *' argument is replaced by a function pointer argument.
10806 Previously 'void *' was abused to point to functions, which works on
10807 many platforms, but is not correct. As these functions are usually
10808 called by macros defined in OpenSSL header files, most source code
10809 should work without changes.
10812 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10813 sections with information on -D... compiler switches used for
10814 compiling the library so that applications can see them. To enable
10815 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10816 must be defined. E.g.,
10817 #define OPENSSL_ALGORITHM_DEFINES
10818 #include <openssl/opensslconf.h>
10819 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10820 [Richard Levitte, Ulf and Bodo Möller]
10822 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10826 *) Change the 'other' type in certificate aux info to a STACK_OF
10827 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10828 the required ASN1 format: arbitrary types determined by an OID.
10831 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10832 argument to 'req'. This is not because the function is newer or
10833 better than others it just uses the work 'NEW' in the certificate
10834 request header lines. Some software needs this.
10837 *) Reorganise password command line arguments: now passwords can be
10838 obtained from various sources. Delete the PEM_cb function and make
10839 it the default behaviour: i.e. if the callback is NULL and the
10840 usrdata argument is not NULL interpret it as a null terminated pass
10841 phrase. If usrdata and the callback are NULL then the pass phrase
10842 is prompted for as usual.
10845 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10846 the support is automatically enabled. The resulting binaries will
10847 autodetect the card and use it if present.
10848 [Ben Laurie and Compaq Inc.]
10850 *) Work around for Netscape hang bug. This sends certificate request
10851 and server done in one record. Since this is perfectly legal in the
10852 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10853 the bugs/SSLv3 entry for more info.
10856 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10859 *) Add -rand argument to smime and pkcs12 applications and read/write
10863 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10866 *) Add command line password options to the remaining applications.
10869 *) Bug fix for BN_div_recp() for numerators with an even number of
10873 *) More tests in bntest.c, and changed test_bn output.
10876 *) ./config recognizes MacOS X now.
10879 *) Bug fix for BN_div() when the first words of num and divisor are
10880 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10883 *) Add support for various broken PKCS#8 formats, and command line
10884 options to produce them.
10887 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10888 get temporary BIGNUMs from a BN_CTX.
10891 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10895 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10896 include a #define from the old name to the new. The original intent
10897 was that statically linked binaries could for example just call
10898 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10899 link with digests. This never worked because SSLeay_add_all_digests()
10900 and SSLeay_add_all_ciphers() were in the same source file so calling
10901 one would link with the other. They are now in separate source files.
10904 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10907 *) Use a less unusual form of the Miller-Rabin primality test (it used
10908 a binary algorithm for exponentiation integrated into the Miller-Rabin
10909 loop, our standard modexp algorithms are faster).
10912 *) Support for the EBCDIC character set completed.
10913 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10915 *) Source code cleanups: use const where appropriate, eliminate casts,
10916 use void * instead of char * in lhash.
10919 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10920 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10921 this the server could overwrite ephemeral keys that the client
10925 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10926 using 50 iterations of the Rabin-Miller test.
10928 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10929 iterations of the Rabin-Miller test as required by the appendix
10930 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10931 As BN_is_prime_fasttest includes trial division, DSA parameter
10932 generation becomes much faster.
10934 This implies a change for the callback functions in DSA_is_prime
10935 and DSA_generate_parameters: The callback function is called once
10936 for each positive witness in the Rabin-Miller test, not just
10937 occasionally in the inner loop; and the parameters to the
10938 callback function now provide an iteration count for the outer
10939 loop rather than for the current invocation of the inner loop.
10940 DSA_generate_parameters additionally can call the callback
10941 function with an 'iteration count' of -1, meaning that a
10942 candidate has passed the trial division test (when q is generated
10943 from an application-provided seed, trial division is skipped).
10946 *) New function BN_is_prime_fasttest that optionally does trial
10947 division before starting the Rabin-Miller test and has
10948 an additional BN_CTX * argument (whereas BN_is_prime always
10949 has to allocate at least one BN_CTX).
10950 'callback(1, -1, cb_arg)' is called when a number has passed the
10951 trial division stage.
10954 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10958 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10961 *) New function BN_pseudo_rand().
10964 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10965 bignum version of BN_from_montgomery() with the working code from
10966 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10970 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10971 made it impossible to use the same SSL_SESSION data structure in
10972 SSL2 clients in multiple threads.
10975 *) The return value of RAND_load_file() no longer counts bytes obtained
10976 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10977 to seed the PRNG (previously an explicit byte count was required).
10978 [Ulf Möller, Bodo Möller]
10980 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10981 used (char *) instead of (void *) and had casts all over the place.
10984 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10987 *) Retain source code compatibility for BN_prime_checks macro:
10988 BN_is_prime(..., BN_prime_checks, ...) now uses
10989 BN_prime_checks_for_size to determine the appropriate number of
10990 Rabin-Miller iterations.
10993 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10994 DH_CHECK_P_NOT_SAFE_PRIME.
10995 (Check if this is true? OpenPGP calls them "strong".)
10998 *) Merge the functionality of "dh" and "gendh" programs into a new program
10999 "dhparam". The old programs are retained for now but will handle DH keys
11000 (instead of parameters) in future.
11003 *) Make the ciphers, s_server and s_client programs check the return values
11004 when a new cipher list is set.
11007 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11008 ciphers. Before when the 56bit ciphers were enabled the sorting was
11011 The syntax for the cipher sorting has been extended to support sorting by
11012 cipher-strength (using the strength_bits hard coded in the tables).
11013 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11015 Fix a bug in the cipher-command parser: when supplying a cipher command
11016 string with an "undefined" symbol (neither command nor alphanumeric
11017 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11018 an error is flagged.
11020 Due to the strength-sorting extension, the code of the
11021 ssl_create_cipher_list() function was completely rearranged. I hope that
11022 the readability was also increased :-)
11023 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11025 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11026 for the first serial number and places 2 in the serial number file. This
11027 avoids problems when the root CA is created with serial number zero and
11028 the first user certificate has the same issuer name and serial number
11032 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11033 the new code. Add documentation for this stuff.
11036 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11037 X509_*() to X509at_*() on the grounds that they don't handle X509
11038 structures and behave in an analogous way to the X509v3 functions:
11039 they shouldn't be called directly but wrapper functions should be used
11042 So we also now have some wrapper functions that call the X509at functions
11043 when passed certificate requests. (TO DO: similar things can be done with
11044 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11045 things. Some of these need some d2i or i2d and print functionality
11046 because they handle more complex structures.)
11049 *) Add missing #ifndefs that caused missing symbols when building libssl
11050 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11051 NO_RSA in ssl/s2*.c.
11052 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11054 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11055 has a return value which indicates the quality of the random data
11056 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11057 error queue. New function RAND_pseudo_bytes() generates output that is
11058 guaranteed to be unique but not unpredictable. RAND_add is like
11059 RAND_seed, but takes an extra argument for an entropy estimate
11060 (RAND_seed always assumes full entropy).
11063 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11064 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11065 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11066 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11067 false-positive rate of at most 2^-80 for random input.
11070 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11073 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11074 in the 0.9.5 release), this returns the chain
11075 from an X509_CTX structure with a dup of the stack and all
11076 the X509 reference counts upped: so the stack will exist
11077 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11080 Also make SSL_SESSION_print() print out the verify return
11084 *) Add manpage for the pkcs12 command. Also change the default
11085 behaviour so MAC iteration counts are used unless the new
11086 -nomaciter option is used. This improves file security and
11087 only older versions of MSIE (4.0 for example) need it.
11090 *) Honor the no-xxx Configure options when creating .DEF files.
11093 *) Add PKCS#10 attributes to field table: challengePassword,
11094 unstructuredName and unstructuredAddress. These are taken from
11095 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11096 international characters are used.
11098 More changes to X509_ATTRIBUTE code: allow the setting of types
11099 based on strings. Remove the 'loc' parameter when adding
11100 attributes because these will be a SET OF encoding which is sorted
11104 *) Initial changes to the 'req' utility to allow request generation
11105 automation. This will allow an application to just generate a template
11106 file containing all the field values and have req construct the
11109 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11110 used all over the place including certificate requests and PKCS#7
11111 structures. They are currently handled manually where necessary with
11112 some primitive wrappers for PKCS#7. The new functions behave in a
11113 manner analogous to the X509 extension functions: they allow
11114 attributes to be looked up by NID and added.
11116 Later something similar to the X509V3 code would be desirable to
11117 automatically handle the encoding, decoding and printing of the
11118 more complex types. The string types like challengePassword can
11119 be handled by the string table functions.
11121 Also modified the multi byte string table handling. Now there is
11122 a 'global mask' which masks out certain types. The table itself
11123 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11124 is useful when for example there is only one permissible type
11125 (as in countryName) and using the mask might result in no valid
11129 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11130 SSL_get_peer_finished to allow applications to obtain the latest
11131 Finished messages sent to the peer or expected from the peer,
11132 respectively. (SSL_get_peer_finished is usually the Finished message
11133 actually received from the peer, otherwise the protocol will be aborted.)
11135 As the Finished message are message digests of the complete handshake
11136 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11137 be used for external authentication procedures when the authentication
11138 provided by SSL/TLS is not desired or is not enough.
11141 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11142 the host supports BWX extension and if Compaq C is present on the
11143 $PATH. Just exploiting of the BWX extension results in 20-30%
11144 performance kick for some algorithms, e.g. DES and RC4 to mention
11145 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11149 *) Add support for MS "fast SGC". This is arguably a violation of the
11150 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11151 weak crypto and after checking the certificate is SGC a second one
11152 with strong crypto. MS SGC stops the first handshake after receiving
11153 the server certificate message and sends a second client hello. Since
11154 a server will typically do all the time consuming operations before
11155 expecting any further messages from the client (server key exchange
11156 is the most expensive) there is little difference between the two.
11158 To get OpenSSL to support MS SGC we have to permit a second client
11159 hello message after we have sent server done. In addition we have to
11160 reset the MAC if we do get this second client hello.
11163 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11164 if a DER encoded private key is RSA or DSA traditional format. Changed
11165 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11166 format DER encoded private key. Newer code should use PKCS#8 format which
11167 has the key type encoded in the ASN1 structure. Added DER private key
11168 support to pkcs8 application.
11171 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11172 ciphersuites has been selected (as required by the SSL 3/TLS 1
11173 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11174 is set, we interpret this as a request to violate the specification
11175 (the worst that can happen is a handshake failure, and 'correct'
11176 behaviour would result in a handshake failure anyway).
11179 *) In SSL_CTX_add_session, take into account that there might be multiple
11180 SSL_SESSION structures with the same session ID (e.g. when two threads
11181 concurrently obtain them from an external cache).
11182 The internal cache can handle only one SSL_SESSION with a given ID,
11183 so if there's a conflict, we now throw out the old one to achieve
11187 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11188 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11189 some routines that use cipher OIDs: some ciphers do not have OIDs
11190 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11194 *) Simplify the trust setting structure and code. Now we just have
11195 two sequences of OIDs for trusted and rejected settings. These will
11196 typically have values the same as the extended key usage extension
11197 and any application specific purposes.
11199 The trust checking code now has a default behaviour: it will just
11200 check for an object with the same NID as the passed id. Functions can
11201 be provided to override either the default behaviour or the behaviour
11202 for a given id. SSL client, server and email already have functions
11203 in place for compatibility: they check the NID and also return "trusted"
11204 if the certificate is self signed.
11207 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11208 traditional format into an EVP_PKEY structure.
11211 *) Add a password callback function PEM_cb() which either prompts for
11212 a password if usr_data is NULL or otherwise assumes it is a null
11213 terminated password. Allow passwords to be passed on command line
11214 environment or config files in a few more utilities.
11217 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11218 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11219 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11220 Update documentation.
11223 *) Support for ASN1 "NULL" type. This could be handled before by using
11224 ASN1_TYPE but there wasn't any function that would try to read a NULL
11225 and produce an error if it couldn't. For compatibility we also have
11226 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11227 don't allocate anything because they don't need to.
11230 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11232 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11234 *) Rebuild of the memory allocation routines used by OpenSSL code and
11235 possibly others as well. The purpose is to make an interface that
11236 provide hooks so anyone can build a separate set of allocation and
11237 deallocation routines to be used by OpenSSL, for example memory
11238 pool implementations, or something else, which was previously hard
11239 since Malloc(), Realloc() and Free() were defined as macros having
11240 the values malloc, realloc and free, respectively (except for Win32
11241 compilations). The same is provided for memory debugging code.
11242 OpenSSL already comes with functionality to find memory leaks, but
11243 this gives people a chance to debug other memory problems.
11245 With these changes, a new set of functions and macros have appeared:
11247 CRYPTO_set_mem_debug_functions() [F]
11248 CRYPTO_get_mem_debug_functions() [F]
11249 CRYPTO_dbg_set_options() [F]
11250 CRYPTO_dbg_get_options() [F]
11251 CRYPTO_malloc_debug_init() [M]
11253 The memory debug functions are NULL by default, unless the library
11254 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11255 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11256 gives the standard debugging functions that come with OpenSSL) or
11257 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11258 provided by the library user) must be used. When the standard
11259 debugging functions are used, CRYPTO_dbg_set_options can be used to
11260 request additional information:
11261 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11262 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11264 Also, things like CRYPTO_set_mem_functions will always give the
11265 expected result (the new set of functions is used for allocation
11266 and deallocation) at all times, regardless of platform and compiler
11269 To finish it up, some functions that were never use in any other
11270 way than through macros have a new API and new semantic:
11272 CRYPTO_dbg_malloc()
11273 CRYPTO_dbg_realloc()
11276 All macros of value have retained their old syntax.
11277 [Richard Levitte and Bodo Moeller]
11279 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11280 ordering of SMIMECapabilities wasn't in "strength order" and there
11281 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11285 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11286 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11287 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11289 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11290 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11291 functionality to handle multipart/signed properly) and a utility
11292 called 'smime' to call all this stuff. This is based on code I
11293 originally wrote for Celo who have kindly allowed it to be
11294 included in OpenSSL.
11297 *) Add variants des_set_key_checked and des_set_key_unchecked of
11298 des_set_key (aka des_key_sched). Global variable des_check_key
11299 decides which of these is called by des_set_key; this way
11300 des_check_key behaves as it always did, but applications and
11301 the library itself, which was buggy for des_check_key == 1,
11302 have a cleaner way to pick the version they need.
11305 *) New function PKCS12_newpass() which changes the password of a
11309 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11310 dynamic mix. In both cases the ids can be used as an index into the
11311 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11312 functions so they accept a list of the field values and the
11313 application doesn't need to directly manipulate the X509_TRUST
11317 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11321 *) Modify the way the V3 extension code looks up extensions. This now
11322 works in a similar way to the object code: we have some "standard"
11323 extensions in a static table which is searched with OBJ_bsearch()
11324 and the application can add dynamic ones if needed. The file
11325 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11326 updated whenever a new extension is added to the core code and kept
11327 in ext_nid order. There is a simple program 'tabtest.c' which checks
11328 this. New extensions are not added too often so this file can readily
11329 be maintained manually.
11331 There are two big advantages in doing things this way. The extensions
11332 can be looked up immediately and no longer need to be "added" using
11333 X509V3_add_standard_extensions(): this function now does nothing.
11334 [Side note: I get *lots* of email saying the extension code doesn't
11335 work because people forget to call this function]
11336 Also no dynamic allocation is done unless new extensions are added:
11337 so if we don't add custom extensions there is no need to call
11338 X509V3_EXT_cleanup().
11341 *) Modify enc utility's salting as follows: make salting the default. Add a
11342 magic header, so unsalted files fail gracefully instead of just decrypting
11343 to garbage. This is because not salting is a big security hole, so people
11344 should be discouraged from doing it.
11347 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11348 digest to be passed on the command line but it only used this
11349 parameter when signing a certificate. Modified so all relevant
11350 operations are affected by the digest parameter including the
11351 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11352 DSA key was used because it didn't fix the digest.
11355 *) Initial certificate chain verify code. Currently tests the untrusted
11356 certificates for consistency with the verify purpose (which is set
11357 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11359 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11360 this is because it will reject chains with invalid extensions whereas
11361 every previous version of OpenSSL and SSLeay made no checks at all.
11363 Trust code: checks the root CA for the relevant trust settings. Trust
11364 settings have an initial value consistent with the verify purpose: e.g.
11365 if the verify purpose is for SSL client use it expects the CA to be
11366 trusted for SSL client use. However the default value can be changed to
11367 permit custom trust settings: one example of this would be to only trust
11368 certificates from a specific "secure" set of CAs.
11370 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11371 which should be used for version portability: especially since the
11372 verify structure is likely to change more often now.
11374 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11375 to set them. If not set then assume SSL clients will verify SSL servers
11378 Two new options to the verify program: -untrusted allows a set of
11379 untrusted certificates to be passed in and -purpose which sets the
11380 intended purpose of the certificate. If a purpose is set then the
11381 new chain verify code is used to check extension consistency.
11384 *) Support for the authority information access extension.
11387 *) Modify RSA and DSA PEM read routines to transparently handle
11388 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11389 public keys in a format compatible with certificate
11390 SubjectPublicKeyInfo structures. Unfortunately there were already
11391 functions called *_PublicKey_* which used various odd formats so
11392 these are retained for compatibility: however the DSA variants were
11393 never in a public release so they have been deleted. Changed dsa/rsa
11394 utilities to handle the new format: note no releases ever handled public
11395 keys so we should be OK.
11397 The primary motivation for this change is to avoid the same fiasco
11398 that dogs private keys: there are several incompatible private key
11399 formats some of which are standard and some OpenSSL specific and
11400 require various evil hacks to allow partial transparent handling and
11401 even then it doesn't work with DER formats. Given the option anything
11402 other than PKCS#8 should be dumped: but the other formats have to
11403 stay in the name of compatibility.
11405 With public keys and the benefit of hindsight one standard format
11406 is used which works with EVP_PKEY, RSA or DSA structures: though
11407 it clearly returns an error if you try to read the wrong kind of key.
11409 Added a -pubkey option to the 'x509' utility to output the public key.
11410 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11411 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11412 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11413 that do the same as the EVP_PKEY_assign_*() except they up the
11414 reference count of the added key (they don't "swallow" the
11418 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11419 CRLs would fail if the file contained no certificates or no CRLs:
11420 added a new function to read in both types and return the number
11421 read: this means that if none are read it will be an error. The
11422 DER versions of the certificate and CRL reader would always fail
11423 because it isn't possible to mix certificates and CRLs in DER format
11424 without choking one or the other routine. Changed this to just read
11425 a certificate: this is the best we can do. Also modified the code
11426 in apps/verify.c to take notice of return codes: it was previously
11427 attempting to read in certificates from NULL pointers and ignoring
11428 any errors: this is one reason why the cert and CRL reader seemed
11429 to work. It doesn't check return codes from the default certificate
11430 routines: these may well fail if the certificates aren't installed.
11433 *) Code to support otherName option in GeneralName.
11436 *) First update to verify code. Change the verify utility
11437 so it warns if it is passed a self signed certificate:
11438 for consistency with the normal behaviour. X509_verify
11439 has been modified to it will now verify a self signed
11440 certificate if *exactly* the same certificate appears
11441 in the store: it was previously impossible to trust a
11442 single self signed certificate. This means that:
11443 openssl verify ss.pem
11444 now gives a warning about a self signed certificate but
11445 openssl verify -CAfile ss.pem ss.pem
11449 *) For servers, store verify_result in SSL_SESSION data structure
11450 (and add it to external session representation).
11451 This is needed when client certificate verifications fails,
11452 but an application-provided verification callback (set by
11453 SSL_CTX_set_cert_verify_callback) allows accepting the session
11454 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11455 but returns 1): When the session is reused, we have to set
11456 ssl->verify_result to the appropriate error code to avoid
11458 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11460 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11461 case in PKCS7_dataInit() where the signed PKCS7 structure
11462 didn't contain any existing data because it was being created.
11463 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11465 *) Add a salt to the key derivation routines in enc.c. This
11466 forms the first 8 bytes of the encrypted file. Also add a
11467 -S option to allow a salt to be input on the command line.
11470 *) New function X509_cmp(). Oddly enough there wasn't a function
11471 to compare two certificates. We do this by working out the SHA1
11472 hash and comparing that. X509_cmp() will be needed by the trust
11476 *) SSL_get1_session() is like SSL_get_session(), but increments
11477 the reference count in the SSL_SESSION returned.
11478 [Geoff Thorpe <geoff@eu.c2.net>]
11480 *) Fix for 'req': it was adding a null to request attributes.
11481 Also change the X509_LOOKUP and X509_INFO code to handle
11482 certificate auxiliary information.
11485 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11489 *) Add the possibility to add extra information to the memory leak
11490 detecting output, to form tracebacks, showing from where each
11491 allocation was originated: CRYPTO_push_info("constant string") adds
11492 the string plus current file name and line number to a per-thread
11493 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11494 is like calling CYRPTO_pop_info() until the stack is empty.
11495 Also updated memory leak detection code to be multi-thread-safe.
11498 *) Add options -text and -noout to pkcs7 utility and delete the
11499 encryption options which never did anything. Update docs.
11502 *) Add options to some of the utilities to allow the pass phrase
11503 to be included on either the command line (not recommended on
11504 OSes like Unix) or read from the environment. Update the
11505 manpages and fix a few bugs.
11508 *) Add a few manpages for some of the openssl commands.
11511 *) Fix the -revoke option in ca. It was freeing up memory twice,
11512 leaking and not finding already revoked certificates.
11515 *) Extensive changes to support certificate auxiliary information.
11516 This involves the use of X509_CERT_AUX structure and X509_AUX
11517 functions. An X509_AUX function such as PEM_read_X509_AUX()
11518 can still read in a certificate file in the usual way but it
11519 will also read in any additional "auxiliary information". By
11520 doing things this way a fair degree of compatibility can be
11521 retained: existing certificates can have this information added
11522 using the new 'x509' options.
11524 Current auxiliary information includes an "alias" and some trust
11525 settings. The trust settings will ultimately be used in enhanced
11526 certificate chain verification routines: currently a certificate
11527 can only be trusted if it is self signed and then it is trusted
11531 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11532 The problem was that one of the replacement routines had not been working
11533 since SSLeay releases. For now the offending routine has been replaced
11534 with non-optimised assembler. Even so, this now gives around 95%
11535 performance improvement for 1024 bit RSA signs.
11538 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11539 handling. Most clients have the effective key size in bits equal to
11540 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11541 A few however don't do this and instead use the size of the decrypted key
11542 to determine the RC2 key length and the AlgorithmIdentifier to determine
11543 the effective key length. In this case the effective key length can still
11544 be 40 bits but the key length can be 168 bits for example. This is fixed
11545 by manually forcing an RC2 key into the EVP_PKEY structure because the
11546 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11547 the key length and effective key length are equal.
11550 *) Add a bunch of functions that should simplify the creation of
11551 X509_NAME structures. Now you should be able to do:
11552 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11553 and have it automatically work out the correct field type and fill in
11554 the structures. The more adventurous can try:
11555 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11556 and it will (hopefully) work out the correct multibyte encoding.
11559 *) Change the 'req' utility to use the new field handling and multibyte
11560 copy routines. Before the DN field creation was handled in an ad hoc
11561 way in req, ca, and x509 which was rather broken and didn't support
11562 BMPStrings or UTF8Strings. Since some software doesn't implement
11563 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11564 using the dirstring_type option. See the new comment in the default
11565 openssl.cnf for more info.
11568 *) Make crypto/rand/md_rand.c more robust:
11569 - Assure unique random numbers after fork().
11570 - Make sure that concurrent threads access the global counter and
11571 md serializably so that we never lose entropy in them
11572 or use exactly the same state in multiple threads.
11573 Access to the large state is not always serializable because
11574 the additional locking could be a performance killer, and
11575 md should be large enough anyway.
11578 *) New file apps/app_rand.c with commonly needed functionality
11579 for handling the random seed file.
11581 Use the random seed file in some applications that previously did not:
11583 dsaparam -genkey (which also ignored its '-rand' option),
11586 x509 (when signing).
11587 Except on systems with /dev/urandom, it is crucial to have a random
11588 seed file at least for key creation, DSA signing, and for DH exchanges;
11589 for RSA signatures we could do without one.
11591 gendh and gendsa (unlike genrsa) used to read only the first byte
11592 of each file listed in the '-rand' option. The function as previously
11593 found in genrsa is now in app_rand.c and is used by all programs
11594 that support '-rand'.
11597 *) In RAND_write_file, use mode 0600 for creating files;
11598 don't just chmod when it may be too late.
11601 *) Report an error from X509_STORE_load_locations
11602 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11605 *) New function ASN1_mbstring_copy() this copies a string in either
11606 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11607 into an ASN1_STRING type. A mask of permissible types is passed
11608 and it chooses the "minimal" type to use or an error if not type
11612 *) Add function equivalents to the various macros in asn1.h. The old
11613 macros are retained with an M_ prefix. Code inside the library can
11614 use the M_ macros. External code (including the openssl utility)
11615 should *NOT* in order to be "shared library friendly".
11618 *) Add various functions that can check a certificate's extensions
11619 to see if it usable for various purposes such as SSL client,
11620 server or S/MIME and CAs of these types. This is currently
11621 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11622 verification. Also added a -purpose flag to x509 utility to
11623 print out all the purposes.
11626 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11630 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11631 for, obtain and decode and extension and obtain its critical flag.
11632 This allows all the necessary extension code to be handled in a
11633 single function call.
11636 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11637 platforms. See crypto/rc4/rc4_enc.c for further details.
11640 *) New -noout option to asn1parse. This causes no output to be produced
11641 its main use is when combined with -strparse and -out to extract data
11642 from a file (which may not be in ASN.1 format).
11645 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11646 when producing the local key id.
11647 [Richard Levitte <levitte@stacken.kth.se>]
11649 *) New option -dhparam in s_server. This allows a DH parameter file to be
11650 stated explicitly. If it is not stated then it tries the first server
11651 certificate file. The previous behaviour hard coded the filename
11655 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11656 a public key to be input or output. For example:
11657 openssl rsa -in key.pem -pubout -out pubkey.pem
11658 Also added necessary DSA public key functions to handle this.
11661 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11662 in the message. This was handled by allowing
11663 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11664 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11666 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11667 to the end of the strings whereas this didn't. This would cause problems
11668 if strings read with d2i_ASN1_bytes() were later modified.
11669 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11671 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11672 data and it contains EOF it will end up returning an error. This is
11673 caused by input 46 bytes long. The cause is due to the way base64
11674 BIOs find the start of base64 encoded data. They do this by trying a
11675 trial decode on each line until they find one that works. When they
11676 do a flag is set and it starts again knowing it can pass all the
11677 data directly through the decoder. Unfortunately it doesn't reset
11678 the context it uses. This means that if EOF is reached an attempt
11679 is made to pass two EOFs through the context and this causes the
11680 resulting error. This can also cause other problems as well. As is
11681 usual with these problems it takes *ages* to find and the fix is
11682 trivial: move one line.
11683 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11685 *) Ugly workaround to get s_client and s_server working under Windows. The
11686 old code wouldn't work because it needed to select() on sockets and the
11687 tty (for keypresses and to see if data could be written). Win32 only
11688 supports select() on sockets so we select() with a 1s timeout on the
11689 sockets and then see if any characters are waiting to be read, if none
11690 are present then we retry, we also assume we can always write data to
11691 the tty. This isn't nice because the code then blocks until we've
11692 received a complete line of data and it is effectively polling the
11693 keyboard at 1s intervals: however it's quite a bit better than not
11694 working at all :-) A dedicated Windows application might handle this
11695 with an event loop for example.
11698 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11699 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11700 will be called when RSA_sign() and RSA_verify() are used. This is useful
11701 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11702 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11703 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11704 This necessitated the support of an extra signature type NID_md5_sha1
11705 for SSL signatures and modifications to the SSL library to use it instead
11706 of calling RSA_public_decrypt() and RSA_private_encrypt().
11709 *) Add new -verify -CAfile and -CApath options to the crl program, these
11710 will lookup a CRL issuers certificate and verify the signature in a
11711 similar way to the verify program. Tidy up the crl program so it
11712 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11713 less strict. It will now permit CRL extensions even if it is not
11714 a V2 CRL: this will allow it to tolerate some broken CRLs.
11717 *) Initialize all non-automatic variables each time one of the openssl
11718 sub-programs is started (this is necessary as they may be started
11719 multiple times from the "OpenSSL>" prompt).
11720 [Lennart Bang, Bodo Moeller]
11722 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11723 removing all other RSA functionality (this is what NO_RSA does). This
11724 is so (for example) those in the US can disable those operations covered
11725 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11729 *) Non-copying interface to BIO pairs.
11730 (still largely untested)
11733 *) New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
11734 ASCII string. This was handled independently in various places before.
11737 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11738 UTF8 strings a character at a time.
11741 *) Use client_version from client hello to select the protocol
11742 (s23_srvr.c) and for RSA client key exchange verification
11743 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11746 *) Add various utility functions to handle SPKACs, these were previously
11747 handled by poking round in the structure internals. Added new function
11748 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11749 print, verify and generate SPKACs. Based on an original idea from
11750 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11753 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11756 *) Allow the config file extension section to be overwritten on the
11757 command line. Based on an original idea from Massimiliano Pala
11758 <madwolf@comune.modena.it>. The new option is called -extensions
11759 and can be applied to ca, req and x509. Also -reqexts to override
11760 the request extensions in req and -crlexts to override the crl extensions
11764 *) Add new feature to the SPKAC handling in ca. Now you can include
11765 the same field multiple times by preceding it by "XXXX." for example:
11768 this is the same syntax as used in the req config file.
11771 *) Allow certificate extensions to be added to certificate requests. These
11772 are specified in a 'req_extensions' option of the req section of the
11773 config file. They can be printed out with the -text option to req but
11774 are otherwise ignored at present.
11777 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11778 data read consists of only the final block it would not decrypted because
11779 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11780 A misplaced 'break' also meant the decrypted final block might not be
11781 copied until the next read.
11784 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11785 a few extra parameters to the DH structure: these will be useful if
11786 for example we want the value of 'q' or implement X9.42 DH.
11789 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11790 provides hooks that allow the default DSA functions or functions on a
11791 "per key" basis to be replaced. This allows hardware acceleration and
11792 hardware key storage to be handled without major modification to the
11793 library. Also added low level modexp hooks and CRYPTO_EX structure and
11794 associated functions.
11797 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11798 as "read only": it can't be written to and the buffer it points to will
11799 not be freed. Reading from a read only BIO is much more efficient than
11800 a normal memory BIO. This was added because there are several times when
11801 an area of memory needs to be read from a BIO. The previous method was
11802 to create a memory BIO and write the data to it, this results in two
11803 copies of the data and an O(n^2) reading algorithm. There is a new
11804 function BIO_new_mem_buf() which creates a read only memory BIO from
11805 an area of memory. Also modified the PKCS#7 routines to use read only
11809 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11810 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11811 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11812 but a retry condition occurred while trying to read the rest.
11815 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11816 NID_pkcs7_encrypted by default: this was wrong since this should almost
11817 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11818 the encrypted data type: this is a more sensible place to put it and it
11819 allows the PKCS#12 code to be tidied up that duplicated this
11823 *) Changed obj_dat.pl script so it takes its input and output files on
11824 the command line. This should avoid shell escape redirection problems
11828 *) Initial support for certificate extension requests, these are included
11829 in things like Xenroll certificate requests. Included functions to allow
11830 extensions to be obtained and added.
11833 *) -crlf option to s_client and s_server for sending newlines as
11834 CRLF (as required by many protocols).
11837 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11839 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11840 [Ralf S. Engelschall]
11842 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11843 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11845 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11849 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11850 DH parameters/keys (q is lost during that conversion, but the resulting
11851 DH parameters contain its length).
11853 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11854 much faster than DH_generate_parameters (which creates parameters
11855 where p = 2*q + 1), and also the smaller q makes DH computations
11856 much more efficient (160-bit exponentiation instead of 1024-bit
11857 exponentiation); so this provides a convenient way to support DHE
11858 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11859 utter importance to use
11860 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11862 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11863 when such DH parameters are used, because otherwise small subgroup
11864 attacks may become possible!
11867 *) Avoid memory leak in i2d_DHparams.
11870 *) Allow the -k option to be used more than once in the enc program:
11871 this allows the same encrypted message to be read by multiple recipients.
11874 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11875 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11876 it will always use the numerical form of the OID, even if it has a short
11880 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11881 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11882 otherwise bn_mod_exp was called. In the case of hardware keys for example
11883 no private key components need be present and it might store extra data
11884 in the RSA structure, which cannot be accessed from bn_mod_exp.
11885 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11886 private key operations.
11889 *) Added support for SPARC Linux.
11892 *) pem_password_cb function type incompatibly changed from
11893 typedef int pem_password_cb(char *buf, int size, int rwflag);
11895 ....(char *buf, int size, int rwflag, void *userdata);
11896 so that applications can pass data to their callbacks:
11897 The PEM[_ASN1]_{read,write}... functions and macros now take an
11898 additional void * argument, which is just handed through whenever
11899 the password callback is called.
11900 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11902 New function SSL_CTX_set_default_passwd_cb_userdata.
11904 Compatibility note: As many C implementations push function arguments
11905 onto the stack in reverse order, the new library version is likely to
11906 interoperate with programs that have been compiled with the old
11907 pem_password_cb definition (PEM_whatever takes some data that
11908 happens to be on the stack as its last argument, and the callback
11909 just ignores this garbage); but there is no guarantee whatsoever that
11912 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11913 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11914 problems not only on Windows, but also on some Unix platforms.
11915 To avoid problematic command lines, these definitions are now in an
11916 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11917 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11920 *) MIPS III/IV assembler module is reimplemented.
11923 *) More DES library cleanups: remove references to srand/rand and
11924 delete an unused file.
11927 *) Add support for the free Netwide assembler (NASM) under Win32,
11928 since not many people have MASM (ml) and it can be hard to obtain.
11929 This is currently experimental but it seems to work OK and pass all
11930 the tests. Check out INSTALL.W32 for info.
11933 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11934 without temporary keys kept an extra copy of the server key,
11935 and connections with temporary keys did not free everything in case
11939 *) New function RSA_check_key and new openssl rsa option -check
11940 for verifying the consistency of RSA keys.
11941 [Ulf Moeller, Bodo Moeller]
11943 *) Various changes to make Win32 compile work:
11944 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11945 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11946 comparison" warnings.
11947 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11950 *) Add a debugging option to PKCS#5 v2 key generation function: when
11951 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11952 derived keys are printed to stderr.
11955 *) Copy the flags in ASN1_STRING_dup().
11956 [Roman E. Pavlov <pre@mo.msk.ru>]
11958 *) The x509 application mishandled signing requests containing DSA
11959 keys when the signing key was also DSA and the parameters didn't match.
11961 It was supposed to omit the parameters when they matched the signing key:
11962 the verifying software was then supposed to automatically use the CA's
11963 parameters if they were absent from the end user certificate.
11965 Omitting parameters is no longer recommended. The test was also
11966 the wrong way round! This was probably due to unusual behaviour in
11967 EVP_cmp_parameters() which returns 1 if the parameters match.
11968 This meant that parameters were omitted when they *didn't* match and
11969 the certificate was useless. Certificates signed with 'ca' didn't have
11971 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11973 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11974 The interface is as follows:
11975 Applications can use
11976 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11977 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11978 "off" is now the default.
11979 The library internally uses
11980 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11981 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11982 to disable memory-checking temporarily.
11984 Some inconsistent states that previously were possible (and were
11985 even the default) are now avoided.
11987 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11988 with each memory chunk allocated; this is occasionally more helpful
11989 than just having a counter.
11991 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11993 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11997 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11998 which largely parallels "options", but is for changing API behaviour,
11999 whereas "options" are about protocol behaviour.
12000 Initial "mode" flags are:
12002 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12003 a single record has been written.
12004 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12005 retries use the same buffer location.
12006 (But all of the contents must be
12010 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12013 *) Fix problems with no-hmac etc.
12014 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12016 *) New functions RSA_get_default_method(), RSA_set_method() and
12017 RSA_get_method(). These allows replacement of RSA_METHODs without having
12018 to mess around with the internals of an RSA structure.
12021 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12022 Also really enable memory leak checks in openssl.c and in some
12024 [Chad C. Mulligan, Bodo Moeller]
12026 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12027 up the length of negative integers. This has now been simplified to just
12028 store the length when it is first determined and use it later, rather
12029 than trying to keep track of where data is copied and updating it to
12031 [Steve Henson, reported by Brien Wheeler
12032 <bwheeler@authentica-security.com>]
12034 *) Add a new function PKCS7_signatureVerify. This allows the verification
12035 of a PKCS#7 signature but with the signing certificate passed to the
12036 function itself. This contrasts with PKCS7_dataVerify which assumes the
12037 certificate is present in the PKCS#7 structure. This isn't always the
12038 case: certificates can be omitted from a PKCS#7 structure and be
12039 distributed by "out of band" means (such as a certificate database).
12042 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12043 function prototypes in pem.h, also change util/mkdef.pl to add the
12044 necessary function names.
12047 *) mk1mf.pl (used by Windows builds) did not properly read the
12048 options set by Configure in the top level Makefile, and Configure
12049 was not even able to write more than one option correctly.
12050 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12053 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12054 file to be loaded from a BIO or FILE pointer. The BIO version will
12055 for example allow memory BIOs to contain config info.
12058 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12059 Whoever hopes to achieve shared-library compatibility across versions
12060 must use this, not the compile-time macro.
12061 (Exercise 0.9.4: Which is the minimum library version required by
12063 Note: All this applies only to multi-threaded programs, others don't
12067 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12068 through a BIO pair triggered the default case, i.e.
12069 SSLerr(...,SSL_R_UNKNOWN_STATE).
12072 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12073 can use the SSL library even if none of the specific BIOs is
12077 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12078 for the encoded length.
12079 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12081 *) Add initial documentation of the X509V3 functions.
12084 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12085 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12086 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12087 secure PKCS#8 private key format with a high iteration count.
12090 *) Fix determination of Perl interpreter: A perl or perl5
12091 _directory_ in $PATH was also accepted as the interpreter.
12092 [Ralf S. Engelschall]
12094 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12095 wrong with it but it was very old and did things like calling
12096 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12097 unusual formatting.
12100 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12101 to use the new extension code.
12104 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12105 with macros. This should make it easier to change their form, add extra
12106 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12110 *) Add to configuration table a new entry that can specify an alternative
12111 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12112 according to Mark Crispin <MRC@Panda.COM>.
12116 *) DES CBC did not update the IV. Weird.
12119 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12120 Changing the behaviour of the former might break existing programs --
12121 where IV updating is needed, des_ncbc_encrypt can be used.
12124 *) When bntest is run from "make test" it drives bc to check its
12125 calculations, as well as internally checking them. If an internal check
12126 fails, it needs to cause bc to give a non-zero result or make test carries
12127 on without noticing the failure. Fixed.
12130 *) DES library cleanups.
12133 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12134 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12135 ciphers. NOTE: although the key derivation function has been verified
12136 against some published test vectors it has not been extensively tested
12137 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12141 *) Instead of "mkdir -p", which is not fully portable, use new
12142 Perl script "util/mkdir-p.pl".
12145 *) Rewrite the way password based encryption (PBE) is handled. It used to
12146 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12147 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12148 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12149 the 'parameter' field of the AlgorithmIdentifier is passed to the
12150 underlying key generation function so it must do its own ASN1 parsing.
12151 This has also changed the EVP_PBE_CipherInit() function which now has a
12152 'parameter' argument instead of literal salt and iteration count values
12153 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12156 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12157 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12158 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12159 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12160 value was just used as a "magic string" and not used directly its
12161 value doesn't matter.
12164 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12168 *) "linux-sparc64" configuration (ultrapenguin).
12169 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12170 "linux-sparc" configuration.
12171 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12173 *) config now generates no-xxx options for missing ciphers.
12176 *) Support the EBCDIC character set (work in progress).
12177 File ebcdic.c not yet included because it has a different license.
12178 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12180 *) Support BS2000/OSD-POSIX.
12181 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12183 *) Make callbacks for key generation use void * instead of char *.
12186 *) Make S/MIME samples compile (not yet tested).
12189 *) Additional typesafe stacks.
12192 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12196 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12198 *) New configuration variant "sco5-gcc".
12200 *) Updated some demos.
12201 [Sean O Riordain, Wade Scholine]
12203 *) Add missing BIO_free at exit of pkcs12 application.
12206 *) Fix memory leak in conf.c.
12209 *) Updates for Win32 to assembler version of MD5.
12212 *) Set #! path to perl in apps/der_chop to where we found it
12213 instead of using a fixed path.
12216 *) SHA library changes for irix64-mips4-cc.
12219 *) Improvements for VMS support.
12223 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12225 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12226 This also avoids the problems with SC4.2 and unpatched SC5.
12227 [Andy Polyakov <appro@fy.chalmers.se>]
12229 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12230 These are required because of the typesafe stack would otherwise break
12231 existing code. If old code used a structure member which used to be STACK
12232 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12233 sk_num or sk_value it would produce an error because the num, data members
12234 are not present in STACK_OF. Now it just produces a warning. sk_set
12235 replaces the old method of assigning a value to sk_value
12236 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12237 that does this will no longer work (and should use sk_set instead) but
12238 this could be regarded as a "questionable" behaviour anyway.
12241 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12242 correctly handle encrypted S/MIME data.
12245 *) Change type of various DES function arguments from des_cblock
12246 (which means, in function argument declarations, pointer to char)
12247 to des_cblock * (meaning pointer to array with 8 char elements),
12248 which allows the compiler to do more typechecking; it was like
12249 that back in SSLeay, but with lots of ugly casts.
12251 Introduce new type const_des_cblock.
12254 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12255 problems: find RecipientInfo structure that matches recipient certificate
12256 and initialise the ASN1 structures properly based on passed cipher.
12259 *) Belatedly make the BN tests actually check the results.
12262 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12263 to and from BNs: it was completely broken. New compilation option
12264 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12265 key elements as negative integers.
12268 *) Reorganize and speed up MD5.
12269 [Andy Polyakov <appro@fy.chalmers.se>]
12272 [Richard Levitte <richard@levitte.org>]
12274 *) New option -out to asn1parse to allow the parsed structure to be
12275 output to a file. This is most useful when combined with the -strparse
12276 option to examine the output of things like OCTET STRINGS.
12279 *) Make SSL library a little more fool-proof by not requiring any longer
12280 that SSL_set_{accept,connect}_state be called before
12281 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12282 in many applications because usually everything *appeared* to work as
12283 intended anyway -- now it really works as intended).
12286 *) Move openssl.cnf out of lib/.
12289 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12290 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12291 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12292 [Ralf S. Engelschall]
12294 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12295 handle PKCS#7 enveloped data properly.
12296 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12298 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12299 copying pointers. The cert_st handling is changed by this in
12300 various ways (and thus what used to be known as ctx->default_cert
12301 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12302 any longer when s->cert does not give us what we need).
12303 ssl_cert_instantiate becomes obsolete by this change.
12304 As soon as we've got the new code right (possibly it already is?),
12305 we have solved a couple of bugs of the earlier code where s->cert
12306 was used as if it could not have been shared with other SSL structures.
12308 Note that using the SSL API in certain dirty ways now will result
12309 in different behaviour than observed with earlier library versions:
12310 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12311 does not influence s as it used to.
12313 In order to clean up things more thoroughly, inside SSL_SESSION
12314 we don't use CERT any longer, but a new structure SESS_CERT
12315 that holds per-session data (if available); currently, this is
12316 the peer's certificate chain and, for clients, the server's certificate
12317 and temporary key. CERT holds only those values that can have
12318 meaningful defaults in an SSL_CTX.
12321 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12322 from the internal representation. Various PKCS#7 fixes: remove some
12323 evil casts and set the enc_dig_alg field properly based on the signing
12327 *) Allow PKCS#12 password to be set from the command line or the
12328 environment. Let 'ca' get its config file name from the environment
12329 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12333 *) Allow certificate policies extension to use an IA5STRING for the
12334 organization field. This is contrary to the PKIX definition but
12335 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12339 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12340 without disallowing inline assembler and the like for non-pedantic builds.
12343 *) Support Borland C++ builder.
12344 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12346 *) Support Mingw32.
12349 *) SHA-1 cleanups and performance enhancements.
12350 [Andy Polyakov <appro@fy.chalmers.se>]
12352 *) Sparc v8plus assembler for the bignum library.
12353 [Andy Polyakov <appro@fy.chalmers.se>]
12355 *) Accept any -xxx and +xxx compiler options in Configure.
12358 *) Update HPUX configuration.
12361 *) Add missing sk_<type>_unshift() function to safestack.h
12362 [Ralf S. Engelschall]
12364 *) New function SSL_CTX_use_certificate_chain_file that sets the
12365 "extra_cert"s in addition to the certificate. (This makes sense
12366 only for "PEM" format files, as chains as a whole are not
12370 *) Support verify_depth from the SSL API.
12371 x509_vfy.c had what can be considered an off-by-one-error:
12372 Its depth (which was not part of the external interface)
12373 was actually counting the number of certificates in a chain;
12374 now it really counts the depth.
12377 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12378 instead of X509err, which often resulted in confusing error
12379 messages since the error codes are not globally unique
12380 (e.g. an alleged error in ssl3_accept when a certificate
12381 didn't match the private key).
12383 *) New function SSL_CTX_set_session_id_context that allows to set a default
12384 value (so that you don't need SSL_set_session_id_context for each
12385 connection using the SSL_CTX).
12388 *) OAEP decoding bug fix.
12391 *) Support INSTALL_PREFIX for package builders, as proposed by
12395 *) New Configure options "threads" and "no-threads". For systems
12396 where the proper compiler options are known (currently Solaris
12397 and Linux), "threads" is the default.
12400 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12403 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12404 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12405 such as /usr/local/bin.
12408 *) "make linux-shared" to build shared libraries.
12409 [Niels Poppe <niels@netbox.org>]
12411 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12414 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12415 extension adding in x509 utility.
12418 *) Remove NOPROTO sections and error code comments.
12421 *) Partial rewrite of the DEF file generator to now parse the ANSI
12425 *) New Configure options --prefix=DIR and --openssldir=DIR.
12428 *) Complete rewrite of the error code script(s). It is all now handled
12429 by one script at the top level which handles error code gathering,
12430 header rewriting and C source file generation. It should be much better
12431 than the old method: it now uses a modified version of Ulf's parser to
12432 read the ANSI prototypes in all header files (thus the old K&R definitions
12433 aren't needed for error creation any more) and do a better job of
12434 translating function codes into names. The old 'ASN1 error code imbedded
12435 in a comment' is no longer necessary and it doesn't use .err files which
12436 have now been deleted. Also the error code call doesn't have to appear all
12437 on one line (which resulted in some large lines...).
12440 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12443 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12444 0 (which usually indicates a closed connection), but continue reading.
12447 *) Fix some race conditions.
12450 *) Add support for CRL distribution points extension. Add Certificate
12451 Policies and CRL distribution points documentation.
12454 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12457 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12458 8 of keying material. Merlin has also confirmed interop with this fix
12459 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12460 [Merlin Hughes <merlin@baltimore.ie>]
12462 *) Fix lots of warnings.
12463 [Richard Levitte <levitte@stacken.kth.se>]
12465 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12466 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12467 [Richard Levitte <levitte@stacken.kth.se>]
12469 *) Fix problems with sizeof(long) == 8.
12470 [Andy Polyakov <appro@fy.chalmers.se>]
12472 *) Change functions to ANSI C.
12475 *) Fix typos in error codes.
12476 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12478 *) Remove defunct assembler files from Configure.
12481 *) SPARC v8 assembler BIGNUM implementation.
12482 [Andy Polyakov <appro@fy.chalmers.se>]
12484 *) Support for Certificate Policies extension: both print and set.
12485 Various additions to support the r2i method this uses.
12488 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12489 return a const string when you are expecting an allocated buffer.
12492 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12493 types DirectoryString and DisplayText.
12496 *) Add code to allow r2i extensions to access the configuration database,
12497 add an LHASH database driver and add several ctx helper functions.
12500 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12501 fail when they extended the size of a BIGNUM.
12504 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12505 support typesafe stack.
12508 *) Fix typo in SSL_[gs]et_options().
12509 [Nils Frostberg <nils@medcom.se>]
12511 *) Delete various functions and files that belonged to the (now obsolete)
12512 old X509V3 handling code.
12515 *) New Configure option "rsaref".
12518 *) Don't auto-generate pem.h.
12521 *) Introduce type-safe ASN.1 SETs.
12524 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12525 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12527 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12528 that links with OpenSSL (well at least cause lots of warnings), but fear
12529 not: the conversion is trivial, and it eliminates loads of evil casts. A
12530 few STACKed things have been converted already. Feel free to convert more.
12531 In the fullness of time, I'll do away with the STACK type altogether.
12534 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12535 specified in <certfile> by updating the entry in the index.txt file.
12536 This way one no longer has to edit the index.txt file manually for
12537 revoking a certificate. The -revoke option does the gory details now.
12538 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12540 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12541 `-text' option at all and this way the `-noout -text' combination was
12542 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12543 [Ralf S. Engelschall]
12545 *) Make sure a corresponding plain text error message exists for the
12546 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12547 verify callback function determined that a certificate was revoked.
12548 [Ralf S. Engelschall]
12550 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12551 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12552 all available ciphers including rc5, which was forgotten until now.
12553 In order to let the testing shell script know which algorithms
12554 are available, a new (up to now undocumented) command
12555 "openssl list-cipher-commands" is used.
12558 *) Bugfix: s_client occasionally would sleep in select() when
12559 it should have checked SSL_pending() first.
12562 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12563 the raw DSA values prior to ASN.1 encoding.
12566 *) Tweaks to Configure
12567 [Niels Poppe <niels@netbox.org>]
12569 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12573 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12576 *) New config option to avoid instructions that are illegal on the 80386.
12577 The default code is faster, but requires at least a 486.
12580 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12581 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12582 same as SSL2_VERSION anyway.
12585 *) New "-showcerts" option for s_client.
12588 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12589 application. Various cleanups and fixes.
12592 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12593 modify error routines to work internally. Add error codes and PBE init
12594 to library startup routines.
12597 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12598 packing functions to asn1 and evp. Changed function names and error
12599 codes along the way.
12602 *) PKCS12 integration: and so it begins... First of several patches to
12603 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12604 objects to objects.h
12607 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12608 and display support for Thawte strong extranet extension.
12611 *) Add LinuxPPC support.
12612 [Jeff Dubrule <igor@pobox.org>]
12614 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12615 bn_div_words in alpha.s.
12616 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12618 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12619 OAEP isn't supported when OpenSSL is built with RSAref.
12620 [Ulf Moeller <ulf@fitug.de>]
12622 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12623 so they no longer are missing under -DNOPROTO.
12624 [Soren S. Jorvang <soren@t.dk>]
12627 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12629 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12630 doesn't work when the session is reused. Coming soon!
12633 *) Fix a security hole, that allows sessions to be reused in the wrong
12634 context thus bypassing client cert protection! All software that uses
12635 client certs and session caches in multiple contexts NEEDS PATCHING to
12636 allow session reuse! A fuller solution is in the works.
12637 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12639 *) Some more source tree cleanups (removed obsolete files
12640 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12641 permission on "config" script to be executable) and a fix for the INSTALL
12643 [Ulf Moeller <ulf@fitug.de>]
12645 *) Remove some legacy and erroneous uses of malloc, free instead of
12647 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12649 *) Make rsa_oaep_test return non-zero on error.
12650 [Ulf Moeller <ulf@fitug.de>]
12652 *) Add support for native Solaris shared libraries. Configure
12653 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12654 if someone would make that last step automatic.
12655 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12657 *) ctx_size was not built with the right compiler during "make links". Fixed.
12660 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12661 except NULL ciphers". This means the default cipher list will no longer
12662 enable NULL ciphers. They need to be specifically enabled e.g. with
12663 the string "DEFAULT:eNULL".
12666 *) Fix to RSA private encryption routines: if p < q then it would
12667 occasionally produce an invalid result. This will only happen with
12668 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12671 *) Be less restrictive and allow also `perl util/perlpath.pl
12672 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12673 because this way one can also use an interpreter named `perl5' (which is
12674 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12675 installed as `perl').
12676 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12678 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12679 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12681 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12682 advapi32.lib to Win32 build and change the pem test comparison
12683 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12684 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12685 and crypto/des/ede_cbcm_enc.c.
12688 *) DES quad checksum was broken on big-endian architectures. Fixed.
12691 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12692 Win32 test batch file so it (might) work again. The Win32 test batch file
12693 is horrible: I feel ill....
12696 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12697 in e_os.h. Audit of header files to check ANSI and non ANSI
12698 sections: 10 functions were absent from non ANSI section and not exported
12699 from Windows DLLs. Fixed up libeay.num for new functions.
12702 *) Make `openssl version' output lines consistent.
12703 [Ralf S. Engelschall]
12705 *) Fix Win32 symbol export lists for BIO functions: Added
12706 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12707 to ms/libeay{16,32}.def.
12708 [Ralf S. Engelschall]
12710 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12711 fine under Unix and passes some trivial tests I've now added. But the
12712 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12713 added to make sure no one expects that this stuff really works in the
12714 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12715 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12717 [Ralf S. Engelschall]
12719 *) Fix the generation of two part addresses in perl.
12720 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12722 *) Add config entry for Linux on MIPS.
12723 [John Tobey <jtobey@channel1.com>]
12725 *) Make links whenever Configure is run, unless we are on Windoze.
12728 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12729 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12733 *) Add a useful kludge to allow package maintainers to specify compiler and
12734 other platforms details on the command line without having to patch the
12735 Configure script everytime: One now can use ``perl Configure
12736 <id>:<details>'', i.e. platform ids are allowed to have details appended
12737 to them (separated by colons). This is treated as there would be a static
12738 pre-configured entry in Configure's %table under key <id> with value
12739 <details> and ``perl Configure <id>'' is called. So, when you want to
12740 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12741 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12742 now, which overrides the FreeBSD-elf entry on-the-fly.
12743 [Ralf S. Engelschall]
12745 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12748 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12749 on the `perl Configure ...' command line. This way one can compile
12750 OpenSSL libraries with Position Independent Code (PIC) which is needed
12751 for linking it into DSOs.
12752 [Ralf S. Engelschall]
12754 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12758 *) Cleaned up the LICENSE document: The official contact for any license
12759 questions now is the OpenSSL core team under openssl-core@openssl.org.
12760 And add a paragraph about the dual-license situation to make sure people
12761 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12762 to the OpenSSL toolkit.
12763 [Ralf S. Engelschall]
12765 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12766 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12767 Additionally cleaned up the `make links' target: Remove unnecessary
12768 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12769 to speed processing and no longer clutter the display with confusing
12770 stuff. Instead only the actually done links are displayed.
12771 [Ralf S. Engelschall]
12773 *) Permit null encryption ciphersuites, used for authentication only. It used
12774 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12775 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12779 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12780 signed attributes when verifying signatures (this would break them),
12781 the detached data encoding was wrong and public keys obtained using
12782 X509_get_pubkey() weren't freed.
12785 *) Add text documentation for the BUFFER functions. Also added a work around
12786 to a Win95 console bug. This was triggered by the password read stuff: the
12787 last character typed gets carried over to the next fread(). If you were
12788 generating a new cert request using 'req' for example then the last
12789 character of the passphrase would be CR which would then enter the first
12793 *) Added the new `Includes OpenSSL Cryptography Software' button as
12794 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12795 button and can be used by applications based on OpenSSL to show the
12796 relationship to the OpenSSL project.
12797 [Ralf S. Engelschall]
12799 *) Remove confusing variables in function signatures in files
12800 ssl/ssl_lib.c and ssl/ssl.h.
12801 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12803 *) Don't install bss_file.c under PREFIX/include/
12804 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12806 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12807 functions that return function pointers and has support for NT specific
12808 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12809 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12810 unsigned to signed types: this was killing the Win32 compile.
12813 *) Add new certificate file to stack functions,
12814 SSL_add_dir_cert_subjects_to_stack() and
12815 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12816 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12817 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12818 This means that Apache-SSL and similar packages don't have to mess around
12819 to add as many CAs as they want to the preferred list.
12822 *) Experiment with doxygen documentation. Currently only partially applied to
12824 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12825 openssl.doxy as the configuration file.
12828 *) Get rid of remaining C++-style comments which strict C compilers hate.
12829 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12831 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12832 compiled in by default: it has problems with large keys.
12835 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12836 DH private keys and/or callback functions which directly correspond to
12837 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12838 is needed for applications which have to configure certificates on a
12839 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12841 For the RSA certificate situation is makes no difference, but
12842 for the DSA certificate situation this fixes the "no shared cipher"
12843 problem where the OpenSSL cipher selection procedure failed because the
12844 temporary keys were not overtaken from the context and the API provided
12845 no way to reconfigure them.
12846 The new functions now let applications reconfigure the stuff and they
12847 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12848 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12849 non-public-API function ssl_cert_instantiate() is used as a helper
12850 function and also to reduce code redundancy inside ssl_rsa.c.
12851 [Ralf S. Engelschall]
12853 *) Move s_server -dcert and -dkey options out of the undocumented feature
12854 area because they are useful for the DSA situation and should be
12855 recognized by the users.
12856 [Ralf S. Engelschall]
12858 *) Fix the cipher decision scheme for export ciphers: the export bits are
12859 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12860 SSL_EXP_MASK. So, the original variable has to be used instead of the
12861 already masked variable.
12862 [Richard Levitte <levitte@stacken.kth.se>]
12864 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12865 [Richard Levitte <levitte@stacken.kth.se>]
12867 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12868 from `int' to `unsigned int' because it's a length and initialized by
12869 EVP_DigestFinal() which expects an `unsigned int *'.
12870 [Richard Levitte <levitte@stacken.kth.se>]
12872 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12873 script. Instead use the usual Shell->Perl transition trick.
12874 [Ralf S. Engelschall]
12876 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12877 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12878 -noout -modulus' as it's already the case for `openssl rsa -noout
12879 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12880 currently the public key is printed (a decision which was already done by
12881 `openssl dsa -modulus' in the past) which serves a similar purpose.
12882 Additionally the NO_RSA no longer completely removes the whole -modulus
12883 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12885 [Ralf S. Engelschall]
12887 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12888 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12889 [Arne Ansper <arne@ats.cyber.ee>]
12891 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12892 to be added. Now both 'req' and 'ca' can use new objects defined in the
12896 *) Add cool BIO that does syslog (or event log on NT).
12897 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12899 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12900 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12901 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12902 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12905 *) Add preliminary config info for new extension code.
12908 *) Make RSA_NO_PADDING really use no padding.
12909 [Ulf Moeller <ulf@fitug.de>]
12911 *) Generate errors when private/public key check is done.
12914 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12915 for some CRL extensions and new objects added.
12918 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12919 key usage extension and fuller support for authority key id.
12922 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12923 padding method for RSA, which is recommended for new applications in PKCS
12924 #1 v2.0 (RFC 2437, October 1998).
12925 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12926 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12927 against Bleichbacher's attack on RSA.
12928 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12931 *) Updates to the new SSL compression code
12932 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12934 *) Fix so that the version number in the master secret, when passed
12935 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12936 (because the server will not accept higher), that the version number
12937 is 0x03,0x01, not 0x03,0x00
12938 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12940 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12941 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12942 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12945 *) Support for RAW extensions where an arbitrary extension can be
12946 created by including its DER encoding. See apps/openssl.cnf for
12950 *) Make sure latest Perl versions don't interpret some generated C array
12951 code as Perl array code in the crypto/err/err_genc.pl script.
12952 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12954 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12955 not many people have the assembler. Various Win32 compilation fixes and
12956 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12957 build instructions.
12960 *) Modify configure script 'Configure' to automatically create crypto/date.h
12961 file under Win32 and also build pem.h from pem.org. New script
12962 util/mkfiles.pl to create the MINFO file on environments that can't do a
12963 'make files': perl util/mkfiles.pl >MINFO should work.
12966 *) Major rework of DES function declarations, in the pursuit of correctness
12967 and purity. As a result, many evil casts evaporated, and some weirdness,
12968 too. You may find this causes warnings in your code. Zapping your evil
12969 casts will probably fix them. Mostly.
12972 *) Fix for a typo in asn1.h. Bug fix to object creation script
12973 obj_dat.pl. It considered a zero in an object definition to mean
12974 "end of object": none of the objects in objects.h have any zeros
12975 so it wasn't spotted.
12976 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12978 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12979 Masking (CBCM). In the absence of test vectors, the best I have been able
12980 to do is check that the decrypt undoes the encrypt, so far. Send me test
12981 vectors if you have them.
12984 *) Correct calculation of key length for export ciphers (too much space was
12985 allocated for null ciphers). This has not been tested!
12988 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12989 message is now correct (it understands "crypto" and "ssl" on its
12990 command line). There is also now an "update" option. This will update
12991 the util/ssleay.num and util/libeay.num files with any new functions.
12993 perl util/mkdef.pl crypto ssl update
12994 it will update them.
12997 *) Overhauled the Perl interface (perl/*):
12998 - ported BN stuff to OpenSSL's different BN library
12999 - made the perl/ source tree CVS-aware
13000 - renamed the package from SSLeay to OpenSSL (the files still contain
13001 their history because I've copied them in the repository)
13002 - removed obsolete files (the test scripts will be replaced
13003 by better Test::Harness variants in the future)
13004 [Ralf S. Engelschall]
13006 *) First cut for a very conservative source tree cleanup:
13007 1. merge various obsolete readme texts into doc/ssleay.txt
13008 where we collect the old documents and readme texts.
13009 2. remove the first part of files where I'm already sure that we no
13010 longer need them because of three reasons: either they are just temporary
13011 files which were left by Eric or they are preserved original files where
13012 I've verified that the diff is also available in the CVS via "cvs diff
13013 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13014 the crypto/md/ stuff).
13015 [Ralf S. Engelschall]
13017 *) More extension code. Incomplete support for subject and issuer alt
13018 name, issuer and authority key id. Change the i2v function parameters
13019 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13020 what that's for :-) Fix to ASN1 macro which messed up
13021 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13024 *) Preliminary support for ENUMERATED type. This is largely copied from the
13028 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13029 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13031 *) Make sure `make rehash' target really finds the `openssl' program.
13032 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13034 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13035 like to hear about it if this slows down other processors.
13038 *) Add CygWin32 platform information to Configure script.
13039 [Alan Batie <batie@aahz.jf.intel.com>]
13041 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13042 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13044 *) New program nseq to manipulate netscape certificate sequences
13047 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13051 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13052 but the BN code had some problems that would cause failures when
13053 doing certificate verification and some other functions.
13054 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13056 *) Add ASN1 and PEM code to support netscape certificate sequences.
13059 *) Add ASN1 and PEM code to support netscape certificate sequences.
13062 *) Add several PKIX and private extended key usage OIDs.
13065 *) Modify the 'ca' program to handle the new extension code. Modify
13066 openssl.cnf for new extension format, add comments.
13069 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13070 and add a sample to openssl.cnf so req -x509 now adds appropriate
13074 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13075 error code, add initial support to X509_print() and x509 application.
13078 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13079 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13080 stuff is currently isolated and isn't even compiled yet.
13083 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13084 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13085 Removed the versions check from X509 routines when loading extensions:
13086 this allows certain broken certificates that don't set the version
13087 properly to be processed.
13090 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13091 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13092 can still be regenerated with "make depend".
13095 *) Spelling mistake in C version of CAST-128.
13096 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13098 *) Changes to the error generation code. The perl script err-code.pl
13099 now reads in the old error codes and retains the old numbers, only
13100 adding new ones if necessary. It also only changes the .err files if new
13101 codes are added. The makefiles have been modified to only insert errors
13102 when needed (to avoid needlessly modifying header files). This is done
13103 by only inserting errors if the .err file is newer than the auto generated
13104 C file. To rebuild all the error codes from scratch (the old behaviour)
13105 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13106 or delete all the .err files.
13109 *) CAST-128 was incorrectly implemented for short keys. The C version has
13110 been fixed, but is untested. The assembler versions are also fixed, but
13111 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13112 to regenerate it if needed.
13113 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13114 Hagino <itojun@kame.net>]
13116 *) File was opened incorrectly in randfile.c.
13117 [Ulf Möller <ulf@fitug.de>]
13119 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13120 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13121 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13122 al: it's just almost always a UTCTime. Note this patch adds new error
13123 codes so do a "make errors" if there are problems.
13126 *) Correct Linux 1 recognition in config.
13127 [Ulf Möller <ulf@fitug.de>]
13129 *) Remove pointless MD5 hash when using DSA keys in ca.
13130 [Anonymous <nobody@replay.com>]
13132 *) Generate an error if given an empty string as a cert directory. Also
13133 generate an error if handed NULL (previously returned 0 to indicate an
13134 error, but didn't set one).
13135 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13137 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13140 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13141 parameters. This was causing a warning which killed off the Win32 compile.
13144 *) Remove C++ style comments from crypto/bn/bn_local.h.
13145 [Neil Costigan <neil.costigan@celocom.com>]
13147 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13148 based on a text string, looking up short and long names and finally
13149 "dot" format. The "dot" format stuff didn't work. Added new function
13150 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13151 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13152 OID is not part of the table.
13155 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13156 X509_LOOKUP_by_alias().
13159 *) Sort openssl functions by name.
13162 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13163 encryption from sample DSA keys (in case anyone is interested the password
13167 *) Make _all_ *_free functions accept a NULL pointer.
13168 [Frans Heymans <fheymans@isaserver.be>]
13170 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13172 [Anonymous <nobody@replay.com>]
13174 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13175 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13177 *) Don't blow it for numeric -newkey arguments to apps/req.
13178 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13180 *) Temp key "for export" tests were wrong in s3_srvr.c.
13181 [Anonymous <nobody@replay.com>]
13183 *) Add prototype for temp key callback functions
13184 SSL_CTX_set_tmp_{rsa,dh}_callback().
13187 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13188 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13191 *) X509_name_add_entry() freed the wrong thing after an error.
13192 [Arne Ansper <arne@ats.cyber.ee>]
13194 *) rsa_eay.c would attempt to free a NULL context.
13195 [Arne Ansper <arne@ats.cyber.ee>]
13197 *) BIO_s_socket() had a broken should_retry() on Windoze.
13198 [Arne Ansper <arne@ats.cyber.ee>]
13200 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13201 [Arne Ansper <arne@ats.cyber.ee>]
13203 *) Make sure the already existing X509_STORE->depth variable is initialized
13204 in X509_STORE_new(), but document the fact that this variable is still
13205 unused in the certificate verification process.
13206 [Ralf S. Engelschall]
13208 *) Fix the various library and apps files to free up pkeys obtained from
13209 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13212 *) Fix reference counting in X509_PUBKEY_get(). This makes
13213 demos/maurice/example2.c work, amongst others, probably.
13214 [Steve Henson and Ben Laurie]
13216 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13217 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13218 are no longer created. This way we have a single and consistent command
13219 line interface `openssl <command>', similar to `cvs <command>'.
13220 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13222 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13223 BIT STRING wrapper always have zero unused bits.
13226 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13229 *) Make the top-level INSTALL documentation easier to understand.
13232 *) Makefiles updated to exit if an error occurs in a sub-directory
13233 make (including if user presses ^C) [Paul Sutton]
13235 *) Make Montgomery context stuff explicit in RSA data structure.
13238 *) Fix build order of pem and err to allow for generated pem.h.
13241 *) Fix renumbering bug in X509_NAME_delete_entry().
13244 *) Enhanced the err-ins.pl script so it makes the error library number
13245 global and can add a library name. This is needed for external ASN1 and
13246 other error libraries.
13249 *) Fixed sk_insert which never worked properly.
13252 *) Fix ASN1 macros so they can handle indefinite length constructed
13253 EXPLICIT tags. Some non standard certificates use these: they can now
13257 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13258 into a single doc/ssleay.txt bundle. This way the information is still
13259 preserved but no longer messes up this directory. Now it's new room for
13260 the new set of documentation files.
13261 [Ralf S. Engelschall]
13263 *) SETs were incorrectly DER encoded. This was a major pain, because they
13264 shared code with SEQUENCEs, which aren't coded the same. This means that
13265 almost everything to do with SETs or SEQUENCEs has either changed name or
13266 number of arguments.
13267 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13269 *) Fix test data to work with the above.
13272 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13273 was already fixed by Eric for 0.9.1 it seems.
13274 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13276 *) Autodetect FreeBSD3.
13279 *) Fix various bugs in Configure. This affects the following platforms:
13283 unixware-2.0-pentium
13287 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13288 before they are needed.
13291 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13295 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13297 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13298 changed SSLeay to OpenSSL in version strings.
13299 [Ralf S. Engelschall]
13301 *) Some fixups to the top-level documents.
13304 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13305 because the symlink to include/ was missing.
13306 [Ralf S. Engelschall]
13308 *) Incorporated the popular no-RSA/DSA-only patches
13309 which allow to compile a RSA-free SSLeay.
13310 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13312 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13313 when "ssleay" is still not found.
13314 [Ralf S. Engelschall]
13316 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13317 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13319 *) Updated the README file.
13320 [Ralf S. Engelschall]
13322 *) Added various .cvsignore files in the CVS repository subdirs
13323 to make a "cvs update" really silent.
13324 [Ralf S. Engelschall]
13326 *) Recompiled the error-definition header files and added
13327 missing symbols to the Win32 linker tables.
13328 [Ralf S. Engelschall]
13330 *) Cleaned up the top-level documents;
13331 o new files: CHANGES and LICENSE
13332 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13333 o merged COPYRIGHT into LICENSE
13334 o removed obsolete TODO file
13335 o renamed MICROSOFT to INSTALL.W32
13336 [Ralf S. Engelschall]
13338 *) Removed dummy files from the 0.9.1b source tree:
13339 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13340 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13341 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13342 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13343 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13344 [Ralf S. Engelschall]
13346 *) Added various platform portability fixes.
13349 *) The Genesis of the OpenSSL rpject:
13350 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13351 Young and Tim J. Hudson created while they were working for C2Net until
13353 [The OpenSSL Project]
13356 Changes between 0.9.0b and 0.9.1b [not released]
13358 *) Updated a few CA certificates under certs/
13361 *) Changed some BIGNUM api stuff.
13364 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13365 DGUX x86, Linux Alpha, etc.
13368 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13369 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13373 *) Add -strparse option to asn1pars program which parses nested
13375 [Dr Stephen Henson <shenson@bigfoot.com>]
13377 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13380 *) DSA fix for "ca" program.
13383 *) Added "-genkey" option to "dsaparam" program.
13386 *) Added RIPE MD160 (rmd160) message digest.
13389 *) Added -a (all) option to "ssleay version" command.
13392 *) Added PLATFORM define which is the id given to Configure.
13395 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13398 *) Extended the ASN.1 parser routines.
13401 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13404 *) Added a BN_CTX to the BN library.
13407 *) Fixed the weak key values in DES library
13410 *) Changed API in EVP library for cipher aliases.
13413 *) Added support for RC2/64bit cipher.
13416 *) Converted the lhash library to the crypto/mem.c functions.
13419 *) Added more recognized ASN.1 object ids.
13422 *) Added more RSA padding checks for SSL/TLS.
13425 *) Added BIO proxy/filter functionality.
13428 *) Added extra_certs to SSL_CTX which can be used
13429 send extra CA certificates to the client in the CA cert chain sending
13430 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13433 *) Now Fortezza is denied in the authentication phase because
13434 this is key exchange mechanism is not supported by SSLeay at all.
13437 *) Additional PKCS1 checks.
13440 *) Support the string "TLSv1" for all TLS v1 ciphers.
13443 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13444 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13447 *) Fixed a few memory leaks.
13450 *) Fixed various code and comment typos.
13453 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13454 bytes sent in the client random.
13455 [Edward Bishop <ebishop@spyglass.com>]