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 *) Added newline escaping functionality to a filename when using openssl dgst.
13 This output format is to replicate the output format found in the '*sum'
14 checksum programs. This aims to preserve backward compatibility.
15 [Matt Eaton, Richard Levitte, and Paul Dale]
17 *) Removed the heartbeat message in DTLS feature, as it has very
18 little usage and doesn't seem to fulfill a valuable purpose.
21 *) Changed the output of 'openssl {digestname} < file' to display the
22 digest name in its output.
25 *) Added a new generic trace API which provides support for enabling
26 instrumentation through trace output. This feature is mainly intended
27 as an aid for developers and is disabled by default. To utilize it,
28 OpenSSL needs to be configured with the `enable-trace` option.
30 If the tracing API is enabled, the application can activate trace output
31 by registering BIOs as trace channels for a number of tracing and debugging
34 The 'openssl' application has been expanded to enable any of the types
35 available via environment variables defined by the user, and serves as
36 one possible example on how to use this functionality.
37 [Richard Levitte & Matthias St. Pierre]
39 *) Added build tests for C++. These are generated files that only do one
40 thing, to include one public OpenSSL head file each. This tests that
41 the public header files can be usefully included in a C++ application.
43 This test isn't enabled by default. It can be enabled with the option
44 'enable-buildtest-c++'.
47 *) Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
50 *) Add KMAC to EVP_MAC.
53 *) Added property based algorithm implementation selection framework to
57 *) Added SCA hardening for modular field inversion in EC_GROUP through
58 a new dedicated field_inv() pointer in EC_METHOD.
59 This also addresses a leakage affecting conversions from projective
60 to affine coordinates.
61 [Billy Bob Brumley, Nicola Tuveri]
63 *) Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
64 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
65 those algorithms that were already supported through the EVP_PKEY API
66 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
67 and scrypt are now wrappers that call EVP_KDF.
70 *) Build devcrypto engine as a dynamic engine.
73 *) Add keyed BLAKE2 to EVP_MAC.
76 *) Fix a bug in the computation of the endpoint-pair shared secret used
77 by DTLS over SCTP. This breaks interoperability with older versions
78 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
79 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
80 interoperability with such broken implementations. However, enabling
81 this switch breaks interoperability with correct implementations.
83 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
84 re-used X509_PUBKEY object if the second PUBKEY is malformed.
87 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
90 *) Change the license to the Apache License v2.0.
93 *) Change the possible version information given with OPENSSL_API_COMPAT.
94 It may be a pre-3.0.0 style numerical version number as it was defined
95 in 1.1.0, and it may also simply take the major version number.
97 Because of the version numbering of pre-3.0.0 releases, the values 0,
98 1 and 2 are equivalent to 0x00908000L (0.9.8), 0x10000000L (1.0.0) and
99 0x10100000L (1.1.0), respectively.
102 *) Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
104 o Major releases (indicated by incrementing the MAJOR release number)
105 may introduce incompatible API/ABI changes.
106 o Minor releases (indicated by incrementing the MINOR release number)
107 may introduce new features but retain API/ABI compatibility.
108 o Patch releases (indicated by incrementing the PATCH number)
109 are intended for bug fixes and other improvements of existing
110 features only (like improving performance or adding documentation)
111 and retain API/ABI compatibility.
114 *) Add support for RFC5297 SIV mode (siv128), including AES-SIV.
117 *) Remove the 'dist' target and add a tarball building script. The
118 'dist' target has fallen out of use, and it shouldn't be
119 necessary to configure just to create a source distribution.
122 *) Recreate the OS390-Unix config target. It no longer relies on a
123 special script like it did for OpenSSL pre-1.1.0.
126 *) Instead of having the source directories listed in Configure, add
127 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
131 *) Add GMAC to EVP_MAC.
134 *) Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
137 *) Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
138 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
139 to facilitate the continued use of MACs through raw private keys in
140 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
143 *) Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
144 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
147 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
148 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
149 are retained for backwards compatibility.
152 *) AES-XTS mode now enforces that its two keys are different to mitigate
153 the attacked described in "Efficient Instantiations of Tweakable
154 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
155 Details of this attack can be obtained from:
156 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
159 *) Rename the object files, i.e. give them other names than in previous
160 versions. Their names now include the name of the final product, as
161 well as its type mnemonic (bin, lib, shlib).
164 *) Added new option for 'openssl list', '-objects', which will display the
165 list of built in objects, i.e. OIDs with names.
168 *) Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
169 improves application performance by removing data copies and providing
170 applications with zero-copy system calls such as sendfile and splice.
173 Changes between 1.1.1a and 1.1.1b [xx XXX xxxx]
175 *) Change the info callback signals for the start and end of a post-handshake
176 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
177 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
178 confused by this and assume that a TLSv1.2 renegotiation has started. This
179 can break KeyUpdate handling. Instead we no longer signal the start and end
180 of a post handshake message exchange (although the messages themselves are
181 still signalled). This could break some applications that were expecting
182 the old signals. However without this KeyUpdate is not usable for many
186 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
188 *) Timing vulnerability in DSA signature generation
190 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
191 timing side channel attack. An attacker could use variations in the signing
192 algorithm to recover the private key.
194 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
198 *) Timing vulnerability in ECDSA signature generation
200 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
201 timing side channel attack. An attacker could use variations in the signing
202 algorithm to recover the private key.
204 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
208 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
209 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
210 of two gigabytes and the error handling improved.
212 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
213 categorized as a normal bug, not a security issue, because the DRBG reseeds
214 automatically and is fully functional even without additional randomness
215 provided by the application.
217 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
219 *) Add a new ClientHello callback. Provides a callback interface that gives
220 the application the ability to adjust the nascent SSL object at the
221 earliest stage of ClientHello processing, immediately after extensions have
222 been collected but before they have been processed. In particular, this
223 callback can adjust the supported TLS versions in response to the contents
227 *) Add SM2 base algorithm support.
230 *) s390x assembly pack: add (improved) hardware-support for the following
231 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
232 aes-cfb/cfb8, aes-ecb.
235 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
236 parameter is no longer accepted, as it leads to a corrupt table. NULL
237 pem_str is reserved for alias entries only.
240 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
241 step for prime curves. The new implementation is based on formulae from
242 differential addition-and-doubling in homogeneous projective coordinates
243 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
244 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
245 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
246 to work in projective coordinates.
247 [Billy Bob Brumley, Nicola Tuveri]
249 *) Change generating and checking of primes so that the error rate of not
250 being prime depends on the intended use based on the size of the input.
251 For larger primes this will result in more rounds of Miller-Rabin.
252 The maximal error rate for primes with more than 1080 bits is lowered
254 [Kurt Roeckx, Annie Yousar]
256 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
259 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
260 moving between systems, and to avoid confusion when a Windows build is
261 done with mingw vs with MSVC. For POSIX installs, there's still a
262 symlink or copy named 'tsget' to avoid that confusion as well.
265 *) Revert blinding in ECDSA sign and instead make problematic addition
266 length-invariant. Switch even to fixed-length Montgomery multiplication.
269 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
270 step for binary curves. The new implementation is based on formulae from
271 differential addition-and-doubling in mixed Lopez-Dahab projective
272 coordinates, modified to independently blind the operands.
273 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
275 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
276 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
277 EC_METHODs to implement their own specialized "ladder step", to take
278 advantage of more favorable coordinate systems or more efficient
279 differential addition-and-doubling algorithms.
280 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
282 *) Modified the random device based seed sources to keep the relevant
283 file descriptors open rather than reopening them on each access.
284 This allows such sources to operate in a chroot() jail without
285 the associated device nodes being available. This behaviour can be
286 controlled using RAND_keep_random_devices_open().
289 *) Numerous side-channel attack mitigations have been applied. This may have
290 performance impacts for some algorithms for the benefit of improved
291 security. Specific changes are noted in this change log by their respective
295 *) AIX shared library support overhaul. Switch to AIX "natural" way of
296 handling shared libraries, which means collecting shared objects of
297 different versions and bitnesses in one common archive. This allows to
298 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
299 doesn't affect the way 3rd party applications are linked, only how
300 multi-version installation is managed.
303 *) Make ec_group_do_inverse_ord() more robust and available to other
304 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
305 mitigations are applied to the fallback BN_mod_inverse().
306 When using this function rather than BN_mod_inverse() directly, new
307 EC cryptosystem implementations are then safer-by-default.
310 *) Add coordinate blinding for EC_POINT and implement projective
311 coordinate blinding for generic prime curves as a countermeasure to
312 chosen point SCA attacks.
313 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
315 *) Add blinding to ECDSA and DSA signatures to protect against side channel
316 attacks discovered by Keegan Ryan (NCC Group).
319 *) Enforce checking in the pkeyutl command line app to ensure that the input
320 length does not exceed the maximum supported digest length when performing
321 a sign, verify or verifyrecover operation.
324 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
325 I/O in combination with something like select() or poll() will hang. This
326 can be turned off again using SSL_CTX_clear_mode().
327 Many applications do not properly handle non-application data records, and
328 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
329 around the problems in those applications, but can also break some.
330 It's recommended to read the manpages about SSL_read(), SSL_write(),
331 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
332 SSL_CTX_set_read_ahead() again.
335 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
336 now allow empty (zero character) pass phrases.
339 *) Apply blinding to binary field modular inversion and remove patent
340 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
343 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
344 binary and prime elliptic curves.
347 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
348 constant time fixed point multiplication.
351 *) Revise elliptic curve scalar multiplication with timing attack
352 defenses: ec_wNAF_mul redirects to a constant time implementation
353 when computing fixed point and variable point multiplication (which
354 in OpenSSL are mostly used with secret scalars in keygen, sign,
355 ECDH derive operations).
356 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
359 *) Updated CONTRIBUTING
362 *) Updated DRBG / RAND to request nonce and additional low entropy
363 randomness from the system.
364 [Matthias St. Pierre]
366 *) Updated 'openssl rehash' to use OpenSSL consistent default.
369 *) Moved the load of the ssl_conf module to libcrypto, which helps
370 loading engines that libssl uses before libssl is initialised.
373 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
376 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
377 [Ingo Schwarze, Rich Salz]
379 *) Added output of accepting IP address and port for 'openssl s_server'
382 *) Added a new API for TLSv1.3 ciphersuites:
383 SSL_CTX_set_ciphersuites()
384 SSL_set_ciphersuites()
387 *) Memory allocation failures consistenly add an error to the error
391 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
392 in libcrypto when run as setuid/setgid.
395 *) Load any config file by default when libssl is used.
398 *) Added new public header file <openssl/rand_drbg.h> and documentation
399 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
400 [Matthias St. Pierre]
402 *) QNX support removed (cannot find contributors to get their approval
403 for the license change).
406 *) TLSv1.3 replay protection for early data has been implemented. See the
407 SSL_read_early_data() man page for further details.
410 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
411 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
412 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
413 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
414 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
415 configuration has been separated out. See the ciphers man page or the
416 SSL_CTX_set_ciphersuites() man page for more information.
419 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
420 in responder mode now supports the new "-multi" option, which
421 spawns the specified number of child processes to handle OCSP
422 requests. The "-timeout" option now also limits the OCSP
423 responder's patience to wait to receive the full client request
424 on a newly accepted connection. Child processes are respawned
425 as needed, and the CA index file is automatically reloaded
426 when changed. This makes it possible to run the "ocsp" responder
427 as a long-running service, making the OpenSSL CA somewhat more
428 feature-complete. In this mode, most diagnostic messages logged
429 after entering the event loop are logged via syslog(3) rather than
433 *) Added support for X448 and Ed448. Heavily based on original work by
437 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
438 objects loaded. This adds the functions OSSL_STORE_expect() and
439 OSSL_STORE_find() as well as needed tools to construct searches and
440 get the search data out of them.
443 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
444 version of OpenSSL should review their configuration settings to ensure
445 that they are still appropriate for TLSv1.3. For further information see:
446 https://wiki.openssl.org/index.php/TLS1.3
449 *) Grand redesign of the OpenSSL random generator
451 The default RAND method now utilizes an AES-CTR DRBG according to
452 NIST standard SP 800-90Ar1. The new random generator is essentially
453 a port of the default random generator from the OpenSSL FIPS 2.0
454 object module. It is a hybrid deterministic random bit generator
455 using an AES-CTR bit stream and which seeds and reseeds itself
456 automatically using trusted system entropy sources.
458 Some of its new features are:
459 o Support for multiple DRBG instances with seed chaining.
460 o The default RAND method makes use of a DRBG.
461 o There is a public and private DRBG instance.
462 o The DRBG instances are fork-safe.
463 o Keep all global DRBG instances on the secure heap if it is enabled.
464 o The public and private DRBG instance are per thread for lock free
466 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
468 *) Changed Configure so it only says what it does and doesn't dump
469 so much data. Instead, ./configdata.pm should be used as a script
470 to display all sorts of configuration data.
473 *) Added processing of "make variables" to Configure.
476 *) Added SHA512/224 and SHA512/256 algorithm support.
479 *) The last traces of Netware support, first removed in 1.1.0, have
483 *) Get rid of Makefile.shared, and in the process, make the processing
484 of certain files (rc.obj, or the .def/.map/.opt files produced from
485 the ordinal files) more visible and hopefully easier to trace and
486 debug (or make silent).
489 *) Make it possible to have environment variable assignments as
490 arguments to config / Configure.
493 *) Add multi-prime RSA (RFC 8017) support.
496 *) Add SM3 implemented according to GB/T 32905-2016
497 [ Jack Lloyd <jack.lloyd@ribose.com>,
498 Ronald Tse <ronald.tse@ribose.com>,
499 Erick Borsboom <erick.borsboom@ribose.com> ]
501 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
502 as documented in RFC6066.
503 Based on a patch from Tomasz Moń
504 [Filipe Raimundo da Silva]
506 *) Add SM4 implemented according to GB/T 32907-2016.
507 [ Jack Lloyd <jack.lloyd@ribose.com>,
508 Ronald Tse <ronald.tse@ribose.com>,
509 Erick Borsboom <erick.borsboom@ribose.com> ]
511 *) Reimplement -newreq-nodes and ERR_error_string_n; the
512 original author does not agree with the license change.
515 *) Add ARIA AEAD TLS support.
518 *) Some macro definitions to support VS6 have been removed. Visual
519 Studio 6 has not worked since 1.1.0
522 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
523 without clearing the errors.
526 *) Add "atfork" functions. If building on a system that without
527 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
528 requirements. The RAND facility now uses/requires this.
534 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
535 not possible to disable entirely. However, it's still possible to
536 disable the console reading UI method, UI_OpenSSL() (use UI_null()
539 To disable, configure with 'no-ui-console'. 'no-ui' is still
540 possible to use as an alias. Check at compile time with the
541 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
542 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
545 *) Add a STORE module, which implements a uniform and URI based reader of
546 stores that can contain keys, certificates, CRLs and numerous other
547 objects. The main API is loosely based on a few stdio functions,
548 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
549 OSSL_STORE_error and OSSL_STORE_close.
550 The implementation uses backends called "loaders" to implement arbitrary
551 URI schemes. There is one built in "loader" for the 'file' scheme.
554 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
555 then adjusted to work on FreeBSD 8.4 as well.
556 Enable by configuring with 'enable-devcryptoeng'. This is done by default
557 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
560 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
561 util/mkerr.pl, which is adapted to allow those prefixes, leading to
562 error code calls like this:
564 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
566 With this change, we claim the namespaces OSSL and OPENSSL in a manner
567 that can be encoded in C. For the foreseeable future, this will only
569 [Richard Levitte and Tim Hudson]
571 *) Removed BSD cryptodev engine.
574 *) Add a build target 'build_all_generated', to build all generated files
575 and only that. This can be used to prepare everything that requires
576 things like perl for a system that lacks perl and then move everything
577 to that system and do the rest of the build there.
580 *) In the UI interface, make it possible to duplicate the user data. This
581 can be used by engines that need to retain the data for a longer time
582 than just the call where this user data is passed.
585 *) Ignore the '-named_curve auto' value for compatibility of applications
587 [Tomas Mraz <tmraz@fedoraproject.org>]
589 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
590 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
591 alerts across multiple records (some of which could be empty). In practice
592 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
593 prohibts this altogether and other libraries (BoringSSL, NSS) do not
594 support this at all. Supporting it adds significant complexity to the
595 record layer, and its removal is unlikely to cause inter-operability
599 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
600 with Z. These are meant to replace LONG and ZLONG and to be size safe.
601 The use of LONG and ZLONG is discouraged and scheduled for deprecation
605 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
606 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
607 [Richard Levitte, Andy Polyakov]
609 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
613 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
614 platform rather than 'mingw'.
617 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
618 success if they are asked to add an object which already exists
619 in the store. This change cascades to other functions which load
620 certificates and CRLs.
623 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
624 facilitate stack unwinding even from assembly subroutines.
627 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
628 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
631 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
632 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
633 which is the minimum version we support.
636 *) Certificate time validation (X509_cmp_time) enforces stricter
637 compliance with RFC 5280. Fractional seconds and timezone offsets
638 are no longer allowed.
641 *) Add support for ARIA
644 *) s_client will now send the Server Name Indication (SNI) extension by
645 default unless the new "-noservername" option is used. The server name is
646 based on the host provided to the "-connect" option unless overridden by
650 *) Add support for SipHash
653 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
654 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
655 prevent issues where no progress is being made and the peer continually
656 sends unrecognised record types, using up resources processing them.
659 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
660 using the algorithm defined in
661 https://www.akkadia.org/drepper/SHA-crypt.txt
664 *) Heartbeat support has been removed; the ABI is changed for now.
665 [Richard Levitte, Rich Salz]
667 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
670 *) The RSA "null" method, which was partially supported to avoid patent
671 issues, has been replaced to always returns NULL.
675 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
677 *) Client DoS due to large DH parameter
679 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
680 malicious server can send a very large prime value to the client. This will
681 cause the client to spend an unreasonably long period of time generating a
682 key for this prime resulting in a hang until the client has finished. This
683 could be exploited in a Denial Of Service attack.
685 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
689 *) Cache timing vulnerability in RSA Key Generation
691 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
692 a cache timing side channel attack. An attacker with sufficient access to
693 mount cache timing attacks during the RSA key generation process could
694 recover the private key.
696 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
697 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
701 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
702 parameter is no longer accepted, as it leads to a corrupt table. NULL
703 pem_str is reserved for alias entries only.
706 *) Revert blinding in ECDSA sign and instead make problematic addition
707 length-invariant. Switch even to fixed-length Montgomery multiplication.
710 *) Change generating and checking of primes so that the error rate of not
711 being prime depends on the intended use based on the size of the input.
712 For larger primes this will result in more rounds of Miller-Rabin.
713 The maximal error rate for primes with more than 1080 bits is lowered
715 [Kurt Roeckx, Annie Yousar]
717 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
720 *) Add blinding to ECDSA and DSA signatures to protect against side channel
721 attacks discovered by Keegan Ryan (NCC Group).
724 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
725 now allow empty (zero character) pass phrases.
728 *) Certificate time validation (X509_cmp_time) enforces stricter
729 compliance with RFC 5280. Fractional seconds and timezone offsets
730 are no longer allowed.
733 *) Fixed a text canonicalisation bug in CMS
735 Where a CMS detached signature is used with text content the text goes
736 through a canonicalisation process first prior to signing or verifying a
737 signature. This process strips trailing space at the end of lines, converts
738 line terminators to CRLF and removes additional trailing line terminators
739 at the end of a file. A bug in the canonicalisation process meant that
740 some characters, such as form-feed, were incorrectly treated as whitespace
741 and removed. This is contrary to the specification (RFC5485). This fix
742 could mean that detached text data signed with an earlier version of
743 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
744 signed with a fixed OpenSSL may fail to verify with an earlier version of
745 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
746 and use the "-binary" flag (for the "cms" command line application) or set
747 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
750 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
752 *) Constructed ASN.1 types with a recursive definition could exceed the stack
754 Constructed ASN.1 types with a recursive definition (such as can be found
755 in PKCS7) could eventually exceed the stack given malicious input with
756 excessive recursion. This could result in a Denial Of Service attack. There
757 are no such structures used within SSL/TLS that come from untrusted sources
758 so this is considered safe.
760 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
765 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
767 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
768 effectively reduced to only comparing the least significant bit of each
769 byte. This allows an attacker to forge messages that would be considered as
770 authenticated in an amount of tries lower than that guaranteed by the
771 security claims of the scheme. The module can only be compiled by the
772 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
774 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
779 *) Add a build target 'build_all_generated', to build all generated files
780 and only that. This can be used to prepare everything that requires
781 things like perl for a system that lacks perl and then move everything
782 to that system and do the rest of the build there.
785 *) Backport SSL_OP_NO_RENGOTIATION
787 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
788 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
789 changes this is no longer possible in 1.1.0. Therefore the new
790 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
791 1.1.0 to provide equivalent functionality.
793 Note that if an application built against 1.1.0h headers (or above) is run
794 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
795 accepted but nothing will happen, i.e. renegotiation will not be prevented.
798 *) Removed the OS390-Unix config target. It relied on a script that doesn't
802 *) rsaz_1024_mul_avx2 overflow bug on x86_64
804 There is an overflow bug in the AVX2 Montgomery multiplication procedure
805 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
806 Analysis suggests that attacks against RSA and DSA as a result of this
807 defect would be very difficult to perform and are not believed likely.
808 Attacks against DH1024 are considered just feasible, because most of the
809 work necessary to deduce information about a private key may be performed
810 offline. The amount of resources required for such an attack would be
811 significant. However, for an attack on TLS to be meaningful, the server
812 would have to share the DH1024 private key among multiple clients, which is
813 no longer an option since CVE-2016-0701.
815 This only affects processors that support the AVX2 but not ADX extensions
816 like Intel Haswell (4th generation).
818 This issue was reported to OpenSSL by David Benjamin (Google). The issue
819 was originally found via the OSS-Fuzz project.
823 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
825 *) bn_sqrx8x_internal carry bug on x86_64
827 There is a carry propagating bug in the x86_64 Montgomery squaring
828 procedure. No EC algorithms are affected. Analysis suggests that attacks
829 against RSA and DSA as a result of this defect would be very difficult to
830 perform and are not believed likely. Attacks against DH are considered just
831 feasible (although very difficult) because most of the work necessary to
832 deduce information about a private key may be performed offline. The amount
833 of resources required for such an attack would be very significant and
834 likely only accessible to a limited number of attackers. An attacker would
835 additionally need online access to an unpatched system using the target
836 private key in a scenario with persistent DH parameters and a private
837 key that is shared between multiple clients.
839 This only affects processors that support the BMI1, BMI2 and ADX extensions
840 like Intel Broadwell (5th generation) and later or AMD Ryzen.
842 This issue was reported to OpenSSL by the OSS-Fuzz project.
846 *) Malformed X.509 IPAddressFamily could cause OOB read
848 If an X.509 certificate has a malformed IPAddressFamily extension,
849 OpenSSL could do a one-byte buffer overread. The most likely result
850 would be an erroneous display of the certificate in text format.
852 This issue was reported to OpenSSL by the OSS-Fuzz project.
856 Changes between 1.1.0e and 1.1.0f [25 May 2017]
858 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
859 platform rather than 'mingw'.
862 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
863 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
864 which is the minimum version we support.
867 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
869 *) Encrypt-Then-Mac renegotiation crash
871 During a renegotiation handshake if the Encrypt-Then-Mac extension is
872 negotiated where it was not in the original handshake (or vice-versa) then
873 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
874 and servers are affected.
876 This issue was reported to OpenSSL by Joe Orton (Red Hat).
880 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
882 *) Truncated packet could crash via OOB read
884 If one side of an SSL/TLS path is running on a 32-bit host and a specific
885 cipher is being used, then a truncated packet can cause that host to
886 perform an out-of-bounds read, usually resulting in a crash.
888 This issue was reported to OpenSSL by Robert Święcki of Google.
892 *) Bad (EC)DHE parameters cause a client crash
894 If a malicious server supplies bad parameters for a DHE or ECDHE key
895 exchange then this can result in the client attempting to dereference a
896 NULL pointer leading to a client crash. This could be exploited in a Denial
899 This issue was reported to OpenSSL by Guido Vranken.
903 *) BN_mod_exp may produce incorrect results on x86_64
905 There is a carry propagating bug in the x86_64 Montgomery squaring
906 procedure. No EC algorithms are affected. Analysis suggests that attacks
907 against RSA and DSA as a result of this defect would be very difficult to
908 perform and are not believed likely. Attacks against DH are considered just
909 feasible (although very difficult) because most of the work necessary to
910 deduce information about a private key may be performed offline. The amount
911 of resources required for such an attack would be very significant and
912 likely only accessible to a limited number of attackers. An attacker would
913 additionally need online access to an unpatched system using the target
914 private key in a scenario with persistent DH parameters and a private
915 key that is shared between multiple clients. For example this can occur by
916 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
917 similar to CVE-2015-3193 but must be treated as a separate problem.
919 This issue was reported to OpenSSL by the OSS-Fuzz project.
923 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
925 *) ChaCha20/Poly1305 heap-buffer-overflow
927 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
928 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
929 crash. This issue is not considered to be exploitable beyond a DoS.
931 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
935 *) CMS Null dereference
937 Applications parsing invalid CMS structures can crash with a NULL pointer
938 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
939 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
940 structure callback if an attempt is made to free certain invalid encodings.
941 Only CHOICE structures using a callback which do not handle NULL value are
944 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
948 *) Montgomery multiplication may produce incorrect results
950 There is a carry propagating bug in the Broadwell-specific Montgomery
951 multiplication procedure that handles input lengths divisible by, but
952 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
953 and DH private keys are impossible. This is because the subroutine in
954 question is not used in operations with the private key itself and an input
955 of the attacker's direct choice. Otherwise the bug can manifest itself as
956 transient authentication and key negotiation failures or reproducible
957 erroneous outcome of public-key operations with specially crafted input.
958 Among EC algorithms only Brainpool P-512 curves are affected and one
959 presumably can attack ECDH key negotiation. Impact was not analyzed in
960 detail, because pre-requisites for attack are considered unlikely. Namely
961 multiple clients have to choose the curve in question and the server has to
962 share the private key among them, neither of which is default behaviour.
963 Even then only clients that chose the curve will be affected.
965 This issue was publicly reported as transient failures and was not
966 initially recognized as a security issue. Thanks to Richard Morgan for
967 providing reproducible case.
971 *) Removed automatic addition of RPATH in shared libraries and executables,
972 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
975 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
977 *) Fix Use After Free for large message sizes
979 The patch applied to address CVE-2016-6307 resulted in an issue where if a
980 message larger than approx 16k is received then the underlying buffer to
981 store the incoming message is reallocated and moved. Unfortunately a
982 dangling pointer to the old location is left which results in an attempt to
983 write to the previously freed location. This is likely to result in a
984 crash, however it could potentially lead to execution of arbitrary code.
986 This issue only affects OpenSSL 1.1.0a.
988 This issue was reported to OpenSSL by Robert Święcki.
992 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
994 *) OCSP Status Request extension unbounded memory growth
996 A malicious client can send an excessively large OCSP Status Request
997 extension. If that client continually requests renegotiation, sending a
998 large OCSP Status Request extension each time, then there will be unbounded
999 memory growth on the server. This will eventually lead to a Denial Of
1000 Service attack through memory exhaustion. Servers with a default
1001 configuration are vulnerable even if they do not support OCSP. Builds using
1002 the "no-ocsp" build time option are not affected.
1004 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1008 *) SSL_peek() hang on empty record
1010 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1011 sends an empty record. This could be exploited by a malicious peer in a
1012 Denial Of Service attack.
1014 This issue was reported to OpenSSL by Alex Gaynor.
1018 *) Excessive allocation of memory in tls_get_message_header() and
1019 dtls1_preprocess_fragment()
1021 A (D)TLS message includes 3 bytes for its length in the header for the
1022 message. This would allow for messages up to 16Mb in length. Messages of
1023 this length are excessive and OpenSSL includes a check to ensure that a
1024 peer is sending reasonably sized messages in order to avoid too much memory
1025 being consumed to service a connection. A flaw in the logic of version
1026 1.1.0 means that memory for the message is allocated too early, prior to
1027 the excessive message length check. Due to way memory is allocated in
1028 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1029 to service a connection. This could lead to a Denial of Service through
1030 memory exhaustion. However, the excessive message length check still takes
1031 place, and this would cause the connection to immediately fail. Assuming
1032 that the application calls SSL_free() on the failed connection in a timely
1033 manner then the 21Mb of allocated memory will then be immediately freed
1034 again. Therefore the excessive memory allocation will be transitory in
1035 nature. This then means that there is only a security impact if:
1037 1) The application does not call SSL_free() in a timely manner in the event
1038 that the connection fails
1040 2) The application is working in a constrained environment where there is
1041 very little free memory
1043 3) The attacker initiates multiple connection attempts such that there are
1044 multiple connections in a state where memory has been allocated for the
1045 connection; SSL_free() has not yet been called; and there is insufficient
1046 memory to service the multiple requests.
1048 Except in the instance of (1) above any Denial Of Service is likely to be
1049 transitory because as soon as the connection fails the memory is
1050 subsequently freed again in the SSL_free() call. However there is an
1051 increased risk during this period of application crashes due to the lack of
1052 memory - which would then mean a more serious Denial of Service.
1054 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1055 (CVE-2016-6307 and CVE-2016-6308)
1058 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1059 had to be removed. Primary reason is that vendor assembler can't
1060 assemble our modules with -KPIC flag. As result it, assembly
1061 support, was not even available as option. But its lack means
1062 lack of side-channel resistant code, which is incompatible with
1063 security by todays standards. Fortunately gcc is readily available
1064 prepackaged option, which we firmly point at...
1067 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1069 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1070 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1071 (to any value) allows Windows user to access PKCS#12 file generated
1072 with Windows CryptoAPI and protected with non-ASCII password, as well
1073 as files generated under UTF-8 locale on Linux also protected with
1077 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1078 have been disabled by default and removed from DEFAULT, just like RC4.
1079 See the RC4 item below to re-enable both.
1082 *) The method for finding the storage location for the Windows RAND seed file
1083 has changed. First we check %RANDFILE%. If that is not set then we check
1084 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1085 all else fails we fall back to C:\.
1088 *) The EVP_EncryptUpdate() function has had its return type changed from void
1089 to int. A return of 0 indicates and error while a return of 1 indicates
1093 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1094 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1095 off the constant time implementation for RSA, DSA and DH have been made
1096 no-ops and deprecated.
1099 *) Windows RAND implementation was simplified to only get entropy by
1100 calling CryptGenRandom(). Various other RAND-related tickets
1102 [Joseph Wylie Yandle, Rich Salz]
1104 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1105 and OPENSSL_LH_, respectively. The old names are available
1106 with API compatibility. They new names are now completely documented.
1109 *) Unify TYPE_up_ref(obj) methods signature.
1110 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1111 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1112 int (instead of void) like all others TYPE_up_ref() methods.
1113 So now these methods also check the return value of CRYPTO_atomic_add(),
1114 and the validity of object reference counter.
1115 [fdasilvayy@gmail.com]
1117 *) With Windows Visual Studio builds, the .pdb files are installed
1118 alongside the installed libraries and executables. For a static
1119 library installation, ossl_static.pdb is the associate compiler
1120 generated .pdb file to be used when linking programs.
1123 *) Remove openssl.spec. Packaging files belong with the packagers.
1126 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1127 recognise x86_64 architectures automatically. You can still decide
1128 to build for a different bitness with the environment variable
1129 KERNEL_BITS (can be 32 or 64), for example:
1131 KERNEL_BITS=32 ./config
1135 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1136 256 bit AES and HMAC with SHA256.
1139 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1142 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1145 *) To enable users to have their own config files and build file templates,
1146 Configure looks in the directory indicated by the environment variable
1147 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1148 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1149 name and is used as is.
1152 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1153 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1154 X509_CERT_FILE_CTX was removed.
1157 *) "shared" builds are now the default. To create only static libraries use
1158 the "no-shared" Configure option.
1161 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1162 All of these option have not worked for some while and are fundamental
1166 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1167 global cleanup functions are no longer required because they are handled
1168 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1169 Explicitly de-initing can cause problems (e.g. where a library that uses
1170 OpenSSL de-inits, but an application is still using it). The affected
1171 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1172 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1173 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1174 COMP_zlib_cleanup().
1177 *) --strict-warnings no longer enables runtime debugging options
1178 such as REF_DEBUG. Instead, debug options are automatically
1179 enabled with '--debug' builds.
1180 [Andy Polyakov, Emilia Käsper]
1182 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1183 have been moved out of the public header files. New functions for managing
1184 these have been added.
1187 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1188 objects have been moved out of the public header files. New
1189 functions for managing these have been added.
1192 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1193 have been moved out of the public header files. New functions for managing
1194 these have been added.
1197 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1198 moved out of the public header files. New functions for managing these
1202 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1205 *) Removed the mk1mf build scripts.
1208 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1209 it is always safe to #include a header now.
1212 *) Removed the aged BC-32 config and all its supporting scripts
1215 *) Removed support for Ultrix, Netware, and OS/2.
1218 *) Add support for HKDF.
1219 [Alessandro Ghedini]
1221 *) Add support for blake2b and blake2s
1224 *) Added support for "pipelining". Ciphers that have the
1225 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1226 encryptions/decryptions simultaneously. There are currently no built-in
1227 ciphers with this property but the expectation is that engines will be able
1228 to offer it to significantly improve throughput. Support has been extended
1229 into libssl so that multiple records for a single connection can be
1230 processed in one go (for >=TLS 1.1).
1233 *) Added the AFALG engine. This is an async capable engine which is able to
1234 offload work to the Linux kernel. In this initial version it only supports
1235 AES128-CBC. The kernel must be version 4.1.0 or greater.
1238 *) OpenSSL now uses a new threading API. It is no longer necessary to
1239 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1240 are two supported threading models: pthreads and windows threads. It is
1241 also possible to configure OpenSSL at compile time for "no-threads". The
1242 old threading API should no longer be used. The functions have been
1243 replaced with "no-op" compatibility macros.
1244 [Alessandro Ghedini, Matt Caswell]
1246 *) Modify behavior of ALPN to invoke callback after SNI/servername
1247 callback, such that updates to the SSL_CTX affect ALPN.
1250 *) Add SSL_CIPHER queries for authentication and key-exchange.
1253 *) Changes to the DEFAULT cipherlist:
1254 - Prefer (EC)DHE handshakes over plain RSA.
1255 - Prefer AEAD ciphers over legacy ciphers.
1256 - Prefer ECDSA over RSA when both certificates are available.
1257 - Prefer TLSv1.2 ciphers/PRF.
1258 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1262 *) Change the ECC default curve list to be this, in order: x25519,
1263 secp256r1, secp521r1, secp384r1.
1266 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1267 disabled by default. They can be re-enabled using the
1268 enable-weak-ssl-ciphers option to Configure.
1271 *) If the server has ALPN configured, but supports no protocols that the
1272 client advertises, send a fatal "no_application_protocol" alert.
1273 This behaviour is SHALL in RFC 7301, though it isn't universally
1274 implemented by other servers.
1277 *) Add X25519 support.
1278 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1279 for public and private key encoding using the format documented in
1280 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1281 key generation and key derivation.
1283 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1287 *) Deprecate SRP_VBASE_get_by_user.
1288 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1289 In order to fix an unavoidable memory leak (CVE-2016-0798),
1290 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1291 seed, even if the seed is configured.
1293 Users should use SRP_VBASE_get1_by_user instead. Note that in
1294 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1295 also that even though configuring the SRP seed attempts to hide
1296 invalid usernames by continuing the handshake with fake
1297 credentials, this behaviour is not constant time and no strong
1298 guarantees are made that the handshake is indistinguishable from
1299 that of a valid user.
1302 *) Configuration change; it's now possible to build dynamic engines
1303 without having to build shared libraries and vice versa. This
1304 only applies to the engines in engines/, those in crypto/engine/
1305 will always be built into libcrypto (i.e. "static").
1307 Building dynamic engines is enabled by default; to disable, use
1308 the configuration option "disable-dynamic-engine".
1310 The only requirements for building dynamic engines are the
1311 presence of the DSO module and building with position independent
1312 code, so they will also automatically be disabled if configuring
1313 with "disable-dso" or "disable-pic".
1315 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1316 are also taken away from openssl/opensslconf.h, as they are
1320 *) Configuration change; if there is a known flag to compile
1321 position independent code, it will always be applied on the
1322 libcrypto and libssl object files, and never on the application
1323 object files. This means other libraries that use routines from
1324 libcrypto / libssl can be made into shared libraries regardless
1325 of how OpenSSL was configured.
1327 If this isn't desirable, the configuration options "disable-pic"
1328 or "no-pic" can be used to disable the use of PIC. This will
1329 also disable building shared libraries and dynamic engines.
1332 *) Removed JPAKE code. It was experimental and has no wide use.
1335 *) The INSTALL_PREFIX Makefile variable has been renamed to
1336 DESTDIR. That makes for less confusion on what this variable
1337 is for. Also, the configuration option --install_prefix is
1341 *) Heartbeat for TLS has been removed and is disabled by default
1342 for DTLS; configure with enable-heartbeats. Code that uses the
1343 old #define's might need to be updated.
1344 [Emilia Käsper, Rich Salz]
1346 *) Rename REF_CHECK to REF_DEBUG.
1349 *) New "unified" build system
1351 The "unified" build system is aimed to be a common system for all
1352 platforms we support. With it comes new support for VMS.
1354 This system builds supports building in a different directory tree
1355 than the source tree. It produces one Makefile (for unix family
1356 or lookalikes), or one descrip.mms (for VMS).
1358 The source of information to make the Makefile / descrip.mms is
1359 small files called 'build.info', holding the necessary
1360 information for each directory with source to compile, and a
1361 template in Configurations, like unix-Makefile.tmpl or
1364 With this change, the library names were also renamed on Windows
1365 and on VMS. They now have names that are closer to the standard
1366 on Unix, and include the major version number, and in certain
1367 cases, the architecture they are built for. See "Notes on shared
1368 libraries" in INSTALL.
1370 We rely heavily on the perl module Text::Template.
1373 *) Added support for auto-initialisation and de-initialisation of the library.
1374 OpenSSL no longer requires explicit init or deinit routines to be called,
1375 except in certain circumstances. See the OPENSSL_init_crypto() and
1376 OPENSSL_init_ssl() man pages for further information.
1379 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1380 "peer" argument is now expected to be a BIO_ADDR object.
1382 *) Rewrite of BIO networking library. The BIO library lacked consistent
1383 support of IPv6, and adding it required some more extensive
1384 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1385 which hold all types of addresses and chains of address information.
1386 It also introduces a new API, with functions like BIO_socket,
1387 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1388 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1389 have been adapted accordingly.
1392 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1396 *) CRIME protection: disable compression by default, even if OpenSSL is
1397 compiled with zlib enabled. Applications can still enable compression
1398 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1399 using the SSL_CONF library to configure compression.
1402 *) The signature of the session callback configured with
1403 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1404 was explicitly marked as 'const unsigned char*' instead of
1408 *) Always DPURIFY. Remove the use of uninitialized memory in the
1409 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1412 *) Removed many obsolete configuration items, including
1413 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1414 MD2_CHAR, MD2_INT, MD2_LONG
1416 IDEA_SHORT, IDEA_LONG
1417 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1418 [Rich Salz, with advice from Andy Polyakov]
1420 *) Many BN internals have been moved to an internal header file.
1421 [Rich Salz with help from Andy Polyakov]
1423 *) Configuration and writing out the results from it has changed.
1424 Files such as Makefile include/openssl/opensslconf.h and are now
1425 produced through general templates, such as Makefile.in and
1426 crypto/opensslconf.h.in and some help from the perl module
1429 Also, the center of configuration information is no longer
1430 Makefile. Instead, Configure produces a perl module in
1431 configdata.pm which holds most of the config data (in the hash
1432 table %config), the target data that comes from the target
1433 configuration in one of the Configurations/*.conf files (in
1437 *) To clarify their intended purposes, the Configure options
1438 --prefix and --openssldir change their semantics, and become more
1439 straightforward and less interdependent.
1441 --prefix shall be used exclusively to give the location INSTALLTOP
1442 where programs, scripts, libraries, include files and manuals are
1443 going to be installed. The default is now /usr/local.
1445 --openssldir shall be used exclusively to give the default
1446 location OPENSSLDIR where certificates, private keys, CRLs are
1447 managed. This is also where the default openssl.cnf gets
1449 If the directory given with this option is a relative path, the
1450 values of both the --prefix value and the --openssldir value will
1451 be combined to become OPENSSLDIR.
1452 The default for --openssldir is INSTALLTOP/ssl.
1454 Anyone who uses --openssldir to specify where OpenSSL is to be
1455 installed MUST change to use --prefix instead.
1458 *) The GOST engine was out of date and therefore it has been removed. An up
1459 to date GOST engine is now being maintained in an external repository.
1460 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1461 support for GOST ciphersuites (these are only activated if a GOST engine
1465 *) EGD is no longer supported by default; use enable-egd when
1467 [Ben Kaduk and Rich Salz]
1469 *) The distribution now has Makefile.in files, which are used to
1470 create Makefile's when Configure is run. *Configure must be run
1471 before trying to build now.*
1474 *) The return value for SSL_CIPHER_description() for error conditions
1478 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1480 Obtaining and performing DNSSEC validation of TLSA records is
1481 the application's responsibility. The application provides
1482 the TLSA records of its choice to OpenSSL, and these are then
1483 used to authenticate the peer.
1485 The TLSA records need not even come from DNS. They can, for
1486 example, be used to implement local end-entity certificate or
1487 trust-anchor "pinning", where the "pin" data takes the form
1488 of TLSA records, which can augment or replace verification
1489 based on the usual WebPKI public certification authorities.
1492 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1493 continues to support deprecated interfaces in default builds.
1494 However, applications are strongly advised to compile their
1495 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1496 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1497 or the 1.1.0 releases.
1499 In environments in which all applications have been ported to
1500 not use any deprecated interfaces OpenSSL's Configure script
1501 should be used with the --api=1.1.0 option to entirely remove
1502 support for the deprecated features from the library and
1503 unconditionally disable them in the installed headers.
1504 Essentially the same effect can be achieved with the "no-deprecated"
1505 argument to Configure, except that this will always restrict
1506 the build to just the latest API, rather than a fixed API
1509 As applications are ported to future revisions of the API,
1510 they should update their compile-time OPENSSL_API_COMPAT define
1511 accordingly, but in most cases should be able to continue to
1512 compile with later releases.
1514 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1515 0x10000000L and 0x00908000L, respectively. However those
1516 versions did not support the OPENSSL_API_COMPAT feature, and
1517 so applications are not typically tested for explicit support
1518 of just the undeprecated features of either release.
1521 *) Add support for setting the minimum and maximum supported protocol.
1522 It can bet set via the SSL_set_min_proto_version() and
1523 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1524 MaxProtocol. It's recommended to use the new APIs to disable
1525 protocols instead of disabling individual protocols using
1526 SSL_set_options() or SSL_CONF's Protocol. This change also
1527 removes support for disabling TLS 1.2 in the OpenSSL TLS
1528 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1531 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1534 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1535 and integrates ECDSA and ECDH functionality into EC. Implementations can
1536 now redirect key generation and no longer need to convert to or from
1539 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1540 include the ec.h header file instead.
1543 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1544 ciphers who are no longer supported and drops support the ephemeral RSA key
1545 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1548 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1549 opaque. For HMAC_CTX, the following constructors and destructors
1552 HMAC_CTX *HMAC_CTX_new(void);
1553 void HMAC_CTX_free(HMAC_CTX *ctx);
1555 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1556 destroy such methods has been added. See EVP_MD_meth_new(3) and
1557 EVP_CIPHER_meth_new(3) for documentation.
1560 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1561 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1562 EVP_MD_CTX_reset() should be called instead to reinitialise
1563 an already created structure.
1564 2) For consistency with the majority of our object creators and
1565 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1566 EVP_MD_CTX_(new|free). The old names are retained as macros
1567 for deprecated builds.
1570 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1571 cryptographic operations to be performed asynchronously as long as an
1572 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1573 further details. Libssl has also had this capability integrated with the
1574 introduction of the new mode SSL_MODE_ASYNC and associated error
1575 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1576 pages. This work was developed in partnership with Intel Corp.
1579 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1580 always enabled now. If you want to disable the support you should
1581 exclude it using the list of supported ciphers. This also means that the
1582 "-no_ecdhe" option has been removed from s_server.
1585 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1586 SSL_{CTX_}set1_curves() which can set a list.
1589 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1590 curve you want to support using SSL_{CTX_}set1_curves().
1593 *) State machine rewrite. The state machine code has been significantly
1594 refactored in order to remove much duplication of code and solve issues
1595 with the old code (see ssl/statem/README for further details). This change
1596 does have some associated API changes. Notably the SSL_state() function
1597 has been removed and replaced by SSL_get_state which now returns an
1598 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1599 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1603 *) All instances of the string "ssleay" in the public API were replaced
1604 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1605 Some error codes related to internal RSA_eay API's were renamed.
1608 *) The demo files in crypto/threads were moved to demo/threads.
1611 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1613 [Matt Caswell, Rich Salz]
1615 *) New ASN.1 embed macro.
1617 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1618 structure is not allocated: it is part of the parent. That is instead of
1626 This reduces memory fragmentation and make it impossible to accidentally
1627 set a mandatory field to NULL.
1629 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1630 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1631 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1635 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1638 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1639 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1640 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1641 DES and RC4 ciphersuites.
1644 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1645 This changes the decoding behaviour for some invalid messages,
1646 though the change is mostly in the more lenient direction, and
1647 legacy behaviour is preserved as much as possible.
1650 *) Fix no-stdio build.
1651 [ David Woodhouse <David.Woodhouse@intel.com> and also
1652 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1654 *) New testing framework
1655 The testing framework has been largely rewritten and is now using
1656 perl and the perl modules Test::Harness and an extended variant of
1657 Test::More called OpenSSL::Test to do its work. All test scripts in
1658 test/ have been rewritten into test recipes, and all direct calls to
1659 executables in test/Makefile have become individual recipes using the
1660 simplified testing OpenSSL::Test::Simple.
1662 For documentation on our testing modules, do:
1664 perldoc test/testlib/OpenSSL/Test/Simple.pm
1665 perldoc test/testlib/OpenSSL/Test.pm
1669 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1670 are used; the latter aborts on memory leaks (usually checked on exit).
1671 Some undocumented "set malloc, etc., hooks" functions were removed
1672 and others were changed. All are now documented.
1675 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1677 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1679 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1680 from RFC4279, RFC4785, RFC5487, RFC5489.
1682 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1683 original RSA_PSK patch.
1686 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1687 era flag was never set throughout the codebase (only read). Also removed
1688 SSL3_FLAGS_POP_BUFFER which was only used if
1689 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1692 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1693 to be "oneline" instead of "compat".
1696 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1697 not aware of clients that still exhibit this bug, and the workaround
1698 hasn't been working properly for a while.
1701 *) The return type of BIO_number_read() and BIO_number_written() as well as
1702 the corresponding num_read and num_write members in the BIO structure has
1703 changed from unsigned long to uint64_t. On platforms where an unsigned
1704 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1708 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1709 OpenSSL without support for them. It also means that maintaining
1710 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1711 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1714 *) Removed support for the two export grade static DH ciphersuites
1715 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1716 were newly added (along with a number of other static DH ciphersuites) to
1717 1.0.2. However the two export ones have *never* worked since they were
1718 introduced. It seems strange in any case to be adding new export
1719 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1722 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1723 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1724 and turned into macros which simply call the new preferred function names
1725 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1726 should use the new names instead. Also as part of this change the ssl23.h
1727 header file has been removed.
1730 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1731 code and the associated standard is no longer considered fit-for-purpose.
1734 *) RT2547 was closed. When generating a private key, try to make the
1735 output file readable only by the owner. This behavior change might
1736 be noticeable when interacting with other software.
1738 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1742 *) Added HTTP GET support to the ocsp command.
1745 *) Changed default digest for the dgst and enc commands from MD5 to
1749 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1752 *) Added support for TLS extended master secret from
1753 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1754 initial patch which was a great help during development.
1757 *) All libssl internal structures have been removed from the public header
1758 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1759 now redundant). Users should not attempt to access internal structures
1760 directly. Instead they should use the provided API functions.
1763 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1764 Access to deprecated functions can be re-enabled by running config with
1765 "enable-deprecated". In addition applications wishing to use deprecated
1766 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1767 will, by default, disable some transitive includes that previously existed
1768 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1771 *) Added support for OCB mode. OpenSSL has been granted a patent license
1772 compatible with the OpenSSL license for use of OCB. Details are available
1773 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1774 for OCB can be removed by calling config with no-ocb.
1777 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1778 compatible client hello.
1781 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1782 done while fixing the error code for the key-too-small case.
1783 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1785 *) CA.sh has been removed; use CA.pl instead.
1788 *) Removed old DES API.
1791 *) Remove various unsupported platforms:
1797 Sinix/ReliantUNIX RM400
1802 16-bit platforms such as WIN16
1805 *) Clean up OPENSSL_NO_xxx #define's
1806 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1807 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1808 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1809 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1810 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1811 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1812 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1813 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1814 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1815 Remove MS_STATIC; it's a relic from platforms <32 bits.
1818 *) Cleaned up dead code
1819 Remove all but one '#ifdef undef' which is to be looked at.
1822 *) Clean up calling of xxx_free routines.
1823 Just like free(), fix most of the xxx_free routines to accept
1824 NULL. Remove the non-null checks from callers. Save much code.
1827 *) Add secure heap for storage of private keys (when possible).
1828 Add BIO_s_secmem(), CBIGNUM, etc.
1829 Contributed by Akamai Technologies under our Corporate CLA.
1832 *) Experimental support for a new, fast, unbiased prime candidate generator,
1833 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1834 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1836 *) New output format NSS in the sess_id command line tool. This allows
1837 exporting the session id and the master key in NSS keylog format.
1838 [Martin Kaiser <martin@kaiser.cx>]
1840 *) Harmonize version and its documentation. -f flag is used to display
1842 [mancha <mancha1@zoho.com>]
1844 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1845 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1846 [mancha <mancha1@zoho.com>]
1848 *) Fix some double frees. These are not thought to be exploitable.
1849 [mancha <mancha1@zoho.com>]
1851 *) A missing bounds check in the handling of the TLS heartbeat extension
1852 can be used to reveal up to 64k of memory to a connected client or
1855 Thanks for Neel Mehta of Google Security for discovering this bug and to
1856 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1857 preparing the fix (CVE-2014-0160)
1858 [Adam Langley, Bodo Moeller]
1860 *) Fix for the attack described in the paper "Recovering OpenSSL
1861 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1862 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1863 http://eprint.iacr.org/2014/140
1865 Thanks to Yuval Yarom and Naomi Benger for discovering this
1866 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1867 [Yuval Yarom and Naomi Benger]
1869 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1870 this fixes a limitation in previous versions of OpenSSL.
1873 *) Experimental encrypt-then-mac support.
1875 Experimental support for encrypt then mac from
1876 draft-gutmann-tls-encrypt-then-mac-02.txt
1878 To enable it set the appropriate extension number (0x42 for the test
1879 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1881 For non-compliant peers (i.e. just about everything) this should have no
1884 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1888 *) Add EVP support for key wrapping algorithms, to avoid problems with
1889 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1890 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1891 algorithms and include tests cases.
1894 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1898 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1899 MGF1 digest and OAEP label.
1902 *) Make openssl verify return errors.
1903 [Chris Palmer <palmer@google.com> and Ben Laurie]
1905 *) New function ASN1_TIME_diff to calculate the difference between two
1906 ASN1_TIME structures or one structure and the current time.
1909 *) Update fips_test_suite to support multiple command line options. New
1910 test to induce all self test errors in sequence and check expected
1914 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1915 sign or verify all in one operation.
1918 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1919 test programs and fips_test_suite. Includes functionality to parse
1920 the minimal script output of fipsalgest.pl directly.
1923 *) Add authorisation parameter to FIPS_module_mode_set().
1926 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1929 *) Use separate DRBG fields for internal and external flags. New function
1930 FIPS_drbg_health_check() to perform on demand health checking. Add
1931 generation tests to fips_test_suite with reduced health check interval to
1932 demonstrate periodic health checking. Add "nodh" option to
1933 fips_test_suite to skip very slow DH test.
1936 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1940 *) More extensive health check for DRBG checking many more failure modes.
1941 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1942 combination: call this in fips_test_suite.
1945 *) Add support for canonical generation of DSA parameter 'g'. See
1948 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1949 POST to handle HMAC cases.
1952 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1953 to return numerical and string versions of the FIPS module number.
1956 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1957 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1958 outside the validated module in the FIPS capable OpenSSL.
1961 *) Minor change to DRBG entropy callback semantics. In some cases
1962 there is no multiple of the block length between min_len and
1963 max_len. Allow the callback to return more than max_len bytes
1964 of entropy but discard any extra: it is the callback's responsibility
1965 to ensure that the extra data discarded does not impact the
1966 requested amount of entropy.
1969 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1970 information in FIPS186-3, SP800-57 and SP800-131A.
1973 *) CCM support via EVP. Interface is very similar to GCM case except we
1974 must supply all data in one chunk (i.e. no update, final) and the
1975 message length must be supplied if AAD is used. Add algorithm test
1979 *) Initial version of POST overhaul. Add POST callback to allow the status
1980 of POST to be monitored and/or failures induced. Modify fips_test_suite
1981 to use callback. Always run all selftests even if one fails.
1984 *) XTS support including algorithm test driver in the fips_gcmtest program.
1985 Note: this does increase the maximum key length from 32 to 64 bytes but
1986 there should be no binary compatibility issues as existing applications
1987 will never use XTS mode.
1990 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1991 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1992 performs algorithm blocking for unapproved PRNG types. Also do not
1993 set PRNG type in FIPS_mode_set(): leave this to the application.
1994 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1995 the standard OpenSSL PRNG: set additional data to a date time vector.
1998 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1999 This shouldn't present any incompatibility problems because applications
2000 shouldn't be using these directly and any that are will need to rethink
2001 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2004 *) Extensive self tests and health checking required by SP800-90 DRBG.
2005 Remove strength parameter from FIPS_drbg_instantiate and always
2006 instantiate at maximum supported strength.
2009 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2012 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2015 *) New function DH_compute_key_padded() to compute a DH key and pad with
2016 leading zeroes if needed: this complies with SP800-56A et al.
2019 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2020 anything, incomplete, subject to change and largely untested at present.
2023 *) Modify fipscanisteronly build option to only build the necessary object
2024 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2027 *) Add experimental option FIPSSYMS to give all symbols in
2028 fipscanister.o and FIPS or fips prefix. This will avoid
2029 conflicts with future versions of OpenSSL. Add perl script
2030 util/fipsas.pl to preprocess assembly language source files
2031 and rename any affected symbols.
2034 *) Add selftest checks and algorithm block of non-fips algorithms in
2035 FIPS mode. Remove DES2 from selftests.
2038 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2039 return internal method without any ENGINE dependencies. Add new
2040 tiny fips sign and verify functions.
2043 *) New build option no-ec2m to disable characteristic 2 code.
2046 *) New build option "fipscanisteronly". This only builds fipscanister.o
2047 and (currently) associated fips utilities. Uses the file Makefile.fips
2048 instead of Makefile.org as the prototype.
2051 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2052 Update fips_gcmtest to use IV generator.
2055 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2056 setting output buffer to NULL. The *Final function must be
2057 called although it will not retrieve any additional data. The tag
2058 can be set or retrieved with a ctrl. The IV length is by default 12
2059 bytes (96 bits) but can be set to an alternative value. If the IV
2060 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2064 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2065 underlying do_cipher function handles all cipher semantics itself
2066 including padding and finalisation. This is useful if (for example)
2067 an ENGINE cipher handles block padding itself. The behaviour of
2068 do_cipher is subtly changed if this flag is set: the return value
2069 is the number of characters written to the output buffer (zero is
2070 no longer an error code) or a negative error code. Also if the
2071 input buffer is NULL and length 0 finalisation should be performed.
2074 *) If a candidate issuer certificate is already part of the constructed
2075 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2078 *) Improve forward-security support: add functions
2080 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2081 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2083 for use by SSL/TLS servers; the callback function will be called whenever a
2084 new session is created, and gets to decide whether the session may be
2085 cached to make it resumable (return 0) or not (return 1). (As by the
2086 SSL/TLS protocol specifications, the session_id sent by the server will be
2087 empty to indicate that the session is not resumable; also, the server will
2088 not generate RFC 4507 (RFC 5077) session tickets.)
2090 A simple reasonable callback implementation is to return is_forward_secure.
2091 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2092 by the SSL/TLS server library, indicating whether it can provide forward
2094 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2096 *) New -verify_name option in command line utilities to set verification
2100 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2101 Add CMAC pkey methods.
2104 *) Experimental renegotiation in s_server -www mode. If the client
2105 browses /reneg connection is renegotiated. If /renegcert it is
2106 renegotiated requesting a certificate.
2109 *) Add an "external" session cache for debugging purposes to s_server. This
2110 should help trace issues which normally are only apparent in deployed
2111 multi-process servers.
2114 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2115 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2116 BIO_set_cipher() and some obscure PEM functions were changed so they
2117 can now return an error. The RAND changes required a change to the
2118 RAND_METHOD structure.
2121 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2122 a gcc attribute to warn if the result of a function is ignored. This
2123 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2124 whose return value is often ignored.
2127 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2128 These allow SCTs (signed certificate timestamps) to be requested and
2129 validated when establishing a connection.
2130 [Rob Percival <robpercival@google.com>]
2132 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2134 *) Prevent padding oracle in AES-NI CBC MAC check
2136 A MITM attacker can use a padding oracle attack to decrypt traffic
2137 when the connection uses an AES CBC cipher and the server support
2140 This issue was introduced as part of the fix for Lucky 13 padding
2141 attack (CVE-2013-0169). The padding check was rewritten to be in
2142 constant time by making sure that always the same bytes are read and
2143 compared against either the MAC or padding bytes. But it no longer
2144 checked that there was enough data to have both the MAC and padding
2147 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2151 *) Fix EVP_EncodeUpdate overflow
2153 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2154 Base64 encoding of binary data. If an attacker is able to supply very large
2155 amounts of input data then a length check can overflow resulting in a heap
2158 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2159 the PEM_write_bio* family of functions. These are mainly used within the
2160 OpenSSL command line applications, so any application which processes data
2161 from an untrusted source and outputs it as a PEM file should be considered
2162 vulnerable to this issue. User applications that call these APIs directly
2163 with large amounts of untrusted data may also be vulnerable.
2165 This issue was reported by Guido Vranken.
2169 *) Fix EVP_EncryptUpdate overflow
2171 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2172 is able to supply very large amounts of input data after a previous call to
2173 EVP_EncryptUpdate() with a partial block then a length check can overflow
2174 resulting in a heap corruption. Following an analysis of all OpenSSL
2175 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2176 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2177 the first called function after an EVP_EncryptInit(), and therefore that
2178 specific call must be safe. The second form is where the length passed to
2179 EVP_EncryptUpdate() can be seen from the code to be some small value and
2180 therefore there is no possibility of an overflow. Since all instances are
2181 one of these two forms, it is believed that there can be no overflows in
2182 internal code due to this problem. It should be noted that
2183 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2184 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2185 of these calls have also been analysed too and it is believed there are no
2186 instances in internal usage where an overflow could occur.
2188 This issue was reported by Guido Vranken.
2192 *) Prevent ASN.1 BIO excessive memory allocation
2194 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2195 a short invalid encoding can cause allocation of large amounts of memory
2196 potentially consuming excessive resources or exhausting memory.
2198 Any application parsing untrusted data through d2i BIO functions is
2199 affected. The memory based functions such as d2i_X509() are *not* affected.
2200 Since the memory based functions are used by the TLS library, TLS
2201 applications are not affected.
2203 This issue was reported by Brian Carpenter.
2209 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2210 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2211 in arbitrary stack data being returned in the buffer.
2213 This issue was reported by Guido Vranken.
2217 *) Modify behavior of ALPN to invoke callback after SNI/servername
2218 callback, such that updates to the SSL_CTX affect ALPN.
2221 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2225 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2226 methods are enabled and ssl2 is disabled the methods return NULL.
2229 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2231 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2232 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2233 provide any "EXPORT" or "LOW" strength ciphers.
2236 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2237 is by default disabled at build-time. Builds that are not configured with
2238 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2239 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2240 will need to explicitly call either of:
2242 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2244 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2246 as appropriate. Even if either of those is used, or the application
2247 explicitly uses the version-specific SSLv2_method() or its client and
2248 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2249 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2250 ciphers, and SSLv2 56-bit DES are no longer available.
2254 *) Fix a double-free in DSA code
2256 A double free bug was discovered when OpenSSL parses malformed DSA private
2257 keys and could lead to a DoS attack or memory corruption for applications
2258 that receive DSA private keys from untrusted sources. This scenario is
2261 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2266 *) Disable SRP fake user seed to address a server memory leak.
2268 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2270 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2271 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2272 was changed to ignore the "fake user" SRP seed, even if the seed
2275 Users should use SRP_VBASE_get1_by_user instead. Note that in
2276 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2277 also that even though configuring the SRP seed attempts to hide
2278 invalid usernames by continuing the handshake with fake
2279 credentials, this behaviour is not constant time and no strong
2280 guarantees are made that the handshake is indistinguishable from
2281 that of a valid user.
2285 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2287 In the BN_hex2bn function the number of hex digits is calculated using an
2288 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2289 large values of |i| this can result in |bn_expand| not allocating any
2290 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2291 field as NULL leading to a subsequent NULL ptr deref. For very large values
2292 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2293 In this case memory is allocated to the internal BIGNUM data field, but it
2294 is insufficiently sized leading to heap corruption. A similar issue exists
2295 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2296 is ever called by user applications with very large untrusted hex/dec data.
2297 This is anticipated to be a rare occurrence.
2299 All OpenSSL internal usage of these functions use data that is not expected
2300 to be untrusted, e.g. config file data or application command line
2301 arguments. If user developed applications generate config file data based
2302 on untrusted data then it is possible that this could also lead to security
2303 consequences. This is also anticipated to be rare.
2305 This issue was reported to OpenSSL by Guido Vranken.
2309 *) Fix memory issues in BIO_*printf functions
2311 The internal |fmtstr| function used in processing a "%s" format string in
2312 the BIO_*printf functions could overflow while calculating the length of a
2313 string and cause an OOB read when printing very long strings.
2315 Additionally the internal |doapr_outch| function can attempt to write to an
2316 OOB memory location (at an offset from the NULL pointer) in the event of a
2317 memory allocation failure. In 1.0.2 and below this could be caused where
2318 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2319 could be in processing a very long "%s" format string. Memory leaks can
2322 The first issue may mask the second issue dependent on compiler behaviour.
2323 These problems could enable attacks where large amounts of untrusted data
2324 is passed to the BIO_*printf functions. If applications use these functions
2325 in this way then they could be vulnerable. OpenSSL itself uses these
2326 functions when printing out human-readable dumps of ASN.1 data. Therefore
2327 applications that print this data could be vulnerable if the data is from
2328 untrusted sources. OpenSSL command line applications could also be
2329 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2330 as command line arguments.
2332 Libssl is not considered directly vulnerable. Additionally certificates etc
2333 received via remote connections via libssl are also unlikely to be able to
2334 trigger these issues because of message size limits enforced within libssl.
2336 This issue was reported to OpenSSL Guido Vranken.
2340 *) Side channel attack on modular exponentiation
2342 A side-channel attack was found which makes use of cache-bank conflicts on
2343 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2344 of RSA keys. The ability to exploit this issue is limited as it relies on
2345 an attacker who has control of code in a thread running on the same
2346 hyper-threaded core as the victim thread which is performing decryptions.
2348 This issue was reported to OpenSSL by Yuval Yarom, The University of
2349 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2350 Nadia Heninger, University of Pennsylvania with more information at
2351 http://cachebleed.info.
2355 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2356 if no keysize is specified with default_bits. This fixes an
2357 omission in an earlier change that changed all RSA/DSA key generation
2358 apps to use 2048 bits by default.
2361 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2362 *) DH small subgroups
2364 Historically OpenSSL only ever generated DH parameters based on "safe"
2365 primes. More recently (in version 1.0.2) support was provided for
2366 generating X9.42 style parameter files such as those required for RFC 5114
2367 support. The primes used in such files may not be "safe". Where an
2368 application is using DH configured with parameters based on primes that are
2369 not "safe" then an attacker could use this fact to find a peer's private
2370 DH exponent. This attack requires that the attacker complete multiple
2371 handshakes in which the peer uses the same private DH exponent. For example
2372 this could be used to discover a TLS server's private DH exponent if it's
2373 reusing the private DH exponent or it's using a static DH ciphersuite.
2375 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2376 TLS. It is not on by default. If the option is not set then the server
2377 reuses the same private DH exponent for the life of the server process and
2378 would be vulnerable to this attack. It is believed that many popular
2379 applications do set this option and would therefore not be at risk.
2381 The fix for this issue adds an additional check where a "q" parameter is
2382 available (as is the case in X9.42 based parameters). This detects the
2383 only known attack, and is the only possible defense for static DH
2384 ciphersuites. This could have some performance impact.
2386 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2387 default and cannot be disabled. This could have some performance impact.
2389 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2393 *) SSLv2 doesn't block disabled ciphers
2395 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2396 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2397 been disabled, provided that the SSLv2 protocol was not also disabled via
2400 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2401 and Sebastian Schinzel.
2405 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2407 *) BN_mod_exp may produce incorrect results on x86_64
2409 There is a carry propagating bug in the x86_64 Montgomery squaring
2410 procedure. No EC algorithms are affected. Analysis suggests that attacks
2411 against RSA and DSA as a result of this defect would be very difficult to
2412 perform and are not believed likely. Attacks against DH are considered just
2413 feasible (although very difficult) because most of the work necessary to
2414 deduce information about a private key may be performed offline. The amount
2415 of resources required for such an attack would be very significant and
2416 likely only accessible to a limited number of attackers. An attacker would
2417 additionally need online access to an unpatched system using the target
2418 private key in a scenario with persistent DH parameters and a private
2419 key that is shared between multiple clients. For example this can occur by
2420 default in OpenSSL DHE based SSL/TLS ciphersuites.
2422 This issue was reported to OpenSSL by Hanno Böck.
2426 *) Certificate verify crash with missing PSS parameter
2428 The signature verification routines will crash with a NULL pointer
2429 dereference if presented with an ASN.1 signature using the RSA PSS
2430 algorithm and absent mask generation function parameter. Since these
2431 routines are used to verify certificate signature algorithms this can be
2432 used to crash any certificate verification operation and exploited in a
2433 DoS attack. Any application which performs certificate verification is
2434 vulnerable including OpenSSL clients and servers which enable client
2437 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2441 *) X509_ATTRIBUTE memory leak
2443 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2444 memory. This structure is used by the PKCS#7 and CMS routines so any
2445 application which reads PKCS#7 or CMS data from untrusted sources is
2446 affected. SSL/TLS is not affected.
2448 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2453 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2454 This changes the decoding behaviour for some invalid messages,
2455 though the change is mostly in the more lenient direction, and
2456 legacy behaviour is preserved as much as possible.
2459 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2461 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2463 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2465 *) Alternate chains certificate forgery
2467 During certificate verification, OpenSSL will attempt to find an
2468 alternative certificate chain if the first attempt to build such a chain
2469 fails. An error in the implementation of this logic can mean that an
2470 attacker could cause certain checks on untrusted certificates to be
2471 bypassed, such as the CA flag, enabling them to use a valid leaf
2472 certificate to act as a CA and "issue" an invalid certificate.
2474 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2478 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2480 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2481 incompatibility in the handling of HMAC. The previous ABI has now been
2485 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2487 *) Malformed ECParameters causes infinite loop
2489 When processing an ECParameters structure OpenSSL enters an infinite loop
2490 if the curve specified is over a specially malformed binary polynomial
2493 This can be used to perform denial of service against any
2494 system which processes public keys, certificate requests or
2495 certificates. This includes TLS clients and TLS servers with
2496 client authentication enabled.
2498 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2502 *) Exploitable out-of-bounds read in X509_cmp_time
2504 X509_cmp_time does not properly check the length of the ASN1_TIME
2505 string and can read a few bytes out of bounds. In addition,
2506 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2509 An attacker can use this to craft malformed certificates and CRLs of
2510 various sizes and potentially cause a segmentation fault, resulting in
2511 a DoS on applications that verify certificates or CRLs. TLS clients
2512 that verify CRLs are affected. TLS clients and servers with client
2513 authentication enabled may be affected if they use custom verification
2516 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2517 independently by Hanno Böck.
2521 *) PKCS7 crash with missing EnvelopedContent
2523 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2524 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2525 with missing content and trigger a NULL pointer dereference on parsing.
2527 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2528 structures from untrusted sources are affected. OpenSSL clients and
2529 servers are not affected.
2531 This issue was reported to OpenSSL by Michal Zalewski (Google).
2535 *) CMS verify infinite loop with unknown hash function
2537 When verifying a signedData message the CMS code can enter an infinite loop
2538 if presented with an unknown hash function OID. This can be used to perform
2539 denial of service against any system which verifies signedData messages using
2541 This issue was reported to OpenSSL by Johannes Bauer.
2545 *) Race condition handling NewSessionTicket
2547 If a NewSessionTicket is received by a multi-threaded client when attempting to
2548 reuse a previous ticket then a race condition can occur potentially leading to
2549 a double free of the ticket data.
2553 *) Only support 256-bit or stronger elliptic curves with the
2554 'ecdh_auto' setting (server) or by default (client). Of supported
2555 curves, prefer P-256 (both).
2558 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2560 *) ClientHello sigalgs DoS fix
2562 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2563 invalid signature algorithms extension a NULL pointer dereference will
2564 occur. This can be exploited in a DoS attack against the server.
2566 This issue was was reported to OpenSSL by David Ramos of Stanford
2569 [Stephen Henson and Matt Caswell]
2571 *) Multiblock corrupted pointer fix
2573 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2574 feature only applies on 64 bit x86 architecture platforms that support AES
2575 NI instructions. A defect in the implementation of "multiblock" can cause
2576 OpenSSL's internal write buffer to become incorrectly set to NULL when
2577 using non-blocking IO. Typically, when the user application is using a
2578 socket BIO for writing, this will only result in a failed connection.
2579 However if some other BIO is used then it is likely that a segmentation
2580 fault will be triggered, thus enabling a potential DoS attack.
2582 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2586 *) Segmentation fault in DTLSv1_listen fix
2588 The DTLSv1_listen function is intended to be stateless and processes the
2589 initial ClientHello from many peers. It is common for user code to loop
2590 over the call to DTLSv1_listen until a valid ClientHello is received with
2591 an associated cookie. A defect in the implementation of DTLSv1_listen means
2592 that state is preserved in the SSL object from one invocation to the next
2593 that can lead to a segmentation fault. Errors processing the initial
2594 ClientHello can trigger this scenario. An example of such an error could be
2595 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2598 This issue was reported to OpenSSL by Per Allansson.
2602 *) Segmentation fault in ASN1_TYPE_cmp fix
2604 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2605 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2606 certificate signature algorithm consistency this can be used to crash any
2607 certificate verification operation and exploited in a DoS attack. Any
2608 application which performs certificate verification is vulnerable including
2609 OpenSSL clients and servers which enable client authentication.
2613 *) Segmentation fault for invalid PSS parameters fix
2615 The signature verification routines will crash with a NULL pointer
2616 dereference if presented with an ASN.1 signature using the RSA PSS
2617 algorithm and invalid parameters. Since these routines are used to verify
2618 certificate signature algorithms this can be used to crash any
2619 certificate verification operation and exploited in a DoS attack. Any
2620 application which performs certificate verification is vulnerable including
2621 OpenSSL clients and servers which enable client authentication.
2623 This issue was was reported to OpenSSL by Brian Carpenter.
2627 *) ASN.1 structure reuse memory corruption fix
2629 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2630 memory corruption via an invalid write. Such reuse is and has been
2631 strongly discouraged and is believed to be rare.
2633 Applications that parse structures containing CHOICE or ANY DEFINED BY
2634 components may be affected. Certificate parsing (d2i_X509 and related
2635 functions) are however not affected. OpenSSL clients and servers are
2640 *) PKCS7 NULL pointer dereferences fix
2642 The PKCS#7 parsing code does not handle missing outer ContentInfo
2643 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2644 missing content and trigger a NULL pointer dereference on parsing.
2646 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2647 otherwise parse PKCS#7 structures from untrusted sources are
2648 affected. OpenSSL clients and servers are not affected.
2650 This issue was reported to OpenSSL by Michal Zalewski (Google).
2654 *) DoS via reachable assert in SSLv2 servers fix
2656 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2657 servers that both support SSLv2 and enable export cipher suites by sending
2658 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2660 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2661 (OpenSSL development team).
2665 *) Empty CKE with client auth and DHE fix
2667 If client auth is used then a server can seg fault in the event of a DHE
2668 ciphersuite being selected and a zero length ClientKeyExchange message
2669 being sent by the client. This could be exploited in a DoS attack.
2673 *) Handshake with unseeded PRNG fix
2675 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2676 with an unseeded PRNG. The conditions are:
2677 - The client is on a platform where the PRNG has not been seeded
2678 automatically, and the user has not seeded manually
2679 - A protocol specific client method version has been used (i.e. not
2680 SSL_client_methodv23)
2681 - A ciphersuite is used that does not require additional random data from
2682 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2684 If the handshake succeeds then the client random that has been used will
2685 have been generated from a PRNG with insufficient entropy and therefore the
2686 output may be predictable.
2688 For example using the following command with an unseeded openssl will
2689 succeed on an unpatched platform:
2691 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2695 *) Use After Free following d2i_ECPrivatekey error fix
2697 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2698 could cause a use after free condition. This, in turn, could cause a double
2699 free in several private key parsing functions (such as d2i_PrivateKey
2700 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2701 for applications that receive EC private keys from untrusted
2702 sources. This scenario is considered rare.
2704 This issue was discovered by the BoringSSL project and fixed in their
2709 *) X509_to_X509_REQ NULL pointer deref fix
2711 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2712 the certificate key is invalid. This function is rarely used in practice.
2714 This issue was discovered by Brian Carpenter.
2718 *) Removed the export ciphers from the DEFAULT ciphers
2721 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2723 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2724 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2725 So far those who have to target multiple platforms would compromise
2726 and argue that binary targeting say ARMv5 would still execute on
2727 ARMv8. "Universal" build resolves this compromise by providing
2728 near-optimal performance even on newer platforms.
2731 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2732 (other platforms pending).
2733 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2735 *) Add support for the SignedCertificateTimestampList certificate and
2736 OCSP response extensions from RFC6962.
2739 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2740 for corner cases. (Certain input points at infinity could lead to
2741 bogus results, with non-infinity inputs mapped to infinity too.)
2744 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2745 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2746 common cases are optimized and there still is room for further
2747 improvements. Vector Permutation AES for Altivec is also added.
2750 *) Add support for little-endian ppc64 Linux target.
2751 [Marcelo Cerri (IBM)]
2753 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2754 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2755 are optimized and there still is room for further improvements.
2756 Both 32- and 64-bit modes are supported.
2757 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2759 *) Improved ARMv7 NEON support.
2762 *) Support for SPARC Architecture 2011 crypto extensions, first
2763 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2764 SHA256/512, MD5, GHASH and modular exponentiation.
2765 [Andy Polyakov, David Miller]
2767 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2769 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2771 *) Support for new and upcoming Intel processors, including AVX2,
2772 BMI and SHA ISA extensions. This includes additional "stitched"
2773 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2776 This work was sponsored by Intel Corp.
2779 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2780 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2781 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2784 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2785 this fixes a limitation in previous versions of OpenSSL.
2788 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2789 MGF1 digest and OAEP label.
2792 *) Add EVP support for key wrapping algorithms, to avoid problems with
2793 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2794 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2795 algorithms and include tests cases.
2798 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2800 [Douglas E. Engert, Steve Henson]
2802 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2803 difference in days and seconds between two tm or ASN1_TIME structures.
2806 *) Add -rev test option to s_server to just reverse order of characters
2807 received by client and send back to server. Also prints an abbreviated
2808 summary of the connection parameters.
2811 *) New option -brief for s_client and s_server to print out a brief summary
2812 of connection parameters.
2815 *) Add callbacks for arbitrary TLS extensions.
2816 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2818 *) New option -crl_download in several openssl utilities to download CRLs
2819 from CRLDP extension in certificates.
2822 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2825 *) New function X509_CRL_diff to generate a delta CRL from the difference
2826 of two full CRLs. Add support to "crl" utility.
2829 *) New functions to set lookup_crls function and to retrieve
2830 X509_STORE from X509_STORE_CTX.
2833 *) Print out deprecated issuer and subject unique ID fields in
2837 *) Extend OCSP I/O functions so they can be used for simple general purpose
2838 HTTP as well as OCSP. New wrapper function which can be used to download
2839 CRLs using the OCSP API.
2842 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2845 *) SSL_CONF* functions. These provide a common framework for application
2846 configuration using configuration files or command lines.
2849 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2850 message callback and prints the results. Needs compile time option
2851 "enable-ssl-trace". New options to s_client and s_server to enable
2855 *) New ctrl and macro to retrieve supported points extensions.
2856 Print out extension in s_server and s_client.
2859 *) New functions to retrieve certificate signature and signature
2863 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2867 *) New Suite B modes for TLS code. These use and enforce the requirements
2868 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2869 only use Suite B curves. The Suite B modes can be set by using the
2870 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2873 *) New chain verification flags for Suite B levels of security. Check
2874 algorithms are acceptable when flags are set in X509_verify_cert.
2877 *) Make tls1_check_chain return a set of flags indicating checks passed
2878 by a certificate chain. Add additional tests to handle client
2879 certificates: checks for matching certificate type and issuer name
2883 *) If an attempt is made to use a signature algorithm not in the peer
2884 preference list abort the handshake. If client has no suitable
2885 signature algorithms in response to a certificate request do not
2886 use the certificate.
2889 *) If server EC tmp key is not in client preference list abort handshake.
2892 *) Add support for certificate stores in CERT structure. This makes it
2893 possible to have different stores per SSL structure or one store in
2894 the parent SSL_CTX. Include distinct stores for certificate chain
2895 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2896 to build and store a certificate chain in CERT structure: returning
2897 an error if the chain cannot be built: this will allow applications
2898 to test if a chain is correctly configured.
2900 Note: if the CERT based stores are not set then the parent SSL_CTX
2901 store is used to retain compatibility with existing behaviour.
2905 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2906 mask based on the current session, check mask when sending client
2907 hello and checking the requested ciphersuite.
2910 *) New ctrls to retrieve and set certificate types in a certificate
2911 request message. Print out received values in s_client. If certificate
2912 types is not set with custom values set sensible values based on
2913 supported signature algorithms.
2916 *) Support for distinct client and server supported signature algorithms.
2919 *) Add certificate callback. If set this is called whenever a certificate
2920 is required by client or server. An application can decide which
2921 certificate chain to present based on arbitrary criteria: for example
2922 supported signature algorithms. Add very simple example to s_server.
2923 This fixes many of the problems and restrictions of the existing client
2924 certificate callback: for example you can now clear an existing
2925 certificate and specify the whole chain.
2928 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2929 the certificate can be used for (if anything). Set valid_flags field
2930 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2931 to have similar checks in it.
2933 Add new "cert_flags" field to CERT structure and include a "strict mode".
2934 This enforces some TLS certificate requirements (such as only permitting
2935 certificate signature algorithms contained in the supported algorithms
2936 extension) which some implementations ignore: this option should be used
2937 with caution as it could cause interoperability issues.
2940 *) Update and tidy signature algorithm extension processing. Work out
2941 shared signature algorithms based on preferences and peer algorithms
2942 and print them out in s_client and s_server. Abort handshake if no
2943 shared signature algorithms.
2946 *) Add new functions to allow customised supported signature algorithms
2947 for SSL and SSL_CTX structures. Add options to s_client and s_server
2951 *) New function SSL_certs_clear() to delete all references to certificates
2952 from an SSL structure. Before this once a certificate had been added
2953 it couldn't be removed.
2956 *) Integrate hostname, email address and IP address checking with certificate
2957 verification. New verify options supporting checking in openssl utility.
2960 *) Fixes and wildcard matching support to hostname and email checking
2961 functions. Add manual page.
2962 [Florian Weimer (Red Hat Product Security Team)]
2964 *) New functions to check a hostname email or IP address against a
2965 certificate. Add options x509 utility to print results of checks against
2969 *) Fix OCSP checking.
2970 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2972 *) Initial experimental support for explicitly trusted non-root CAs.
2973 OpenSSL still tries to build a complete chain to a root but if an
2974 intermediate CA has a trust setting included that is used. The first
2975 setting is used: whether to trust (e.g., -addtrust option to the x509
2979 *) Add -trusted_first option which attempts to find certificates in the
2980 trusted store even if an untrusted chain is also supplied.
2983 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2984 platform support for Linux and Android.
2987 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2990 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2991 When in FIPS mode the approved implementations are used as normal,
2992 when not in FIPS mode the internal unapproved versions are used instead.
2993 This means that the FIPS capable OpenSSL isn't forced to use the
2994 (often lower performance) FIPS implementations outside FIPS mode.
2997 *) Transparently support X9.42 DH parameters when calling
2998 PEM_read_bio_DHparameters. This means existing applications can handle
2999 the new parameter format automatically.
3002 *) Initial experimental support for X9.42 DH parameter format: mainly
3003 to support use of 'q' parameter for RFC5114 parameters.
3006 *) Add DH parameters from RFC5114 including test data to dhtest.
3009 *) Support for automatic EC temporary key parameter selection. If enabled
3010 the most preferred EC parameters are automatically used instead of
3011 hardcoded fixed parameters. Now a server just has to call:
3012 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3013 support ECDH and use the most appropriate parameters.
3016 *) Enhance and tidy EC curve and point format TLS extension code. Use
3017 static structures instead of allocation if default values are used.
3018 New ctrls to set curves we wish to support and to retrieve shared curves.
3019 Print out shared curves in s_server. New options to s_server and s_client
3020 to set list of supported curves.
3023 *) New ctrls to retrieve supported signature algorithms and
3024 supported curve values as an array of NIDs. Extend openssl utility
3025 to print out received values.
3028 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3029 between NIDs and the more common NIST names such as "P-256". Enhance
3030 ecparam utility and ECC method to recognise the NIST names for curves.
3033 *) Enhance SSL/TLS certificate chain handling to support different
3034 chains for each certificate instead of one chain in the parent SSL_CTX.
3037 *) Support for fixed DH ciphersuite client authentication: where both
3038 server and client use DH certificates with common parameters.
3041 *) Support for fixed DH ciphersuites: those requiring DH server
3045 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3047 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3048 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3049 X509_CINF_get_signature were reverted post internal team review.
3051 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3053 *) Build fixes for the Windows and OpenVMS platforms
3054 [Matt Caswell and Richard Levitte]
3056 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3058 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3059 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3060 dereference. This could lead to a Denial Of Service attack. Thanks to
3061 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3065 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3066 dtls1_buffer_record function under certain conditions. In particular this
3067 could occur if an attacker sent repeated DTLS records with the same
3068 sequence number but for the next epoch. The memory leak could be exploited
3069 by an attacker in a Denial of Service attack through memory exhaustion.
3070 Thanks to Chris Mueller for reporting this issue.
3074 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3075 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3076 method would be set to NULL which could later result in a NULL pointer
3077 dereference. Thanks to Frank Schmirler for reporting this issue.
3081 *) Abort handshake if server key exchange message is omitted for ephemeral
3084 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3085 reporting this issue.
3089 *) Remove non-export ephemeral RSA code on client and server. This code
3090 violated the TLS standard by allowing the use of temporary RSA keys in
3091 non-export ciphersuites and could be used by a server to effectively
3092 downgrade the RSA key length used to a value smaller than the server
3093 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3094 INRIA or reporting this issue.
3098 *) Fixed issue where DH client certificates are accepted without verification.
3099 An OpenSSL server will accept a DH certificate for client authentication
3100 without the certificate verify message. This effectively allows a client to
3101 authenticate without the use of a private key. This only affects servers
3102 which trust a client certificate authority which issues certificates
3103 containing DH keys: these are extremely rare and hardly ever encountered.
3104 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3109 *) Ensure that the session ID context of an SSL is updated when its
3110 SSL_CTX is updated via SSL_set_SSL_CTX.
3112 The session ID context is typically set from the parent SSL_CTX,
3113 and can vary with the CTX.
3116 *) Fix various certificate fingerprint issues.
3118 By using non-DER or invalid encodings outside the signed portion of a
3119 certificate the fingerprint can be changed without breaking the signature.
3120 Although no details of the signed portion of the certificate can be changed
3121 this can cause problems with some applications: e.g. those using the
3122 certificate fingerprint for blacklists.
3124 1. Reject signatures with non zero unused bits.
3126 If the BIT STRING containing the signature has non zero unused bits reject
3127 the signature. All current signature algorithms require zero unused bits.
3129 2. Check certificate algorithm consistency.
3131 Check the AlgorithmIdentifier inside TBS matches the one in the
3132 certificate signature. NB: this will result in signature failure
3133 errors for some broken certificates.
3135 Thanks to Konrad Kraszewski from Google for reporting this issue.
3137 3. Check DSA/ECDSA signatures use DER.
3139 Re-encode DSA/ECDSA signatures and compare with the original received
3140 signature. Return an error if there is a mismatch.
3142 This will reject various cases including garbage after signature
3143 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3144 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3145 (negative or with leading zeroes).
3147 Further analysis was conducted and fixes were developed by Stephen Henson
3148 of the OpenSSL core team.
3153 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3154 results on some platforms, including x86_64. This bug occurs at random
3155 with a very low probability, and is not known to be exploitable in any
3156 way, though its exact impact is difficult to determine. Thanks to Pieter
3157 Wuille (Blockstream) who reported this issue and also suggested an initial
3158 fix. Further analysis was conducted by the OpenSSL development team and
3159 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3160 the OpenSSL core team.
3164 *) Do not resume sessions on the server if the negotiated protocol
3165 version does not match the session's version. Resuming with a different
3166 version, while not strictly forbidden by the RFC, is of questionable
3167 sanity and breaks all known clients.
3168 [David Benjamin, Emilia Käsper]
3170 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3171 early CCS messages during renegotiation. (Note that because
3172 renegotiation is encrypted, this early CCS was not exploitable.)
3175 *) Tighten client-side session ticket handling during renegotiation:
3176 ensure that the client only accepts a session ticket if the server sends
3177 the extension anew in the ServerHello. Previously, a TLS client would
3178 reuse the old extension state and thus accept a session ticket if one was
3179 announced in the initial ServerHello.
3181 Similarly, ensure that the client requires a session ticket if one
3182 was advertised in the ServerHello. Previously, a TLS client would
3183 ignore a missing NewSessionTicket message.
3186 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3188 *) SRTP Memory Leak.
3190 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3191 sends a carefully crafted handshake message, to cause OpenSSL to fail
3192 to free up to 64k of memory causing a memory leak. This could be
3193 exploited in a Denial Of Service attack. This issue affects OpenSSL
3194 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3195 whether SRTP is used or configured. Implementations of OpenSSL that
3196 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3198 The fix was developed by the OpenSSL team.
3202 *) Session Ticket Memory Leak.
3204 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3205 integrity of that ticket is first verified. In the event of a session
3206 ticket integrity check failing, OpenSSL will fail to free memory
3207 causing a memory leak. By sending a large number of invalid session
3208 tickets an attacker could exploit this issue in a Denial Of Service
3213 *) Build option no-ssl3 is incomplete.
3215 When OpenSSL is configured with "no-ssl3" as a build option, servers
3216 could accept and complete a SSL 3.0 handshake, and clients could be
3217 configured to send them.
3219 [Akamai and the OpenSSL team]
3221 *) Add support for TLS_FALLBACK_SCSV.
3222 Client applications doing fallback retries should call
3223 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3225 [Adam Langley, Bodo Moeller]
3227 *) Add additional DigestInfo checks.
3229 Re-encode DigestInto in DER and check against the original when
3230 verifying RSA signature: this will reject any improperly encoded
3231 DigestInfo structures.
3233 Note: this is a precautionary measure and no attacks are currently known.
3237 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3239 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3240 SRP code can be overrun an internal buffer. Add sanity check that
3241 g, A, B < N to SRP code.
3243 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3244 Group for discovering this issue.
3248 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3249 TLS 1.0 instead of higher protocol versions when the ClientHello message
3250 is badly fragmented. This allows a man-in-the-middle attacker to force a
3251 downgrade to TLS 1.0 even if both the server and the client support a
3252 higher protocol version, by modifying the client's TLS records.
3254 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3255 researching this issue.
3259 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3260 to a denial of service attack. A malicious server can crash the client
3261 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3262 ciphersuite and sending carefully crafted handshake messages.
3264 Thanks to Felix Gröbert (Google) for discovering and researching this
3269 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3270 to leak memory. This can be exploited through a Denial of Service attack.
3271 Thanks to Adam Langley for discovering and researching this issue.
3275 *) An attacker can force openssl to consume large amounts of memory whilst
3276 processing DTLS handshake messages. This can be exploited through a
3277 Denial of Service attack.
3278 Thanks to Adam Langley for discovering and researching this issue.
3282 *) An attacker can force an error condition which causes openssl to crash
3283 whilst processing DTLS packets due to memory being freed twice. This
3284 can be exploited through a Denial of Service attack.
3285 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3290 *) If a multithreaded client connects to a malicious server using a resumed
3291 session and the server sends an ec point format extension it could write
3292 up to 255 bytes to freed memory.
3294 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3299 *) A malicious server can crash an OpenSSL client with a null pointer
3300 dereference (read) by specifying an SRP ciphersuite even though it was not
3301 properly negotiated with the client. This can be exploited through a
3302 Denial of Service attack.
3304 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3305 discovering and researching this issue.
3309 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3310 X509_name_oneline, X509_name_print_ex et al. to leak some information
3311 from the stack. Applications may be affected if they echo pretty printing
3312 output to the attacker.
3314 Thanks to Ivan Fratric (Google) for discovering this issue.
3316 [Emilia Käsper, and Steve Henson]
3318 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3319 for corner cases. (Certain input points at infinity could lead to
3320 bogus results, with non-infinity inputs mapped to infinity too.)
3323 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3325 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3326 handshake can force the use of weak keying material in OpenSSL
3327 SSL/TLS clients and servers.
3329 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3330 researching this issue. (CVE-2014-0224)
3331 [KIKUCHI Masashi, Steve Henson]
3333 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3334 OpenSSL DTLS client the code can be made to recurse eventually crashing
3337 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3339 [Imre Rad, Steve Henson]
3341 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3342 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3343 client or server. This is potentially exploitable to run arbitrary
3344 code on a vulnerable client or server.
3346 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3347 [Jüri Aedla, Steve Henson]
3349 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3350 are subject to a denial of service attack.
3352 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3353 this issue. (CVE-2014-3470)
3354 [Felix Gröbert, Ivan Fratric, Steve Henson]
3356 *) Harmonize version and its documentation. -f flag is used to display
3358 [mancha <mancha1@zoho.com>]
3360 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3361 in i2d_ECPrivateKey.
3362 [mancha <mancha1@zoho.com>]
3364 *) Fix some double frees. These are not thought to be exploitable.
3365 [mancha <mancha1@zoho.com>]
3367 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3369 *) A missing bounds check in the handling of the TLS heartbeat extension
3370 can be used to reveal up to 64k of memory to a connected client or
3373 Thanks for Neel Mehta of Google Security for discovering this bug and to
3374 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3375 preparing the fix (CVE-2014-0160)
3376 [Adam Langley, Bodo Moeller]
3378 *) Fix for the attack described in the paper "Recovering OpenSSL
3379 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3380 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3381 http://eprint.iacr.org/2014/140
3383 Thanks to Yuval Yarom and Naomi Benger for discovering this
3384 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3385 [Yuval Yarom and Naomi Benger]
3387 *) TLS pad extension: draft-agl-tls-padding-03
3389 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3390 TLS client Hello record length value would otherwise be > 255 and
3391 less that 512 pad with a dummy extension containing zeroes so it
3392 is at least 512 bytes long.
3394 [Adam Langley, Steve Henson]
3396 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3398 *) Fix for TLS record tampering bug. A carefully crafted invalid
3399 handshake could crash OpenSSL with a NULL pointer exception.
3400 Thanks to Anton Johansson for reporting this issues.
3403 *) Keep original DTLS digest and encryption contexts in retransmission
3404 structures so we can use the previous session parameters if they need
3405 to be resent. (CVE-2013-6450)
3408 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3409 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3410 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3411 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3412 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3413 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3414 [Rob Stradling, Adam Langley]
3416 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3418 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3419 supporting platforms or when small records were transferred.
3420 [Andy Polyakov, Steve Henson]
3422 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3424 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3426 This addresses the flaw in CBC record processing discovered by
3427 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3428 at: http://www.isg.rhul.ac.uk/tls/
3430 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3431 Security Group at Royal Holloway, University of London
3432 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3433 Emilia Käsper for the initial patch.
3435 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3437 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3438 ciphersuites which can be exploited in a denial of service attack.
3439 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3440 and detecting this bug and to Wolfgang Ettlinger
3441 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3445 *) Return an error when checking OCSP signatures when key is NULL.
3446 This fixes a DoS attack. (CVE-2013-0166)
3449 *) Make openssl verify return errors.
3450 [Chris Palmer <palmer@google.com> and Ben Laurie]
3452 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3453 the right response is stapled. Also change SSL_get_certificate()
3454 so it returns the certificate actually sent.
3455 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3456 [Rob Stradling <rob.stradling@comodo.com>]
3458 *) Fix possible deadlock when decoding public keys.
3461 *) Don't use TLS 1.0 record version number in initial client hello
3465 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3467 *) Sanity check record length before skipping explicit IV in TLS
3468 1.2, 1.1 and DTLS to fix DoS attack.
3470 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3471 fuzzing as a service testing platform.
3475 *) Initialise tkeylen properly when encrypting CMS messages.
3476 Thanks to Solar Designer of Openwall for reporting this issue.
3479 *) In FIPS mode don't try to use composite ciphers as they are not
3483 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3485 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3486 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3487 mean any application compiled against OpenSSL 1.0.0 headers setting
3488 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3489 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3490 0x10000000L Any application which was previously compiled against
3491 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3492 will need to be recompiled as a result. Letting be results in
3493 inability to disable specifically TLS 1.1 and in client context,
3494 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3497 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3498 disable just protocol X, but all protocols above X *if* there are
3499 protocols *below* X still enabled. In more practical terms it means
3500 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3501 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3502 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3506 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3508 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3509 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3510 in CRYPTO_realloc_clean.
3512 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3513 issue and to Adam Langley <agl@chromium.org> for fixing it.
3515 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3517 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3520 *) Workarounds for some broken servers that "hang" if a client hello
3521 record length exceeds 255 bytes.
3523 1. Do not use record version number > TLS 1.0 in initial client
3524 hello: some (but not all) hanging servers will now work.
3525 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3526 the number of ciphers sent in the client hello. This should be
3527 set to an even number, such as 50, for example by passing:
3528 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3529 Most broken servers should now work.
3530 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3531 TLS 1.2 client support entirely.
3534 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3537 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3539 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3540 STRING form instead of a DigestInfo.
3543 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3544 and the RSA_sign/RSA_verify functions. This was made more apparent when
3545 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3546 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3547 the correct format in RSA_verify so both forms transparently work.
3550 *) Some servers which support TLS 1.0 can choke if we initially indicate
3551 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3552 encrypted premaster secret. As a workaround use the maximum permitted
3553 client version in client hello, this should keep such servers happy
3554 and still work with previous versions of OpenSSL.
3557 *) Add support for TLS/DTLS heartbeats.
3558 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3560 *) Add support for SCTP.
3561 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3563 *) Improved PRNG seeding for VOS.
3564 [Paul Green <Paul.Green@stratus.com>]
3566 *) Extensive assembler packs updates, most notably:
3568 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3569 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3570 - x86_64: bit-sliced AES implementation;
3571 - ARM: NEON support, contemporary platforms optimizations;
3572 - s390x: z196 support;
3573 - *: GHASH and GF(2^m) multiplication implementations;
3577 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3578 (removal of unnecessary code)
3579 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3581 *) Add TLS key material exporter from RFC 5705.
3584 *) Add DTLS-SRTP negotiation from RFC 5764.
3587 *) Add Next Protocol Negotiation,
3588 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3589 disabled with a no-npn flag to config or Configure. Code donated
3591 [Adam Langley <agl@google.com> and Ben Laurie]
3593 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3594 NIST-P256, NIST-P521, with constant-time single point multiplication on
3595 typical inputs. Compiler support for the nonstandard type __uint128_t is
3596 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3597 Code made available under Apache License version 2.0.
3599 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3600 line to include this in your build of OpenSSL, and run "make depend" (or
3601 "make update"). This enables the following EC_METHODs:
3603 EC_GFp_nistp224_method()
3604 EC_GFp_nistp256_method()
3605 EC_GFp_nistp521_method()
3607 EC_GROUP_new_by_curve_name() will automatically use these (while
3608 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3610 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3612 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3613 all platforms. Move ssize_t definition from e_os.h to the public
3614 header file e_os2.h as it now appears in public header file cms.h
3617 *) New -sigopt option to the ca, req and x509 utilities. Additional
3618 signature parameters can be passed using this option and in
3622 *) Add RSA PSS signing function. This will generate and set the
3623 appropriate AlgorithmIdentifiers for PSS based on those in the
3624 corresponding EVP_MD_CTX structure. No application support yet.
3627 *) Support for companion algorithm specific ASN1 signing routines.
3628 New function ASN1_item_sign_ctx() signs a pre-initialised
3629 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3630 the appropriate parameters.
3633 *) Add new algorithm specific ASN1 verification initialisation function
3634 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3635 handling will be the same no matter what EVP_PKEY_METHOD is used.
3636 Add a PSS handler to support verification of PSS signatures: checked
3637 against a number of sample certificates.
3640 *) Add signature printing for PSS. Add PSS OIDs.
3641 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3643 *) Add algorithm specific signature printing. An individual ASN1 method
3644 can now print out signatures instead of the standard hex dump.
3646 More complex signatures (e.g. PSS) can print out more meaningful
3647 information. Include DSA version that prints out the signature
3651 *) Password based recipient info support for CMS library: implementing
3655 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3656 neatly separates the code into cipher and PBE sections and is required
3657 for some algorithms that split PBES2 into separate pieces (such as
3658 password based CMS).
3661 *) Session-handling fixes:
3662 - Fix handling of connections that are resuming with a session ID,
3663 but also support Session Tickets.
3664 - Fix a bug that suppressed issuing of a new ticket if the client
3665 presented a ticket with an expired session.
3666 - Try to set the ticket lifetime hint to something reasonable.
3667 - Make tickets shorter by excluding irrelevant information.
3668 - On the client side, don't ignore renewed tickets.
3669 [Adam Langley, Bodo Moeller (Google)]
3671 *) Fix PSK session representation.
3674 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3676 This work was sponsored by Intel.
3679 *) Add GCM support to TLS library. Some custom code is needed to split
3680 the IV between the fixed (from PRF) and explicit (from TLS record)
3681 portions. This adds all GCM ciphersuites supported by RFC5288 and
3682 RFC5289. Generalise some AES* cipherstrings to include GCM and
3683 add a special AESGCM string for GCM only.
3686 *) Expand range of ctrls for AES GCM. Permit setting invocation
3687 field on decrypt and retrieval of invocation field only on encrypt.
3690 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3691 As required by RFC5289 these ciphersuites cannot be used if for
3692 versions of TLS earlier than 1.2.
3695 *) For FIPS capable OpenSSL interpret a NULL default public key method
3696 as unset and return the appropriate default but do *not* set the default.
3697 This means we can return the appropriate method in applications that
3698 switch between FIPS and non-FIPS modes.
3701 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3702 ENGINE is used then we cannot handle that in the FIPS module so we
3703 keep original code iff non-FIPS operations are allowed.
3706 *) Add -attime option to openssl utilities.
3707 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3709 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3712 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3713 FIPS EC methods unconditionally for now.
3716 *) New build option no-ec2m to disable characteristic 2 code.
3719 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3720 all cases can be covered as some introduce binary incompatibilities.
3723 *) Redirect RSA operations to FIPS module including keygen,
3724 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3727 *) Add similar low level API blocking to ciphers.
3730 *) Low level digest APIs are not approved in FIPS mode: any attempt
3731 to use these will cause a fatal error. Applications that *really* want
3732 to use them can use the private_* version instead.
3735 *) Redirect cipher operations to FIPS module for FIPS builds.
3738 *) Redirect digest operations to FIPS module for FIPS builds.
3741 *) Update build system to add "fips" flag which will link in fipscanister.o
3742 for static and shared library builds embedding a signature if needed.
3745 *) Output TLS supported curves in preference order instead of numerical
3746 order. This is currently hardcoded for the highest order curves first.
3747 This should be configurable so applications can judge speed vs strength.
3750 *) Add TLS v1.2 server support for client authentication.
3753 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3757 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3758 FIPS modules versions.
3761 *) Add TLS v1.2 client side support for client authentication. Keep cache
3762 of handshake records longer as we don't know the hash algorithm to use
3763 until after the certificate request message is received.
3766 *) Initial TLS v1.2 client support. Add a default signature algorithms
3767 extension including all the algorithms we support. Parse new signature
3768 format in client key exchange. Relax some ECC signing restrictions for
3769 TLS v1.2 as indicated in RFC5246.
3772 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3773 to new signature format when needed using client digest preference.
3774 All server ciphersuites should now work correctly in TLS v1.2. No client
3775 support yet and no support for client certificates.
3778 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3779 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3780 ciphersuites. At present only RSA key exchange ciphersuites work with
3781 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3782 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3783 and version checking.
3786 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3787 with this defined it will not be affected by any changes to ssl internal
3788 structures. Add several utility functions to allow openssl application
3789 to work with OPENSSL_NO_SSL_INTERN defined.
3792 *) A long standing patch to add support for SRP from EdelWeb (Peter
3793 Sylvester and Christophe Renou) was integrated.
3794 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3795 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3798 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3801 *) Permit abbreviated handshakes when renegotiating using the function
3802 SSL_renegotiate_abbreviated().
3803 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3805 *) Add call to ENGINE_register_all_complete() to
3806 ENGINE_load_builtin_engines(), so some implementations get used
3807 automatically instead of needing explicit application support.
3810 *) Add support for TLS key exporter as described in RFC5705.
3811 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3813 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3814 a few changes are required:
3816 Add SSL_OP_NO_TLSv1_1 flag.
3817 Add TLSv1_1 methods.
3818 Update version checking logic to handle version 1.1.
3819 Add explicit IV handling (ported from DTLS code).
3820 Add command line options to s_client/s_server.
3823 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3825 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3826 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3827 content decryption and always return the same error. Note: this attack
3828 needs on average 2^20 messages so it only affects automated senders. The
3829 old behaviour can be re-enabled in the CMS code by setting the
3830 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3831 an MMA defence is not necessary.
3832 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3833 this issue. (CVE-2012-0884)
3836 *) Fix CVE-2011-4619: make sure we really are receiving a
3837 client hello before rejecting multiple SGC restarts. Thanks to
3838 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3841 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3843 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3844 Thanks to Antonio Martin, Enterprise Secure Access Research and
3845 Development, Cisco Systems, Inc. for discovering this bug and
3846 preparing a fix. (CVE-2012-0050)
3849 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3851 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3852 of the Vaudenay padding oracle attack on CBC mode encryption
3853 which enables an efficient plaintext recovery attack against
3854 the OpenSSL implementation of DTLS. Their attack exploits timing
3855 differences arising during decryption processing. A research
3856 paper describing this attack can be found at:
3857 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3858 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3859 Security Group at Royal Holloway, University of London
3860 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3861 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3862 for preparing the fix. (CVE-2011-4108)
3863 [Robin Seggelmann, Michael Tuexen]
3865 *) Clear bytes used for block padding of SSL 3.0 records.
3867 [Adam Langley (Google)]
3869 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3870 Kadianakis <desnacked@gmail.com> for discovering this issue and
3871 Adam Langley for preparing the fix. (CVE-2011-4619)
3872 [Adam Langley (Google)]
3874 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3875 [Andrey Kulikov <amdeich@gmail.com>]
3877 *) Prevent malformed RFC3779 data triggering an assertion failure.
3878 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3879 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3880 [Rob Austein <sra@hactrn.net>]
3882 *) Improved PRNG seeding for VOS.
3883 [Paul Green <Paul.Green@stratus.com>]
3885 *) Fix ssl_ciph.c set-up race.
3886 [Adam Langley (Google)]
3888 *) Fix spurious failures in ecdsatest.c.
3889 [Emilia Käsper (Google)]
3891 *) Fix the BIO_f_buffer() implementation (which was mixing different
3892 interpretations of the '..._len' fields).
3893 [Adam Langley (Google)]
3895 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3896 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3897 threads won't reuse the same blinding coefficients.
3899 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3900 lock to call BN_BLINDING_invert_ex, and avoids one use of
3901 BN_BLINDING_update for each BN_BLINDING structure (previously,
3902 the last update always remained unused).
3903 [Emilia Käsper (Google)]
3905 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3906 [Bob Buckholz (Google)]
3908 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3910 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3911 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3912 [Kaspar Brand <ossl@velox.ch>]
3914 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3915 for multi-threaded use of ECDH. (CVE-2011-3210)
3916 [Adam Langley (Google)]
3918 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3921 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3922 signature public key algorithm by using OID xref utilities instead.
3923 Before this you could only use some ECC ciphersuites with SHA1 only.
3926 *) Add protection against ECDSA timing attacks as mentioned in the paper
3927 by Billy Bob Brumley and Nicola Tuveri, see:
3929 http://eprint.iacr.org/2011/232.pdf
3931 [Billy Bob Brumley and Nicola Tuveri]
3933 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3935 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3936 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3938 *) Fix bug in string printing code: if *any* escaping is enabled we must
3939 escape the escape character (backslash) or the resulting string is
3943 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3945 *) Disable code workaround for ancient and obsolete Netscape browsers
3946 and servers: an attacker can use it in a ciphersuite downgrade attack.
3947 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3950 *) Fixed J-PAKE implementation error, originally discovered by
3951 Sebastien Martini, further info and confirmation from Stefan
3952 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3955 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3957 *) Fix extension code to avoid race conditions which can result in a buffer
3958 overrun vulnerability: resumed sessions must not be modified as they can
3959 be shared by multiple threads. CVE-2010-3864
3962 *) Fix WIN32 build system to correctly link an ENGINE directory into
3966 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3968 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3970 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3972 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3974 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3975 context. The operation can be customised via the ctrl mechanism in
3976 case ENGINEs want to include additional functionality.
3979 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3982 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3983 output hashes compatible with older versions of OpenSSL.
3984 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3986 *) Fix compression algorithm handling: if resuming a session use the
3987 compression algorithm of the resumed session instead of determining
3988 it from client hello again. Don't allow server to change algorithm.
3991 *) Add load_crls() function to apps tidying load_certs() too. Add option
3992 to verify utility to allow additional CRLs to be included.
3995 *) Update OCSP request code to permit adding custom headers to the request:
3996 some responders need this.
3999 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4001 [Julia Lawall <julia@diku.dk>]
4003 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4004 needlessly dereferenced structures, used obsolete functions and
4005 didn't handle all updated verify codes correctly.
4008 *) Disable MD2 in the default configuration.
4011 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4012 indicate the initial BIO being pushed or popped. This makes it possible
4013 to determine whether the BIO is the one explicitly called or as a result
4014 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4015 it handles reference counts correctly and doesn't zero out the I/O bio
4016 when it is not being explicitly popped. WARNING: applications which
4017 included workarounds for the old buggy behaviour will need to be modified
4018 or they could free up already freed BIOs.
4021 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4022 renaming to all platforms (within the 0.9.8 branch, this was
4023 done conditionally on Netware platforms to avoid a name clash).
4024 [Guenter <lists@gknw.net>]
4026 *) Add ECDHE and PSK support to DTLS.
4027 [Michael Tuexen <tuexen@fh-muenster.de>]
4029 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4033 *) Add "missing" function EVP_MD_flags() (without this the only way to
4034 retrieve a digest flags is by accessing the structure directly. Update
4035 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4036 or cipher is registered as in the "from" argument. Print out all
4037 registered digests in the dgst usage message instead of manually
4038 attempting to work them out.
4041 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4042 this allows the use of compression and extensions. Change default cipher
4043 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4044 by default unless an application cipher string requests it.
4047 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4048 key ids to find matching certificates and keys but some PKCS#12 files
4049 don't follow the (somewhat unwritten) rules and this strategy fails.
4050 Now just gather all certificates together and the first private key
4051 then look for the first certificate that matches the key.
4054 *) Support use of registered digest and cipher names for dgst and cipher
4055 commands instead of having to add each one as a special case. So now
4062 openssl dgst -sha256 foo
4064 and this works for ENGINE based algorithms too.
4068 *) Update Gost ENGINE to support parameter files.
4069 [Victor B. Wagner <vitus@cryptocom.ru>]
4071 *) Support GeneralizedTime in ca utility.
4072 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4074 *) Enhance the hash format used for certificate directory links. The new
4075 form uses the canonical encoding (meaning equivalent names will work
4076 even if they aren't identical) and uses SHA1 instead of MD5. This form
4077 is incompatible with the older format and as a result c_rehash should
4078 be used to rebuild symbolic links.
4081 *) Make PKCS#8 the default write format for private keys, replacing the
4082 traditional format. This form is standardised, more secure and doesn't
4083 include an implicit MD5 dependency.
4086 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4087 committed to OpenSSL should pass this lot as a minimum.
4090 *) Add session ticket override functionality for use by EAP-FAST.
4091 [Jouni Malinen <j@w1.fi>]
4093 *) Modify HMAC functions to return a value. Since these can be implemented
4094 in an ENGINE errors can occur.
4097 *) Type-checked OBJ_bsearch_ex.
4100 *) Type-checked OBJ_bsearch. Also some constification necessitated
4101 by type-checking. Still to come: TXT_DB, bsearch(?),
4102 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4106 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4107 seconds to a tm structure directly, instead of going through OS
4108 specific date routines. This avoids any issues with OS routines such
4109 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4110 and X509_time_adj_ex() to cover the extended range. The existing
4111 X509_time_adj() is still usable and will no longer have any date issues.
4114 *) Delta CRL support. New use deltas option which will attempt to locate
4115 and search any appropriate delta CRLs available.
4117 This work was sponsored by Google.
4120 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4121 code and add additional score elements. Validate alternate CRL paths
4122 as part of the CRL checking and indicate a new error "CRL path validation
4123 error" in this case. Applications wanting additional details can use
4124 the verify callback and check the new "parent" field. If this is not
4125 NULL CRL path validation is taking place. Existing applications won't
4126 see this because it requires extended CRL support which is off by
4129 This work was sponsored by Google.
4132 *) Support for freshest CRL extension.
4134 This work was sponsored by Google.
4137 *) Initial indirect CRL support. Currently only supported in the CRLs
4138 passed directly and not via lookup. Process certificate issuer
4139 CRL entry extension and lookup CRL entries by bother issuer name
4140 and serial number. Check and process CRL issuer entry in IDP extension.
4142 This work was sponsored by Google.
4145 *) Add support for distinct certificate and CRL paths. The CRL issuer
4146 certificate is validated separately in this case. Only enabled if
4147 an extended CRL support flag is set: this flag will enable additional
4148 CRL functionality in future.
4150 This work was sponsored by Google.
4153 *) Add support for policy mappings extension.
4155 This work was sponsored by Google.
4158 *) Fixes to pathlength constraint, self issued certificate handling,
4159 policy processing to align with RFC3280 and PKITS tests.
4161 This work was sponsored by Google.
4164 *) Support for name constraints certificate extension. DN, email, DNS
4165 and URI types are currently supported.
4167 This work was sponsored by Google.
4170 *) To cater for systems that provide a pointer-based thread ID rather
4171 than numeric, deprecate the current numeric thread ID mechanism and
4172 replace it with a structure and associated callback type. This
4173 mechanism allows a numeric "hash" to be extracted from a thread ID in
4174 either case, and on platforms where pointers are larger than 'long',
4175 mixing is done to help ensure the numeric 'hash' is usable even if it
4176 can't be guaranteed unique. The default mechanism is to use "&errno"
4177 as a pointer-based thread ID to distinguish between threads.
4179 Applications that want to provide their own thread IDs should now use
4180 CRYPTO_THREADID_set_callback() to register a callback that will call
4181 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4183 Note that ERR_remove_state() is now deprecated, because it is tied
4184 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4185 to free the current thread's error state should be replaced by
4186 ERR_remove_thread_state(NULL).
4188 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4189 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4190 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4191 application was previously providing a numeric thread callback that
4192 was inappropriate for distinguishing threads, then uniqueness might
4193 have been obtained with &errno that happened immediately in the
4194 intermediate development versions of OpenSSL; this is no longer the
4195 case, the numeric thread callback will now override the automatic use
4197 [Geoff Thorpe, with help from Bodo Moeller]
4199 *) Initial support for different CRL issuing certificates. This covers a
4200 simple case where the self issued certificates in the chain exist and
4201 the real CRL issuer is higher in the existing chain.
4203 This work was sponsored by Google.
4206 *) Removed effectively defunct crypto/store from the build.
4209 *) Revamp of STACK to provide stronger type-checking. Still to come:
4210 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4211 ASN1_STRING, CONF_VALUE.
4214 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4215 RAM on SSL connections. This option can save about 34k per idle SSL.
4218 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4219 STACK, TXT_DB, bsearch, qsort.
4222 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4223 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4224 support for data, signedData, compressedData, digestedData and
4225 encryptedData, envelopedData types included. Scripts to check against
4226 RFC4134 examples draft and interop and consistency checks of many
4227 content types and variants.
4230 *) Add options to enc utility to support use of zlib compression BIO.
4233 *) Extend mk1mf to support importing of options and assembly language
4234 files from Configure script, currently only included in VC-WIN32.
4235 The assembly language rules can now optionally generate the source
4236 files from the associated perl scripts.
4239 *) Implement remaining functionality needed to support GOST ciphersuites.
4240 Interop testing has been performed using CryptoPro implementations.
4241 [Victor B. Wagner <vitus@cryptocom.ru>]
4243 *) s390x assembler pack.
4246 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4250 *) Implement Opaque PRF Input TLS extension as specified in
4251 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4252 official specification yet and no extension type assignment by
4253 IANA exists, this extension (for now) will have to be explicitly
4254 enabled when building OpenSSL by providing the extension number
4255 to use. For example, specify an option
4257 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4259 to the "config" or "Configure" script to enable the extension,
4260 assuming extension number 0x9527 (which is a completely arbitrary
4261 and unofficial assignment based on the MD5 hash of the Internet
4262 Draft). Note that by doing so, you potentially lose
4263 interoperability with other TLS implementations since these might
4264 be using the same extension number for other purposes.
4266 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4267 opaque PRF input value to use in the handshake. This will create
4268 an internal copy of the length-'len' string at 'src', and will
4269 return non-zero for success.
4271 To get more control and flexibility, provide a callback function
4274 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4275 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4279 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4282 Callback function 'cb' will be called in handshakes, and is
4283 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4284 Argument 'arg' is for application purposes (the value as given to
4285 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4286 be provided to the callback function). The callback function
4287 has to return non-zero to report success: usually 1 to use opaque
4288 PRF input just if possible, or 2 to enforce use of the opaque PRF
4289 input. In the latter case, the library will abort the handshake
4290 if opaque PRF input is not successfully negotiated.
4292 Arguments 'peerinput' and 'len' given to the callback function
4293 will always be NULL and 0 in the case of a client. A server will
4294 see the client's opaque PRF input through these variables if
4295 available (NULL and 0 otherwise). Note that if the server
4296 provides an opaque PRF input, the length must be the same as the
4297 length of the client's opaque PRF input.
4299 Note that the callback function will only be called when creating
4300 a new session (session resumption can resume whatever was
4301 previously negotiated), and will not be called in SSL 2.0
4302 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4303 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4304 for applications that need to enforce opaque PRF input.
4308 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4311 [Victor B. Wagner <vitus@cryptocom.ru>]
4313 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4314 RFC4507bis. The encrypted ticket format is an encrypted encoded
4315 SSL_SESSION structure, that way new session features are automatically
4318 If a client application caches session in an SSL_SESSION structure
4319 support is transparent because tickets are now stored in the encoded
4322 The SSL_CTX structure automatically generates keys for ticket
4323 protection in servers so again support should be possible
4324 with no application modification.
4326 If a client or server wishes to disable RFC4507 support then the option
4327 SSL_OP_NO_TICKET can be set.
4329 Add a TLS extension debugging callback to allow the contents of any client
4330 or server extensions to be examined.
4332 This work was sponsored by Google.
4335 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4336 OpenSSL should now compile cleanly on gcc 4.2
4337 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4339 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4340 support including streaming MAC support: this is required for GOST
4341 ciphersuite support.
4342 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4344 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4345 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4346 to output in BER and PEM format.
4349 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4350 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4351 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4352 ENGINE support for HMAC keys which are unextractable. New -mac and
4353 -macopt options to dgst utility.
4356 *) New option -sigopt to dgst utility. Update dgst to use
4357 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4358 alternative signing parameters such as X9.31 or PSS in the dgst
4362 *) Change ssl_cipher_apply_rule(), the internal function that does
4363 the work each time a ciphersuite string requests enabling
4364 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4365 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4366 the order of disabled ciphersuites such that those ciphersuites
4367 that most recently went from enabled to disabled not only stay
4368 in order with respect to each other, but also have higher priority
4369 than other disabled ciphersuites the next time ciphersuites are
4372 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4373 the same ciphersuites as with "HIGH" alone, but in a specific
4374 order where the PSK ciphersuites come first (since they are the
4375 most recently disabled ciphersuites when "HIGH" is parsed).
4377 Also, change ssl_create_cipher_list() (using this new
4378 functionality) such that between otherwise identical
4379 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4383 *) Change ssl_create_cipher_list() so that it automatically
4384 arranges the ciphersuites in reasonable order before starting
4385 to process the rule string. Thus, the definition for "DEFAULT"
4386 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4387 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4388 This makes it much easier to arrive at a reasonable default order
4389 in applications for which anonymous ciphers are OK (meaning
4390 that you can't actually use DEFAULT).
4391 [Bodo Moeller; suggested by Victor Duchovni]
4393 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4394 processing) into multiple integers instead of setting
4395 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4396 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4397 (These masks as well as the individual bit definitions are hidden
4398 away into the non-exported interface ssl/ssl_locl.h, so this
4399 change to the definition of the SSL_CIPHER structure shouldn't
4400 affect applications.) This give us more bits for each of these
4401 categories, so there is no longer a need to coagulate AES128 and
4402 AES256 into a single algorithm bit, and to coagulate Camellia128
4403 and Camellia256 into a single algorithm bit, which has led to all
4406 Thus, among other things, the kludge introduced in 0.9.7m and
4407 0.9.8e for masking out AES256 independently of AES128 or masking
4408 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4410 With the change, we also introduce new ciphersuite aliases that
4411 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4415 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4416 Use the leftmost N bytes of the signature input if the input is
4417 larger than the prime q (with N being the size in bytes of q).
4420 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4421 it yet and it is largely untested.
4424 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4427 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4428 some compilers (gcc 4.2 and later) reject their use. Safestack is
4429 reimplemented. Update ASN1 to avoid use of legacy functions.
4432 *) Win32/64 targets are linked with Winsock2.
4435 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4436 to external functions. This can be used to increase CRL handling
4437 efficiency especially when CRLs are very large by (for example) storing
4438 the CRL revoked certificates in a database.
4441 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4442 new CRLs added to a directory can be used. New command line option
4443 -verify_return_error to s_client and s_server. This causes real errors
4444 to be returned by the verify callback instead of carrying on no matter
4445 what. This reflects the way a "real world" verify callback would behave.
4448 *) GOST engine, supporting several GOST algorithms and public key formats.
4449 Kindly donated by Cryptocom.
4452 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4453 partitioned by DP are handled but no indirect CRL or reason partitioning
4454 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4455 selected via a scoring technique which handles IDP and AKID in CRLs.
4458 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4459 will ultimately be used for all verify operations: this will remove the
4460 X509_STORE dependency on certificate verification and allow alternative
4461 lookup methods. X509_STORE based implementations of these two callbacks.
4464 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4465 Modify get_crl() to find a valid (unexpired) CRL if possible.
4468 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4469 this would be called X509_CRL_cmp() but that name is already used by
4470 a function that just compares CRL issuer names. Cache several CRL
4471 extensions in X509_CRL structure and cache CRLDP in X509.
4474 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4475 this maps equivalent X509_NAME structures into a consistent structure.
4476 Name comparison can then be performed rapidly using memcmp().
4479 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4483 *) Allow digests to supply their own micalg string for S/MIME type using
4484 the ctrl EVP_MD_CTRL_MICALG.
4487 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4488 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4489 ctrl. It can then customise the structure before and/or after signing
4493 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4494 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4495 to free up any added signature OIDs.
4498 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4499 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4500 digest and cipher tables. New options added to openssl utility:
4501 list-message-digest-algorithms and list-cipher-algorithms.
4504 *) Change the array representation of binary polynomials: the list
4505 of degrees of non-zero coefficients is now terminated with -1.
4506 Previously it was terminated with 0, which was also part of the
4507 value; thus, the array representation was not applicable to
4508 polynomials where t^0 has coefficient zero. This change makes
4509 the array representation useful in a more general context.
4512 *) Various modifications and fixes to SSL/TLS cipher string
4513 handling. For ECC, the code now distinguishes between fixed ECDH
4514 with RSA certificates on the one hand and with ECDSA certificates
4515 on the other hand, since these are separate ciphersuites. The
4516 unused code for Fortezza ciphersuites has been removed.
4518 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4519 (not "ECDHE"). For consistency with the code for DH
4520 certificates, use of ECDH certificates is now considered ECDH
4521 authentication, not RSA or ECDSA authentication (the latter is
4522 merely the CA's signing algorithm and not actively used in the
4525 The temporary ciphersuite alias "ECCdraft" is no longer
4526 available, and ECC ciphersuites are no longer excluded from "ALL"
4527 and "DEFAULT". The following aliases now exist for RFC 4492
4528 ciphersuites, most of these by analogy with the DH case:
4530 kECDHr - ECDH cert, signed with RSA
4531 kECDHe - ECDH cert, signed with ECDSA
4532 kECDH - ECDH cert (signed with either RSA or ECDSA)
4533 kEECDH - ephemeral ECDH
4534 ECDH - ECDH cert or ephemeral ECDH
4540 AECDH - anonymous ECDH
4541 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4545 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4546 Use correct micalg parameters depending on digest(s) in signed message.
4549 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4550 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4553 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4554 an engine to register a method. Add ENGINE lookups for methods and
4555 functional reference processing.
4558 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4559 EVP_{Sign,Verify}* which allow an application to customise the signature
4563 *) New -resign option to smime utility. This adds one or more signers
4564 to an existing PKCS#7 signedData structure. Also -md option to use an
4565 alternative message digest algorithm for signing.
4568 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4569 create PKCS7 structures containing multiple signers. Update smime
4570 application to support multiple signers.
4573 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4577 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4578 Reorganize PBE internals to lookup from a static table using NIDs,
4579 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4580 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4581 PRF which will be automatically used with PBES2.
4584 *) Replace the algorithm specific calls to generate keys in "req" with the
4588 *) Update PKCS#7 enveloped data routines to use new API. This is now
4589 supported by any public key method supporting the encrypt operation. A
4590 ctrl is added to allow the public key algorithm to examine or modify
4591 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4595 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4596 a default digest type to use. In most cases this will be SHA1 but some
4597 algorithms (such as GOST) need to specify an alternative digest. The
4598 return value indicates how strong the preference is 1 means optional and
4599 2 is mandatory (that is it is the only supported type). Modify
4600 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4601 use the default md. Update openssl utilities to use the default digest
4602 type for signing if it is not explicitly indicated.
4605 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4606 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4607 signing method from the key type. This effectively removes the link
4608 between digests and public key types.
4611 *) Add an OID cross reference table and utility functions. Its purpose is to
4612 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4613 rsaEncryption. This will allow some of the algorithm specific hackery
4614 needed to use the correct OID to be removed.
4617 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4618 structures for PKCS7_sign(). They are now set up by the relevant public
4622 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4625 *) Add support for key derivation (agreement) in the API, DH method and
4629 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4630 public and private key formats. As a side effect these add additional
4631 command line functionality not previously available: DSA signatures can be
4632 generated and verified using pkeyutl and DH key support and generation in
4637 [Oliver Tappe <zooey@hirschkaefer.de>]
4639 *) New make target "install_html_docs" installs HTML renditions of the
4641 [Oliver Tappe <zooey@hirschkaefer.de>]
4643 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4644 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4645 support key and parameter generation and add initial key generation
4646 functionality for RSA.
4649 *) Add functions for main EVP_PKEY_method operations. The undocumented
4650 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4651 EVP_PKEY_{encrypt,decrypt}_old.
4654 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4655 key API, doesn't do much yet.
4658 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4659 public key algorithms. New option to openssl utility:
4660 "list-public-key-algorithms" to print out info.
4663 *) Implement the Supported Elliptic Curves Extension for
4664 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4667 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4668 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4671 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4672 utilities such as rsa, dsa, dsaparam etc except they process any key
4676 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4677 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4678 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4682 *) Initial support for pluggable public key ASN1.
4683 De-spaghettify the public key ASN1 handling. Move public and private
4684 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4685 algorithm specific handling to a single module within the relevant
4686 algorithm directory. Add functions to allow (near) opaque processing
4687 of public and private key structures.
4690 *) Implement the Supported Point Formats Extension for
4691 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4694 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4695 for the psk identity [hint] and the psk callback functions to the
4696 SSL_SESSION, SSL and SSL_CTX structure.
4699 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4703 SSL_CTX_use_psk_identity_hint
4704 SSL_get_psk_identity_hint
4705 SSL_get_psk_identity
4706 SSL_use_psk_identity_hint
4708 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4710 *) Add RFC 3161 compliant time stamp request creation, response generation
4711 and response verification functionality.
4712 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4714 *) Add initial support for TLS extensions, specifically for the server_name
4715 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4716 have new members for a host name. The SSL data structure has an
4717 additional member SSL_CTX *initial_ctx so that new sessions can be
4718 stored in that context to allow for session resumption, even after the
4719 SSL has been switched to a new SSL_CTX in reaction to a client's
4720 server_name extension.
4722 New functions (subject to change):
4724 SSL_get_servername()
4725 SSL_get_servername_type()
4728 New CTRL codes and macros (subject to change):
4730 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4731 - SSL_CTX_set_tlsext_servername_callback()
4732 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4733 - SSL_CTX_set_tlsext_servername_arg()
4734 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4736 openssl s_client has a new '-servername ...' option.
4738 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4739 '-key2 ...', '-servername_fatal' (subject to change). This allows
4740 testing the HostName extension for a specific single host name ('-cert'
4741 and '-key' remain fallbacks for handshakes without HostName
4742 negotiation). If the unrecognized_name alert has to be sent, this by
4743 default is a warning; it becomes fatal with the '-servername_fatal'
4746 [Peter Sylvester, Remy Allais, Christophe Renou]
4748 *) Whirlpool hash implementation is added.
4751 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4752 bn(64,32). Because of instruction set limitations it doesn't have
4753 any negative impact on performance. This was done mostly in order
4754 to make it possible to share assembler modules, such as bn_mul_mont
4755 implementations, between 32- and 64-bit builds without hassle.
4758 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4759 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4763 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4764 dedicated Montgomery multiplication procedure, is introduced.
4765 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4766 "64-bit" performance on certain 32-bit targets.
4769 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4770 in SSL structures. New SSL ctrl to set maximum send fragment size.
4771 Save memory by setting the I/O buffer sizes dynamically instead of
4772 using the maximum available value.
4775 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4776 in addition to the text details.
4779 *) Very, very preliminary EXPERIMENTAL support for printing of general
4780 ASN1 structures. This currently produces rather ugly output and doesn't
4781 handle several customised structures at all.
4784 *) Integrated support for PVK file format and some related formats such
4785 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4786 these in the 'rsa' and 'dsa' utilities.
4789 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4792 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4793 place for the (very old) "NETSCAPE" format certificates which are now
4794 handled using new ASN1 code equivalents.
4797 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4798 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4799 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4802 *) Modify CRL distribution points extension code to print out previously
4803 unsupported fields. Enhance extension setting code to allow setting of
4807 *) Add print and set support for Issuing Distribution Point CRL extension.
4810 *) Change 'Configure' script to enable Camellia by default.
4813 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4815 *) When rejecting SSL/TLS records due to an incorrect version number, never
4816 update s->server with a new major version number. As of
4817 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4818 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4819 the previous behavior could result in a read attempt at NULL when
4820 receiving specific incorrect SSL/TLS records once record payload
4821 protection is active. (CVE-2010-0740)
4822 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4824 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4825 could be crashed if the relevant tables were not present (e.g. chrooted).
4826 [Tomas Hoger <thoger@redhat.com>]
4828 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4830 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4831 [Martin Olsson, Neel Mehta]
4833 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4834 accommodate for stack sorting, always a write lock!).
4837 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4838 excessive delays in the RAND_poll(): over a minute. As a workaround
4839 include a time check in the inner Heap32Next loop too.
4842 *) The code that handled flushing of data in SSL/TLS originally used the
4843 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4844 the problem outlined in PR#1949. The fix suggested there however can
4845 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4846 of Apache). So instead simplify the code to flush unconditionally.
4847 This should be fine since flushing with no data to flush is a no op.
4850 *) Handle TLS versions 2.0 and later properly and correctly use the
4851 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4852 off ancient servers have a habit of sticking around for a while...
4855 *) Modify compression code so it frees up structures without using the
4856 ex_data callbacks. This works around a problem where some applications
4857 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4858 restarting) then use compression (e.g. SSL with compression) later.
4859 This results in significant per-connection memory leaks and
4860 has caused some security issues including CVE-2008-1678 and
4864 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4865 change when encrypting or decrypting.
4868 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4869 connect and renegotiate with servers which do not support RI.
4870 Until RI is more widely deployed this option is enabled by default.
4873 *) Add "missing" ssl ctrls to clear options and mode.
4876 *) If client attempts to renegotiate and doesn't support RI respond with
4877 a no_renegotiation alert as required by RFC5746. Some renegotiating
4878 TLS clients will continue a connection gracefully when they receive
4879 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4880 waiting for a server hello which it will never receive. Now we treat a
4881 received no_renegotiation alert as a fatal error. This is because
4882 applications requesting a renegotiation might well expect it to succeed
4883 and would have no code in place to handle the server denying it so the
4884 only safe thing to do is to terminate the connection.
4887 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4888 peer supports secure renegotiation and 0 otherwise. Print out peer
4889 renegotiation support in s_client/s_server.
4892 *) Replace the highly broken and deprecated SPKAC certification method with
4893 the updated NID creation version. This should correctly handle UTF8.
4896 *) Implement RFC5746. Re-enable renegotiation but require the extension
4897 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4898 turns out to be a bad idea. It has been replaced by
4899 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4900 SSL_CTX_set_options(). This is really not recommended unless you
4901 know what you are doing.
4902 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4904 *) Fixes to stateless session resumption handling. Use initial_ctx when
4905 issuing and attempting to decrypt tickets in case it has changed during
4906 servername handling. Use a non-zero length session ID when attempting
4907 stateless session resumption: this makes it possible to determine if
4908 a resumption has occurred immediately after receiving server hello
4909 (several places in OpenSSL subtly assume this) instead of later in
4913 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4914 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4915 fixes for a few places where the return code is not checked
4917 [Julia Lawall <julia@diku.dk>]
4919 *) Add --strict-warnings option to Configure script to include devteam
4920 warnings in other configurations.
4923 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4924 makes it possible to install openssl libraries in locations which
4925 have names other than "lib", for example "/usr/lib64" which some
4927 [Steve Henson, based on patch from Jeremy Utley]
4929 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4930 X690 8.9.12 and can produce some misleading textual output of OIDs.
4931 [Steve Henson, reported by Dan Kaminsky]
4933 *) Delete MD2 from algorithm tables. This follows the recommendation in
4934 several standards that it is not used in new applications due to
4935 several cryptographic weaknesses. For binary compatibility reasons
4936 the MD2 API is still compiled in by default.
4939 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4943 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4944 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4946 [Guenter <lists@gknw.net>]
4948 *) Fix the server certificate chain building code to use X509_verify_cert(),
4949 it used to have an ad-hoc builder which was unable to cope with anything
4950 other than a simple chain.
4951 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4953 *) Don't check self signed certificate signatures in X509_verify_cert()
4954 by default (a flag can override this): it just wastes time without
4955 adding any security. As a useful side effect self signed root CAs
4956 with non-FIPS digests are now usable in FIPS mode.
4959 *) In dtls1_process_out_of_seq_message() the check if the current message
4960 is already buffered was missing. For every new message was memory
4961 allocated, allowing an attacker to perform an denial of service attack
4962 with sending out of seq handshake messages until there is no memory
4963 left. Additionally every future message was buffered, even if the
4964 sequence number made no sense and would be part of another handshake.
4965 So only messages with sequence numbers less than 10 in advance will be
4966 buffered. (CVE-2009-1378)
4967 [Robin Seggelmann, discovered by Daniel Mentz]
4969 *) Records are buffered if they arrive with a future epoch to be
4970 processed after finishing the corresponding handshake. There is
4971 currently no limitation to this buffer allowing an attacker to perform
4972 a DOS attack with sending records with future epochs until there is no
4973 memory left. This patch adds the pqueue_size() function to determine
4974 the size of a buffer and limits the record buffer to 100 entries.
4976 [Robin Seggelmann, discovered by Daniel Mentz]
4978 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4979 parent structure is freed. (CVE-2009-1379)
4982 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4983 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4986 [Ilya O. <vrghost@gmail.com>]
4988 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4990 *) Disable renegotiation completely - this fixes a severe security
4991 problem (CVE-2009-3555) at the cost of breaking all
4992 renegotiation. Renegotiation can be re-enabled by setting
4993 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4994 run-time. This is really not recommended unless you know what
4998 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5000 *) Don't set val to NULL when freeing up structures, it is freed up by
5001 underlying code. If sizeof(void *) > sizeof(long) this can result in
5002 zeroing past the valid field. (CVE-2009-0789)
5003 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5005 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5006 checked correctly. This would allow some invalid signed attributes to
5007 appear to verify correctly. (CVE-2009-0591)
5008 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5010 *) Reject UniversalString and BMPString types with invalid lengths. This
5011 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5012 a legal length. (CVE-2009-0590)
5015 *) Set S/MIME signing as the default purpose rather than setting it
5016 unconditionally. This allows applications to override it at the store
5020 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5021 to handle some structures.
5024 *) Improve efficiency of mem_gets: don't search whole buffer each time
5026 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5028 *) New -hex option for openssl rand.
5031 *) Print out UTF8String and NumericString when parsing ASN1.
5034 *) Support NumericString type for name components.
5037 *) Allow CC in the environment to override the automatically chosen
5038 compiler. Note that nothing is done to ensure flags work with the
5042 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5044 *) Properly check EVP_VerifyFinal() and similar return values
5046 [Ben Laurie, Bodo Moeller, Google Security Team]
5048 *) Enable TLS extensions by default.
5051 *) Allow the CHIL engine to be loaded, whether the application is
5052 multithreaded or not. (This does not release the developer from the
5053 obligation to set up the dynamic locking callbacks.)
5054 [Sander Temme <sander@temme.net>]
5056 *) Use correct exit code if there is an error in dgst command.
5057 [Steve Henson; problem pointed out by Roland Dirlewanger]
5059 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5060 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5063 *) Add experimental JPAKE support, including demo authentication in
5064 s_client and s_server.
5067 *) Set the comparison function in v3_addr_canonize().
5068 [Rob Austein <sra@hactrn.net>]
5070 *) Add support for XMPP STARTTLS in s_client.
5071 [Philip Paeps <philip@freebsd.org>]
5073 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5074 to ensure that even with this option, only ciphersuites in the
5075 server's preference list will be accepted. (Note that the option
5076 applies only when resuming a session, so the earlier behavior was
5077 just about the algorithm choice for symmetric cryptography.)
5080 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5082 *) Fix NULL pointer dereference if a DTLS server received
5083 ChangeCipherSpec as first record (CVE-2009-1386).
5086 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5087 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5090 *) The fix in 0.9.8c that supposedly got rid of unsafe
5091 double-checked locking was incomplete for RSA blinding,
5092 addressing just one layer of what turns out to have been
5093 doubly unsafe triple-checked locking.
5095 So now fix this for real by retiring the MONT_HELPER macro
5096 in crypto/rsa/rsa_eay.c.
5098 [Bodo Moeller; problem pointed out by Marius Schilder]
5100 *) Various precautionary measures:
5102 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5104 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5105 (NB: This would require knowledge of the secret session ticket key
5106 to exploit, in which case you'd be SOL either way.)
5108 - Change bn_nist.c so that it will properly handle input BIGNUMs
5109 outside the expected range.
5111 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5114 [Neel Mehta, Bodo Moeller]
5116 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5117 the load fails. Useful for distros.
5118 [Ben Laurie and the FreeBSD team]
5120 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5123 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5126 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5128 This work was sponsored by Logica.
5131 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5132 keystores. Support for SSL/TLS client authentication too.
5133 Not compiled unless enable-capieng specified to Configure.
5135 This work was sponsored by Logica.
5138 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5139 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5140 attribute creation routines such as certificate requests and PKCS#12
5144 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5146 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5147 handshake which could lead to a client crash as found using the
5148 Codenomicon TLS test suite (CVE-2008-1672)
5149 [Steve Henson, Mark Cox]
5151 *) Fix double free in TLS server name extensions which could lead to
5152 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5155 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5157 Clear the error queue to ensure that error entries left from
5158 older function calls do not interfere with the correct operation.
5159 [Lutz Jaenicke, Erik de Castro Lopo]
5161 *) Remove root CA certificates of commercial CAs:
5163 The OpenSSL project does not recommend any specific CA and does not
5164 have any policy with respect to including or excluding any CA.
5165 Therefore it does not make any sense to ship an arbitrary selection
5166 of root CA certificates with the OpenSSL software.
5169 *) RSA OAEP patches to fix two separate invalid memory reads.
5170 The first one involves inputs when 'lzero' is greater than
5171 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5172 before the beginning of from). The second one involves inputs where
5173 the 'db' section contains nothing but zeroes (there is a one-byte
5174 invalid read after the end of 'db').
5175 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5177 *) Partial backport from 0.9.9-dev:
5179 Introduce bn_mul_mont (dedicated Montgomery multiplication
5180 procedure) as a candidate for BIGNUM assembler implementation.
5181 While 0.9.9-dev uses assembler for various architectures, only
5182 x86_64 is available by default here in the 0.9.8 branch, and
5183 32-bit x86 is available through a compile-time setting.
5185 To try the 32-bit x86 assembler implementation, use Configure
5186 option "enable-montasm" (which exists only for this backport).
5188 As "enable-montasm" for 32-bit x86 disclaims code stability
5189 anyway, in this constellation we activate additional code
5190 backported from 0.9.9-dev for further performance improvements,
5191 namely BN_from_montgomery_word. (To enable this otherwise,
5192 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5194 [Andy Polyakov (backport partially by Bodo Moeller)]
5196 *) Add TLS session ticket callback. This allows an application to set
5197 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5198 values. This is useful for key rollover for example where several key
5199 sets may exist with different names.
5202 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5203 This was broken until now in 0.9.8 releases, such that the only way
5204 a registered ENGINE could be used (assuming it initialises
5205 successfully on the host) was to explicitly set it as the default
5206 for the relevant algorithms. This is in contradiction with 0.9.7
5207 behaviour and the documentation. With this fix, when an ENGINE is
5208 registered into a given algorithm's table of implementations, the
5209 'uptodate' flag is reset so that auto-discovery will be used next
5210 time a new context for that algorithm attempts to select an
5212 [Ian Lister (tweaked by Geoff Thorpe)]
5214 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5215 implementation in the following ways:
5217 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5220 Lack of BER streaming support means one pass streaming processing is
5221 only supported if data is detached: setting the streaming flag is
5222 ignored for embedded content.
5224 CMS support is disabled by default and must be explicitly enabled
5225 with the enable-cms configuration option.
5228 *) Update the GMP engine glue to do direct copies between BIGNUM and
5229 mpz_t when openssl and GMP use the same limb size. Otherwise the
5230 existing "conversion via a text string export" trick is still used.
5231 [Paul Sheer <paulsheer@gmail.com>]
5233 *) Zlib compression BIO. This is a filter BIO which compressed and
5234 uncompresses any data passed through it.
5237 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5238 RFC3394 compatible AES key wrapping.
5241 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5242 sets string data without copying. X509_ALGOR_set0() and
5243 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5244 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5245 from an X509_ATTRIBUTE structure optionally checking it occurs only
5246 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5250 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5251 to get the expected BN_FLG_CONSTTIME behavior.
5252 [Bodo Moeller (Google)]
5256 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5257 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5258 - added some more tests to do_tests.pl
5259 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5260 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5261 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5262 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5263 - various changes to netware.pl to enable gcc-cross builds on Win32
5265 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5266 - various changes to fix missing prototype warnings
5267 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5268 - added AES, WHIRLPOOL and CPUID assembler code to build files
5269 - added missing AES assembler make rules to mk1mf.pl
5270 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5271 [Guenter Knauf <eflash@gmx.net>]
5273 *) Implement certificate status request TLS extension defined in RFC3546.
5274 A client can set the appropriate parameters and receive the encoded
5275 OCSP response via a callback. A server can query the supplied parameters
5276 and set the encoded OCSP response in the callback. Add simplified examples
5277 to s_client and s_server.
5280 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5282 *) Fix various bugs:
5283 + Binary incompatibility of ssl_ctx_st structure
5284 + DTLS interoperation with non-compliant servers
5285 + Don't call get_session_cb() without proposed session
5286 + Fix ia64 assembler code
5287 [Andy Polyakov, Steve Henson]
5289 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5291 *) DTLS Handshake overhaul. There were longstanding issues with
5292 OpenSSL DTLS implementation, which were making it impossible for
5293 RFC 4347 compliant client to communicate with OpenSSL server.
5294 Unfortunately just fixing these incompatibilities would "cut off"
5295 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5296 server keeps tolerating non RFC compliant syntax. The opposite is
5297 not true, 0.9.8f client can not communicate with earlier server.
5298 This update even addresses CVE-2007-4995.
5301 *) Changes to avoid need for function casts in OpenSSL: some compilers
5302 (gcc 4.2 and later) reject their use.
5303 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5306 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5307 RFC4507bis. The encrypted ticket format is an encrypted encoded
5308 SSL_SESSION structure, that way new session features are automatically
5311 If a client application caches session in an SSL_SESSION structure
5312 support is transparent because tickets are now stored in the encoded
5315 The SSL_CTX structure automatically generates keys for ticket
5316 protection in servers so again support should be possible
5317 with no application modification.
5319 If a client or server wishes to disable RFC4507 support then the option
5320 SSL_OP_NO_TICKET can be set.
5322 Add a TLS extension debugging callback to allow the contents of any client
5323 or server extensions to be examined.
5325 This work was sponsored by Google.
5328 *) Add initial support for TLS extensions, specifically for the server_name
5329 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5330 have new members for a host name. The SSL data structure has an
5331 additional member SSL_CTX *initial_ctx so that new sessions can be
5332 stored in that context to allow for session resumption, even after the
5333 SSL has been switched to a new SSL_CTX in reaction to a client's
5334 server_name extension.
5336 New functions (subject to change):
5338 SSL_get_servername()
5339 SSL_get_servername_type()
5342 New CTRL codes and macros (subject to change):
5344 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5345 - SSL_CTX_set_tlsext_servername_callback()
5346 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5347 - SSL_CTX_set_tlsext_servername_arg()
5348 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5350 openssl s_client has a new '-servername ...' option.
5352 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5353 '-key2 ...', '-servername_fatal' (subject to change). This allows
5354 testing the HostName extension for a specific single host name ('-cert'
5355 and '-key' remain fallbacks for handshakes without HostName
5356 negotiation). If the unrecognized_name alert has to be sent, this by
5357 default is a warning; it becomes fatal with the '-servername_fatal'
5360 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5362 *) Add AES and SSE2 assembly language support to VC++ build.
5365 *) Mitigate attack on final subtraction in Montgomery reduction.
5368 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5369 (which previously caused an internal error).
5372 *) Squeeze another 10% out of IGE mode when in != out.
5375 *) AES IGE mode speedup.
5376 [Dean Gaudet (Google)]
5378 *) Add the Korean symmetric 128-bit cipher SEED (see
5379 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5380 add SEED ciphersuites from RFC 4162:
5382 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5383 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5384 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5385 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5387 To minimize changes between patchlevels in the OpenSSL 0.9.8
5388 series, SEED remains excluded from compilation unless OpenSSL
5389 is configured with 'enable-seed'.
5390 [KISA, Bodo Moeller]
5392 *) Mitigate branch prediction attacks, which can be practical if a
5393 single processor is shared, allowing a spy process to extract
5394 information. For detailed background information, see
5395 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5396 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5397 and Necessary Software Countermeasures"). The core of the change
5398 are new versions BN_div_no_branch() and
5399 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5400 respectively, which are slower, but avoid the security-relevant
5401 conditional branches. These are automatically called by BN_div()
5402 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5403 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5404 remove a conditional branch.
5406 BN_FLG_CONSTTIME is the new name for the previous
5407 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5408 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5409 in the exponent causes BN_mod_exp_mont() to use the alternative
5410 implementation in BN_mod_exp_mont_consttime().) The old name
5411 remains as a deprecated alias.
5413 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5414 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5415 constant-time implementations for more than just exponentiation.
5416 Here too the old name is kept as a deprecated alias.
5418 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5419 the BN_BLINDING structure gets an independent copy of the
5420 modulus. This means that the previous "BIGNUM *m" argument to
5421 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5422 essentially becomes "const BIGNUM *m", although we can't actually
5423 change this in the header file before 0.9.9. It allows
5424 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5425 enable BN_FLG_CONSTTIME.
5427 [Matthew D Wood (Intel Corp)]
5429 *) In the SSL/TLS server implementation, be strict about session ID
5430 context matching (which matters if an application uses a single
5431 external cache for different purposes). Previously,
5432 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5433 set. This did ensure strict client verification, but meant that,
5434 with applications using a single external cache for quite
5435 different requirements, clients could circumvent ciphersuite
5436 restrictions for a given session ID context by starting a session
5437 in a different context.
5440 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5441 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5442 authentication-only ciphersuites.
5445 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5446 not complete and could lead to a possible single byte overflow
5447 (CVE-2007-5135) [Ben Laurie]
5449 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5451 *) Since AES128 and AES256 (and similarly Camellia128 and
5452 Camellia256) share a single mask bit in the logic of
5453 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5454 kludge to work properly if AES128 is available and AES256 isn't
5455 (or if Camellia128 is available and Camellia256 isn't).
5458 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5459 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5460 When a point or a seed is encoded in a BIT STRING, we need to
5461 prevent the removal of trailing zero bits to get the proper DER
5462 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5463 of a NamedBitList, for which trailing 0 bits need to be removed.)
5466 *) Have SSL/TLS server implementation tolerate "mismatched" record
5467 protocol version while receiving ClientHello even if the
5468 ClientHello is fragmented. (The server can't insist on the
5469 particular protocol version it has chosen before the ServerHello
5470 message has informed the client about his choice.)
5473 *) Add RFC 3779 support.
5474 [Rob Austein for ARIN, Ben Laurie]
5476 *) Load error codes if they are not already present instead of using a
5477 static variable. This allows them to be cleanly unloaded and reloaded.
5478 Improve header file function name parsing.
5481 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5482 or CAPABILITY handshake as required by RFCs.
5485 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5487 *) Introduce limits to prevent malicious keys being able to
5488 cause a denial of service. (CVE-2006-2940)
5489 [Steve Henson, Bodo Moeller]
5491 *) Fix ASN.1 parsing of certain invalid structures that can result
5492 in a denial of service. (CVE-2006-2937) [Steve Henson]
5494 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5495 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5497 *) Fix SSL client code which could crash if connecting to a
5498 malicious SSLv2 server. (CVE-2006-4343)
5499 [Tavis Ormandy and Will Drewry, Google Security Team]
5501 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5502 match only those. Before that, "AES256-SHA" would be interpreted
5503 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5504 the same strength classification in 0.9.7h) as we currently only
5505 have a single AES bit in the ciphersuite description bitmap.
5506 That change, however, also applied to ciphersuite strings such as
5507 "RC4-MD5" that intentionally matched multiple ciphersuites --
5508 namely, SSL 2.0 ciphersuites in addition to the more common ones
5509 from SSL 3.0/TLS 1.0.
5511 So we change the selection algorithm again: Naming an explicit
5512 ciphersuite selects this one ciphersuite, and any other similar
5513 ciphersuite (same bitmap) from *other* protocol versions.
5514 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5515 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5517 Since SSL 2.0 does not have any ciphersuites for which the
5518 128/256 bit distinction would be relevant, this works for now.
5519 The proper fix will be to use different bits for AES128 and
5520 AES256, which would have avoided the problems from the beginning;
5521 however, bits are scarce, so we can only do this in a new release
5522 (not just a patchlevel) when we can change the SSL_CIPHER
5523 definition to split the single 'unsigned long mask' bitmap into
5524 multiple values to extend the available space.
5528 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5530 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5531 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5533 *) Add AES IGE and biIGE modes.
5536 *) Change the Unix randomness entropy gathering to use poll() when
5537 possible instead of select(), since the latter has some
5538 undesirable limitations.
5539 [Darryl Miles via Richard Levitte and Bodo Moeller]
5541 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5542 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5543 cannot be implicitly activated as part of, e.g., the "AES" alias.
5544 However, please upgrade to OpenSSL 0.9.9[-dev] for
5545 non-experimental use of the ECC ciphersuites to get TLS extension
5546 support, which is required for curve and point format negotiation
5547 to avoid potential handshake problems.
5550 *) Disable rogue ciphersuites:
5552 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5553 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5554 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5556 The latter two were purportedly from
5557 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5560 Also deactivate the remaining ciphersuites from
5561 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5562 unofficial, and the ID has long expired.
5565 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5566 dual-core machines) and other potential thread-safety issues.
5569 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5570 versions), which is now available for royalty-free use
5571 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5572 Also, add Camellia TLS ciphersuites from RFC 4132.
5574 To minimize changes between patchlevels in the OpenSSL 0.9.8
5575 series, Camellia remains excluded from compilation unless OpenSSL
5576 is configured with 'enable-camellia'.
5579 *) Disable the padding bug check when compression is in use. The padding
5580 bug check assumes the first packet is of even length, this is not
5581 necessarily true if compression is enabled and can result in false
5582 positives causing handshake failure. The actual bug test is ancient
5583 code so it is hoped that implementations will either have fixed it by
5584 now or any which still have the bug do not support compression.
5587 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5589 *) When applying a cipher rule check to see if string match is an explicit
5590 cipher suite and only match that one cipher suite if it is.
5593 *) Link in manifests for VC++ if needed.
5594 [Austin Ziegler <halostatue@gmail.com>]
5596 *) Update support for ECC-based TLS ciphersuites according to
5597 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5598 TLS extensions, which are supported starting with the 0.9.9
5599 branch, not in the OpenSSL 0.9.8 branch).
5602 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5603 opaque EVP_CIPHER_CTX handling.
5606 *) Fixes and enhancements to zlib compression code. We now only use
5607 "zlib1.dll" and use the default __cdecl calling convention on Win32
5608 to conform with the standards mentioned here:
5609 http://www.zlib.net/DLL_FAQ.txt
5610 Static zlib linking now works on Windows and the new --with-zlib-include
5611 --with-zlib-lib options to Configure can be used to supply the location
5612 of the headers and library. Gracefully handle case where zlib library
5616 *) Several fixes and enhancements to the OID generation code. The old code
5617 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5618 handle numbers larger than ULONG_MAX, truncated printing and had a
5619 non standard OBJ_obj2txt() behaviour.
5622 *) Add support for building of engines under engine/ as shared libraries
5623 under VC++ build system.
5626 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5627 Hopefully, we will not see any false combination of paths any more.
5630 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5632 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5633 (part of SSL_OP_ALL). This option used to disable the
5634 countermeasure against man-in-the-middle protocol-version
5635 rollback in the SSL 2.0 server implementation, which is a bad
5636 idea. (CVE-2005-2969)
5638 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5639 for Information Security, National Institute of Advanced Industrial
5640 Science and Technology [AIST], Japan)]
5642 *) Add two function to clear and return the verify parameter flags.
5645 *) Keep cipherlists sorted in the source instead of sorting them at
5646 runtime, thus removing the need for a lock.
5649 *) Avoid some small subgroup attacks in Diffie-Hellman.
5650 [Nick Mathewson and Ben Laurie]
5652 *) Add functions for well-known primes.
5655 *) Extended Windows CE support.
5656 [Satoshi Nakamura and Andy Polyakov]
5658 *) Initialize SSL_METHOD structures at compile time instead of during
5659 runtime, thus removing the need for a lock.
5662 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5663 attempting to decrypt each encrypted key in turn. Add support to
5667 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5669 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5672 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5675 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5676 key into the same file any more.
5679 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5682 *) Add -utf8 command line and config file option to 'ca'.
5683 [Stefan <stf@udoma.org]
5685 *) Removed the macro des_crypt(), as it seems to conflict with some
5686 libraries. Use DES_crypt().
5689 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5690 involves renaming the source and generated shared-libs for
5691 both. The engines will accept the corrected or legacy ids
5692 ('ncipher' and '4758_cca' respectively) when binding. NB,
5693 this only applies when building 'shared'.
5694 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5696 *) Add attribute functions to EVP_PKEY structure. Modify
5697 PKCS12_create() to recognize a CSP name attribute and
5698 use it. Make -CSP option work again in pkcs12 utility.
5701 *) Add new functionality to the bn blinding code:
5702 - automatic re-creation of the BN_BLINDING parameters after
5703 a fixed number of uses (currently 32)
5704 - add new function for parameter creation
5705 - introduce flags to control the update behaviour of the
5706 BN_BLINDING parameters
5707 - hide BN_BLINDING structure
5708 Add a second BN_BLINDING slot to the RSA structure to improve
5709 performance when a single RSA object is shared among several
5713 *) Add support for DTLS.
5714 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5716 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5717 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5720 *) Remove buggy and incomplete DH cert support from
5721 ssl/ssl_rsa.c and ssl/s3_both.c
5724 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5725 the apps/openssl applications.
5728 *) Compile clean with "-Wall -Wmissing-prototypes
5729 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5730 DEBUG_SAFESTACK must also be set.
5733 *) Change ./Configure so that certain algorithms can be disabled by default.
5734 The new counterpiece to "no-xxx" is "enable-xxx".
5736 The patented RC5 and MDC2 algorithms will now be disabled unless
5737 "enable-rc5" and "enable-mdc2", respectively, are specified.
5739 (IDEA remains enabled despite being patented. This is because IDEA
5740 is frequently required for interoperability, and there is no license
5741 fee for non-commercial use. As before, "no-idea" can be used to
5742 avoid this algorithm.)
5746 *) Add processing of proxy certificates (see RFC 3820). This work was
5747 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5748 EGEE (Enabling Grids for E-science in Europe).
5751 *) RC4 performance overhaul on modern architectures/implementations, such
5752 as Intel P4, IA-64 and AMD64.
5755 *) New utility extract-section.pl. This can be used specify an alternative
5756 section number in a pod file instead of having to treat each file as
5757 a separate case in Makefile. This can be done by adding two lines to the
5760 =for comment openssl_section:XXX
5762 The blank line is mandatory.
5766 *) New arguments -certform, -keyform and -pass for s_client and s_server
5767 to allow alternative format key and certificate files and passphrase
5771 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5772 update associated structures and add various utility functions.
5774 Add new policy related verify parameters, include policy checking in
5775 standard verify code. Enhance 'smime' application with extra parameters
5776 to support policy checking and print out.
5779 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5780 Nehemiah processors. These extensions support AES encryption in hardware
5781 as well as RNG (though RNG support is currently disabled).
5782 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5784 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5787 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5788 [Andy Polyakov and a number of other people]
5790 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5791 implementation contributed by IBM.
5792 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5794 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5795 exponent rather than 'unsigned long'. There is a corresponding change to
5796 the new 'rsa_keygen' element of the RSA_METHOD structure.
5797 [Jelte Jansen, Geoff Thorpe]
5799 *) Functionality for creating the initial serial number file is now
5800 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5802 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5803 number file to 1, which is bound to cause problems. To avoid
5804 the problems while respecting compatibility between different 0.9.7
5805 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5806 CA.pl for serial number initialization. With the new release 0.9.8,
5807 we can fix the problem directly in the 'ca' utility.)
5810 *) Reduced header interdependencies by declaring more opaque objects in
5811 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5812 give fewer recursive includes, which could break lazy source code - so
5813 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5814 developers should define this symbol when building and using openssl to
5815 ensure they track the recommended behaviour, interfaces, [etc], but
5816 backwards-compatible behaviour prevails when this isn't defined.
5819 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5822 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5823 This will generate a random key of the appropriate length based on the
5824 cipher context. The EVP_CIPHER can provide its own random key generation
5825 routine to support keys of a specific form. This is used in the des and
5826 3des routines to generate a key of the correct parity. Update S/MIME
5827 code to use new functions and hence generate correct parity DES keys.
5828 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5829 valid (weak or incorrect parity).
5832 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5833 as looking them up. This is useful when the verified structure may contain
5834 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5835 present unless the new PKCS7_NO_CRL flag is asserted.
5838 *) Extend ASN1 oid configuration module. It now additionally accepts the
5841 shortName = some long name, 1.2.3.4
5844 *) Reimplemented the BN_CTX implementation. There is now no more static
5845 limitation on the number of variables it can handle nor the depth of the
5846 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5847 information can now expand as required, and rather than having a single
5848 static array of bignums, BN_CTX now uses a linked-list of such arrays
5849 allowing it to expand on demand whilst maintaining the usefulness of
5850 BN_CTX's "bundling".
5853 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5854 to allow all RSA operations to function using a single BN_CTX.
5857 *) Preliminary support for certificate policy evaluation and checking. This
5858 is initially intended to pass the tests outlined in "Conformance Testing
5859 of Relying Party Client Certificate Path Processing Logic" v1.07.
5862 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5863 remained unused and not that useful. A variety of other little bignum
5864 tweaks and fixes have also been made continuing on from the audit (see
5868 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5869 associated ASN1, EVP and SSL functions and old ASN1 macros.
5872 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5873 and this should never fail. So the return value from the use of
5874 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5875 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5878 *) BN_CTX_get() should return zero-valued bignums, providing the same
5879 initialised value as BN_new().
5880 [Geoff Thorpe, suggested by Ulf Möller]
5882 *) Support for inhibitAnyPolicy certificate extension.
5885 *) An audit of the BIGNUM code is underway, for which debugging code is
5886 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5887 is considered valid when processing BIGNUMs, and causes execution to
5888 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5889 further steps are taken to deliberately pollute unused data in BIGNUM
5890 structures to try and expose faulty code further on. For now, openssl will
5891 (in its default mode of operation) continue to tolerate the inconsistent
5892 forms that it has tolerated in the past, but authors and packagers should
5893 consider trying openssl and their own applications when compiled with
5894 these debugging symbols defined. It will help highlight potential bugs in
5895 their own code, and will improve the test coverage for OpenSSL itself. At
5896 some point, these tighter rules will become openssl's default to improve
5897 maintainability, though the assert()s and other overheads will remain only
5898 in debugging configurations. See bn.h for more details.
5899 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5901 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5902 that can only be obtained through BN_CTX_new() (which implicitly
5903 initialises it). The presence of this function only made it possible
5904 to overwrite an existing structure (and cause memory leaks).
5907 *) Because of the callback-based approach for implementing LHASH as a
5908 template type, lh_insert() adds opaque objects to hash-tables and
5909 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5910 to clean up those corresponding objects before destroying the hash table
5911 (and losing the object pointers). So some over-zealous constifications in
5912 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5913 objects as "const" and the lh_doall[_arg] callback wrappers are not
5914 prototyped to have "const" restrictions on the object pointers they are
5915 given (and so aren't required to cast them away any more).
5918 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5919 (speed) prefers to use its own implementation. The two implementations
5920 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5921 its object type properly exposed (MS_TM) instead of casting to/from "char
5922 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5923 aren't necessarily the greatest nomenclatures - but this is what was used
5924 internally to the implementation so I've used that for now.
5927 *) Ensure that deprecated functions do not get compiled when
5928 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5929 the self-tests were still using deprecated key-generation functions so
5930 these have been updated also.
5933 *) Reorganise PKCS#7 code to separate the digest location functionality
5934 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5935 New function PKCS7_set_digest() to set the digest type for PKCS#7
5936 digestedData type. Add additional code to correctly generate the
5937 digestedData type and add support for this type in PKCS7 initialization
5941 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5942 structure of type "other".
5945 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5946 sure the loop does correctly stop and breaking ("division by zero")
5947 modulus operations are not performed. The (pre-generated) prime
5948 table crypto/bn/bn_prime.h was already correct, but it could not be
5949 re-generated on some platforms because of the "division by zero"
5950 situation in the script.
5951 [Ralf S. Engelschall]
5953 *) Update support for ECC-based TLS ciphersuites according to
5954 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5955 SHA-1 now is only used for "small" curves (where the
5956 representation of a field element takes up to 24 bytes); for
5957 larger curves, the field element resulting from ECDH is directly
5958 used as premaster secret.
5959 [Douglas Stebila (Sun Microsystems Laboratories)]
5961 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5962 curve secp160r1 to the tests.
5963 [Douglas Stebila (Sun Microsystems Laboratories)]
5965 *) Add the possibility to load symbols globally with DSO.
5966 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5968 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5969 control of the error stack.
5972 *) Add support for STORE in ENGINE.
5975 *) Add the STORE type. The intention is to provide a common interface
5976 to certificate and key stores, be they simple file-based stores, or
5977 HSM-type store, or LDAP stores, or...
5978 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5981 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5982 pass a list of arguments to any function as well as provide a way
5983 for a function to pass data back to the caller.
5986 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5987 works like BUF_strdup() but can be used to duplicate a portion of
5988 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5992 *) Add the function sk_find_ex() which works like sk_find(), but will
5993 return an index to an element even if an exact match couldn't be
5994 found. The index is guaranteed to point at the element where the
5995 searched-for key would be inserted to preserve sorting order.
5998 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5999 takes an extra flags argument for optional functionality. Currently,
6000 the following flags are defined:
6002 OBJ_BSEARCH_VALUE_ON_NOMATCH
6003 This one gets OBJ_bsearch_ex() to return a pointer to the first
6004 element where the comparing function returns a negative or zero
6007 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6008 This one gets OBJ_bsearch_ex() to return a pointer to the first
6009 element where the comparing function returns zero. This is useful
6010 if there are more than one element where the comparing function
6014 *) Make it possible to create self-signed certificates with 'openssl ca'
6015 in such a way that the self-signed certificate becomes part of the
6016 CA database and uses the same mechanisms for serial number generation
6017 as all other certificate signing. The new flag '-selfsign' enables
6018 this functionality. Adapt CA.sh and CA.pl.in.
6021 *) Add functionality to check the public key of a certificate request
6022 against a given private. This is useful to check that a certificate
6023 request can be signed by that key (self-signing).
6026 *) Make it possible to have multiple active certificates with the same
6027 subject in the CA index file. This is done only if the keyword
6028 'unique_subject' is set to 'no' in the main CA section (default
6029 if 'CA_default') of the configuration file. The value is saved
6030 with the database itself in a separate index attribute file,
6031 named like the index file with '.attr' appended to the name.
6034 *) Generate multi-valued AVAs using '+' notation in config files for
6038 *) Support for nameConstraints certificate extension.
6041 *) Support for policyConstraints certificate extension.
6044 *) Support for policyMappings certificate extension.
6047 *) Make sure the default DSA_METHOD implementation only uses its
6048 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6049 and change its own handlers to be NULL so as to remove unnecessary
6050 indirection. This lets alternative implementations fallback to the
6051 default implementation more easily.
6054 *) Support for directoryName in GeneralName related extensions
6058 *) Make it possible to link applications using Makefile.shared.
6059 Make that possible even when linking against static libraries!
6062 *) Support for single pass processing for S/MIME signing. This now
6063 means that S/MIME signing can be done from a pipe, in addition
6064 cleartext signing (multipart/signed type) is effectively streaming
6065 and the signed data does not need to be all held in memory.
6067 This is done with a new flag PKCS7_STREAM. When this flag is set
6068 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6069 is done after the data is output (and digests calculated) in
6070 SMIME_write_PKCS7().
6073 *) Add full support for -rpath/-R, both in shared libraries and
6074 applications, at least on the platforms where it's known how
6078 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6079 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6080 will now compute a table of multiples of the generator that
6081 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6082 faster (notably in the case of a single point multiplication,
6083 scalar * generator).
6084 [Nils Larsch, Bodo Moeller]
6086 *) IPv6 support for certificate extensions. The various extensions
6087 which use the IP:a.b.c.d can now take IPv6 addresses using the
6088 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6092 *) Added an ENGINE that implements RSA by performing private key
6093 exponentiations with the GMP library. The conversions to and from
6094 GMP's mpz_t format aren't optimised nor are any montgomery forms
6095 cached, and on x86 it appears OpenSSL's own performance has caught up.
6096 However there are likely to be other architectures where GMP could
6097 provide a boost. This ENGINE is not built in by default, but it can be
6098 specified at Configure time and should be accompanied by the necessary
6099 linker additions, eg;
6100 ./config -DOPENSSL_USE_GMP -lgmp
6103 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6104 testing availability of engines with "-t" - the old behaviour is
6105 produced by increasing the feature's verbosity with "-tt".
6108 *) ECDSA routines: under certain error conditions uninitialized BN objects
6109 could be freed. Solution: make sure initialization is performed early
6110 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6114 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6115 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6116 software implementations. For DSA and DH, parameter generation can
6117 also be overridden by providing the appropriate method callbacks.
6120 *) Change the "progress" mechanism used in key-generation and
6121 primality testing to functions that take a new BN_GENCB pointer in
6122 place of callback/argument pairs. The new API functions have "_ex"
6123 postfixes and the older functions are reimplemented as wrappers for
6124 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6125 declarations of the old functions to help (graceful) attempts to
6126 migrate to the new functions. Also, the new key-generation API
6127 functions operate on a caller-supplied key-structure and return
6128 success/failure rather than returning a key or NULL - this is to
6129 help make "keygen" another member function of RSA_METHOD etc.
6131 Example for using the new callback interface:
6133 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6137 BN_GENCB_set(&my_cb, my_callback, my_arg);
6139 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6140 /* For the meaning of a, b in calls to my_callback(), see the
6141 * documentation of the function that calls the callback.
6142 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6143 * my_callback should return 1 if it wants BN_is_prime_ex()
6144 * to continue, or 0 to stop.
6149 *) Change the ZLIB compression method to be stateful, and make it
6150 available to TLS with the number defined in
6151 draft-ietf-tls-compression-04.txt.
6154 *) Add the ASN.1 structures and functions for CertificatePair, which
6155 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6157 CertificatePair ::= SEQUENCE {
6158 forward [0] Certificate OPTIONAL,
6159 reverse [1] Certificate OPTIONAL,
6160 -- at least one of the pair shall be present -- }
6162 Also implement the PEM functions to read and write certificate
6163 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6165 This needed to be defined, mostly for the sake of the LDAP
6166 attribute crossCertificatePair, but may prove useful elsewhere as
6170 *) Make it possible to inhibit symlinking of shared libraries in
6171 Makefile.shared, for Cygwin's sake.
6174 *) Extend the BIGNUM API by creating a function
6175 void BN_set_negative(BIGNUM *a, int neg);
6176 and a macro that behave like
6177 int BN_is_negative(const BIGNUM *a);
6179 to avoid the need to access 'a->neg' directly in applications.
6182 *) Implement fast modular reduction for pseudo-Mersenne primes
6183 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6184 EC_GROUP_new_curve_GFp() will now automatically use this
6186 [Nils Larsch <nla@trustcenter.de>]
6188 *) Add new lock type (CRYPTO_LOCK_BN).
6191 *) Change the ENGINE framework to automatically load engines
6192 dynamically from specific directories unless they could be
6193 found to already be built in or loaded. Move all the
6194 current engines except for the cryptodev one to a new
6196 The engines in engines/ are built as shared libraries if
6197 the "shared" options was given to ./Configure or ./config.
6198 Otherwise, they are inserted in libcrypto.a.
6199 /usr/local/ssl/engines is the default directory for dynamic
6200 engines, but that can be overridden at configure time through
6201 the usual use of --prefix and/or --openssldir, and at run
6202 time with the environment variable OPENSSL_ENGINES.
6203 [Geoff Thorpe and Richard Levitte]
6205 *) Add Makefile.shared, a helper makefile to build shared
6206 libraries. Adapt Makefile.org.
6209 *) Add version info to Win32 DLLs.
6210 [Peter 'Luna' Runestig" <peter@runestig.com>]
6212 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6213 can be added using this API to created arbitrary PKCS#12
6214 files while avoiding the low level API.
6216 New options to PKCS12_create(), key or cert can be NULL and
6217 will then be omitted from the output file. The encryption
6218 algorithm NIDs can be set to -1 for no encryption, the mac
6219 iteration count can be set to 0 to omit the mac.
6221 Enhance pkcs12 utility by making the -nokeys and -nocerts
6222 options work when creating a PKCS#12 file. New option -nomac
6223 to omit the mac, NONE can be set for an encryption algorithm.
6224 New code is modified to use the enhanced PKCS12_create()
6225 instead of the low level API.
6228 *) Extend ASN1 encoder to support indefinite length constructed
6229 encoding. This can output sequences tags and octet strings in
6230 this form. Modify pk7_asn1.c to support indefinite length
6231 encoding. This is experimental and needs additional code to
6232 be useful, such as an ASN1 bio and some enhanced streaming
6235 Extend template encode functionality so that tagging is passed
6236 down to the template encoder.
6239 *) Let 'openssl req' fail if an argument to '-newkey' is not
6240 recognized instead of using RSA as a default.
6243 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6244 As these are not official, they are not included in "ALL";
6245 the "ECCdraft" ciphersuite group alias can be used to select them.
6246 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6248 *) Add ECDH engine support.
6249 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6251 *) Add ECDH in new directory crypto/ecdh/.
6252 [Douglas Stebila (Sun Microsystems Laboratories)]
6254 *) Let BN_rand_range() abort with an error after 100 iterations
6255 without success (which indicates a broken PRNG).
6258 *) Change BN_mod_sqrt() so that it verifies that the input value
6259 is really the square of the return value. (Previously,
6260 BN_mod_sqrt would show GIGO behaviour.)
6263 *) Add named elliptic curves over binary fields from X9.62, SECG,
6264 and WAP/WTLS; add OIDs that were still missing.
6266 [Sheueling Chang Shantz and Douglas Stebila
6267 (Sun Microsystems Laboratories)]
6269 *) Extend the EC library for elliptic curves over binary fields
6270 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6273 EC_GF2m_simple_method
6277 EC_GROUP_new_curve_GF2m
6278 EC_GROUP_set_curve_GF2m
6279 EC_GROUP_get_curve_GF2m
6280 EC_POINT_set_affine_coordinates_GF2m
6281 EC_POINT_get_affine_coordinates_GF2m
6282 EC_POINT_set_compressed_coordinates_GF2m
6284 Point compression for binary fields is disabled by default for
6285 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6288 As binary polynomials are represented as BIGNUMs, various members
6289 of the EC_GROUP and EC_POINT data structures can be shared
6290 between the implementations for prime fields and binary fields;
6291 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6292 are essentially identical to their ..._GFp counterparts.
6293 (For simplicity, the '..._GFp' prefix has been dropped from
6294 various internal method names.)
6296 An internal 'field_div' method (similar to 'field_mul' and
6297 'field_sqr') has been added; this is used only for binary fields.
6299 [Sheueling Chang Shantz and Douglas Stebila
6300 (Sun Microsystems Laboratories)]
6302 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6303 through methods ('mul', 'precompute_mult').
6305 The generic implementations (now internally called 'ec_wNAF_mul'
6306 and 'ec_wNAF_precomputed_mult') remain the default if these
6307 methods are undefined.
6309 [Sheueling Chang Shantz and Douglas Stebila
6310 (Sun Microsystems Laboratories)]
6312 *) New function EC_GROUP_get_degree, which is defined through
6313 EC_METHOD. For curves over prime fields, this returns the bit
6314 length of the modulus.
6316 [Sheueling Chang Shantz and Douglas Stebila
6317 (Sun Microsystems Laboratories)]
6319 *) New functions EC_GROUP_dup, EC_POINT_dup.
6320 (These simply call ..._new and ..._copy).
6322 [Sheueling Chang Shantz and Douglas Stebila
6323 (Sun Microsystems Laboratories)]
6325 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6326 Polynomials are represented as BIGNUMs (where the sign bit is not
6327 used) in the following functions [macros]:
6330 BN_GF2m_sub [= BN_GF2m_add]
6331 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6332 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6333 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6335 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6336 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6337 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6338 BN_GF2m_cmp [= BN_ucmp]
6340 (Note that only the 'mod' functions are actually for fields GF(2^m).
6341 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6343 For some functions, an the irreducible polynomial defining a
6344 field can be given as an 'unsigned int[]' with strictly
6345 decreasing elements giving the indices of those bits that are set;
6346 i.e., p[] represents the polynomial
6347 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6349 p[0] > p[1] > ... > p[k] = 0.
6350 This applies to the following functions:
6355 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6356 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6358 BN_GF2m_mod_sqrt_arr
6359 BN_GF2m_mod_solve_quad_arr
6363 Conversion can be performed by the following functions:
6368 bntest.c has additional tests for binary polynomial arithmetic.
6370 Two implementations for BN_GF2m_mod_div() are available.
6371 The default algorithm simply uses BN_GF2m_mod_inv() and
6372 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6373 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6374 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6376 [Sheueling Chang Shantz and Douglas Stebila
6377 (Sun Microsystems Laboratories)]
6379 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6380 functionality is disabled at compile-time.
6381 [Douglas Stebila <douglas.stebila@sun.com>]
6383 *) Change default behaviour of 'openssl asn1parse' so that more
6384 information is visible when viewing, e.g., a certificate:
6386 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6387 mode the content of non-printable OCTET STRINGs is output in a
6388 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6389 avoid the appearance of a printable string.
6390 [Nils Larsch <nla@trustcenter.de>]
6392 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6394 EC_GROUP_set_asn1_flag()
6395 EC_GROUP_get_asn1_flag()
6396 EC_GROUP_set_point_conversion_form()
6397 EC_GROUP_get_point_conversion_form()
6398 These control ASN1 encoding details:
6399 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6400 has been set to OPENSSL_EC_NAMED_CURVE.
6401 - Points are encoded in uncompressed form by default; options for
6402 asn1_for are as for point2oct, namely
6403 POINT_CONVERSION_COMPRESSED
6404 POINT_CONVERSION_UNCOMPRESSED
6405 POINT_CONVERSION_HYBRID
6407 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6410 EC_GROUP_get0_seed()
6411 EC_GROUP_get_seed_len()
6412 This is used only for ASN1 purposes (so far).
6413 [Nils Larsch <nla@trustcenter.de>]
6415 *) Add 'field_type' member to EC_METHOD, which holds the NID
6416 of the appropriate field type OID. The new function
6417 EC_METHOD_get_field_type() returns this value.
6418 [Nils Larsch <nla@trustcenter.de>]
6423 EC_POINT_point2hex()
6424 EC_POINT_hex2point()
6425 providing useful interfaces to EC_POINT_point2oct() and
6426 EC_POINT_oct2point().
6427 [Nils Larsch <nla@trustcenter.de>]
6429 *) Change internals of the EC library so that the functions
6430 EC_GROUP_set_generator()
6431 EC_GROUP_get_generator()
6432 EC_GROUP_get_order()
6433 EC_GROUP_get_cofactor()
6434 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6435 to methods, which would lead to unnecessary code duplication when
6436 adding different types of curves.
6437 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6439 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6440 arithmetic, and such that modified wNAFs are generated
6441 (which avoid length expansion in many cases).
6444 *) Add a function EC_GROUP_check_discriminant() (defined via
6445 EC_METHOD) that verifies that the curve discriminant is non-zero.
6447 Add a function EC_GROUP_check() that makes some sanity tests
6448 on a EC_GROUP, its generator and order. This includes
6449 EC_GROUP_check_discriminant().
6450 [Nils Larsch <nla@trustcenter.de>]
6452 *) Add ECDSA in new directory crypto/ecdsa/.
6454 Add applications 'openssl ecparam' and 'openssl ecdsa'
6455 (these are based on 'openssl dsaparam' and 'openssl dsa').
6457 ECDSA support is also included in various other files across the
6458 library. Most notably,
6459 - 'openssl req' now has a '-newkey ecdsa:file' option;
6460 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6461 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6462 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6463 them suitable for ECDSA where domain parameters must be
6464 extracted before the specific public key;
6465 - ECDSA engine support has been added.
6466 [Nils Larsch <nla@trustcenter.de>]
6468 *) Include some named elliptic curves, and add OIDs from X9.62,
6469 SECG, and WAP/WTLS. Each curve can be obtained from the new
6471 EC_GROUP_new_by_curve_name(),
6472 and the list of available named curves can be obtained with
6473 EC_get_builtin_curves().
6474 Also add a 'curve_name' member to EC_GROUP objects, which can be
6476 EC_GROUP_set_curve_name()
6477 EC_GROUP_get_curve_name()
6478 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6480 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6481 was actually never needed) and in BN_mul(). The removal in BN_mul()
6482 required a small change in bn_mul_part_recursive() and the addition
6483 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6484 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6485 bn_sub_words() and bn_add_words() except they take arrays with
6489 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6491 *) Cleanse PEM buffers before freeing them since they may contain
6493 [Benjamin Bennett <ben@psc.edu>]
6495 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6496 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6497 authentication-only ciphersuites.
6500 *) Since AES128 and AES256 share a single mask bit in the logic of
6501 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6502 kludge to work properly if AES128 is available and AES256 isn't.
6505 *) Expand security boundary to match 1.1.1 module.
6508 *) Remove redundant features: hash file source, editing of test vectors
6509 modify fipsld to use external fips_premain.c signature.
6512 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6513 run algorithm test programs.
6516 *) Make algorithm test programs more tolerant of whitespace.
6519 *) Have SSL/TLS server implementation tolerate "mismatched" record
6520 protocol version while receiving ClientHello even if the
6521 ClientHello is fragmented. (The server can't insist on the
6522 particular protocol version it has chosen before the ServerHello
6523 message has informed the client about his choice.)
6526 *) Load error codes if they are not already present instead of using a
6527 static variable. This allows them to be cleanly unloaded and reloaded.
6530 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6532 *) Introduce limits to prevent malicious keys being able to
6533 cause a denial of service. (CVE-2006-2940)
6534 [Steve Henson, Bodo Moeller]
6536 *) Fix ASN.1 parsing of certain invalid structures that can result
6537 in a denial of service. (CVE-2006-2937) [Steve Henson]
6539 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6540 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6542 *) Fix SSL client code which could crash if connecting to a
6543 malicious SSLv2 server. (CVE-2006-4343)
6544 [Tavis Ormandy and Will Drewry, Google Security Team]
6546 *) Change ciphersuite string processing so that an explicit
6547 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6548 will no longer include "AES128-SHA"), and any other similar
6549 ciphersuite (same bitmap) from *other* protocol versions (so that
6550 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6551 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6552 changes from 0.9.8b and 0.9.8d.
6555 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6557 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6558 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6560 *) Change the Unix randomness entropy gathering to use poll() when
6561 possible instead of select(), since the latter has some
6562 undesirable limitations.
6563 [Darryl Miles via Richard Levitte and Bodo Moeller]
6565 *) Disable rogue ciphersuites:
6567 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6568 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6569 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6571 The latter two were purportedly from
6572 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6575 Also deactivate the remaining ciphersuites from
6576 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6577 unofficial, and the ID has long expired.
6580 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6581 dual-core machines) and other potential thread-safety issues.
6584 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6586 *) Adapt fipsld and the build system to link against the validated FIPS
6587 module in FIPS mode.
6590 *) Fixes for VC++ 2005 build under Windows.
6593 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6594 from a Windows bash shell such as MSYS. It is autodetected from the
6595 "config" script when run from a VC++ environment. Modify standard VC++
6596 build to use fipscanister.o from the GNU make build.
6599 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6601 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6602 The value now differs depending on if you build for FIPS or not.
6603 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6604 safely run with a non-FIPSed libcrypto, as it may crash because of
6605 the difference induced by this change.
6608 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6610 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6611 (part of SSL_OP_ALL). This option used to disable the
6612 countermeasure against man-in-the-middle protocol-version
6613 rollback in the SSL 2.0 server implementation, which is a bad
6614 idea. (CVE-2005-2969)
6616 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6617 for Information Security, National Institute of Advanced Industrial
6618 Science and Technology [AIST], Japan)]
6620 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6621 mainly for FIPS compliance and not fully integrated at this stage.
6624 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6625 the exponentiation using a fixed-length exponent. (Otherwise,
6626 the information leaked through timing could expose the secret key
6627 after many signatures; cf. Bleichenbacher's attack on DSA with
6631 *) Make a new fixed-window mod_exp implementation the default for
6632 RSA, DSA, and DH private-key operations so that the sequence of
6633 squares and multiplies and the memory access pattern are
6634 independent of the particular secret key. This will mitigate
6635 cache-timing and potential related attacks.
6637 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6638 and this is automatically used by BN_mod_exp_mont() if the new flag
6639 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6640 will use this BN flag for private exponents unless the flag
6641 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6642 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6644 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6646 *) Change the client implementation for SSLv23_method() and
6647 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6648 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6649 (Previously, the SSL 2.0 backwards compatible Client Hello
6650 message format would be used even with SSL_OP_NO_SSLv2.)
6653 *) Add support for smime-type MIME parameter in S/MIME messages which some
6657 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6658 a threadsafe manner. Modify rsa code to use new function and add calls
6659 to dsa and dh code (which had race conditions before).
6662 *) Include the fixed error library code in the C error file definitions
6663 instead of fixing them up at runtime. This keeps the error code
6664 structures constant.
6667 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6669 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6672 *) Fixes for newer kerberos headers. NB: the casts are needed because
6673 the 'length' field is signed on one version and unsigned on another
6674 with no (?) obvious way to tell the difference, without these VC++
6675 complains. Also the "definition" of FAR (blank) is no longer included
6676 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6677 some needed definitions.
6680 *) Undo Cygwin change.
6683 *) Added support for proxy certificates according to RFC 3820.
6684 Because they may be a security thread to unaware applications,
6685 they must be explicitly allowed in run-time. See
6686 docs/HOWTO/proxy_certificates.txt for further information.
6689 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6691 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6692 server and client random values. Previously
6693 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6694 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6696 This change has negligible security impact because:
6698 1. Server and client random values still have 24 bytes of pseudo random
6701 2. Server and client random values are sent in the clear in the initial
6704 3. The master secret is derived using the premaster secret (48 bytes in
6705 size for static RSA ciphersuites) as well as client server and random
6708 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6711 [Stephen Henson, reported by UK NISCC]
6713 *) Use Windows randomness collection on Cygwin.
6716 *) Fix hang in EGD/PRNGD query when communication socket is closed
6717 prematurely by EGD/PRNGD.
6718 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6720 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6723 *) Back-port of selected performance improvements from development
6724 branch, as well as improved support for PowerPC platforms.
6727 *) Add lots of checks for memory allocation failure, error codes to indicate
6728 failure and freeing up memory if a failure occurs.
6729 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6731 *) Add new -passin argument to dgst.
6734 *) Perform some character comparisons of different types in X509_NAME_cmp:
6735 this is needed for some certificates that re-encode DNs into UTF8Strings
6736 (in violation of RFC3280) and can't or won't issue name rollover
6740 *) Make an explicit check during certificate validation to see that
6741 the CA setting in each certificate on the chain is correct. As a
6742 side effect always do the following basic checks on extensions,
6743 not just when there's an associated purpose to the check:
6745 - if there is an unhandled critical extension (unless the user
6746 has chosen to ignore this fault)
6747 - if the path length has been exceeded (if one is set at all)
6748 - that certain extensions fit the associated purpose (if one has
6752 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6754 *) Avoid a race condition when CRLs are checked in a multi threaded
6755 environment. This would happen due to the reordering of the revoked
6756 entries during signature checking and serial number lookup. Now the
6757 encoding is cached and the serial number sort performed under a lock.
6758 Add new STACK function sk_is_sorted().
6761 *) Add Delta CRL to the extension code.
6764 *) Various fixes to s3_pkt.c so alerts are sent properly.
6765 [David Holmes <d.holmes@f5.com>]
6767 *) Reduce the chances of duplicate issuer name and serial numbers (in
6768 violation of RFC3280) using the OpenSSL certificate creation utilities.
6769 This is done by creating a random 64 bit value for the initial serial
6770 number when a serial number file is created or when a self signed
6771 certificate is created using 'openssl req -x509'. The initial serial
6772 number file is created using 'openssl x509 -next_serial' in CA.pl
6773 rather than being initialized to 1.
6776 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6778 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6779 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6780 [Joe Orton, Steve Henson]
6782 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6784 [Joe Orton, Steve Henson]
6786 *) Make it possible to have multiple active certificates with the same
6787 subject in the CA index file. This is done only if the keyword
6788 'unique_subject' is set to 'no' in the main CA section (default
6789 if 'CA_default') of the configuration file. The value is saved
6790 with the database itself in a separate index attribute file,
6791 named like the index file with '.attr' appended to the name.
6794 *) X509 verify fixes. Disable broken certificate workarounds when
6795 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6796 keyUsage extension present. Don't accept CRLs with unhandled critical
6797 extensions: since verify currently doesn't process CRL extensions this
6798 rejects a CRL with *any* critical extensions. Add new verify error codes
6802 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6803 A clarification of RFC2560 will require the use of OCTET STRINGs and
6804 some implementations cannot handle the current raw format. Since OpenSSL
6805 copies and compares OCSP nonces as opaque blobs without any attempt at
6806 parsing them this should not create any compatibility issues.
6809 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6810 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6811 this HMAC (and other) operations are several times slower than OpenSSL
6815 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6816 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6818 *) Use the correct content when signing type "other".
6821 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6823 *) Fix various bugs revealed by running the NISCC test suite:
6825 Stop out of bounds reads in the ASN1 code when presented with
6826 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6828 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6830 If verify callback ignores invalid public key errors don't try to check
6831 certificate signature with the NULL public key.
6835 *) New -ignore_err option in ocsp application to stop the server
6836 exiting on the first error in a request.
6839 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6840 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6844 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6845 extra data after the compression methods not only for TLS 1.0
6846 but also for SSL 3.0 (as required by the specification).
6847 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6849 *) Change X509_certificate_type() to mark the key as exported/exportable
6850 when it's 512 *bits* long, not 512 bytes.
6853 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6854 blocks during encryption.
6857 *) Various fixes to base64 BIO and non blocking I/O. On write
6858 flushes were not handled properly if the BIO retried. On read
6859 data was not being buffered properly and had various logic bugs.
6860 This also affects blocking I/O when the data being decoded is a
6864 *) Various S/MIME bugfixes and compatibility changes:
6865 output correct application/pkcs7 MIME type if
6866 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6867 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6868 of files as .eml work). Correctly handle very long lines in MIME
6872 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6874 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6875 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6876 a protocol version number mismatch like a decryption error
6877 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6880 *) Turn on RSA blinding by default in the default implementation
6881 to avoid a timing attack. Applications that don't want it can call
6882 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6883 They would be ill-advised to do so in most cases.
6884 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6886 *) Change RSA blinding code so that it works when the PRNG is not
6887 seeded (in this case, the secret RSA exponent is abused as
6888 an unpredictable seed -- if it is not unpredictable, there
6889 is no point in blinding anyway). Make RSA blinding thread-safe
6890 by remembering the creator's thread ID in rsa->blinding and
6891 having all other threads use local one-time blinding factors
6892 (this requires more computation than sharing rsa->blinding, but
6893 avoids excessive locking; and if an RSA object is not shared
6894 between threads, blinding will still be very fast).
6897 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6898 ENGINE as defaults for all supported algorithms irrespective of
6899 the 'flags' parameter. 'flags' is now honoured, so applications
6900 should make sure they are passing it correctly.
6903 *) Target "mingw" now allows native Windows code to be generated in
6904 the Cygwin environment as well as with the MinGW compiler.
6907 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6909 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6910 via timing by performing a MAC computation even if incorrect
6911 block cipher padding has been found. This is a countermeasure
6912 against active attacks where the attacker has to distinguish
6913 between bad padding and a MAC verification error. (CVE-2003-0078)
6915 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6916 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6917 Martin Vuagnoux (EPFL, Ilion)]
6919 *) Make the no-err option work as intended. The intention with no-err
6920 is not to have the whole error stack handling routines removed from
6921 libcrypto, it's only intended to remove all the function name and
6922 reason texts, thereby removing some of the footprint that may not
6923 be interesting if those errors aren't displayed anyway.
6925 NOTE: it's still possible for any application or module to have it's
6926 own set of error texts inserted. The routines are there, just not
6927 used by default when no-err is given.
6930 *) Add support for FreeBSD on IA64.
6931 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6933 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6934 Kerberos function mit_des_cbc_cksum(). Before this change,
6935 the value returned by DES_cbc_cksum() was like the one from
6936 mit_des_cbc_cksum(), except the bytes were swapped.
6937 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6939 *) Allow an application to disable the automatic SSL chain building.
6940 Before this a rather primitive chain build was always performed in
6941 ssl3_output_cert_chain(): an application had no way to send the
6942 correct chain if the automatic operation produced an incorrect result.
6944 Now the chain builder is disabled if either:
6946 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6948 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6950 The reasoning behind this is that an application would not want the
6951 auto chain building to take place if extra chain certificates are
6952 present and it might also want a means of sending no additional
6953 certificates (for example the chain has two certificates and the
6957 *) Add the possibility to build without the ENGINE framework.
6958 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6960 *) Under Win32 gmtime() can return NULL: check return value in
6961 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6964 *) DSA routines: under certain error conditions uninitialized BN objects
6965 could be freed. Solution: make sure initialization is performed early
6966 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6967 Nils Larsch <nla@trustcenter.de> via PR#459)
6970 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6971 checked on reconnect on the client side, therefore session resumption
6972 could still fail with a "ssl session id is different" error. This
6973 behaviour is masked when SSL_OP_ALL is used due to
6974 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6975 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6976 followup to PR #377.
6979 *) IA-32 assembler support enhancements: unified ELF targets, support
6980 for SCO/Caldera platforms, fix for Cygwin shared build.
6983 *) Add support for FreeBSD on sparc64. As a consequence, support for
6984 FreeBSD on non-x86 processors is separate from x86 processors on
6985 the config script, much like the NetBSD support.
6986 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6988 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6990 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6993 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6994 code (06) was taken as the first octet of the session ID and the last
6995 octet was ignored consequently. As a result SSLv2 client side session
6996 caching could not have worked due to the session ID mismatch between
6998 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7002 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7003 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7007 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7008 seems that in spite of existing for more than a year, many application
7009 author have done nothing to provide the necessary callbacks, which
7010 means that this particular engine will not work properly anywhere.
7011 This is a very unfortunate situation which forces us, in the name
7012 of usability, to give the hw_ncipher.c a static lock, which is part
7014 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7015 appear in 0.9.8 or later. We EXPECT application authors to have
7016 dealt properly with this when 0.9.8 is released (unless we actually
7017 make such changes in the libcrypto locking code that changes will
7018 have to be made anyway).
7021 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7022 octets have been read, EOF or an error occurs. Without this change
7023 some truncated ASN1 structures will not produce an error.
7026 *) Disable Heimdal support, since it hasn't been fully implemented.
7027 Still give the possibility to force the use of Heimdal, but with
7028 warnings and a request that patches get sent to openssl-dev.
7031 *) Add the VC-CE target, introduce the WINCE sysname, and add
7032 INSTALL.WCE and appropriate conditionals to make it build.
7033 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7035 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7036 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7037 edit numbers of the version.
7038 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7040 *) Introduce safe string copy and catenation functions
7041 (BUF_strlcpy() and BUF_strlcat()).
7042 [Ben Laurie (CHATS) and Richard Levitte]
7044 *) Avoid using fixed-size buffers for one-line DNs.
7045 [Ben Laurie (CHATS)]
7047 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7048 resizing buffers containing secrets, and use where appropriate.
7049 [Ben Laurie (CHATS)]
7051 *) Avoid using fixed size buffers for configuration file location.
7052 [Ben Laurie (CHATS)]
7054 *) Avoid filename truncation for various CA files.
7055 [Ben Laurie (CHATS)]
7057 *) Use sizeof in preference to magic numbers.
7058 [Ben Laurie (CHATS)]
7060 *) Avoid filename truncation in cert requests.
7061 [Ben Laurie (CHATS)]
7063 *) Add assertions to check for (supposedly impossible) buffer
7065 [Ben Laurie (CHATS)]
7067 *) Don't cache truncated DNS entries in the local cache (this could
7068 potentially lead to a spoofing attack).
7069 [Ben Laurie (CHATS)]
7071 *) Fix various buffers to be large enough for hex/decimal
7072 representations in a platform independent manner.
7073 [Ben Laurie (CHATS)]
7075 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7076 resizing buffers containing secrets, and use where appropriate.
7077 [Ben Laurie (CHATS)]
7079 *) Add BIO_indent() to avoid much slightly worrying code to do
7081 [Ben Laurie (CHATS)]
7083 *) Convert sprintf()/BIO_puts() to BIO_printf().
7084 [Ben Laurie (CHATS)]
7086 *) buffer_gets() could terminate with the buffer only half
7088 [Ben Laurie (CHATS)]
7090 *) Add assertions to prevent user-supplied crypto functions from
7091 overflowing internal buffers by having large block sizes, etc.
7092 [Ben Laurie (CHATS)]
7094 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7096 [Ben Laurie (CHATS)]
7098 *) Eliminate unused copy of key in RC4.
7099 [Ben Laurie (CHATS)]
7101 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7102 [Ben Laurie (CHATS)]
7104 *) Fix off-by-one error in EGD path.
7105 [Ben Laurie (CHATS)]
7107 *) If RANDFILE path is too long, ignore instead of truncating.
7108 [Ben Laurie (CHATS)]
7110 *) Eliminate unused and incorrectly sized X.509 structure
7112 [Ben Laurie (CHATS)]
7114 *) Eliminate unused and dangerous function knumber().
7115 [Ben Laurie (CHATS)]
7117 *) Eliminate unused and dangerous structure, KSSL_ERR.
7118 [Ben Laurie (CHATS)]
7120 *) Protect against overlong session ID context length in an encoded
7121 session object. Since these are local, this does not appear to be
7123 [Ben Laurie (CHATS)]
7125 *) Change from security patch (see 0.9.6e below) that did not affect
7126 the 0.9.6 release series:
7128 Remote buffer overflow in SSL3 protocol - an attacker could
7129 supply an oversized master key in Kerberos-enabled versions.
7131 [Ben Laurie (CHATS)]
7133 *) Change the SSL kerb5 codes to match RFC 2712.
7136 *) Make -nameopt work fully for req and add -reqopt switch.
7137 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7139 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7140 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7142 *) Make sure tests can be performed even if the corresponding algorithms
7143 have been removed entirely. This was also the last step to make
7144 OpenSSL compilable with DJGPP under all reasonable conditions.
7145 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7147 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7148 to allow version independent disabling of normally unselected ciphers,
7149 which may be activated as a side-effect of selecting a single cipher.
7151 (E.g., cipher list string "RSA" enables ciphersuites that are left
7152 out of "ALL" because they do not provide symmetric encryption.
7153 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7154 [Lutz Jaenicke, Bodo Moeller]
7156 *) Add appropriate support for separate platform-dependent build
7157 directories. The recommended way to make a platform-dependent
7158 build directory is the following (tested on Linux), maybe with
7161 # Place yourself outside of the OpenSSL source tree. In
7162 # this example, the environment variable OPENSSL_SOURCE
7163 # is assumed to contain the absolute OpenSSL source directory.
7164 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7165 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7166 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7167 mkdir -p `dirname $F`
7168 ln -s $OPENSSL_SOURCE/$F $F
7171 To be absolutely sure not to disturb the source tree, a "make clean"
7172 is a good thing. If it isn't successful, don't worry about it,
7173 it probably means the source directory is very clean.
7176 *) Make sure any ENGINE control commands make local copies of string
7177 pointers passed to them whenever necessary. Otherwise it is possible
7178 the caller may have overwritten (or deallocated) the original string
7179 data when a later ENGINE operation tries to use the stored values.
7180 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7182 *) Improve diagnostics in file reading and command-line digests.
7183 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7185 *) Add AES modes CFB and OFB to the object database. Correct an
7186 error in AES-CFB decryption.
7189 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7190 allows existing EVP_CIPHER_CTX structures to be reused after
7191 calling EVP_*Final(). This behaviour is used by encryption
7192 BIOs and some applications. This has the side effect that
7193 applications must explicitly clean up cipher contexts with
7194 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7197 *) Check the values of dna and dnb in bn_mul_recursive before calling
7198 bn_mul_comba (a non zero value means the a or b arrays do not contain
7199 n2 elements) and fallback to bn_mul_normal if either is not zero.
7202 *) Fix escaping of non-ASCII characters when using the -subj option
7203 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7206 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7207 form for "surname", serialNumber has no short form.
7208 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7209 therefore remove "mail" short name for "internet 7".
7210 The OID for unique identifiers in X509 certificates is
7211 x500UniqueIdentifier, not uniqueIdentifier.
7212 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7215 *) Add an "init" command to the ENGINE config module and auto initialize
7216 ENGINEs. Without any "init" command the ENGINE will be initialized
7217 after all ctrl commands have been executed on it. If init=1 the
7218 ENGINE is initialized at that point (ctrls before that point are run
7219 on the uninitialized ENGINE and after on the initialized one). If
7220 init=0 then the ENGINE will not be initialized at all.
7223 *) Fix the 'app_verify_callback' interface so that the user-defined
7224 argument is actually passed to the callback: In the
7225 SSL_CTX_set_cert_verify_callback() prototype, the callback
7226 declaration has been changed from
7229 int (*cb)(X509_STORE_CTX *,void *);
7230 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7231 i=s->ctx->app_verify_callback(&ctx)
7232 has been changed into
7233 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7235 To update applications using SSL_CTX_set_cert_verify_callback(),
7236 a dummy argument can be added to their callback functions.
7237 [D. K. Smetters <smetters@parc.xerox.com>]
7239 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7240 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7242 *) Add and OPENSSL_LOAD_CONF define which will cause
7243 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7244 This allows older applications to transparently support certain
7245 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7246 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7247 load the config file and OPENSSL_add_all_algorithms_conf() which will
7248 always load it have also been added.
7251 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7252 Adjust NIDs and EVP layer.
7253 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7255 *) Config modules support in openssl utility.
7257 Most commands now load modules from the config file,
7258 though in a few (such as version) this isn't done
7259 because it couldn't be used for anything.
7261 In the case of ca and req the config file used is
7262 the same as the utility itself: that is the -config
7263 command line option can be used to specify an
7267 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7268 use "openssl_conf" if filename is NULL use default openssl config file.
7271 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7272 config section name. Add a new flag to tolerate a missing config file
7273 and move code to CONF_modules_load_file().
7276 *) Support for crypto accelerator cards from Accelerated Encryption
7277 Processing, www.aep.ie. (Use engine 'aep')
7278 The support was copied from 0.9.6c [engine] and adapted/corrected
7279 to work with the new engine framework.
7280 [AEP Inc. and Richard Levitte]
7282 *) Support for SureWare crypto accelerator cards from Baltimore
7283 Technologies. (Use engine 'sureware')
7284 The support was copied from 0.9.6c [engine] and adapted
7285 to work with the new engine framework.
7288 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7289 make the newer ENGINE framework commands for the CHIL engine work.
7290 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7292 *) Make it possible to produce shared libraries on ReliantUNIX.
7293 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7295 *) Add the configuration target debug-linux-ppro.
7296 Make 'openssl rsa' use the general key loading routines
7297 implemented in apps.c, and make those routines able to
7298 handle the key format FORMAT_NETSCAPE and the variant
7300 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7302 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7303 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7305 *) Add -keyform to rsautl, and document -engine.
7306 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7308 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7309 BIO_R_NO_SUCH_FILE error code rather than the generic
7310 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7313 *) Add new functions
7315 ERR_peek_last_error_line
7316 ERR_peek_last_error_line_data.
7317 These are similar to
7320 ERR_peek_error_line_data,
7321 but report on the latest error recorded rather than the first one
7322 still in the error queue.
7323 [Ben Laurie, Bodo Moeller]
7325 *) default_algorithms option in ENGINE config module. This allows things
7327 default_algorithms = ALL
7328 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7331 *) Preliminary ENGINE config module.
7334 *) New experimental application configuration code.
7337 *) Change the AES code to follow the same name structure as all other
7338 symmetric ciphers, and behave the same way. Move everything to
7339 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7340 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7342 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7343 [Ben Laurie and Theo de Raadt]
7345 *) Add option to output public keys in req command.
7346 [Massimiliano Pala madwolf@openca.org]
7348 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7349 (up to about 10% better than before for P-192 and P-224).
7352 *) New functions/macros
7354 SSL_CTX_set_msg_callback(ctx, cb)
7355 SSL_CTX_set_msg_callback_arg(ctx, arg)
7356 SSL_set_msg_callback(ssl, cb)
7357 SSL_set_msg_callback_arg(ssl, arg)
7359 to request calling a callback function
7361 void cb(int write_p, int version, int content_type,
7362 const void *buf, size_t len, SSL *ssl, void *arg)
7364 whenever a protocol message has been completely received
7365 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7366 protocol version according to which the SSL library interprets
7367 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7368 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7369 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7370 specification (change_cipher_spec(20), alert(21), handshake(22)).
7371 'buf' and 'len' point to the actual message, 'ssl' to the
7372 SSL object, and 'arg' is the application-defined value set by
7373 SSL[_CTX]_set_msg_callback_arg().
7375 'openssl s_client' and 'openssl s_server' have new '-msg' options
7376 to enable a callback that displays all protocol messages.
7379 *) Change the shared library support so shared libraries are built as
7380 soon as the corresponding static library is finished, and thereby get
7381 openssl and the test programs linked against the shared library.
7382 This still only happens when the keyword "shard" has been given to
7383 the configuration scripts.
7385 NOTE: shared library support is still an experimental thing, and
7386 backward binary compatibility is still not guaranteed.
7387 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7389 *) Add support for Subject Information Access extension.
7390 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7392 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7393 additional bytes when new memory had to be allocated, not just
7394 when reusing an existing buffer.
7397 *) New command line and configuration option 'utf8' for the req command.
7398 This allows field values to be specified as UTF8 strings.
7401 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7402 runs for the former and machine-readable output for the latter.
7405 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7406 of the e-mail address in the DN (i.e., it will go into a certificate
7407 extension only). The new configuration file option 'email_in_dn = no'
7408 has the same effect.
7409 [Massimiliano Pala madwolf@openca.org]
7411 *) Change all functions with names starting with des_ to be starting
7412 with DES_ instead. Add wrappers that are compatible with libdes,
7413 but are named _ossl_old_des_*. Finally, add macros that map the
7414 des_* symbols to the corresponding _ossl_old_des_* if libdes
7415 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7416 desired, the des_* symbols will be mapped to DES_*, with one
7419 Since we provide two compatibility mappings, the user needs to
7420 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7421 compatibility is desired. The default (i.e., when that macro
7422 isn't defined) is OpenSSL 0.9.6c compatibility.
7424 There are also macros that enable and disable the support of old
7425 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7426 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7427 are defined, the default will apply: to support the old des routines.
7429 In either case, one must include openssl/des.h to get the correct
7430 definitions. Do not try to just include openssl/des_old.h, that
7433 NOTE: This is a major break of an old API into a new one. Software
7434 authors are encouraged to switch to the DES_ style functions. Some
7435 time in the future, des_old.h and the libdes compatibility functions
7436 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7437 default), and then completely removed.
7440 *) Test for certificates which contain unsupported critical extensions.
7441 If such a certificate is found during a verify operation it is
7442 rejected by default: this behaviour can be overridden by either
7443 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7444 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7445 X509_supported_extension() has also been added which returns 1 if a
7446 particular extension is supported.
7449 *) Modify the behaviour of EVP cipher functions in similar way to digests
7450 to retain compatibility with existing code.
7453 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7454 compatibility with existing code. In particular the 'ctx' parameter does
7455 not have to be to be initialized before the call to EVP_DigestInit() and
7456 it is tidied up after a call to EVP_DigestFinal(). New function
7457 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7458 EVP_MD_CTX_copy() changed to not require the destination to be
7459 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7460 requires the destination to be valid.
7462 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7463 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7466 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7467 so that complete 'Handshake' protocol structures are kept in memory
7468 instead of overwriting 'msg_type' and 'length' with 'body' data.
7471 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7472 [Massimo Santin via Richard Levitte]
7474 *) Major restructuring to the underlying ENGINE code. This includes
7475 reduction of linker bloat, separation of pure "ENGINE" manipulation
7476 (initialisation, etc) from functionality dealing with implementations
7477 of specific crypto interfaces. This change also introduces integrated
7478 support for symmetric ciphers and digest implementations - so ENGINEs
7479 can now accelerate these by providing EVP_CIPHER and EVP_MD
7480 implementations of their own. This is detailed in crypto/engine/README
7481 as it couldn't be adequately described here. However, there are a few
7482 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7483 were changed in the original introduction of ENGINE code have now
7484 reverted back - the hooking from this code to ENGINE is now a good
7485 deal more passive and at run-time, operations deal directly with
7486 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7487 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7488 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7489 they were not being used by the framework as there is no concept of a
7490 BIGNUM_METHOD and they could not be generalised to the new
7491 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7492 ENGINE_cpy() has been removed as it cannot be consistently defined in
7496 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7499 *) Change mkdef.pl to sort symbols that get the same entry number,
7500 and make sure the automatically generated functions ERR_load_*
7501 become part of libeay.num as well.
7504 *) New function SSL_renegotiate_pending(). This returns true once
7505 renegotiation has been requested (either SSL_renegotiate() call
7506 or HelloRequest/ClientHello received from the peer) and becomes
7507 false once a handshake has been completed.
7508 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7509 sends a HelloRequest, but does not ensure that a handshake takes
7510 place. SSL_renegotiate_pending() is useful for checking if the
7511 client has followed the request.)
7514 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7515 By default, clients may request session resumption even during
7516 renegotiation (if session ID contexts permit); with this option,
7517 session resumption is possible only in the first handshake.
7519 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7520 more bits available for options that should not be part of
7521 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7524 *) Add some demos for certificate and certificate request creation.
7527 *) Make maximum certificate chain size accepted from the peer application
7528 settable (SSL*_get/set_max_cert_list()), as proposed by
7529 "Douglas E. Engert" <deengert@anl.gov>.
7532 *) Add support for shared libraries for Unixware-7
7533 (Boyd Lynn Gerber <gerberb@zenez.com>).
7536 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7537 be done prior to destruction. Use this to unload error strings from
7538 ENGINEs that load their own error strings. NB: This adds two new API
7539 functions to "get" and "set" this destroy handler in an ENGINE.
7542 *) Alter all existing ENGINE implementations (except "openssl" and
7543 "openbsd") to dynamically instantiate their own error strings. This
7544 makes them more flexible to be built both as statically-linked ENGINEs
7545 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7546 Also, add stub code to each that makes building them as self-contained
7547 shared-libraries easier (see README.ENGINE).
7550 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7551 implementations into applications that are completely implemented in
7552 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7553 commands that can be used to configure what shared-library to load and
7554 to control aspects of the way it is handled. Also, made an update to
7555 the README.ENGINE file that brings its information up-to-date and
7556 provides some information and instructions on the "dynamic" ENGINE
7557 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7560 *) Make it possible to unload ranges of ERR strings with a new
7561 "ERR_unload_strings" function.
7564 *) Add a copy() function to EVP_MD.
7567 *) Make EVP_MD routines take a context pointer instead of just the
7568 md_data void pointer.
7571 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7572 that the digest can only process a single chunk of data
7573 (typically because it is provided by a piece of
7574 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7575 is only going to provide a single chunk of data, and hence the
7576 framework needn't accumulate the data for oneshot drivers.
7579 *) As with "ERR", make it possible to replace the underlying "ex_data"
7580 functions. This change also alters the storage and management of global
7581 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7582 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7583 index counters. The API functions that use this state have been changed
7584 to take a "class_index" rather than pointers to the class's local STACK
7585 and counter, and there is now an API function to dynamically create new
7586 classes. This centralisation allows us to (a) plug a lot of the
7587 thread-safety problems that existed, and (b) makes it possible to clean
7588 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7589 such data would previously have always leaked in application code and
7590 workarounds were in place to make the memory debugging turn a blind eye
7591 to it. Application code that doesn't use this new function will still
7592 leak as before, but their memory debugging output will announce it now
7593 rather than letting it slide.
7595 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7596 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7597 has a return value to indicate success or failure.
7600 *) Make it possible to replace the underlying "ERR" functions such that the
7601 global state (2 LHASH tables and 2 locks) is only used by the "default"
7602 implementation. This change also adds two functions to "get" and "set"
7603 the implementation prior to it being automatically set the first time
7604 any other ERR function takes place. Ie. an application can call "get",
7605 pass the return value to a module it has just loaded, and that module
7606 can call its own "set" function using that value. This means the
7607 module's "ERR" operations will use (and modify) the error state in the
7608 application and not in its own statically linked copy of OpenSSL code.
7611 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7612 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7613 the operation, and provides a more encapsulated way for external code
7614 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7615 to use these functions rather than manually incrementing the counts.
7617 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7620 *) Add EVP test program.
7623 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7626 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7627 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7628 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7629 These allow a CRL to be built without having to access X509_CRL fields
7630 directly. Modify 'ca' application to use new functions.
7633 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7634 bug workarounds. Rollback attack detection is a security feature.
7635 The problem will only arise on OpenSSL servers when TLSv1 is not
7636 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7637 Software authors not wanting to support TLSv1 will have special reasons
7638 for their choice and can explicitly enable this option.
7639 [Bodo Moeller, Lutz Jaenicke]
7641 *) Rationalise EVP so it can be extended: don't include a union of
7642 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7643 (similar to those existing for EVP_CIPHER_CTX).
7648 EVP_MD_CTX_init(&md); /* new function call */
7649 EVP_DigestInit(&md, EVP_sha1());
7650 EVP_DigestUpdate(&md, in, len);
7651 EVP_DigestFinal(&md, out, NULL);
7652 EVP_MD_CTX_cleanup(&md); /* new function call */
7656 *) Make DES key schedule conform to the usual scheme, as well as
7657 correcting its structure. This means that calls to DES functions
7658 now have to pass a pointer to a des_key_schedule instead of a
7659 plain des_key_schedule (which was actually always a pointer
7662 des_key_schedule ks;
7664 des_set_key_checked(..., &ks);
7665 des_ncbc_encrypt(..., &ks, ...);
7667 (Note that a later change renames 'des_...' into 'DES_...'.)
7670 *) Initial reduction of linker bloat: the use of some functions, such as
7671 PEM causes large amounts of unused functions to be linked in due to
7672 poor organisation. For example pem_all.c contains every PEM function
7673 which has a knock on effect of linking in large amounts of (unused)
7674 ASN1 code. Grouping together similar functions and splitting unrelated
7675 functions prevents this.
7678 *) Cleanup of EVP macros.
7681 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7682 correct _ecb suffix.
7685 *) Add initial OCSP responder support to ocsp application. The
7686 revocation information is handled using the text based index
7687 use by the ca application. The responder can either handle
7688 requests generated internally, supplied in files (for example
7689 via a CGI script) or using an internal minimal server.
7692 *) Add configuration choices to get zlib compression for TLS.
7695 *) Changes to Kerberos SSL for RFC 2712 compliance:
7696 1. Implemented real KerberosWrapper, instead of just using
7697 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7698 2. Implemented optional authenticator field of KerberosWrapper.
7700 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7701 and authenticator structs; see crypto/krb5/.
7703 Generalized Kerberos calls to support multiple Kerberos libraries.
7704 [Vern Staats <staatsvr@asc.hpc.mil>,
7705 Jeffrey Altman <jaltman@columbia.edu>
7706 via Richard Levitte]
7708 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7709 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7710 values for each of the key sizes rather than having just
7711 parameters (and 'speed' generating keys each time).
7714 *) Speed up EVP routines.
7717 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7718 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7719 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7720 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7722 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7723 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7724 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7727 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7729 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7732 *) Added the OS2-EMX target.
7733 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7735 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7736 to support NCONF routines in extension code. New function CONF_set_nconf()
7737 to allow functions which take an NCONF to also handle the old LHASH
7738 structure: this means that the old CONF compatible routines can be
7739 retained (in particular wrt extensions) without having to duplicate the
7740 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7743 *) Enhance the general user interface with mechanisms for inner control
7744 and with possibilities to have yes/no kind of prompts.
7747 *) Change all calls to low level digest routines in the library and
7748 applications to use EVP. Add missing calls to HMAC_cleanup() and
7749 don't assume HMAC_CTX can be copied using memcpy().
7750 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7752 *) Add the possibility to control engines through control names but with
7753 arbitrary arguments instead of just a string.
7754 Change the key loaders to take a UI_METHOD instead of a callback
7755 function pointer. NOTE: this breaks binary compatibility with earlier
7756 versions of OpenSSL [engine].
7757 Adapt the nCipher code for these new conditions and add a card insertion
7761 *) Enhance the general user interface with mechanisms to better support
7762 dialog box interfaces, application-defined prompts, the possibility
7763 to use defaults (for example default passwords from somewhere else)
7764 and interrupts/cancellations.
7767 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7768 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7771 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7772 tidy up some unnecessarily weird code in 'sk_new()').
7773 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7775 *) Change the key loading routines for ENGINEs to use the same kind
7776 callback (pem_password_cb) as all other routines that need this
7780 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7781 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7782 than this minimum value is recommended.
7785 *) New random seeder for OpenVMS, using the system process statistics
7786 that are easily reachable.
7789 *) Windows apparently can't transparently handle global
7790 variables defined in DLLs. Initialisations such as:
7792 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7794 won't compile. This is used by the any applications that need to
7795 declare their own ASN1 modules. This was fixed by adding the option
7796 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7797 needed for static libraries under Win32.
7800 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7801 setting of purpose and trust fields. New X509_STORE trust and
7802 purpose functions and tidy up setting in other SSL functions.
7805 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7806 structure. These are inherited by X509_STORE_CTX when it is
7807 initialised. This allows various defaults to be set in the
7808 X509_STORE structure (such as flags for CRL checking and custom
7809 purpose or trust settings) for functions which only use X509_STORE_CTX
7810 internally such as S/MIME.
7812 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7813 trust settings if they are not set in X509_STORE. This allows X509_STORE
7814 purposes and trust (in S/MIME for example) to override any set by default.
7816 Add command line options for CRL checking to smime, s_client and s_server
7820 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7821 are set then the CRL is looked up in the X509_STORE structure and
7822 its validity and signature checked, then if the certificate is found
7823 in the CRL the verify fails with a revoked error.
7825 Various new CRL related callbacks added to X509_STORE_CTX structure.
7827 Command line options added to 'verify' application to support this.
7829 This needs some additional work, such as being able to handle multiple
7830 CRLs with different times, extension based lookup (rather than just
7831 by subject name) and ultimately more complete V2 CRL extension
7835 *) Add a general user interface API (crypto/ui/). This is designed
7836 to replace things like des_read_password and friends (backward
7837 compatibility functions using this new API are provided).
7838 The purpose is to remove prompting functions from the DES code
7839 section as well as provide for prompting through dialog boxes in
7840 a window system and the like.
7843 *) Add "ex_data" support to ENGINE so implementations can add state at a
7844 per-structure level rather than having to store it globally.
7847 *) Make it possible for ENGINE structures to be copied when retrieved by
7848 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7849 This causes the "original" ENGINE structure to act like a template,
7850 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7851 operational state can be localised to each ENGINE structure, despite the
7852 fact they all share the same "methods". New ENGINE structures returned in
7853 this case have no functional references and the return value is the single
7854 structural reference. This matches the single structural reference returned
7855 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7859 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7860 needs to match any other type at all we need to manually clear the
7864 *) Changes to the "openssl engine" utility to include;
7865 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7866 about an ENGINE's available control commands.
7867 - executing control commands from command line arguments using the
7868 '-pre' and '-post' switches. '-post' is only used if '-t' is
7869 specified and the ENGINE is successfully initialised. The syntax for
7870 the individual commands are colon-separated, for example;
7871 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7874 *) New dynamic control command support for ENGINEs. ENGINEs can now
7875 declare their own commands (numbers), names (strings), descriptions,
7876 and input types for run-time discovery by calling applications. A
7877 subset of these commands are implicitly classed as "executable"
7878 depending on their input type, and only these can be invoked through
7879 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7880 can be based on user input, config files, etc). The distinction is
7881 that "executable" commands cannot return anything other than a boolean
7882 result and can only support numeric or string input, whereas some
7883 discoverable commands may only be for direct use through
7884 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7885 pointers, or other custom uses. The "executable" commands are to
7886 support parameterisations of ENGINE behaviour that can be
7887 unambiguously defined by ENGINEs and used consistently across any
7888 OpenSSL-based application. Commands have been added to all the
7889 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7890 control over shared-library paths without source code alterations.
7893 *) Changed all ENGINE implementations to dynamically allocate their
7894 ENGINEs rather than declaring them statically. Apart from this being
7895 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7896 this also allows the implementations to compile without using the
7897 internal engine_int.h header.
7900 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7901 'const' value. Any code that should be able to modify a RAND_METHOD
7902 should already have non-const pointers to it (ie. they should only
7903 modify their own ones).
7906 *) Made a variety of little tweaks to the ENGINE code.
7907 - "atalla" and "ubsec" string definitions were moved from header files
7908 to C code. "nuron" string definitions were placed in variables
7909 rather than hard-coded - allowing parameterisation of these values
7910 later on via ctrl() commands.
7911 - Removed unused "#if 0"'d code.
7912 - Fixed engine list iteration code so it uses ENGINE_free() to release
7913 structural references.
7914 - Constified the RAND_METHOD element of ENGINE structures.
7915 - Constified various get/set functions as appropriate and added
7916 missing functions (including a catch-all ENGINE_cpy that duplicates
7917 all ENGINE values onto a new ENGINE except reference counts/state).
7918 - Removed NULL parameter checks in get/set functions. Setting a method
7919 or function to NULL is a way of cancelling out a previously set
7920 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7921 and doesn't justify the extra error symbols and code.
7922 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7923 flags from engine_int.h to engine.h.
7924 - Changed prototypes for ENGINE handler functions (init(), finish(),
7925 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7928 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7929 to the algorithm using long division. The binary algorithm can be
7930 used only if the modulus is odd. On 32-bit systems, it is faster
7931 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7932 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7933 up to 450 bits. In 64-bit environments, the binary algorithm
7934 appears to be advantageous for much longer moduli; here we use it
7935 for moduli up to 2048 bits.
7938 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7939 could not support the combine flag in choice fields.
7942 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7943 extensions from a certificate request to the certificate.
7946 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7947 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7948 file: this allows the display of the certificate about to be
7949 signed to be customised, to allow certain fields to be included
7950 or excluded and extension details. The old system didn't display
7951 multicharacter strings properly, omitted fields not in the policy
7952 and couldn't display additional details such as extensions.
7955 *) Function EC_POINTs_mul for multiple scalar multiplication
7956 of an arbitrary number of elliptic curve points
7957 \sum scalars[i]*points[i],
7958 optionally including the generator defined for the EC_GROUP:
7959 scalar*generator + \sum scalars[i]*points[i].
7961 EC_POINT_mul is a simple wrapper function for the typical case
7962 that the point list has just one item (besides the optional
7966 *) First EC_METHODs for curves over GF(p):
7968 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7969 operations and provides various method functions that can also
7970 operate with faster implementations of modular arithmetic.
7972 EC_GFp_mont_method() reuses most functions that are part of
7973 EC_GFp_simple_method, but uses Montgomery arithmetic.
7975 [Bodo Moeller; point addition and point doubling
7976 implementation directly derived from source code provided by
7977 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7979 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7980 crypto/ec/ec_lib.c):
7982 Curves are EC_GROUP objects (with an optional group generator)
7983 based on EC_METHODs that are built into the library.
7985 Points are EC_POINT objects based on EC_GROUP objects.
7987 Most of the framework would be able to handle curves over arbitrary
7988 finite fields, but as there are no obvious types for fields other
7989 than GF(p), some functions are limited to that for now.
7992 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7993 that the file contains a complete HTTP response.
7996 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7997 change the def and num file printf format specifier from "%-40sXXX"
7998 to "%-39s XXX". The latter will always guarantee a space after the
7999 field while the former will cause them to run together if the field
8000 is 40 of more characters long.
8003 *) Constify the cipher and digest 'method' functions and structures
8004 and modify related functions to take constant EVP_MD and EVP_CIPHER
8008 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8009 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8012 *) Modify EVP_Digest*() routines so they now return values. Although the
8013 internal software routines can never fail additional hardware versions
8017 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8019 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8020 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8023 ERR_R_NESTED_ASN1_ERROR
8025 ERR_R_MISSING_ASN1_EOS
8026 were 4 .. 9, conflicting with
8027 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8029 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8030 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8032 Add new error code 'ERR_R_INTERNAL_ERROR'.
8035 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8039 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8040 sets the subject name for a new request or supersedes the
8041 subject name in a given request. Formats that can be parsed are
8042 'CN=Some Name, OU=myOU, C=IT'
8044 'CN=Some Name/OU=myOU/C=IT'.
8046 Add options '-batch' and '-verbose' to 'openssl req'.
8047 [Massimiliano Pala <madwolf@hackmasters.net>]
8049 *) Introduce the possibility to access global variables through
8050 functions on platform were that's the best way to handle exporting
8051 global variables in shared libraries. To enable this functionality,
8052 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8053 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8054 is normally done by Configure or something similar).
8056 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8057 in the source file (foo.c) like this:
8059 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8060 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8062 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8063 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8065 OPENSSL_DECLARE_GLOBAL(int,foo);
8066 #define foo OPENSSL_GLOBAL_REF(foo)
8067 OPENSSL_DECLARE_GLOBAL(double,bar);
8068 #define bar OPENSSL_GLOBAL_REF(bar)
8070 The #defines are very important, and therefore so is including the
8071 header file everywhere where the defined globals are used.
8073 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8074 of ASN.1 items, but that structure is a bit different.
8076 The largest change is in util/mkdef.pl which has been enhanced with
8077 better and easier to understand logic to choose which symbols should
8078 go into the Windows .def files as well as a number of fixes and code
8079 cleanup (among others, algorithm keywords are now sorted
8080 lexicographically to avoid constant rewrites).
8083 *) In BN_div() keep a copy of the sign of 'num' before writing the
8084 result to 'rm' because if rm==num the value will be overwritten
8085 and produce the wrong result if 'num' is negative: this caused
8086 problems with BN_mod() and BN_nnmod().
8089 *) Function OCSP_request_verify(). This checks the signature on an
8090 OCSP request and verifies the signer certificate. The signer
8091 certificate is just checked for a generic purpose and OCSP request
8095 *) Add OCSP_check_validity() function to check the validity of OCSP
8096 responses. OCSP responses are prepared in real time and may only
8097 be a few seconds old. Simply checking that the current time lies
8098 between thisUpdate and nextUpdate max reject otherwise valid responses
8099 caused by either OCSP responder or client clock inaccuracy. Instead
8100 we allow thisUpdate and nextUpdate to fall within a certain period of
8101 the current time. The age of the response can also optionally be
8102 checked. Two new options -validity_period and -status_age added to
8106 *) If signature or public key algorithm is unrecognized print out its
8107 OID rather that just UNKNOWN.
8110 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8111 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8112 ID to be generated from the issuer certificate alone which can then be
8113 passed to OCSP_id_issuer_cmp().
8116 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8117 ASN1 modules to export functions returning ASN1_ITEM pointers
8118 instead of the ASN1_ITEM structures themselves. This adds several
8119 new macros which allow the underlying ASN1 function/structure to
8120 be accessed transparently. As a result code should not use ASN1_ITEM
8121 references directly (such as &X509_it) but instead use the relevant
8122 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8123 use of the new ASN1 code on platforms where exporting structures
8124 is problematical (for example in shared libraries) but exporting
8125 functions returning pointers to structures is not.
8128 *) Add support for overriding the generation of SSL/TLS session IDs.
8129 These callbacks can be registered either in an SSL_CTX or per SSL.
8130 The purpose of this is to allow applications to control, if they wish,
8131 the arbitrary values chosen for use as session IDs, particularly as it
8132 can be useful for session caching in multiple-server environments. A
8133 command-line switch for testing this (and any client code that wishes
8134 to use such a feature) has been added to "s_server".
8135 [Geoff Thorpe, Lutz Jaenicke]
8137 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8138 of the form '#if defined(...) || defined(...) || ...' and
8139 '#if !defined(...) && !defined(...) && ...'. This also avoids
8140 the growing number of special cases it was previously handling.
8143 *) Make all configuration macros available for application by making
8144 sure they are available in opensslconf.h, by giving them names starting
8145 with "OPENSSL_" to avoid conflicts with other packages and by making
8146 sure e_os2.h will cover all platform-specific cases together with
8148 Additionally, it is now possible to define configuration/platform-
8149 specific names (called "system identities"). In the C code, these
8150 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8151 macro with the name beginning with "OPENSSL_SYS_", which is determined
8152 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8156 *) New option -set_serial to 'req' and 'x509' this allows the serial
8157 number to use to be specified on the command line. Previously self
8158 signed certificates were hard coded with serial number 0 and the
8159 CA options of 'x509' had to use a serial number in a file which was
8163 *) New options to 'ca' utility to support V2 CRL entry extensions.
8164 Currently CRL reason, invalidity date and hold instruction are
8165 supported. Add new CRL extensions to V3 code and some new objects.
8168 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8169 disable standard block padding (aka PKCS#5 padding) in the EVP
8170 API, which was previously mandatory. This means that the data is
8171 not padded in any way and so the total length much be a multiple
8172 of the block size, otherwise an error occurs.
8175 *) Initial (incomplete) OCSP SSL support.
8178 *) New function OCSP_parse_url(). This splits up a URL into its host,
8179 port and path components: primarily to parse OCSP URLs. New -url
8180 option to ocsp utility.
8183 *) New nonce behavior. The return value of OCSP_check_nonce() now
8184 reflects the various checks performed. Applications can decide
8185 whether to tolerate certain situations such as an absent nonce
8186 in a response when one was present in a request: the ocsp application
8187 just prints out a warning. New function OCSP_add1_basic_nonce()
8188 this is to allow responders to include a nonce in a response even if
8189 the request is nonce-less.
8192 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8193 skipped when using openssl x509 multiple times on a single input file,
8194 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8197 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8198 set string type: to handle setting ASN1_TIME structures. Fix ca
8199 utility to correctly initialize revocation date of CRLs.
8202 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8203 the clients preferred ciphersuites and rather use its own preferences.
8204 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8205 Internet Explorer by ensuring unchanged hash method during stepup.
8206 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8209 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8210 to aes and add a new 'exist' option to print out symbols that don't
8214 *) Additional options to ocsp utility to allow flags to be set and
8215 additional certificates supplied.
8218 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8219 OCSP client a number of certificate to only verify the response
8223 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8224 handle the new API. Currently only ECB, CBC modes supported. Add new
8227 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8228 Encryption Standard (AES) Ciphersuites for Transport Layer
8229 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8230 not enabled by default and were not part of the "ALL" ciphersuite
8231 alias because they were not yet official; they could be
8232 explicitly requested by specifying the "AESdraft" ciphersuite
8233 group alias. In the final release of OpenSSL 0.9.7, the group
8234 alias is called "AES" and is part of "ALL".)
8235 [Ben Laurie, Steve Henson, Bodo Moeller]
8237 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8238 request to response.
8241 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8242 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8243 extract information from a certificate request. OCSP_response_create()
8244 creates a response and optionally adds a basic response structure.
8245 OCSP_basic_add1_status() adds a complete single response to a basic
8246 response and returns the OCSP_SINGLERESP structure just added (to allow
8247 extensions to be included for example). OCSP_basic_add1_cert() adds a
8248 certificate to a basic response and OCSP_basic_sign() signs a basic
8249 response with various flags. New helper functions ASN1_TIME_check()
8250 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8251 (converts ASN1_TIME to GeneralizedTime).
8254 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8255 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8256 structure from a certificate. X509_pubkey_digest() digests the public_key
8257 contents: this is used in various key identifiers.
8260 *) Make sk_sort() tolerate a NULL argument.
8261 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8263 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8264 passed by the function are trusted implicitly. If any of them signed the
8265 response then it is assumed to be valid and is not verified.
8268 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8269 to data. This was previously part of the PKCS7 ASN1 code. This
8270 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8271 [Steve Henson, reported by Kenneth R. Robinette
8272 <support@securenetterm.com>]
8274 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8275 routines: without these tracing memory leaks is very painful.
8276 Fix leaks in PKCS12 and PKCS7 routines.
8279 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8280 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8281 effectively meant GeneralizedTime would never be used. Now it
8282 is initialised to -1 but X509_time_adj() now has to check the value
8283 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8284 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8285 [Steve Henson, reported by Kenneth R. Robinette
8286 <support@securenetterm.com>]
8288 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8289 result in a zero length in the ASN1_INTEGER structure which was
8290 not consistent with the structure when d2i_ASN1_INTEGER() was used
8291 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8292 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8293 where it did not print out a minus for negative ASN1_INTEGER.
8296 *) Add summary printout to ocsp utility. The various functions which
8297 convert status values to strings have been renamed to:
8298 OCSP_response_status_str(), OCSP_cert_status_str() and
8299 OCSP_crl_reason_str() and are no longer static. New options
8300 to verify nonce values and to disable verification. OCSP response
8301 printout format cleaned up.
8304 *) Add additional OCSP certificate checks. These are those specified
8305 in RFC2560. This consists of two separate checks: the CA of the
8306 certificate being checked must either be the OCSP signer certificate
8307 or the issuer of the OCSP signer certificate. In the latter case the
8308 OCSP signer certificate must contain the OCSP signing extended key
8309 usage. This check is performed by attempting to match the OCSP
8310 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8311 in the OCSP_CERTID structures of the response.
8314 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8315 and related routines. This uses the standard OpenSSL certificate
8316 verify routines to perform initial checks (just CA validity) and
8317 to obtain the certificate chain. Then additional checks will be
8318 performed on the chain. Currently the root CA is checked to see
8319 if it is explicitly trusted for OCSP signing. This is used to set
8320 a root CA as a global signing root: that is any certificate that
8321 chains to that CA is an acceptable OCSP signing certificate.
8324 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8325 extensions from a separate configuration file.
8326 As when reading extensions from the main configuration file,
8327 the '-extensions ...' option may be used for specifying the
8329 [Massimiliano Pala <madwolf@comune.modena.it>]
8331 *) New OCSP utility. Allows OCSP requests to be generated or
8332 read. The request can be sent to a responder and the output
8333 parsed, outputed or printed in text form. Not complete yet:
8334 still needs to check the OCSP response validity.
8337 *) New subcommands for 'openssl ca':
8338 'openssl ca -status <serial>' prints the status of the cert with
8339 the given serial number (according to the index file).
8340 'openssl ca -updatedb' updates the expiry status of certificates
8342 [Massimiliano Pala <madwolf@comune.modena.it>]
8344 *) New '-newreq-nodes' command option to CA.pl. This is like
8345 '-newreq', but calls 'openssl req' with the '-nodes' option
8346 so that the resulting key is not encrypted.
8347 [Damien Miller <djm@mindrot.org>]
8349 *) New configuration for the GNU Hurd.
8350 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8352 *) Initial code to implement OCSP basic response verify. This
8353 is currently incomplete. Currently just finds the signer's
8354 certificate and verifies the signature on the response.
8357 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8358 value of OPENSSLDIR. This is available via the new '-d' option
8359 to 'openssl version', and is also included in 'openssl version -a'.
8362 *) Allowing defining memory allocation callbacks that will be given
8363 file name and line number information in additional arguments
8364 (a const char* and an int). The basic functionality remains, as
8365 well as the original possibility to just replace malloc(),
8366 realloc() and free() by functions that do not know about these
8367 additional arguments. To register and find out the current
8368 settings for extended allocation functions, the following
8369 functions are provided:
8371 CRYPTO_set_mem_ex_functions
8372 CRYPTO_set_locked_mem_ex_functions
8373 CRYPTO_get_mem_ex_functions
8374 CRYPTO_get_locked_mem_ex_functions
8376 These work the same way as CRYPTO_set_mem_functions and friends.
8377 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8378 extended allocation function is enabled.
8379 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8380 a conventional allocation function is enabled.
8381 [Richard Levitte, Bodo Moeller]
8383 *) Finish off removing the remaining LHASH function pointer casts.
8384 There should no longer be any prototype-casting required when using
8385 the LHASH abstraction, and any casts that remain are "bugs". See
8386 the callback types and macros at the head of lhash.h for details
8387 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8390 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8391 If /dev/[u]random devices are not available or do not return enough
8392 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8394 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8395 /etc/entropy will be queried once each in this sequence, querying stops
8396 when enough entropy was collected without querying more sockets.
8399 *) Change the Unix RAND_poll() variant to be able to poll several
8400 random devices, as specified by DEVRANDOM, until a sufficient amount
8401 of data has been collected. We spend at most 10 ms on each file
8402 (select timeout) and read in non-blocking mode. DEVRANDOM now
8403 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8404 (previously it was just the string "/dev/urandom"), so on typical
8405 platforms the 10 ms delay will never occur.
8406 Also separate out the Unix variant to its own file, rand_unix.c.
8407 For VMS, there's a currently-empty rand_vms.c.
8410 *) Move OCSP client related routines to ocsp_cl.c. These
8411 provide utility functions which an application needing
8412 to issue a request to an OCSP responder and analyse the
8413 response will typically need: as opposed to those which an
8414 OCSP responder itself would need which will be added later.
8416 OCSP_request_sign() signs an OCSP request with an API similar
8417 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8418 response. OCSP_response_get1_basic() extracts basic response
8419 from response. OCSP_resp_find_status(): finds and extracts status
8420 information from an OCSP_CERTID structure (which will be created
8421 when the request structure is built). These are built from lower
8422 level functions which work on OCSP_SINGLERESP structures but
8423 won't normally be used unless the application wishes to examine
8424 extensions in the OCSP response for example.
8426 Replace nonce routines with a pair of functions.
8427 OCSP_request_add1_nonce() adds a nonce value and optionally
8428 generates a random value. OCSP_check_nonce() checks the
8429 validity of the nonce in an OCSP response.
8432 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8433 This doesn't copy the supplied OCSP_CERTID and avoids the
8434 need to free up the newly created id. Change return type
8435 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8436 This can then be used to add extensions to the request.
8437 Deleted OCSP_request_new(), since most of its functionality
8438 is now in OCSP_REQUEST_new() (and the case insensitive name
8439 clash) apart from the ability to set the request name which
8440 will be added elsewhere.
8443 *) Update OCSP API. Remove obsolete extensions argument from
8444 various functions. Extensions are now handled using the new
8445 OCSP extension code. New simple OCSP HTTP function which
8446 can be used to send requests and parse the response.
8449 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8450 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8451 uses the special reorder version of SET OF to sort the attributes
8452 and reorder them to match the encoded order. This resolves a long
8453 standing problem: a verify on a PKCS7 structure just after signing
8454 it used to fail because the attribute order did not match the
8455 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8456 it uses the received order. This is necessary to tolerate some broken
8457 software that does not order SET OF. This is handled by encoding
8458 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8459 to produce the required SET OF.
8462 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8463 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8464 files to get correct declarations of the ASN.1 item variables.
8467 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8468 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8469 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8470 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8471 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8472 ASN1_ITEM and no wrapper functions.
8475 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8476 replace the old function pointer based I/O routines. Change most of
8477 the *_d2i_bio() and *_d2i_fp() functions to use these.
8480 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8481 lines, recognize more "algorithms" that can be deselected, and make
8482 it complain about algorithm deselection that isn't recognised.
8485 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8486 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8487 to use new functions. Add NO_ASN1_OLD which can be set to remove
8488 some old style ASN1 functions: this can be used to determine if old
8489 code will still work when these eventually go away.
8492 *) New extension functions for OCSP structures, these follow the
8493 same conventions as certificates and CRLs.
8496 *) New function X509V3_add1_i2d(). This automatically encodes and
8497 adds an extension. Its behaviour can be customised with various
8498 flags to append, replace or delete. Various wrappers added for
8499 certificates and CRLs.
8502 *) Fix to avoid calling the underlying ASN1 print routine when
8503 an extension cannot be parsed. Correct a typo in the
8504 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8507 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8508 entries for variables.
8511 *) Add functionality to apps/openssl.c for detecting locking
8512 problems: As the program is single-threaded, all we have
8513 to do is register a locking callback using an array for
8514 storing which locks are currently held by the program.
8517 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8518 SSL_get_ex_data_X509_STORE_idx(), which is used in
8519 ssl_verify_cert_chain() and thus can be called at any time
8520 during TLS/SSL handshakes so that thread-safety is essential.
8521 Unfortunately, the ex_data design is not at all suited
8522 for multi-threaded use, so it probably should be abolished.
8525 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8526 [Broadcom, tweaked and integrated by Geoff Thorpe]
8528 *) Move common extension printing code to new function
8529 X509V3_print_extensions(). Reorganise OCSP print routines and
8530 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8533 *) New function X509_signature_print() to remove duplication in some
8537 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8538 set (this was treated exactly the same as SET OF previously). This
8539 is used to reorder the STACK representing the structure to match the
8540 encoding. This will be used to get round a problem where a PKCS7
8541 structure which was signed could not be verified because the STACK
8542 order did not reflect the encoded order.
8545 *) Reimplement the OCSP ASN1 module using the new code.
8548 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8549 for its ASN1 operations. The old style function pointers still exist
8550 for now but they will eventually go away.
8553 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8554 completely replaces the old ASN1 functionality with a table driven
8555 encoder and decoder which interprets an ASN1_ITEM structure describing
8556 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8557 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8558 has also been converted to the new form.
8561 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8562 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8563 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8564 for negative moduli.
8567 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8568 of not touching the result's sign bit.
8571 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8575 *) Changed the LHASH code to use prototypes for callbacks, and created
8576 macros to declare and implement thin (optionally static) functions
8577 that provide type-safety and avoid function pointer casting for the
8578 type-specific callbacks.
8581 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8583 [Veers Staats <staatsvr@asc.hpc.mil>,
8584 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8586 *) Reformat the FAQ so the different questions and answers can be divided
8587 in sections depending on the subject.
8590 *) Have the zlib compression code load ZLIB.DLL dynamically under
8594 *) New function BN_mod_sqrt for computing square roots modulo a prime
8595 (using the probabilistic Tonelli-Shanks algorithm unless
8596 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8597 be handled deterministically).
8598 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8600 *) Make BN_mod_inverse faster by explicitly handling small quotients
8601 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8602 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8605 *) New function BN_kronecker.
8608 *) Fix BN_gcd so that it works on negative inputs; the result is
8609 positive unless both parameters are zero.
8610 Previously something reasonably close to an infinite loop was
8611 possible because numbers could be growing instead of shrinking
8612 in the implementation of Euclid's algorithm.
8615 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8616 sign of the number in question.
8618 Fix BN_is_word(a,w) to work correctly for w == 0.
8620 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8621 because its test if the absolute value of 'a' equals 'w'.
8622 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8623 it exists mostly for use in the implementations of BN_is_zero(),
8624 BN_is_one(), and BN_is_word().
8627 *) New function BN_swap.
8630 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8631 the exponentiation functions are more likely to produce reasonable
8632 results on negative inputs.
8635 *) Change BN_mod_mul so that the result is always non-negative.
8636 Previously, it could be negative if one of the factors was negative;
8637 I don't think anyone really wanted that behaviour.
8640 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8641 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8642 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8643 and add new functions:
8652 BN_mod_lshift1_quick
8656 These functions always generate non-negative results.
8658 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8659 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8661 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8662 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8663 be reduced modulo m.
8664 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8667 The following entry accidentally appeared in the CHANGES file
8668 distributed with OpenSSL 0.9.7. The modifications described in
8669 it do *not* apply to OpenSSL 0.9.7.
8671 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8672 was actually never needed) and in BN_mul(). The removal in BN_mul()
8673 required a small change in bn_mul_part_recursive() and the addition
8674 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8675 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8676 bn_sub_words() and bn_add_words() except they take arrays with
8681 *) In 'openssl passwd', verify passwords read from the terminal
8682 unless the '-salt' option is used (which usually means that
8683 verification would just waste user's time since the resulting
8684 hash is going to be compared with some given password hash)
8685 or the new '-noverify' option is used.
8687 This is an incompatible change, but it does not affect
8688 non-interactive use of 'openssl passwd' (passwords on the command
8689 line, '-stdin' option, '-in ...' option) and thus should not
8693 *) Remove all references to RSAref, since there's no more need for it.
8696 *) Make DSO load along a path given through an environment variable
8697 (SHLIB_PATH) with shl_load().
8700 *) Constify the ENGINE code as a result of BIGNUM constification.
8701 Also constify the RSA code and most things related to it. In a
8702 few places, most notable in the depth of the ASN.1 code, ugly
8703 casts back to non-const were required (to be solved at a later
8707 *) Make it so the openssl application has all engines loaded by default.
8710 *) Constify the BIGNUM routines a little more.
8713 *) Add the following functions:
8715 ENGINE_load_cswift()
8717 ENGINE_load_atalla()
8719 ENGINE_load_builtin_engines()
8721 That way, an application can itself choose if external engines that
8722 are built-in in OpenSSL shall ever be used or not. The benefit is
8723 that applications won't have to be linked with libdl or other dso
8724 libraries unless it's really needed.
8726 Changed 'openssl engine' to load all engines on demand.
8727 Changed the engine header files to avoid the duplication of some
8728 declarations (they differed!).
8731 *) 'openssl engine' can now list capabilities.
8734 *) Better error reporting in 'openssl engine'.
8737 *) Never call load_dh_param(NULL) in s_server.
8740 *) Add engine application. It can currently list engines by name and
8741 identity, and test if they are actually available.
8744 *) Improve RPM specification file by forcing symbolic linking and making
8745 sure the installed documentation is also owned by root.root.
8746 [Damien Miller <djm@mindrot.org>]
8748 *) Give the OpenSSL applications more possibilities to make use of
8749 keys (public as well as private) handled by engines.
8752 *) Add OCSP code that comes from CertCo.
8755 *) Add VMS support for the Rijndael code.
8758 *) Added untested support for Nuron crypto accelerator.
8761 *) Add support for external cryptographic devices. This code was
8762 previously distributed separately as the "engine" branch.
8763 [Geoff Thorpe, Richard Levitte]
8765 *) Rework the filename-translation in the DSO code. It is now possible to
8766 have far greater control over how a "name" is turned into a filename
8767 depending on the operating environment and any oddities about the
8768 different shared library filenames on each system.
8771 *) Support threads on FreeBSD-elf in Configure.
8774 *) Fix for SHA1 assembly problem with MASM: it produces
8775 warnings about corrupt line number information when assembling
8776 with debugging information. This is caused by the overlapping
8778 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8781 NCONF_get_number() has no error checking at all. As a replacement,
8782 NCONF_get_number_e() is defined (_e for "error checking") and is
8783 promoted strongly. The old NCONF_get_number is kept around for
8784 binary backward compatibility.
8785 Make it possible for methods to load from something other than a BIO,
8786 by providing a function pointer that is given a name instead of a BIO.
8787 For example, this could be used to load configuration data from an
8791 *) Fix for non blocking accept BIOs. Added new I/O special reason
8792 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8793 with non blocking I/O was not possible because no retry code was
8794 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8798 *) Added the beginnings of Rijndael support.
8801 *) Fix for bug in DirectoryString mask setting. Add support for
8802 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8803 to allow certificate printing to more controllable, additional
8804 'certopt' option to 'x509' to allow new printing options to be
8808 *) Clean old EAY MD5 hack from e_os.h.
8811 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8813 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8814 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8815 [Joe Orton, Steve Henson]
8817 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8819 *) Fix additional bug revealed by the NISCC test suite:
8821 Stop bug triggering large recursion when presented with
8822 certain ASN.1 tags (CVE-2003-0851)
8825 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8827 *) Fix various bugs revealed by running the NISCC test suite:
8829 Stop out of bounds reads in the ASN1 code when presented with
8830 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8832 If verify callback ignores invalid public key errors don't try to check
8833 certificate signature with the NULL public key.
8837 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8838 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8842 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8843 extra data after the compression methods not only for TLS 1.0
8844 but also for SSL 3.0 (as required by the specification).
8845 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8847 *) Change X509_certificate_type() to mark the key as exported/exportable
8848 when it's 512 *bits* long, not 512 bytes.
8851 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8853 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8854 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8855 a protocol version number mismatch like a decryption error
8856 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8859 *) Turn on RSA blinding by default in the default implementation
8860 to avoid a timing attack. Applications that don't want it can call
8861 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8862 They would be ill-advised to do so in most cases.
8863 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8865 *) Change RSA blinding code so that it works when the PRNG is not
8866 seeded (in this case, the secret RSA exponent is abused as
8867 an unpredictable seed -- if it is not unpredictable, there
8868 is no point in blinding anyway). Make RSA blinding thread-safe
8869 by remembering the creator's thread ID in rsa->blinding and
8870 having all other threads use local one-time blinding factors
8871 (this requires more computation than sharing rsa->blinding, but
8872 avoids excessive locking; and if an RSA object is not shared
8873 between threads, blinding will still be very fast).
8876 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8878 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8879 via timing by performing a MAC computation even if incorrect
8880 block cipher padding has been found. This is a countermeasure
8881 against active attacks where the attacker has to distinguish
8882 between bad padding and a MAC verification error. (CVE-2003-0078)
8884 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8885 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8886 Martin Vuagnoux (EPFL, Ilion)]
8888 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8890 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8891 memory from it's contents. This is done with a counter that will
8892 place alternating values in each byte. This can be used to solve
8893 two issues: 1) the removal of calls to memset() by highly optimizing
8894 compilers, and 2) cleansing with other values than 0, since those can
8895 be read through on certain media, for example a swap space on disk.
8898 *) Bugfix: client side session caching did not work with external caching,
8899 because the session->cipher setting was not restored when reloading
8900 from the external cache. This problem was masked, when
8901 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8902 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8905 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8906 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8907 [Zeev Lieber <zeev-l@yahoo.com>]
8909 *) Undo an undocumented change introduced in 0.9.6e which caused
8910 repeated calls to OpenSSL_add_all_ciphers() and
8911 OpenSSL_add_all_digests() to be ignored, even after calling
8915 *) Change the default configuration reader to deal with last line not
8916 being properly terminated.
8919 *) Change X509_NAME_cmp() so it applies the special rules on handling
8920 DN values that are of type PrintableString, as well as RDNs of type
8921 emailAddress where the value has the type ia5String.
8922 [stefank@valicert.com via Richard Levitte]
8924 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8925 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8926 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8927 the bitwise-OR of the two for use by the majority of applications
8928 wanting this behaviour, and update the docs. The documented
8929 behaviour and actual behaviour were inconsistent and had been
8930 changing anyway, so this is more a bug-fix than a behavioural
8932 [Geoff Thorpe, diagnosed by Nadav Har'El]
8934 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8935 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8938 *) Fix initialization code race conditions in
8939 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8940 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8941 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8942 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8943 ssl2_get_cipher_by_char(),
8944 ssl3_get_cipher_by_char().
8945 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8947 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8948 the cached sessions are flushed, as the remove_cb() might use ex_data
8949 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8950 (see [openssl.org #212]).
8951 [Geoff Thorpe, Lutz Jaenicke]
8953 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8954 length, instead of the encoding length to d2i_ASN1_OBJECT.
8957 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8959 *) [In 0.9.6g-engine release:]
8960 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8961 [Lynn Gazis <lgazis@rainbow.com>]
8963 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8965 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8966 and get fix the header length calculation.
8967 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8968 Alon Kantor <alonk@checkpoint.com> (and others),
8971 *) Use proper error handling instead of 'assertions' in buffer
8972 overflow checks added in 0.9.6e. This prevents DoS (the
8973 assertions could call abort()).
8974 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8976 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8978 *) Add various sanity checks to asn1_get_length() to reject
8979 the ASN1 length bytes if they exceed sizeof(long), will appear
8980 negative or the content length exceeds the length of the
8982 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8984 *) Fix cipher selection routines: ciphers without encryption had no flags
8985 for the cipher strength set and where therefore not handled correctly
8986 by the selection routines (PR #130).
8989 *) Fix EVP_dsa_sha macro.
8993 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8994 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8995 that was added in OpenSSL 0.9.6d.
8997 As the countermeasure turned out to be incompatible with some
8998 broken SSL implementations, the new option is part of SSL_OP_ALL.
8999 SSL_OP_ALL is usually employed when compatibility with weird SSL
9000 implementations is desired (e.g. '-bugs' option to 's_client' and
9001 's_server'), so the new option is automatically set in many
9005 *) Changes in security patch:
9007 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9008 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9009 Air Force Materiel Command, USAF, under agreement number
9012 *) Add various sanity checks to asn1_get_length() to reject
9013 the ASN1 length bytes if they exceed sizeof(long), will appear
9014 negative or the content length exceeds the length of the
9015 supplied buffer. (CVE-2002-0659)
9016 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9018 *) Assertions for various potential buffer overflows, not known to
9020 [Ben Laurie (CHATS)]
9022 *) Various temporary buffers to hold ASCII versions of integers were
9023 too small for 64 bit platforms. (CVE-2002-0655)
9024 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9026 *) Remote buffer overflow in SSL3 protocol - an attacker could
9027 supply an oversized session ID to a client. (CVE-2002-0656)
9028 [Ben Laurie (CHATS)]
9030 *) Remote buffer overflow in SSL2 protocol - an attacker could
9031 supply an oversized client master key. (CVE-2002-0656)
9032 [Ben Laurie (CHATS)]
9034 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9036 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9037 encoded as NULL) with id-dsa-with-sha1.
9038 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9040 *) Check various X509_...() return values in apps/req.c.
9041 [Nils Larsch <nla@trustcenter.de>]
9043 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9044 an end-of-file condition would erroneously be flagged, when the CRLF
9045 was just at the end of a processed block. The bug was discovered when
9046 processing data through a buffering memory BIO handing the data to a
9047 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9048 <ptsekov@syntrex.com> and Nedelcho Stanev.
9051 *) Implement a countermeasure against a vulnerability recently found
9052 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9053 before application data chunks to avoid the use of known IVs
9054 with data potentially chosen by the attacker.
9057 *) Fix length checks in ssl3_get_client_hello().
9060 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9061 to prevent ssl3_read_internal() from incorrectly assuming that
9062 ssl3_read_bytes() found application data while handshake
9063 processing was enabled when in fact s->s3->in_read_app_data was
9064 merely automatically cleared during the initial handshake.
9065 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9067 *) Fix object definitions for Private and Enterprise: they were not
9068 recognized in their shortname (=lowercase) representation. Extend
9069 obj_dat.pl to issue an error when using undefined keywords instead
9070 of silently ignoring the problem (Svenning Sorensen
9071 <sss@sss.dnsalias.net>).
9074 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9075 generators, i.e. generators other than 2 and 5. (Previously, the
9076 code did not properly initialise the 'add' and 'rem' values to
9077 BN_generate_prime().)
9079 In the new general case, we do not insist that 'generator' is
9080 actually a primitive root: This requirement is rather pointless;
9081 a generator of the order-q subgroup is just as good, if not
9085 *) Map new X509 verification errors to alerts. Discovered and submitted by
9086 Tom Wu <tom@arcot.com>.
9089 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9090 returning non-zero before the data has been completely received
9091 when using non-blocking I/O.
9092 [Bodo Moeller; problem pointed out by John Hughes]
9094 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9095 [Ben Laurie, Lutz Jaenicke]
9097 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9098 Yoram Zahavi <YoramZ@gilian.com>).
9101 *) Add information about CygWin 1.3 and on, and preserve proper
9102 configuration for the versions before that.
9103 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9105 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9106 check whether we deal with a copy of a session and do not delete from
9107 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9108 <izhar@checkpoint.com>.
9111 *) Do not store session data into the internal session cache, if it
9112 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9113 flag is set). Proposed by Aslam <aslam@funk.com>.
9116 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9120 *) [In 0.9.6d-engine release:]
9121 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9122 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9124 *) Add the configuration target linux-s390x.
9125 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9127 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9128 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9129 variable as an indication that a ClientHello message has been
9130 received. As the flag value will be lost between multiple
9131 invocations of ssl3_accept when using non-blocking I/O, the
9132 function may not be aware that a handshake has actually taken
9133 place, thus preventing a new session from being added to the
9136 To avoid this problem, we now set s->new_session to 2 instead of
9137 using a local variable.
9138 [Lutz Jaenicke, Bodo Moeller]
9140 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9141 if the SSL_R_LENGTH_MISMATCH error is detected.
9142 [Geoff Thorpe, Bodo Moeller]
9144 *) New 'shared_ldflag' column in Configure platform table.
9147 *) Fix EVP_CIPHER_mode macro.
9148 ["Dan S. Camper" <dan@bti.net>]
9150 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9151 type, we must throw them away by setting rr->length to 0.
9152 [D P Chang <dpc@qualys.com>]
9154 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9156 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9157 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9158 worked incorrectly for those cases where range = 10..._2 and
9159 3*range is two bits longer than range.)
9162 *) Only add signing time to PKCS7 structures if it is not already
9166 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9167 OBJ_ld_ce should be OBJ_id_ce.
9168 Also some ip-pda OIDs in crypto/objects/objects.txt were
9169 incorrect (cf. RFC 3039).
9170 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9172 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9173 returns early because it has nothing to do.
9174 [Andy Schneider <andy.schneider@bjss.co.uk>]
9176 *) [In 0.9.6c-engine release:]
9177 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9178 [Andy Schneider <andy.schneider@bjss.co.uk>]
9180 *) [In 0.9.6c-engine release:]
9181 Add support for Cryptographic Appliance's keyserver technology.
9182 (Use engine 'keyclient')
9183 [Cryptographic Appliances and Geoff Thorpe]
9185 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9186 is called via tools/c89.sh because arguments have to be
9187 rearranged (all '-L' options must appear before the first object
9189 [Richard Shapiro <rshapiro@abinitio.com>]
9191 *) [In 0.9.6c-engine release:]
9192 Add support for Broadcom crypto accelerator cards, backported
9194 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9196 *) [In 0.9.6c-engine release:]
9197 Add support for SureWare crypto accelerator cards from
9198 Baltimore Technologies. (Use engine 'sureware')
9199 [Baltimore Technologies and Mark Cox]
9201 *) [In 0.9.6c-engine release:]
9202 Add support for crypto accelerator cards from Accelerated
9203 Encryption Processing, www.aep.ie. (Use engine 'aep')
9204 [AEP Inc. and Mark Cox]
9206 *) Add a configuration entry for gcc on UnixWare.
9207 [Gary Benson <gbenson@redhat.com>]
9209 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9210 messages are stored in a single piece (fixed-length part and
9211 variable-length part combined) and fix various bugs found on the way.
9214 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9215 instead. BIO_gethostbyname() does not know what timeouts are
9216 appropriate, so entries would stay in cache even when they have
9218 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9220 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9221 faced with a pathologically small ClientHello fragment that does
9222 not contain client_version: Instead of aborting with an error,
9223 simply choose the highest available protocol version (i.e.,
9224 TLS 1.0 unless it is disabled). In practice, ClientHello
9225 messages are never sent like this, but this change gives us
9226 strictly correct behaviour at least for TLS.
9229 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9230 never resets s->method to s->ctx->method when called from within
9231 one of the SSL handshake functions.
9232 [Bodo Moeller; problem pointed out by Niko Baric]
9234 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9235 (sent using the client's version number) if client_version is
9236 smaller than the protocol version in use. Also change
9237 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9238 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9239 the client will at least see that alert.
9242 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9246 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9247 client receives HelloRequest while in a handshake.
9248 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9250 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9251 should end in 'break', not 'goto end' which circumvents various
9252 cleanups done in state SSL_ST_OK. But session related stuff
9253 must be disabled for SSL_ST_OK in the case that we just sent a
9256 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9257 before just sending a HelloRequest.
9258 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9260 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9261 reveal whether illegal block cipher padding was found or a MAC
9262 verification error occurred. (Neither SSLerr() codes nor alerts
9263 are directly visible to potential attackers, but the information
9264 may leak via logfiles.)
9266 Similar changes are not required for the SSL 2.0 implementation
9267 because the number of padding bytes is sent in clear for SSL 2.0,
9268 and the extra bytes are just ignored. However ssl/s2_pkt.c
9269 failed to verify that the purported number of padding bytes is in
9273 *) Add OpenUNIX-8 support including shared libraries
9274 (Boyd Lynn Gerber <gerberb@zenez.com>).
9277 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9278 'wristwatch attack' using huge encoding parameters (cf.
9279 James H. Manger's CRYPTO 2001 paper). Note that the
9280 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9281 encoding parameters and hence was not vulnerable.
9284 *) BN_sqr() bug fix.
9285 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9287 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9288 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9289 followed by modular reduction.
9290 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9292 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9293 equivalent based on BN_pseudo_rand() instead of BN_rand().
9296 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9297 This function was broken, as the check for a new client hello message
9298 to handle SGC did not allow these large messages.
9299 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9302 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9305 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9306 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9309 *) Rework the configuration and shared library support for Tru64 Unix.
9310 The configuration part makes use of modern compiler features and
9311 still retains old compiler behavior for those that run older versions
9312 of the OS. The shared library support part includes a variant that
9313 uses the RPATH feature, and is available through the special
9314 configuration target "alpha-cc-rpath", which will never be selected
9316 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9318 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9319 with the same message size as in ssl3_get_certificate_request().
9320 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9321 messages might inadvertently be reject as too long.
9322 [Petr Lampa <lampa@fee.vutbr.cz>]
9324 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9327 *) Modified SSL library such that the verify_callback that has been set
9328 specificly for an SSL object with SSL_set_verify() is actually being
9329 used. Before the change, a verify_callback set with this function was
9330 ignored and the verify_callback() set in the SSL_CTX at the time of
9331 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9332 to allow the necessary settings.
9335 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9336 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9337 done automatically (in contradiction to the requirements of the C
9338 standard). This made problems when used from OpenSSH.
9341 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9342 dh->length and always used
9344 BN_rand_range(priv_key, dh->p).
9346 BN_rand_range() is not necessary for Diffie-Hellman, and this
9347 specific range makes Diffie-Hellman unnecessarily inefficient if
9348 dh->length (recommended exponent length) is much smaller than the
9349 length of dh->p. We could use BN_rand_range() if the order of
9350 the subgroup was stored in the DH structure, but we only have
9355 BN_rand(priv_key, l, ...)
9357 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9363 RSA_eay_public_encrypt
9364 RSA_eay_private_decrypt
9365 RSA_eay_private_encrypt (signing)
9366 RSA_eay_public_decrypt (signature verification)
9368 (default implementations for RSA_public_encrypt,
9369 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9370 always reject numbers >= n.
9373 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9374 to synchronize access to 'locking_thread'. This is necessary on
9375 systems where access to 'locking_thread' (an 'unsigned long'
9376 variable) is not atomic.
9379 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9380 *before* setting the 'crypto_lock_rand' flag. The previous code had
9381 a race condition if 0 is a valid thread ID.
9382 [Travis Vitek <vitek@roguewave.com>]
9384 *) Add support for shared libraries under Irix.
9385 [Albert Chin-A-Young <china@thewrittenword.com>]
9387 *) Add configuration option to build on Linux on both big-endian and
9389 [Ralf Baechle <ralf@uni-koblenz.de>]
9391 *) Add the possibility to create shared libraries on HP-UX.
9394 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9396 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9397 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9398 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9399 PRNG state recovery was possible based on the output of
9400 one PRNG request appropriately sized to gain knowledge on
9401 'md' followed by enough consecutive 1-byte PRNG requests
9402 to traverse all of 'state'.
9404 1. When updating 'md_local' (the current thread's copy of 'md')
9405 during PRNG output generation, hash all of the previous
9406 'md_local' value, not just the half used for PRNG output.
9408 2. Make the number of bytes from 'state' included into the hash
9409 independent from the number of PRNG bytes requested.
9411 The first measure alone would be sufficient to avoid
9412 Markku-Juhani's attack. (Actually it had never occurred
9413 to me that the half of 'md_local' used for chaining was the
9414 half from which PRNG output bytes were taken -- I had always
9415 assumed that the secret half would be used.) The second
9416 measure makes sure that additional data from 'state' is never
9417 mixed into 'md_local' in small portions; this heuristically
9418 further strengthens the PRNG.
9421 *) Fix crypto/bn/asm/mips3.s.
9424 *) When only the key is given to "enc", the IV is undefined. Print out
9425 an error message in this case.
9428 *) Handle special case when X509_NAME is empty in X509 printing routines.
9431 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9432 positive and less than q.
9435 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9436 used: it isn't thread safe and the add_lock_callback should handle
9438 [Paul Rose <Paul.Rose@bridge.com>]
9440 *) Verify that incoming data obeys the block size in
9441 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9445 [Ulf Möller, Bodo Möller]
9447 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9448 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9449 when fixing the server behaviour for backwards-compatible 'client
9450 hello' messages. (Note that the attack is impractical against
9451 SSL 3.0 and TLS 1.0 anyway because length and version checking
9452 means that the probability of guessing a valid ciphertext is
9453 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9456 Before 0.9.5, the countermeasure (hide the error by generating a
9457 random 'decryption result') did not work properly because
9458 ERR_clear_error() was missing, meaning that SSL_get_error() would
9459 detect the supposedly ignored error.
9461 Both problems are now fixed.
9464 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9465 (previously it was 1024).
9468 *) Fix for compatibility mode trust settings: ignore trust settings
9469 unless some valid trust or reject settings are present.
9472 *) Fix for blowfish EVP: its a variable length cipher.
9475 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9476 parameters in DSA public key structures and return an error in the
9477 DSA routines if parameters are absent.
9480 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9481 in the current directory if neither $RANDFILE nor $HOME was set.
9482 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9483 caused some confusion to Windows users who haven't defined $HOME.
9484 Thus RAND_file_name() is changed again: e_os.h can define a
9485 DEFAULT_HOME, which will be used if $HOME is not set.
9486 For Windows, we use "C:"; on other platforms, we still require
9487 environment variables.
9489 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9490 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9491 having multiple threads call RAND_poll() concurrently.
9494 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9495 combination of a flag and a thread ID variable.
9496 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9497 flag), *other* threads can enter ssleay_add_bytes without obeying
9498 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9499 that they do not hold after the first thread unsets add_do_not_lock).
9502 *) Change bctest again: '-x' expressions are not available in all
9506 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9508 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9509 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9511 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9512 the default extension for executables, if any. Also, make the perl
9513 scripts that use symlink() to test if it really exists and use "cp"
9514 if it doesn't. All this made OpenSSL compilable and installable in
9518 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9519 If SEQUENCE is length is indefinite just set c->slen to the total
9520 amount of data available.
9521 [Steve Henson, reported by shige@FreeBSD.org]
9522 [This change does not apply to 0.9.7.]
9524 *) Change bctest to avoid here-documents inside command substitution
9525 (workaround for FreeBSD /bin/sh bug).
9526 For compatibility with Ultrix, avoid shell functions (introduced
9527 in the bctest version that searches along $PATH).
9530 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9531 with des_encrypt() defined on some operating systems, like Solaris
9535 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9536 On the Importance of Eliminating Errors in Cryptographic
9537 Computations, J. Cryptology 14 (2001) 2, 101-119,
9538 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9541 *) MIPS assembler BIGNUM division bug fix.
9544 *) Disabled incorrect Alpha assembler code.
9547 *) Fix PKCS#7 decode routines so they correctly update the length
9548 after reading an EOC for the EXPLICIT tag.
9550 [This change does not apply to 0.9.7.]
9552 *) Fix bug in PKCS#12 key generation routines. This was triggered
9553 if a 3DES key was generated with a 0 initial byte. Include
9554 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9555 (but broken) behaviour.
9558 *) Enhance bctest to search for a working bc along $PATH and print
9560 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9562 *) Fix memory leaks in err.c: free err_data string if necessary;
9563 don't write to the wrong index in ERR_set_error_data.
9566 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9570 *) Replace rdtsc with _emit statements for VC++ version 5.
9571 [Jeremy Cooper <jeremy@baymoo.org>]
9573 *) Make it possible to reuse SSLv2 sessions.
9576 *) In copy_email() check for >= 0 as a return value for
9577 X509_NAME_get_index_by_NID() since 0 is a valid index.
9578 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9580 *) Avoid coredump with unsupported or invalid public keys by checking if
9581 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9582 PKCS7_verify() fails with non detached data.
9585 *) Don't use getenv in library functions when run as setuid/setgid.
9586 New function OPENSSL_issetugid().
9589 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9590 due to incorrect handling of multi-threading:
9592 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9594 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9596 3. Count how many times MemCheck_off() has been called so that
9597 nested use can be treated correctly. This also avoids
9598 inband-signalling in the previous code (which relied on the
9599 assumption that thread ID 0 is impossible).
9602 *) Add "-rand" option also to s_client and s_server.
9605 *) Fix CPU detection on Irix 6.x.
9606 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9607 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9609 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9612 [This change does not apply to 0.9.7.]
9614 *) Use the cached encoding of an X509_NAME structure rather than
9615 copying it. This is apparently the reason for the libsafe "errors"
9616 but the code is actually correct.
9619 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9620 Bleichenbacher's DSA attack.
9621 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9622 to be set and top=0 forces the highest bit to be set; top=-1 is new
9623 and leaves the highest bit random.
9624 [Ulf Moeller, Bodo Moeller]
9626 *) In the NCONF_...-based implementations for CONF_... queries
9627 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9628 a temporary CONF structure with the data component set to NULL
9629 (which gives segmentation faults in lh_retrieve).
9630 Instead, use NULL for the CONF pointer in CONF_get_string and
9631 CONF_get_number (which may use environment variables) and directly
9632 return NULL from CONF_get_section.
9635 *) Fix potential buffer overrun for EBCDIC.
9638 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9639 keyUsage if basicConstraints absent for a CA.
9642 *) Make SMIME_write_PKCS7() write mail header values with a format that
9643 is more generally accepted (no spaces before the semicolon), since
9644 some programs can't parse those values properly otherwise. Also make
9645 sure BIO's that break lines after each write do not create invalid
9649 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9650 macros previously used would not encode an empty SEQUENCE OF
9651 and break the signature.
9653 [This change does not apply to 0.9.7.]
9655 *) Zero the premaster secret after deriving the master secret in
9659 *) Add some EVP_add_digest_alias registrations (as found in
9660 OpenSSL_add_all_digests()) to SSL_library_init()
9661 aka OpenSSL_add_ssl_algorithms(). This provides improved
9662 compatibility with peers using X.509 certificates
9663 with unconventional AlgorithmIdentifier OIDs.
9666 *) Fix for Irix with NO_ASM.
9667 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9669 *) ./config script fixes.
9670 [Ulf Moeller, Richard Levitte]
9672 *) Fix 'openssl passwd -1'.
9675 *) Change PKCS12_key_gen_asc() so it can cope with non null
9676 terminated strings whose length is passed in the passlen
9677 parameter, for example from PEM callbacks. This was done
9678 by adding an extra length parameter to asc2uni().
9679 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9681 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9682 call failed, free the DSA structure.
9685 *) Fix to uni2asc() to cope with zero length Unicode strings.
9686 These are present in some PKCS#12 files.
9689 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9690 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9691 when writing a 32767 byte record.
9692 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9694 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9695 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9697 (RSA objects have a reference count access to which is protected
9698 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9699 so they are meant to be shared between threads.)
9700 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9701 "Reddie, Steven" <Steven.Reddie@ca.com>]
9703 *) Fix a deadlock in CRYPTO_mem_leaks().
9706 *) Use better test patterns in bntest.
9709 *) rand_win.c fix for Borland C.
9712 *) BN_rshift bugfix for n == 0.
9715 *) Add a 'bctest' script that checks for some known 'bc' bugs
9716 so that 'make test' does not abort just because 'bc' is broken.
9719 *) Store verify_result within SSL_SESSION also for client side to
9720 avoid potential security hole. (Re-used sessions on the client side
9721 always resulted in verify_result==X509_V_OK, not using the original
9722 result of the server certificate verification.)
9725 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9726 SSL3_RT_APPLICATION_DATA, return 0.
9727 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9731 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9732 releases, have been re-implemented by renaming the previous
9733 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9734 and ssl3_read_internal, respectively, and adding 'peek' parameters
9735 to them. The new ssl[23]_{read,peek} functions are calls to
9736 ssl[23]_read_internal with the 'peek' flag set appropriately.
9737 A 'peek' parameter has also been added to ssl3_read_bytes, which
9738 does the actual work for ssl3_read_internal.
9741 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9742 the method-specific "init()" handler. Also clean up ex_data after
9743 calling the method-specific "finish()" handler. Previously, this was
9744 happening the other way round.
9747 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9748 The previous value, 12, was not always sufficient for BN_mod_exp().
9751 *) Make sure that shared libraries get the internal name engine with
9752 the full version number and not just 0. This should mark the
9753 shared libraries as not backward compatible. Of course, this should
9754 be changed again when we can guarantee backward binary compatibility.
9757 *) Fix typo in get_cert_by_subject() in by_dir.c
9758 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9760 *) Rework the system to generate shared libraries:
9762 - Make note of the expected extension for the shared libraries and
9763 if there is a need for symbolic links from for example libcrypto.so.0
9764 to libcrypto.so.0.9.7. There is extended info in Configure for
9767 - Make as few rebuilds of the shared libraries as possible.
9769 - Still avoid linking the OpenSSL programs with the shared libraries.
9771 - When installing, install the shared libraries separately from the
9775 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9777 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9778 and not in SSL_clear because the latter is also used by the
9779 accept/connect functions; previously, the settings made by
9780 SSL_set_read_ahead would be lost during the handshake.
9781 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9783 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9784 Previously, it would create entries for disabled algorithms no
9788 *) Added several new manual pages for SSL_* function.
9791 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9793 *) In ssl23_get_client_hello, generate an error message when faced
9794 with an initial SSL 3.0/TLS record that is too small to contain the
9795 first two bytes of the ClientHello message, i.e. client_version.
9796 (Note that this is a pathologic case that probably has never happened
9797 in real life.) The previous approach was to use the version number
9798 from the record header as a substitute; but our protocol choice
9799 should not depend on that one because it is not authenticated
9800 by the Finished messages.
9803 *) More robust randomness gathering functions for Windows.
9804 [Jeffrey Altman <jaltman@columbia.edu>]
9806 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9807 not set then we don't setup the error code for issuer check errors
9808 to avoid possibly overwriting other errors which the callback does
9809 handle. If an application does set the flag then we assume it knows
9810 what it is doing and can handle the new informational codes
9814 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9815 a general "ANY" type, as such it should be able to decode anything
9816 including tagged types. However it didn't check the class so it would
9817 wrongly interpret tagged types in the same way as their universal
9818 counterpart and unknown types were just rejected. Changed so that the
9819 tagged and unknown types are handled in the same way as a SEQUENCE:
9820 that is the encoding is stored intact. There is also a new type
9821 "V_ASN1_OTHER" which is used when the class is not universal, in this
9822 case we have no idea what the actual type is so we just lump them all
9826 *) On VMS, stdout may very well lead to a file that is written to
9827 in a record-oriented fashion. That means that every write() will
9828 write a separate record, which will be read separately by the
9829 programs trying to read from it. This can be very confusing.
9831 The solution is to put a BIO filter in the way that will buffer
9832 text until a linefeed is reached, and then write everything a
9833 line at a time, so every record written will be an actual line,
9834 not chunks of lines and not (usually doesn't happen, but I've
9835 seen it once) several lines in one record. BIO_f_linebuffer() is
9838 Currently, it's a VMS-only method, because that's where it has
9839 been tested well enough.
9842 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9843 it can return incorrect results.
9844 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9845 but it was in 0.9.6-beta[12].)
9848 *) Disable the check for content being present when verifying detached
9849 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9850 include zero length content when signing messages.
9853 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9854 BIO_ctrl (for BIO pairs).
9857 *) Add DSO method for VMS.
9860 *) Bug fix: Montgomery multiplication could produce results with the
9864 *) Add RPM specification openssl.spec and modify it to build three
9865 packages. The default package contains applications, application
9866 documentation and run-time libraries. The devel package contains
9867 include files, static libraries and function documentation. The
9868 doc package contains the contents of the doc directory. The original
9869 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9872 *) Add a large number of documentation files for many SSL routines.
9873 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9875 *) Add a configuration entry for Sony News 4.
9876 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9878 *) Don't set the two most significant bits to one when generating a
9879 random number < q in the DSA library.
9882 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9883 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9884 the underlying transport is blocking) if a handshake took place.
9885 (The default behaviour is needed by applications such as s_client
9886 and s_server that use select() to determine when to use SSL_read;
9887 but for applications that know in advance when to expect data, it
9888 just makes things more complicated.)
9891 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9895 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9896 work better on such systems.
9897 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9899 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9900 Update PKCS12_parse() so it copies the friendlyName and the
9901 keyid to the certificates aux info.
9904 *) Fix bug in PKCS7_verify() which caused an infinite loop
9905 if there was more than one signature.
9906 [Sven Uszpelkat <su@celocom.de>]
9908 *) Major change in util/mkdef.pl to include extra information
9909 about each symbol, as well as presenting variables as well
9910 as functions. This change means that there's n more need
9911 to rebuild the .num files when some algorithms are excluded.
9914 *) Allow the verify time to be set by an application,
9915 rather than always using the current time.
9918 *) Phase 2 verify code reorganisation. The certificate
9919 verify code now looks up an issuer certificate by a
9920 number of criteria: subject name, authority key id
9921 and key usage. It also verifies self signed certificates
9922 by the same criteria. The main comparison function is
9923 X509_check_issued() which performs these checks.
9925 Lot of changes were necessary in order to support this
9926 without completely rewriting the lookup code.
9928 Authority and subject key identifier are now cached.
9930 The LHASH 'certs' is X509_STORE has now been replaced
9931 by a STACK_OF(X509_OBJECT). This is mainly because an
9932 LHASH can't store or retrieve multiple objects with
9933 the same hash value.
9935 As a result various functions (which were all internal
9936 use only) have changed to handle the new X509_STORE
9937 structure. This will break anything that messed round
9938 with X509_STORE internally.
9940 The functions X509_STORE_add_cert() now checks for an
9941 exact match, rather than just subject name.
9943 The X509_STORE API doesn't directly support the retrieval
9944 of multiple certificates matching a given criteria, however
9945 this can be worked round by performing a lookup first
9946 (which will fill the cache with candidate certificates)
9947 and then examining the cache for matches. This is probably
9948 the best we can do without throwing out X509_LOOKUP
9949 entirely (maybe later...).
9951 The X509_VERIFY_CTX structure has been enhanced considerably.
9953 All certificate lookup operations now go via a get_issuer()
9954 callback. Although this currently uses an X509_STORE it
9955 can be replaced by custom lookups. This is a simple way
9956 to bypass the X509_STORE hackery necessary to make this
9957 work and makes it possible to use more efficient techniques
9958 in future. A very simple version which uses a simple
9959 STACK for its trusted certificate store is also provided
9960 using X509_STORE_CTX_trusted_stack().
9962 The verify_cb() and verify() callbacks now have equivalents
9963 in the X509_STORE_CTX structure.
9965 X509_STORE_CTX also has a 'flags' field which can be used
9966 to customise the verify behaviour.
9969 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9970 excludes S/MIME capabilities.
9973 *) When a certificate request is read in keep a copy of the
9974 original encoding of the signed data and use it when outputting
9975 again. Signatures then use the original encoding rather than
9976 a decoded, encoded version which may cause problems if the
9977 request is improperly encoded.
9980 *) For consistency with other BIO_puts implementations, call
9981 buffer_write(b, ...) directly in buffer_puts instead of calling
9984 In BIO_puts, increment b->num_write as in BIO_write.
9985 [Peter.Sylvester@EdelWeb.fr]
9987 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9988 BN_zero, we may not return a BIGNUM with an array consisting of
9992 *) Avoid calling abort() from within the library when problems are
9993 detected, except if preprocessor symbols have been defined
9994 (such as REF_CHECK, BN_DEBUG etc.).
9997 *) New openssl application 'rsautl'. This utility can be
9998 used for low level RSA operations. DER public key
9999 BIO/fp routines also added.
10002 *) New Configure entry and patches for compiling on QNX 4.
10003 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10005 *) A demo state-machine implementation was sponsored by
10006 Nuron (http://www.nuron.com/) and is now available in
10007 demos/state_machine.
10010 *) New options added to the 'dgst' utility for signature
10011 generation and verification.
10014 *) Unrecognized PKCS#7 content types are now handled via a
10015 catch all ASN1_TYPE structure. This allows unsupported
10016 types to be stored as a "blob" and an application can
10017 encode and decode it manually.
10020 *) Fix various signed/unsigned issues to make a_strex.c
10021 compile under VC++.
10022 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10024 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10025 length if passed a buffer. ASN1_INTEGER_to_BN failed
10026 if passed a NULL BN and its argument was negative.
10027 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10029 *) Modification to PKCS#7 encoding routines to output definite
10030 length encoding. Since currently the whole structures are in
10031 memory there's not real point in using indefinite length
10032 constructed encoding. However if OpenSSL is compiled with
10033 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10036 *) Added BIO_vprintf() and BIO_vsnprintf().
10039 *) Added more prefixes to parse for in the strings written
10040 through a logging bio, to cover all the levels that are available
10041 through syslog. The prefixes are now:
10043 PANIC, EMERG, EMR => LOG_EMERG
10044 ALERT, ALR => LOG_ALERT
10045 CRIT, CRI => LOG_CRIT
10046 ERROR, ERR => LOG_ERR
10047 WARNING, WARN, WAR => LOG_WARNING
10048 NOTICE, NOTE, NOT => LOG_NOTICE
10049 INFO, INF => LOG_INFO
10050 DEBUG, DBG => LOG_DEBUG
10052 and as before, if none of those prefixes are present at the
10053 beginning of the string, LOG_ERR is chosen.
10055 On Win32, the LOG_* levels are mapped according to this:
10057 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10058 LOG_WARNING => EVENTLOG_WARNING_TYPE
10059 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10063 *) Made it possible to reconfigure with just the configuration
10064 argument "reconf" or "reconfigure". The command line arguments
10065 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10066 and are retrieved from there when reconfiguring.
10069 *) MD4 implemented.
10070 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10072 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10075 *) The obj_dat.pl script was messing up the sorting of object
10076 names. The reason was that it compared the quoted version
10077 of strings as a result "OCSP" > "OCSP Signing" because
10078 " > SPACE. Changed script to store unquoted versions of
10079 names and add quotes on output. It was also omitting some
10080 names from the lookup table if they were given a default
10081 value (that is if SN is missing it is given the same
10082 value as LN and vice versa), these are now added on the
10083 grounds that if an object has a name we should be able to
10084 look it up. Finally added warning output when duplicate
10085 short or long names are found.
10088 *) Changes needed for Tandem NSK.
10089 [Scott Uroff <scott@xypro.com>]
10091 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10092 RSA_padding_check_SSLv23(), special padding was never detected
10093 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10094 version rollback attacks was not effective.
10096 In s23_clnt.c, don't use special rollback-attack detection padding
10097 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10098 client; similarly, in s23_srvr.c, don't do the rollback check if
10099 SSL 2.0 is the only protocol enabled in the server.
10102 *) Make it possible to get hexdumps of unprintable data with 'openssl
10103 asn1parse'. By implication, the functions ASN1_parse_dump() and
10104 BIO_dump_indent() are added.
10107 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10108 these print out strings and name structures based on various
10109 flags including RFC2253 support and proper handling of
10110 multibyte characters. Added options to the 'x509' utility
10111 to allow the various flags to be set.
10114 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10115 Also change the functions X509_cmp_current_time() and
10116 X509_gmtime_adj() work with an ASN1_TIME structure,
10117 this will enable certificates using GeneralizedTime in validity
10118 dates to be checked.
10121 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10122 negative public key encodings) on by default,
10123 NO_NEG_PUBKEY_BUG can be set to disable it.
10126 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10127 content octets. An i2c_ASN1_OBJECT is unnecessary because
10128 the encoding can be trivially obtained from the structure.
10131 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10132 not read locks (CRYPTO_r_[un]lock).
10135 *) A first attempt at creating official support for shared
10136 libraries through configuration. I've kept it so the
10137 default is static libraries only, and the OpenSSL programs
10138 are always statically linked for now, but there are
10139 preparations for dynamic linking in place.
10140 This has been tested on Linux and Tru64.
10143 *) Randomness polling function for Win9x, as described in:
10144 Peter Gutmann, Software Generation of Practically Strong
10148 *) Fix so PRNG is seeded in req if using an already existing
10152 *) New options to smime application. -inform and -outform
10153 allow alternative formats for the S/MIME message including
10154 PEM and DER. The -content option allows the content to be
10155 specified separately. This should allow things like Netscape
10156 form signing output easier to verify.
10159 *) Fix the ASN1 encoding of tags using the 'long form'.
10162 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10163 STRING types. These convert content octets to and from the
10164 underlying type. The actual tag and length octets are
10165 already assumed to have been read in and checked. These
10166 are needed because all other string types have virtually
10167 identical handling apart from the tag. By having versions
10168 of the ASN1 functions that just operate on content octets
10169 IMPLICIT tagging can be handled properly. It also allows
10170 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10171 and ASN1_INTEGER are identical apart from the tag.
10174 *) Change the handling of OID objects as follows:
10176 - New object identifiers are inserted in objects.txt, following
10177 the syntax given in objects.README.
10178 - objects.pl is used to process obj_mac.num and create a new
10180 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10183 This is currently kind of a hack, and the perl code in objects.pl
10184 isn't very elegant, but it works as I intended. The simplest way
10185 to check that it worked correctly is to look in obj_dat.h and
10186 check the array nid_objs and make sure the objects haven't moved
10187 around (this is important!). Additions are OK, as well as
10188 consistent name changes.
10191 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10194 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10195 The given file adds to whatever has already been seeded into the
10196 random pool through the RANDFILE configuration file option or
10197 environment variable, or the default random state file.
10200 *) mkstack.pl now sorts each macro group into lexical order.
10201 Previously the output order depended on the order the files
10202 appeared in the directory, resulting in needless rewriting
10206 *) Patches to make OpenSSL compile under Win32 again. Mostly
10207 work arounds for the VC++ problem that it treats func() as
10208 func(void). Also stripped out the parts of mkdef.pl that
10209 added extra typesafe functions: these no longer exist.
10212 *) Reorganisation of the stack code. The macros are now all
10213 collected in safestack.h . Each macro is defined in terms of
10214 a "stack macro" of the form SKM_<name>(type, a, b). The
10215 DEBUG_SAFESTACK is now handled in terms of function casts,
10216 this has the advantage of retaining type safety without the
10217 use of additional functions. If DEBUG_SAFESTACK is not defined
10218 then the non typesafe macros are used instead. Also modified the
10219 mkstack.pl script to handle the new form. Needs testing to see
10220 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10221 the default if no major problems. Similar behaviour for ASN1_SET_OF
10222 and PKCS12_STACK_OF.
10225 *) When some versions of IIS use the 'NET' form of private key the
10226 key derivation algorithm is different. Normally MD5(password) is
10227 used as a 128 bit RC4 key. In the modified case
10228 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10229 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10230 as the old Netscape_RSA functions except they have an additional
10231 'sgckey' parameter which uses the modified algorithm. Also added
10232 an -sgckey command line option to the rsa utility. Thanks to
10233 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10234 algorithm to openssl-dev.
10237 *) The evp_local.h macros were using 'c.##kname' which resulted in
10238 invalid expansion on some systems (SCO 5.0.5 for example).
10239 Corrected to 'c.kname'.
10240 [Phillip Porch <root@theporch.com>]
10242 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10243 a STACK of email addresses from a certificate or request, these look
10244 in the subject name and the subject alternative name extensions and
10245 omit any duplicate addresses.
10248 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10249 This makes DSA verification about 2 % faster.
10252 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10253 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10254 plus overhead for 1024 bit moduli).
10255 This makes exponentiations about 0.5 % faster for 1024 bit
10256 exponents (as measured by "openssl speed rsa2048").
10259 *) Rename memory handling macros to avoid conflicts with other
10261 Malloc => OPENSSL_malloc
10262 Malloc_locked => OPENSSL_malloc_locked
10263 Realloc => OPENSSL_realloc
10264 Free => OPENSSL_free
10267 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10268 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10271 *) CygWin32 support.
10272 [John Jarvie <jjarvie@newsguy.com>]
10274 *) The type-safe stack code has been rejigged. It is now only compiled
10275 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10276 by default all type-specific stack functions are "#define"d back to
10277 standard stack functions. This results in more streamlined output
10278 but retains the type-safety checking possibilities of the original
10282 *) The STACK code has been cleaned up, and certain type declarations
10283 that didn't make a lot of sense have been brought in line. This has
10284 also involved a cleanup of sorts in safestack.h to more correctly
10285 map type-safe stack functions onto their plain stack counterparts.
10286 This work has also resulted in a variety of "const"ifications of
10287 lots of the code, especially "_cmp" operations which should normally
10288 be prototyped with "const" parameters anyway.
10291 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10292 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10293 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10294 where all of 'md' is used each time the PRNG is used, but 'state'
10295 is used only indexed by a cyclic counter. As entropy may not be
10296 well distributed from the beginning, 'md' is important as a
10297 chaining variable. However, the output function chains only half
10298 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10299 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10300 in all of 'state' being rewritten, with the new values depending
10301 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10304 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10305 the handshake is continued after ssl_verify_cert_chain();
10306 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10307 can lead to 'unexplainable' connection aborts later.
10308 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10310 *) Major EVP API cipher revision.
10311 Add hooks for extra EVP features. This allows various cipher
10312 parameters to be set in the EVP interface. Support added for variable
10313 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10314 setting of RC2 and RC5 parameters.
10316 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10319 Remove lots of duplicated code from the EVP library. For example *every*
10320 cipher init() function handles the 'iv' in the same way according to the
10321 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10322 for CFB and OFB modes they zero ctx->num.
10324 New functionality allows removal of S/MIME code RC2 hack.
10326 Most of the routines have the same form and so can be declared in terms
10329 By shifting this to the top level EVP_CipherInit() it can be removed from
10330 all individual ciphers. If the cipher wants to handle IVs or keys
10331 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10334 Change lots of functions like EVP_EncryptUpdate() to now return a
10335 value: although software versions of the algorithms cannot fail
10336 any installed hardware versions can.
10339 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10340 this option is set, tolerate broken clients that send the negotiated
10341 protocol version number instead of the requested protocol version
10345 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10346 i.e. non-zero for export ciphersuites, zero otherwise.
10347 Previous versions had this flag inverted, inconsistent with
10348 rsa_tmp_cb (..._TMP_RSA_CB).
10349 [Bodo Moeller; problem reported by Amit Chopra]
10351 *) Add missing DSA library text string. Work around for some IIS
10352 key files with invalid SEQUENCE encoding.
10355 *) Add a document (doc/standards.txt) that list all kinds of standards
10356 and so on that are implemented in OpenSSL.
10359 *) Enhance c_rehash script. Old version would mishandle certificates
10360 with the same subject name hash and wouldn't handle CRLs at all.
10361 Added -fingerprint option to crl utility, to support new c_rehash
10365 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10368 *) Fix for SSL server purpose checking. Server checking was
10369 rejecting certificates which had extended key usage present
10370 but no ssl client purpose.
10371 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10373 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10374 is a little unclear about how a blank password is handled.
10375 Since the password in encoded as a BMPString with terminating
10376 double NULL a zero length password would end up as just the
10377 double NULL. However no password at all is different and is
10378 handled differently in the PKCS#12 key generation code. NS
10379 treats a blank password as zero length. MSIE treats it as no
10380 password on export: but it will try both on import. We now do
10381 the same: PKCS12_parse() tries zero length and no password if
10382 the password is set to "" or NULL (NULL is now a valid password:
10383 it wasn't before) as does the pkcs12 application.
10386 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10387 perror when PEM_read_bio_X509_REQ fails, the error message must
10388 be obtained from the error queue.
10391 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10392 it in ERR_remove_state if appropriate, and change ERR_get_state
10393 accordingly to avoid race conditions (this is necessary because
10394 thread_hash is no longer constant once set).
10397 *) Bugfix for linux-elf makefile.one.
10400 *) RSA_get_default_method() will now cause a default
10401 RSA_METHOD to be chosen if one doesn't exist already.
10402 Previously this was only set during a call to RSA_new()
10403 or RSA_new_method(NULL) meaning it was possible for
10404 RSA_get_default_method() to return NULL.
10407 *) Added native name translation to the existing DSO code
10408 that will convert (if the flag to do so is set) filenames
10409 that are sufficiently small and have no path information
10410 into a canonical native form. Eg. "blah" converted to
10411 "libblah.so" or "blah.dll" etc.
10414 *) New function ERR_error_string_n(e, buf, len) which is like
10415 ERR_error_string(e, buf), but writes at most 'len' bytes
10416 including the 0 terminator. For ERR_error_string_n, 'buf'
10418 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10420 *) CONF library reworked to become more general. A new CONF
10421 configuration file reader "class" is implemented as well as a
10422 new functions (NCONF_*, for "New CONF") to handle it. The now
10423 old CONF_* functions are still there, but are reimplemented to
10424 work in terms of the new functions. Also, a set of functions
10425 to handle the internal storage of the configuration data is
10426 provided to make it easier to write new configuration file
10427 reader "classes" (I can definitely see something reading a
10428 configuration file in XML format, for example), called _CONF_*,
10429 or "the configuration storage API"...
10431 The new configuration file reading functions are:
10433 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10434 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10436 NCONF_default, NCONF_WIN32
10438 NCONF_dump_fp, NCONF_dump_bio
10440 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10441 NCONF_new creates a new CONF object. This works in the same way
10442 as other interfaces in OpenSSL, like the BIO interface.
10443 NCONF_dump_* dump the internal storage of the configuration file,
10444 which is useful for debugging. All other functions take the same
10445 arguments as the old CONF_* functions wth the exception of the
10446 first that must be a `CONF *' instead of a `LHASH *'.
10448 To make it easer to use the new classes with the old CONF_* functions,
10449 the function CONF_set_default_method is provided.
10452 *) Add '-tls1' option to 'openssl ciphers', which was already
10453 mentioned in the documentation but had not been implemented.
10454 (This option is not yet really useful because even the additional
10455 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10458 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10459 OpenSSL-based applications) load shared libraries and bind to
10460 them in a portable way.
10461 [Geoff Thorpe, with contributions from Richard Levitte]
10463 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10465 *) Make sure _lrotl and _lrotr are only used with MSVC.
10467 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10468 (the default implementation of RAND_status).
10470 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10471 to '-clrext' (= clear extensions), as intended and documented.
10472 [Bodo Moeller; inconsistency pointed out by Michael Attili
10473 <attili@amaxo.com>]
10475 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10476 was larger than the MD block size.
10477 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10479 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10480 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10481 using the passed key: if the passed key was a private key the result
10482 of X509_print(), for example, would be to print out all the private key
10486 *) des_quad_cksum() byte order bug fix.
10487 [Ulf Möller, using the problem description in krb4-0.9.7, where
10488 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10490 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10492 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10494 *) For easily testing in shell scripts whether some command
10495 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10496 returns with exit code 0 iff no command of the given name is available.
10497 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10498 the output goes to stdout and nothing is printed to stderr.
10499 Additional arguments are always ignored.
10501 Since for each cipher there is a command of the same name,
10502 the 'no-cipher' compilation switches can be tested this way.
10504 ('openssl no-XXX' is not able to detect pseudo-commands such
10505 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10508 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10511 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10512 is set; it will be thrown away anyway because each handshake creates
10514 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10515 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10516 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10517 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10520 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10521 'Q' and 'R' lose their special meanings (quit/renegotiate).
10522 This is part of what -quiet does; unlike -quiet, -ign_eof
10523 does not suppress any output.
10526 *) Add compatibility options to the purpose and trust code. The
10527 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10528 accepts a certificate or CA, this was the previous behaviour,
10529 with all the associated security issues.
10531 X509_TRUST_COMPAT is the old trust behaviour: only and
10532 automatically trust self signed roots in certificate store. A
10533 new trust setting X509_TRUST_DEFAULT is used to specify that
10534 a purpose has no associated trust setting and it should instead
10535 use the value in the default purpose.
10538 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10539 and fix a memory leak.
10542 *) In util/mkerr.pl (which implements 'make errors'), preserve
10543 reason strings from the previous version of the .c file, as
10544 the default to have only downcase letters (and digits) in
10545 automatically generated reasons codes is not always appropriate.
10548 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10549 using strerror. Previously, ERR_reason_error_string() returned
10550 library names as reason strings for SYSerr; but SYSerr is a special
10551 case where small numbers are errno values, not library numbers.
10554 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10555 converts DSA parameters into DH parameters. (When creating parameters,
10556 DSA_generate_parameters is used.)
10559 *) Include 'length' (recommended exponent length) in C code generated
10560 by 'openssl dhparam -C'.
10563 *) The second argument to set_label in perlasm was already being used
10564 so couldn't be used as a "file scope" flag. Moved to third argument
10568 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10569 instead of RAND_bytes for encryption IVs and salts.
10572 *) Include RAND_status() into RAND_METHOD instead of implementing
10573 it only for md_rand.c Otherwise replacing the PRNG by calling
10574 RAND_set_rand_method would be impossible.
10577 *) Don't let DSA_generate_key() enter an infinite loop if the random
10578 number generation fails.
10581 *) New 'rand' application for creating pseudo-random output.
10584 *) Added configuration support for Linux/IA64
10585 [Rolf Haberrecker <rolf@suse.de>]
10587 *) Assembler module support for Mingw32.
10590 *) Shared library support for HPUX (in shlib/).
10591 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10593 *) Shared library support for Solaris gcc.
10594 [Lutz Behnke <behnke@trustcenter.de>]
10596 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10598 *) PKCS7_encrypt() was adding text MIME headers twice because they
10599 were added manually and by SMIME_crlf_copy().
10602 *) In bntest.c don't call BN_rand with zero bits argument.
10603 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10605 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10606 case was implemented. This caused BN_div_recp() to fail occasionally.
10609 *) Add an optional second argument to the set_label() in the perl
10610 assembly language builder. If this argument exists and is set
10611 to 1 it signals that the assembler should use a symbol whose
10612 scope is the entire file, not just the current function. This
10613 is needed with MASM which uses the format label:: for this scope.
10614 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10616 *) Change the ASN1 types so they are typedefs by default. Before
10617 almost all types were #define'd to ASN1_STRING which was causing
10618 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10622 *) Change names of new functions to the new get1/get0 naming
10623 convention: After 'get1', the caller owns a reference count
10624 and has to call ..._free; 'get0' returns a pointer to some
10625 data structure without incrementing reference counters.
10626 (Some of the existing 'get' functions increment a reference
10627 counter, some don't.)
10628 Similarly, 'set1' and 'add1' functions increase reference
10629 counters or duplicate objects.
10632 *) Allow for the possibility of temp RSA key generation failure:
10633 the code used to assume it always worked and crashed on failure.
10636 *) Fix potential buffer overrun problem in BIO_printf().
10637 [Ulf Möller, using public domain code by Patrick Powell; problem
10638 pointed out by David Sacerdote <das33@cornell.edu>]
10640 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10641 RAND_egd() and RAND_status(). In the command line application,
10642 the EGD socket can be specified like a seed file using RANDFILE
10646 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10647 Some CAs (e.g. Verisign) distribute certificates in this form.
10650 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10651 list to exclude them. This means that no special compilation option
10652 is needed to use anonymous DH: it just needs to be included in the
10656 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10657 EVP_MD_type. The old functionality is available in a new macro called
10658 EVP_MD_md(). Change code that uses it and update docs.
10661 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10662 where the 'void *' argument is replaced by a function pointer argument.
10663 Previously 'void *' was abused to point to functions, which works on
10664 many platforms, but is not correct. As these functions are usually
10665 called by macros defined in OpenSSL header files, most source code
10666 should work without changes.
10669 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10670 sections with information on -D... compiler switches used for
10671 compiling the library so that applications can see them. To enable
10672 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10673 must be defined. E.g.,
10674 #define OPENSSL_ALGORITHM_DEFINES
10675 #include <openssl/opensslconf.h>
10676 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10677 [Richard Levitte, Ulf and Bodo Möller]
10679 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10683 *) Change the 'other' type in certificate aux info to a STACK_OF
10684 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10685 the required ASN1 format: arbitrary types determined by an OID.
10688 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10689 argument to 'req'. This is not because the function is newer or
10690 better than others it just uses the work 'NEW' in the certificate
10691 request header lines. Some software needs this.
10694 *) Reorganise password command line arguments: now passwords can be
10695 obtained from various sources. Delete the PEM_cb function and make
10696 it the default behaviour: i.e. if the callback is NULL and the
10697 usrdata argument is not NULL interpret it as a null terminated pass
10698 phrase. If usrdata and the callback are NULL then the pass phrase
10699 is prompted for as usual.
10702 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10703 the support is automatically enabled. The resulting binaries will
10704 autodetect the card and use it if present.
10705 [Ben Laurie and Compaq Inc.]
10707 *) Work around for Netscape hang bug. This sends certificate request
10708 and server done in one record. Since this is perfectly legal in the
10709 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10710 the bugs/SSLv3 entry for more info.
10713 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10716 *) Add -rand argument to smime and pkcs12 applications and read/write
10720 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10723 *) Add command line password options to the remaining applications.
10726 *) Bug fix for BN_div_recp() for numerators with an even number of
10730 *) More tests in bntest.c, and changed test_bn output.
10733 *) ./config recognizes MacOS X now.
10736 *) Bug fix for BN_div() when the first words of num and divisor are
10737 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10740 *) Add support for various broken PKCS#8 formats, and command line
10741 options to produce them.
10744 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10745 get temporary BIGNUMs from a BN_CTX.
10748 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10752 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10753 include a #define from the old name to the new. The original intent
10754 was that statically linked binaries could for example just call
10755 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10756 link with digests. This never worked because SSLeay_add_all_digests()
10757 and SSLeay_add_all_ciphers() were in the same source file so calling
10758 one would link with the other. They are now in separate source files.
10761 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10764 *) Use a less unusual form of the Miller-Rabin primality test (it used
10765 a binary algorithm for exponentiation integrated into the Miller-Rabin
10766 loop, our standard modexp algorithms are faster).
10769 *) Support for the EBCDIC character set completed.
10770 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10772 *) Source code cleanups: use const where appropriate, eliminate casts,
10773 use void * instead of char * in lhash.
10776 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10777 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10778 this the server could overwrite ephemeral keys that the client
10782 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10783 using 50 iterations of the Rabin-Miller test.
10785 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10786 iterations of the Rabin-Miller test as required by the appendix
10787 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10788 As BN_is_prime_fasttest includes trial division, DSA parameter
10789 generation becomes much faster.
10791 This implies a change for the callback functions in DSA_is_prime
10792 and DSA_generate_parameters: The callback function is called once
10793 for each positive witness in the Rabin-Miller test, not just
10794 occasionally in the inner loop; and the parameters to the
10795 callback function now provide an iteration count for the outer
10796 loop rather than for the current invocation of the inner loop.
10797 DSA_generate_parameters additionally can call the callback
10798 function with an 'iteration count' of -1, meaning that a
10799 candidate has passed the trial division test (when q is generated
10800 from an application-provided seed, trial division is skipped).
10803 *) New function BN_is_prime_fasttest that optionally does trial
10804 division before starting the Rabin-Miller test and has
10805 an additional BN_CTX * argument (whereas BN_is_prime always
10806 has to allocate at least one BN_CTX).
10807 'callback(1, -1, cb_arg)' is called when a number has passed the
10808 trial division stage.
10811 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10815 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10818 *) New function BN_pseudo_rand().
10821 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10822 bignum version of BN_from_montgomery() with the working code from
10823 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10827 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10828 made it impossible to use the same SSL_SESSION data structure in
10829 SSL2 clients in multiple threads.
10832 *) The return value of RAND_load_file() no longer counts bytes obtained
10833 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10834 to seed the PRNG (previously an explicit byte count was required).
10835 [Ulf Möller, Bodo Möller]
10837 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10838 used (char *) instead of (void *) and had casts all over the place.
10841 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10844 *) Retain source code compatibility for BN_prime_checks macro:
10845 BN_is_prime(..., BN_prime_checks, ...) now uses
10846 BN_prime_checks_for_size to determine the appropriate number of
10847 Rabin-Miller iterations.
10850 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10851 DH_CHECK_P_NOT_SAFE_PRIME.
10852 (Check if this is true? OpenPGP calls them "strong".)
10855 *) Merge the functionality of "dh" and "gendh" programs into a new program
10856 "dhparam". The old programs are retained for now but will handle DH keys
10857 (instead of parameters) in future.
10860 *) Make the ciphers, s_server and s_client programs check the return values
10861 when a new cipher list is set.
10864 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10865 ciphers. Before when the 56bit ciphers were enabled the sorting was
10868 The syntax for the cipher sorting has been extended to support sorting by
10869 cipher-strength (using the strength_bits hard coded in the tables).
10870 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10872 Fix a bug in the cipher-command parser: when supplying a cipher command
10873 string with an "undefined" symbol (neither command nor alphanumeric
10874 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10875 an error is flagged.
10877 Due to the strength-sorting extension, the code of the
10878 ssl_create_cipher_list() function was completely rearranged. I hope that
10879 the readability was also increased :-)
10880 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10882 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10883 for the first serial number and places 2 in the serial number file. This
10884 avoids problems when the root CA is created with serial number zero and
10885 the first user certificate has the same issuer name and serial number
10889 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10890 the new code. Add documentation for this stuff.
10893 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10894 X509_*() to X509at_*() on the grounds that they don't handle X509
10895 structures and behave in an analogous way to the X509v3 functions:
10896 they shouldn't be called directly but wrapper functions should be used
10899 So we also now have some wrapper functions that call the X509at functions
10900 when passed certificate requests. (TO DO: similar things can be done with
10901 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10902 things. Some of these need some d2i or i2d and print functionality
10903 because they handle more complex structures.)
10906 *) Add missing #ifndefs that caused missing symbols when building libssl
10907 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10908 NO_RSA in ssl/s2*.c.
10909 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10911 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10912 has a return value which indicates the quality of the random data
10913 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10914 error queue. New function RAND_pseudo_bytes() generates output that is
10915 guaranteed to be unique but not unpredictable. RAND_add is like
10916 RAND_seed, but takes an extra argument for an entropy estimate
10917 (RAND_seed always assumes full entropy).
10920 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10921 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10922 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10923 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10924 false-positive rate of at most 2^-80 for random input.
10927 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10930 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10931 in the 0.9.5 release), this returns the chain
10932 from an X509_CTX structure with a dup of the stack and all
10933 the X509 reference counts upped: so the stack will exist
10934 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10937 Also make SSL_SESSION_print() print out the verify return
10941 *) Add manpage for the pkcs12 command. Also change the default
10942 behaviour so MAC iteration counts are used unless the new
10943 -nomaciter option is used. This improves file security and
10944 only older versions of MSIE (4.0 for example) need it.
10947 *) Honor the no-xxx Configure options when creating .DEF files.
10950 *) Add PKCS#10 attributes to field table: challengePassword,
10951 unstructuredName and unstructuredAddress. These are taken from
10952 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10953 international characters are used.
10955 More changes to X509_ATTRIBUTE code: allow the setting of types
10956 based on strings. Remove the 'loc' parameter when adding
10957 attributes because these will be a SET OF encoding which is sorted
10961 *) Initial changes to the 'req' utility to allow request generation
10962 automation. This will allow an application to just generate a template
10963 file containing all the field values and have req construct the
10966 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10967 used all over the place including certificate requests and PKCS#7
10968 structures. They are currently handled manually where necessary with
10969 some primitive wrappers for PKCS#7. The new functions behave in a
10970 manner analogous to the X509 extension functions: they allow
10971 attributes to be looked up by NID and added.
10973 Later something similar to the X509V3 code would be desirable to
10974 automatically handle the encoding, decoding and printing of the
10975 more complex types. The string types like challengePassword can
10976 be handled by the string table functions.
10978 Also modified the multi byte string table handling. Now there is
10979 a 'global mask' which masks out certain types. The table itself
10980 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10981 is useful when for example there is only one permissible type
10982 (as in countryName) and using the mask might result in no valid
10986 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10987 SSL_get_peer_finished to allow applications to obtain the latest
10988 Finished messages sent to the peer or expected from the peer,
10989 respectively. (SSL_get_peer_finished is usually the Finished message
10990 actually received from the peer, otherwise the protocol will be aborted.)
10992 As the Finished message are message digests of the complete handshake
10993 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10994 be used for external authentication procedures when the authentication
10995 provided by SSL/TLS is not desired or is not enough.
10998 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10999 the host supports BWX extension and if Compaq C is present on the
11000 $PATH. Just exploiting of the BWX extension results in 20-30%
11001 performance kick for some algorithms, e.g. DES and RC4 to mention
11002 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11006 *) Add support for MS "fast SGC". This is arguably a violation of the
11007 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11008 weak crypto and after checking the certificate is SGC a second one
11009 with strong crypto. MS SGC stops the first handshake after receiving
11010 the server certificate message and sends a second client hello. Since
11011 a server will typically do all the time consuming operations before
11012 expecting any further messages from the client (server key exchange
11013 is the most expensive) there is little difference between the two.
11015 To get OpenSSL to support MS SGC we have to permit a second client
11016 hello message after we have sent server done. In addition we have to
11017 reset the MAC if we do get this second client hello.
11020 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11021 if a DER encoded private key is RSA or DSA traditional format. Changed
11022 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11023 format DER encoded private key. Newer code should use PKCS#8 format which
11024 has the key type encoded in the ASN1 structure. Added DER private key
11025 support to pkcs8 application.
11028 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11029 ciphersuites has been selected (as required by the SSL 3/TLS 1
11030 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11031 is set, we interpret this as a request to violate the specification
11032 (the worst that can happen is a handshake failure, and 'correct'
11033 behaviour would result in a handshake failure anyway).
11036 *) In SSL_CTX_add_session, take into account that there might be multiple
11037 SSL_SESSION structures with the same session ID (e.g. when two threads
11038 concurrently obtain them from an external cache).
11039 The internal cache can handle only one SSL_SESSION with a given ID,
11040 so if there's a conflict, we now throw out the old one to achieve
11044 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11045 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11046 some routines that use cipher OIDs: some ciphers do not have OIDs
11047 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11051 *) Simplify the trust setting structure and code. Now we just have
11052 two sequences of OIDs for trusted and rejected settings. These will
11053 typically have values the same as the extended key usage extension
11054 and any application specific purposes.
11056 The trust checking code now has a default behaviour: it will just
11057 check for an object with the same NID as the passed id. Functions can
11058 be provided to override either the default behaviour or the behaviour
11059 for a given id. SSL client, server and email already have functions
11060 in place for compatibility: they check the NID and also return "trusted"
11061 if the certificate is self signed.
11064 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11065 traditional format into an EVP_PKEY structure.
11068 *) Add a password callback function PEM_cb() which either prompts for
11069 a password if usr_data is NULL or otherwise assumes it is a null
11070 terminated password. Allow passwords to be passed on command line
11071 environment or config files in a few more utilities.
11074 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11075 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11076 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11077 Update documentation.
11080 *) Support for ASN1 "NULL" type. This could be handled before by using
11081 ASN1_TYPE but there wasn't any function that would try to read a NULL
11082 and produce an error if it couldn't. For compatibility we also have
11083 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11084 don't allocate anything because they don't need to.
11087 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11089 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11091 *) Rebuild of the memory allocation routines used by OpenSSL code and
11092 possibly others as well. The purpose is to make an interface that
11093 provide hooks so anyone can build a separate set of allocation and
11094 deallocation routines to be used by OpenSSL, for example memory
11095 pool implementations, or something else, which was previously hard
11096 since Malloc(), Realloc() and Free() were defined as macros having
11097 the values malloc, realloc and free, respectively (except for Win32
11098 compilations). The same is provided for memory debugging code.
11099 OpenSSL already comes with functionality to find memory leaks, but
11100 this gives people a chance to debug other memory problems.
11102 With these changes, a new set of functions and macros have appeared:
11104 CRYPTO_set_mem_debug_functions() [F]
11105 CRYPTO_get_mem_debug_functions() [F]
11106 CRYPTO_dbg_set_options() [F]
11107 CRYPTO_dbg_get_options() [F]
11108 CRYPTO_malloc_debug_init() [M]
11110 The memory debug functions are NULL by default, unless the library
11111 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11112 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11113 gives the standard debugging functions that come with OpenSSL) or
11114 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11115 provided by the library user) must be used. When the standard
11116 debugging functions are used, CRYPTO_dbg_set_options can be used to
11117 request additional information:
11118 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11119 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11121 Also, things like CRYPTO_set_mem_functions will always give the
11122 expected result (the new set of functions is used for allocation
11123 and deallocation) at all times, regardless of platform and compiler
11126 To finish it up, some functions that were never use in any other
11127 way than through macros have a new API and new semantic:
11129 CRYPTO_dbg_malloc()
11130 CRYPTO_dbg_realloc()
11133 All macros of value have retained their old syntax.
11134 [Richard Levitte and Bodo Moeller]
11136 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11137 ordering of SMIMECapabilities wasn't in "strength order" and there
11138 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11142 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11143 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11144 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11146 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11147 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11148 functionality to handle multipart/signed properly) and a utility
11149 called 'smime' to call all this stuff. This is based on code I
11150 originally wrote for Celo who have kindly allowed it to be
11151 included in OpenSSL.
11154 *) Add variants des_set_key_checked and des_set_key_unchecked of
11155 des_set_key (aka des_key_sched). Global variable des_check_key
11156 decides which of these is called by des_set_key; this way
11157 des_check_key behaves as it always did, but applications and
11158 the library itself, which was buggy for des_check_key == 1,
11159 have a cleaner way to pick the version they need.
11162 *) New function PKCS12_newpass() which changes the password of a
11166 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11167 dynamic mix. In both cases the ids can be used as an index into the
11168 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11169 functions so they accept a list of the field values and the
11170 application doesn't need to directly manipulate the X509_TRUST
11174 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11178 *) Modify the way the V3 extension code looks up extensions. This now
11179 works in a similar way to the object code: we have some "standard"
11180 extensions in a static table which is searched with OBJ_bsearch()
11181 and the application can add dynamic ones if needed. The file
11182 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11183 updated whenever a new extension is added to the core code and kept
11184 in ext_nid order. There is a simple program 'tabtest.c' which checks
11185 this. New extensions are not added too often so this file can readily
11186 be maintained manually.
11188 There are two big advantages in doing things this way. The extensions
11189 can be looked up immediately and no longer need to be "added" using
11190 X509V3_add_standard_extensions(): this function now does nothing.
11191 [Side note: I get *lots* of email saying the extension code doesn't
11192 work because people forget to call this function]
11193 Also no dynamic allocation is done unless new extensions are added:
11194 so if we don't add custom extensions there is no need to call
11195 X509V3_EXT_cleanup().
11198 *) Modify enc utility's salting as follows: make salting the default. Add a
11199 magic header, so unsalted files fail gracefully instead of just decrypting
11200 to garbage. This is because not salting is a big security hole, so people
11201 should be discouraged from doing it.
11204 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11205 digest to be passed on the command line but it only used this
11206 parameter when signing a certificate. Modified so all relevant
11207 operations are affected by the digest parameter including the
11208 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11209 DSA key was used because it didn't fix the digest.
11212 *) Initial certificate chain verify code. Currently tests the untrusted
11213 certificates for consistency with the verify purpose (which is set
11214 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11216 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11217 this is because it will reject chains with invalid extensions whereas
11218 every previous version of OpenSSL and SSLeay made no checks at all.
11220 Trust code: checks the root CA for the relevant trust settings. Trust
11221 settings have an initial value consistent with the verify purpose: e.g.
11222 if the verify purpose is for SSL client use it expects the CA to be
11223 trusted for SSL client use. However the default value can be changed to
11224 permit custom trust settings: one example of this would be to only trust
11225 certificates from a specific "secure" set of CAs.
11227 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11228 which should be used for version portability: especially since the
11229 verify structure is likely to change more often now.
11231 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11232 to set them. If not set then assume SSL clients will verify SSL servers
11235 Two new options to the verify program: -untrusted allows a set of
11236 untrusted certificates to be passed in and -purpose which sets the
11237 intended purpose of the certificate. If a purpose is set then the
11238 new chain verify code is used to check extension consistency.
11241 *) Support for the authority information access extension.
11244 *) Modify RSA and DSA PEM read routines to transparently handle
11245 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11246 public keys in a format compatible with certificate
11247 SubjectPublicKeyInfo structures. Unfortunately there were already
11248 functions called *_PublicKey_* which used various odd formats so
11249 these are retained for compatibility: however the DSA variants were
11250 never in a public release so they have been deleted. Changed dsa/rsa
11251 utilities to handle the new format: note no releases ever handled public
11252 keys so we should be OK.
11254 The primary motivation for this change is to avoid the same fiasco
11255 that dogs private keys: there are several incompatible private key
11256 formats some of which are standard and some OpenSSL specific and
11257 require various evil hacks to allow partial transparent handling and
11258 even then it doesn't work with DER formats. Given the option anything
11259 other than PKCS#8 should be dumped: but the other formats have to
11260 stay in the name of compatibility.
11262 With public keys and the benefit of hindsight one standard format
11263 is used which works with EVP_PKEY, RSA or DSA structures: though
11264 it clearly returns an error if you try to read the wrong kind of key.
11266 Added a -pubkey option to the 'x509' utility to output the public key.
11267 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11268 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11269 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11270 that do the same as the EVP_PKEY_assign_*() except they up the
11271 reference count of the added key (they don't "swallow" the
11275 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11276 CRLs would fail if the file contained no certificates or no CRLs:
11277 added a new function to read in both types and return the number
11278 read: this means that if none are read it will be an error. The
11279 DER versions of the certificate and CRL reader would always fail
11280 because it isn't possible to mix certificates and CRLs in DER format
11281 without choking one or the other routine. Changed this to just read
11282 a certificate: this is the best we can do. Also modified the code
11283 in apps/verify.c to take notice of return codes: it was previously
11284 attempting to read in certificates from NULL pointers and ignoring
11285 any errors: this is one reason why the cert and CRL reader seemed
11286 to work. It doesn't check return codes from the default certificate
11287 routines: these may well fail if the certificates aren't installed.
11290 *) Code to support otherName option in GeneralName.
11293 *) First update to verify code. Change the verify utility
11294 so it warns if it is passed a self signed certificate:
11295 for consistency with the normal behaviour. X509_verify
11296 has been modified to it will now verify a self signed
11297 certificate if *exactly* the same certificate appears
11298 in the store: it was previously impossible to trust a
11299 single self signed certificate. This means that:
11300 openssl verify ss.pem
11301 now gives a warning about a self signed certificate but
11302 openssl verify -CAfile ss.pem ss.pem
11306 *) For servers, store verify_result in SSL_SESSION data structure
11307 (and add it to external session representation).
11308 This is needed when client certificate verifications fails,
11309 but an application-provided verification callback (set by
11310 SSL_CTX_set_cert_verify_callback) allows accepting the session
11311 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11312 but returns 1): When the session is reused, we have to set
11313 ssl->verify_result to the appropriate error code to avoid
11315 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11317 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11318 case in PKCS7_dataInit() where the signed PKCS7 structure
11319 didn't contain any existing data because it was being created.
11320 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11322 *) Add a salt to the key derivation routines in enc.c. This
11323 forms the first 8 bytes of the encrypted file. Also add a
11324 -S option to allow a salt to be input on the command line.
11327 *) New function X509_cmp(). Oddly enough there wasn't a function
11328 to compare two certificates. We do this by working out the SHA1
11329 hash and comparing that. X509_cmp() will be needed by the trust
11333 *) SSL_get1_session() is like SSL_get_session(), but increments
11334 the reference count in the SSL_SESSION returned.
11335 [Geoff Thorpe <geoff@eu.c2.net>]
11337 *) Fix for 'req': it was adding a null to request attributes.
11338 Also change the X509_LOOKUP and X509_INFO code to handle
11339 certificate auxiliary information.
11342 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11346 *) Add the possibility to add extra information to the memory leak
11347 detecting output, to form tracebacks, showing from where each
11348 allocation was originated: CRYPTO_push_info("constant string") adds
11349 the string plus current file name and line number to a per-thread
11350 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11351 is like calling CYRPTO_pop_info() until the stack is empty.
11352 Also updated memory leak detection code to be multi-thread-safe.
11355 *) Add options -text and -noout to pkcs7 utility and delete the
11356 encryption options which never did anything. Update docs.
11359 *) Add options to some of the utilities to allow the pass phrase
11360 to be included on either the command line (not recommended on
11361 OSes like Unix) or read from the environment. Update the
11362 manpages and fix a few bugs.
11365 *) Add a few manpages for some of the openssl commands.
11368 *) Fix the -revoke option in ca. It was freeing up memory twice,
11369 leaking and not finding already revoked certificates.
11372 *) Extensive changes to support certificate auxiliary information.
11373 This involves the use of X509_CERT_AUX structure and X509_AUX
11374 functions. An X509_AUX function such as PEM_read_X509_AUX()
11375 can still read in a certificate file in the usual way but it
11376 will also read in any additional "auxiliary information". By
11377 doing things this way a fair degree of compatibility can be
11378 retained: existing certificates can have this information added
11379 using the new 'x509' options.
11381 Current auxiliary information includes an "alias" and some trust
11382 settings. The trust settings will ultimately be used in enhanced
11383 certificate chain verification routines: currently a certificate
11384 can only be trusted if it is self signed and then it is trusted
11388 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11389 The problem was that one of the replacement routines had not been working
11390 since SSLeay releases. For now the offending routine has been replaced
11391 with non-optimised assembler. Even so, this now gives around 95%
11392 performance improvement for 1024 bit RSA signs.
11395 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11396 handling. Most clients have the effective key size in bits equal to
11397 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11398 A few however don't do this and instead use the size of the decrypted key
11399 to determine the RC2 key length and the AlgorithmIdentifier to determine
11400 the effective key length. In this case the effective key length can still
11401 be 40 bits but the key length can be 168 bits for example. This is fixed
11402 by manually forcing an RC2 key into the EVP_PKEY structure because the
11403 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11404 the key length and effective key length are equal.
11407 *) Add a bunch of functions that should simplify the creation of
11408 X509_NAME structures. Now you should be able to do:
11409 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11410 and have it automatically work out the correct field type and fill in
11411 the structures. The more adventurous can try:
11412 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11413 and it will (hopefully) work out the correct multibyte encoding.
11416 *) Change the 'req' utility to use the new field handling and multibyte
11417 copy routines. Before the DN field creation was handled in an ad hoc
11418 way in req, ca, and x509 which was rather broken and didn't support
11419 BMPStrings or UTF8Strings. Since some software doesn't implement
11420 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11421 using the dirstring_type option. See the new comment in the default
11422 openssl.cnf for more info.
11425 *) Make crypto/rand/md_rand.c more robust:
11426 - Assure unique random numbers after fork().
11427 - Make sure that concurrent threads access the global counter and
11428 md serializably so that we never lose entropy in them
11429 or use exactly the same state in multiple threads.
11430 Access to the large state is not always serializable because
11431 the additional locking could be a performance killer, and
11432 md should be large enough anyway.
11435 *) New file apps/app_rand.c with commonly needed functionality
11436 for handling the random seed file.
11438 Use the random seed file in some applications that previously did not:
11440 dsaparam -genkey (which also ignored its '-rand' option),
11443 x509 (when signing).
11444 Except on systems with /dev/urandom, it is crucial to have a random
11445 seed file at least for key creation, DSA signing, and for DH exchanges;
11446 for RSA signatures we could do without one.
11448 gendh and gendsa (unlike genrsa) used to read only the first byte
11449 of each file listed in the '-rand' option. The function as previously
11450 found in genrsa is now in app_rand.c and is used by all programs
11451 that support '-rand'.
11454 *) In RAND_write_file, use mode 0600 for creating files;
11455 don't just chmod when it may be too late.
11458 *) Report an error from X509_STORE_load_locations
11459 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11462 *) New function ASN1_mbstring_copy() this copies a string in either
11463 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11464 into an ASN1_STRING type. A mask of permissible types is passed
11465 and it chooses the "minimal" type to use or an error if not type
11469 *) Add function equivalents to the various macros in asn1.h. The old
11470 macros are retained with an M_ prefix. Code inside the library can
11471 use the M_ macros. External code (including the openssl utility)
11472 should *NOT* in order to be "shared library friendly".
11475 *) Add various functions that can check a certificate's extensions
11476 to see if it usable for various purposes such as SSL client,
11477 server or S/MIME and CAs of these types. This is currently
11478 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11479 verification. Also added a -purpose flag to x509 utility to
11480 print out all the purposes.
11483 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11487 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11488 for, obtain and decode and extension and obtain its critical flag.
11489 This allows all the necessary extension code to be handled in a
11490 single function call.
11493 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11494 platforms. See crypto/rc4/rc4_enc.c for further details.
11497 *) New -noout option to asn1parse. This causes no output to be produced
11498 its main use is when combined with -strparse and -out to extract data
11499 from a file (which may not be in ASN.1 format).
11502 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11503 when producing the local key id.
11504 [Richard Levitte <levitte@stacken.kth.se>]
11506 *) New option -dhparam in s_server. This allows a DH parameter file to be
11507 stated explicitly. If it is not stated then it tries the first server
11508 certificate file. The previous behaviour hard coded the filename
11512 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11513 a public key to be input or output. For example:
11514 openssl rsa -in key.pem -pubout -out pubkey.pem
11515 Also added necessary DSA public key functions to handle this.
11518 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11519 in the message. This was handled by allowing
11520 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11521 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11523 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11524 to the end of the strings whereas this didn't. This would cause problems
11525 if strings read with d2i_ASN1_bytes() were later modified.
11526 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11528 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11529 data and it contains EOF it will end up returning an error. This is
11530 caused by input 46 bytes long. The cause is due to the way base64
11531 BIOs find the start of base64 encoded data. They do this by trying a
11532 trial decode on each line until they find one that works. When they
11533 do a flag is set and it starts again knowing it can pass all the
11534 data directly through the decoder. Unfortunately it doesn't reset
11535 the context it uses. This means that if EOF is reached an attempt
11536 is made to pass two EOFs through the context and this causes the
11537 resulting error. This can also cause other problems as well. As is
11538 usual with these problems it takes *ages* to find and the fix is
11539 trivial: move one line.
11540 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11542 *) Ugly workaround to get s_client and s_server working under Windows. The
11543 old code wouldn't work because it needed to select() on sockets and the
11544 tty (for keypresses and to see if data could be written). Win32 only
11545 supports select() on sockets so we select() with a 1s timeout on the
11546 sockets and then see if any characters are waiting to be read, if none
11547 are present then we retry, we also assume we can always write data to
11548 the tty. This isn't nice because the code then blocks until we've
11549 received a complete line of data and it is effectively polling the
11550 keyboard at 1s intervals: however it's quite a bit better than not
11551 working at all :-) A dedicated Windows application might handle this
11552 with an event loop for example.
11555 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11556 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11557 will be called when RSA_sign() and RSA_verify() are used. This is useful
11558 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11559 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11560 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11561 This necessitated the support of an extra signature type NID_md5_sha1
11562 for SSL signatures and modifications to the SSL library to use it instead
11563 of calling RSA_public_decrypt() and RSA_private_encrypt().
11566 *) Add new -verify -CAfile and -CApath options to the crl program, these
11567 will lookup a CRL issuers certificate and verify the signature in a
11568 similar way to the verify program. Tidy up the crl program so it
11569 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11570 less strict. It will now permit CRL extensions even if it is not
11571 a V2 CRL: this will allow it to tolerate some broken CRLs.
11574 *) Initialize all non-automatic variables each time one of the openssl
11575 sub-programs is started (this is necessary as they may be started
11576 multiple times from the "OpenSSL>" prompt).
11577 [Lennart Bang, Bodo Moeller]
11579 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11580 removing all other RSA functionality (this is what NO_RSA does). This
11581 is so (for example) those in the US can disable those operations covered
11582 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11586 *) Non-copying interface to BIO pairs.
11587 (still largely untested)
11590 *) New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
11591 ASCII string. This was handled independently in various places before.
11594 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11595 UTF8 strings a character at a time.
11598 *) Use client_version from client hello to select the protocol
11599 (s23_srvr.c) and for RSA client key exchange verification
11600 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11603 *) Add various utility functions to handle SPKACs, these were previously
11604 handled by poking round in the structure internals. Added new function
11605 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11606 print, verify and generate SPKACs. Based on an original idea from
11607 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11610 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11613 *) Allow the config file extension section to be overwritten on the
11614 command line. Based on an original idea from Massimiliano Pala
11615 <madwolf@comune.modena.it>. The new option is called -extensions
11616 and can be applied to ca, req and x509. Also -reqexts to override
11617 the request extensions in req and -crlexts to override the crl extensions
11621 *) Add new feature to the SPKAC handling in ca. Now you can include
11622 the same field multiple times by preceding it by "XXXX." for example:
11625 this is the same syntax as used in the req config file.
11628 *) Allow certificate extensions to be added to certificate requests. These
11629 are specified in a 'req_extensions' option of the req section of the
11630 config file. They can be printed out with the -text option to req but
11631 are otherwise ignored at present.
11634 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11635 data read consists of only the final block it would not decrypted because
11636 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11637 A misplaced 'break' also meant the decrypted final block might not be
11638 copied until the next read.
11641 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11642 a few extra parameters to the DH structure: these will be useful if
11643 for example we want the value of 'q' or implement X9.42 DH.
11646 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11647 provides hooks that allow the default DSA functions or functions on a
11648 "per key" basis to be replaced. This allows hardware acceleration and
11649 hardware key storage to be handled without major modification to the
11650 library. Also added low level modexp hooks and CRYPTO_EX structure and
11651 associated functions.
11654 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11655 as "read only": it can't be written to and the buffer it points to will
11656 not be freed. Reading from a read only BIO is much more efficient than
11657 a normal memory BIO. This was added because there are several times when
11658 an area of memory needs to be read from a BIO. The previous method was
11659 to create a memory BIO and write the data to it, this results in two
11660 copies of the data and an O(n^2) reading algorithm. There is a new
11661 function BIO_new_mem_buf() which creates a read only memory BIO from
11662 an area of memory. Also modified the PKCS#7 routines to use read only
11666 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11667 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11668 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11669 but a retry condition occurred while trying to read the rest.
11672 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11673 NID_pkcs7_encrypted by default: this was wrong since this should almost
11674 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11675 the encrypted data type: this is a more sensible place to put it and it
11676 allows the PKCS#12 code to be tidied up that duplicated this
11680 *) Changed obj_dat.pl script so it takes its input and output files on
11681 the command line. This should avoid shell escape redirection problems
11685 *) Initial support for certificate extension requests, these are included
11686 in things like Xenroll certificate requests. Included functions to allow
11687 extensions to be obtained and added.
11690 *) -crlf option to s_client and s_server for sending newlines as
11691 CRLF (as required by many protocols).
11694 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11696 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11697 [Ralf S. Engelschall]
11699 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11700 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11702 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11706 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11707 DH parameters/keys (q is lost during that conversion, but the resulting
11708 DH parameters contain its length).
11710 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11711 much faster than DH_generate_parameters (which creates parameters
11712 where p = 2*q + 1), and also the smaller q makes DH computations
11713 much more efficient (160-bit exponentiation instead of 1024-bit
11714 exponentiation); so this provides a convenient way to support DHE
11715 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11716 utter importance to use
11717 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11719 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11720 when such DH parameters are used, because otherwise small subgroup
11721 attacks may become possible!
11724 *) Avoid memory leak in i2d_DHparams.
11727 *) Allow the -k option to be used more than once in the enc program:
11728 this allows the same encrypted message to be read by multiple recipients.
11731 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11732 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11733 it will always use the numerical form of the OID, even if it has a short
11737 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11738 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11739 otherwise bn_mod_exp was called. In the case of hardware keys for example
11740 no private key components need be present and it might store extra data
11741 in the RSA structure, which cannot be accessed from bn_mod_exp.
11742 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11743 private key operations.
11746 *) Added support for SPARC Linux.
11749 *) pem_password_cb function type incompatibly changed from
11750 typedef int pem_password_cb(char *buf, int size, int rwflag);
11752 ....(char *buf, int size, int rwflag, void *userdata);
11753 so that applications can pass data to their callbacks:
11754 The PEM[_ASN1]_{read,write}... functions and macros now take an
11755 additional void * argument, which is just handed through whenever
11756 the password callback is called.
11757 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11759 New function SSL_CTX_set_default_passwd_cb_userdata.
11761 Compatibility note: As many C implementations push function arguments
11762 onto the stack in reverse order, the new library version is likely to
11763 interoperate with programs that have been compiled with the old
11764 pem_password_cb definition (PEM_whatever takes some data that
11765 happens to be on the stack as its last argument, and the callback
11766 just ignores this garbage); but there is no guarantee whatsoever that
11769 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11770 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11771 problems not only on Windows, but also on some Unix platforms.
11772 To avoid problematic command lines, these definitions are now in an
11773 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11774 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11777 *) MIPS III/IV assembler module is reimplemented.
11780 *) More DES library cleanups: remove references to srand/rand and
11781 delete an unused file.
11784 *) Add support for the free Netwide assembler (NASM) under Win32,
11785 since not many people have MASM (ml) and it can be hard to obtain.
11786 This is currently experimental but it seems to work OK and pass all
11787 the tests. Check out INSTALL.W32 for info.
11790 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11791 without temporary keys kept an extra copy of the server key,
11792 and connections with temporary keys did not free everything in case
11796 *) New function RSA_check_key and new openssl rsa option -check
11797 for verifying the consistency of RSA keys.
11798 [Ulf Moeller, Bodo Moeller]
11800 *) Various changes to make Win32 compile work:
11801 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11802 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11803 comparison" warnings.
11804 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11807 *) Add a debugging option to PKCS#5 v2 key generation function: when
11808 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11809 derived keys are printed to stderr.
11812 *) Copy the flags in ASN1_STRING_dup().
11813 [Roman E. Pavlov <pre@mo.msk.ru>]
11815 *) The x509 application mishandled signing requests containing DSA
11816 keys when the signing key was also DSA and the parameters didn't match.
11818 It was supposed to omit the parameters when they matched the signing key:
11819 the verifying software was then supposed to automatically use the CA's
11820 parameters if they were absent from the end user certificate.
11822 Omitting parameters is no longer recommended. The test was also
11823 the wrong way round! This was probably due to unusual behaviour in
11824 EVP_cmp_parameters() which returns 1 if the parameters match.
11825 This meant that parameters were omitted when they *didn't* match and
11826 the certificate was useless. Certificates signed with 'ca' didn't have
11828 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11830 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11831 The interface is as follows:
11832 Applications can use
11833 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11834 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11835 "off" is now the default.
11836 The library internally uses
11837 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11838 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11839 to disable memory-checking temporarily.
11841 Some inconsistent states that previously were possible (and were
11842 even the default) are now avoided.
11844 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11845 with each memory chunk allocated; this is occasionally more helpful
11846 than just having a counter.
11848 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11850 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11854 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11855 which largely parallels "options", but is for changing API behaviour,
11856 whereas "options" are about protocol behaviour.
11857 Initial "mode" flags are:
11859 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11860 a single record has been written.
11861 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11862 retries use the same buffer location.
11863 (But all of the contents must be
11867 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11870 *) Fix problems with no-hmac etc.
11871 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11873 *) New functions RSA_get_default_method(), RSA_set_method() and
11874 RSA_get_method(). These allows replacement of RSA_METHODs without having
11875 to mess around with the internals of an RSA structure.
11878 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11879 Also really enable memory leak checks in openssl.c and in some
11881 [Chad C. Mulligan, Bodo Moeller]
11883 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11884 up the length of negative integers. This has now been simplified to just
11885 store the length when it is first determined and use it later, rather
11886 than trying to keep track of where data is copied and updating it to
11888 [Steve Henson, reported by Brien Wheeler
11889 <bwheeler@authentica-security.com>]
11891 *) Add a new function PKCS7_signatureVerify. This allows the verification
11892 of a PKCS#7 signature but with the signing certificate passed to the
11893 function itself. This contrasts with PKCS7_dataVerify which assumes the
11894 certificate is present in the PKCS#7 structure. This isn't always the
11895 case: certificates can be omitted from a PKCS#7 structure and be
11896 distributed by "out of band" means (such as a certificate database).
11899 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11900 function prototypes in pem.h, also change util/mkdef.pl to add the
11901 necessary function names.
11904 *) mk1mf.pl (used by Windows builds) did not properly read the
11905 options set by Configure in the top level Makefile, and Configure
11906 was not even able to write more than one option correctly.
11907 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11910 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11911 file to be loaded from a BIO or FILE pointer. The BIO version will
11912 for example allow memory BIOs to contain config info.
11915 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11916 Whoever hopes to achieve shared-library compatibility across versions
11917 must use this, not the compile-time macro.
11918 (Exercise 0.9.4: Which is the minimum library version required by
11920 Note: All this applies only to multi-threaded programs, others don't
11924 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11925 through a BIO pair triggered the default case, i.e.
11926 SSLerr(...,SSL_R_UNKNOWN_STATE).
11929 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11930 can use the SSL library even if none of the specific BIOs is
11934 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11935 for the encoded length.
11936 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11938 *) Add initial documentation of the X509V3 functions.
11941 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11942 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11943 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11944 secure PKCS#8 private key format with a high iteration count.
11947 *) Fix determination of Perl interpreter: A perl or perl5
11948 _directory_ in $PATH was also accepted as the interpreter.
11949 [Ralf S. Engelschall]
11951 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11952 wrong with it but it was very old and did things like calling
11953 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11954 unusual formatting.
11957 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11958 to use the new extension code.
11961 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11962 with macros. This should make it easier to change their form, add extra
11963 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11967 *) Add to configuration table a new entry that can specify an alternative
11968 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11969 according to Mark Crispin <MRC@Panda.COM>.
11973 *) DES CBC did not update the IV. Weird.
11976 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11977 Changing the behaviour of the former might break existing programs --
11978 where IV updating is needed, des_ncbc_encrypt can be used.
11981 *) When bntest is run from "make test" it drives bc to check its
11982 calculations, as well as internally checking them. If an internal check
11983 fails, it needs to cause bc to give a non-zero result or make test carries
11984 on without noticing the failure. Fixed.
11987 *) DES library cleanups.
11990 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11991 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11992 ciphers. NOTE: although the key derivation function has been verified
11993 against some published test vectors it has not been extensively tested
11994 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11998 *) Instead of "mkdir -p", which is not fully portable, use new
11999 Perl script "util/mkdir-p.pl".
12002 *) Rewrite the way password based encryption (PBE) is handled. It used to
12003 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12004 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12005 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12006 the 'parameter' field of the AlgorithmIdentifier is passed to the
12007 underlying key generation function so it must do its own ASN1 parsing.
12008 This has also changed the EVP_PBE_CipherInit() function which now has a
12009 'parameter' argument instead of literal salt and iteration count values
12010 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12013 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12014 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12015 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12016 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12017 value was just used as a "magic string" and not used directly its
12018 value doesn't matter.
12021 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12025 *) "linux-sparc64" configuration (ultrapenguin).
12026 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12027 "linux-sparc" configuration.
12028 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12030 *) config now generates no-xxx options for missing ciphers.
12033 *) Support the EBCDIC character set (work in progress).
12034 File ebcdic.c not yet included because it has a different license.
12035 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12037 *) Support BS2000/OSD-POSIX.
12038 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12040 *) Make callbacks for key generation use void * instead of char *.
12043 *) Make S/MIME samples compile (not yet tested).
12046 *) Additional typesafe stacks.
12049 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12053 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12055 *) New configuration variant "sco5-gcc".
12057 *) Updated some demos.
12058 [Sean O Riordain, Wade Scholine]
12060 *) Add missing BIO_free at exit of pkcs12 application.
12063 *) Fix memory leak in conf.c.
12066 *) Updates for Win32 to assembler version of MD5.
12069 *) Set #! path to perl in apps/der_chop to where we found it
12070 instead of using a fixed path.
12073 *) SHA library changes for irix64-mips4-cc.
12076 *) Improvements for VMS support.
12080 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12082 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12083 This also avoids the problems with SC4.2 and unpatched SC5.
12084 [Andy Polyakov <appro@fy.chalmers.se>]
12086 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12087 These are required because of the typesafe stack would otherwise break
12088 existing code. If old code used a structure member which used to be STACK
12089 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12090 sk_num or sk_value it would produce an error because the num, data members
12091 are not present in STACK_OF. Now it just produces a warning. sk_set
12092 replaces the old method of assigning a value to sk_value
12093 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12094 that does this will no longer work (and should use sk_set instead) but
12095 this could be regarded as a "questionable" behaviour anyway.
12098 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12099 correctly handle encrypted S/MIME data.
12102 *) Change type of various DES function arguments from des_cblock
12103 (which means, in function argument declarations, pointer to char)
12104 to des_cblock * (meaning pointer to array with 8 char elements),
12105 which allows the compiler to do more typechecking; it was like
12106 that back in SSLeay, but with lots of ugly casts.
12108 Introduce new type const_des_cblock.
12111 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12112 problems: find RecipientInfo structure that matches recipient certificate
12113 and initialise the ASN1 structures properly based on passed cipher.
12116 *) Belatedly make the BN tests actually check the results.
12119 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12120 to and from BNs: it was completely broken. New compilation option
12121 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12122 key elements as negative integers.
12125 *) Reorganize and speed up MD5.
12126 [Andy Polyakov <appro@fy.chalmers.se>]
12129 [Richard Levitte <richard@levitte.org>]
12131 *) New option -out to asn1parse to allow the parsed structure to be
12132 output to a file. This is most useful when combined with the -strparse
12133 option to examine the output of things like OCTET STRINGS.
12136 *) Make SSL library a little more fool-proof by not requiring any longer
12137 that SSL_set_{accept,connect}_state be called before
12138 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12139 in many applications because usually everything *appeared* to work as
12140 intended anyway -- now it really works as intended).
12143 *) Move openssl.cnf out of lib/.
12146 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12147 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12148 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12149 [Ralf S. Engelschall]
12151 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12152 handle PKCS#7 enveloped data properly.
12153 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12155 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12156 copying pointers. The cert_st handling is changed by this in
12157 various ways (and thus what used to be known as ctx->default_cert
12158 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12159 any longer when s->cert does not give us what we need).
12160 ssl_cert_instantiate becomes obsolete by this change.
12161 As soon as we've got the new code right (possibly it already is?),
12162 we have solved a couple of bugs of the earlier code where s->cert
12163 was used as if it could not have been shared with other SSL structures.
12165 Note that using the SSL API in certain dirty ways now will result
12166 in different behaviour than observed with earlier library versions:
12167 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12168 does not influence s as it used to.
12170 In order to clean up things more thoroughly, inside SSL_SESSION
12171 we don't use CERT any longer, but a new structure SESS_CERT
12172 that holds per-session data (if available); currently, this is
12173 the peer's certificate chain and, for clients, the server's certificate
12174 and temporary key. CERT holds only those values that can have
12175 meaningful defaults in an SSL_CTX.
12178 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12179 from the internal representation. Various PKCS#7 fixes: remove some
12180 evil casts and set the enc_dig_alg field properly based on the signing
12184 *) Allow PKCS#12 password to be set from the command line or the
12185 environment. Let 'ca' get its config file name from the environment
12186 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12190 *) Allow certificate policies extension to use an IA5STRING for the
12191 organization field. This is contrary to the PKIX definition but
12192 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12196 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12197 without disallowing inline assembler and the like for non-pedantic builds.
12200 *) Support Borland C++ builder.
12201 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12203 *) Support Mingw32.
12206 *) SHA-1 cleanups and performance enhancements.
12207 [Andy Polyakov <appro@fy.chalmers.se>]
12209 *) Sparc v8plus assembler for the bignum library.
12210 [Andy Polyakov <appro@fy.chalmers.se>]
12212 *) Accept any -xxx and +xxx compiler options in Configure.
12215 *) Update HPUX configuration.
12218 *) Add missing sk_<type>_unshift() function to safestack.h
12219 [Ralf S. Engelschall]
12221 *) New function SSL_CTX_use_certificate_chain_file that sets the
12222 "extra_cert"s in addition to the certificate. (This makes sense
12223 only for "PEM" format files, as chains as a whole are not
12227 *) Support verify_depth from the SSL API.
12228 x509_vfy.c had what can be considered an off-by-one-error:
12229 Its depth (which was not part of the external interface)
12230 was actually counting the number of certificates in a chain;
12231 now it really counts the depth.
12234 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12235 instead of X509err, which often resulted in confusing error
12236 messages since the error codes are not globally unique
12237 (e.g. an alleged error in ssl3_accept when a certificate
12238 didn't match the private key).
12240 *) New function SSL_CTX_set_session_id_context that allows to set a default
12241 value (so that you don't need SSL_set_session_id_context for each
12242 connection using the SSL_CTX).
12245 *) OAEP decoding bug fix.
12248 *) Support INSTALL_PREFIX for package builders, as proposed by
12252 *) New Configure options "threads" and "no-threads". For systems
12253 where the proper compiler options are known (currently Solaris
12254 and Linux), "threads" is the default.
12257 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12260 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12261 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12262 such as /usr/local/bin.
12265 *) "make linux-shared" to build shared libraries.
12266 [Niels Poppe <niels@netbox.org>]
12268 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12271 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12272 extension adding in x509 utility.
12275 *) Remove NOPROTO sections and error code comments.
12278 *) Partial rewrite of the DEF file generator to now parse the ANSI
12282 *) New Configure options --prefix=DIR and --openssldir=DIR.
12285 *) Complete rewrite of the error code script(s). It is all now handled
12286 by one script at the top level which handles error code gathering,
12287 header rewriting and C source file generation. It should be much better
12288 than the old method: it now uses a modified version of Ulf's parser to
12289 read the ANSI prototypes in all header files (thus the old K&R definitions
12290 aren't needed for error creation any more) and do a better job of
12291 translating function codes into names. The old 'ASN1 error code imbedded
12292 in a comment' is no longer necessary and it doesn't use .err files which
12293 have now been deleted. Also the error code call doesn't have to appear all
12294 on one line (which resulted in some large lines...).
12297 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12300 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12301 0 (which usually indicates a closed connection), but continue reading.
12304 *) Fix some race conditions.
12307 *) Add support for CRL distribution points extension. Add Certificate
12308 Policies and CRL distribution points documentation.
12311 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12314 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12315 8 of keying material. Merlin has also confirmed interop with this fix
12316 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12317 [Merlin Hughes <merlin@baltimore.ie>]
12319 *) Fix lots of warnings.
12320 [Richard Levitte <levitte@stacken.kth.se>]
12322 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12323 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12324 [Richard Levitte <levitte@stacken.kth.se>]
12326 *) Fix problems with sizeof(long) == 8.
12327 [Andy Polyakov <appro@fy.chalmers.se>]
12329 *) Change functions to ANSI C.
12332 *) Fix typos in error codes.
12333 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12335 *) Remove defunct assembler files from Configure.
12338 *) SPARC v8 assembler BIGNUM implementation.
12339 [Andy Polyakov <appro@fy.chalmers.se>]
12341 *) Support for Certificate Policies extension: both print and set.
12342 Various additions to support the r2i method this uses.
12345 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12346 return a const string when you are expecting an allocated buffer.
12349 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12350 types DirectoryString and DisplayText.
12353 *) Add code to allow r2i extensions to access the configuration database,
12354 add an LHASH database driver and add several ctx helper functions.
12357 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12358 fail when they extended the size of a BIGNUM.
12361 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12362 support typesafe stack.
12365 *) Fix typo in SSL_[gs]et_options().
12366 [Nils Frostberg <nils@medcom.se>]
12368 *) Delete various functions and files that belonged to the (now obsolete)
12369 old X509V3 handling code.
12372 *) New Configure option "rsaref".
12375 *) Don't auto-generate pem.h.
12378 *) Introduce type-safe ASN.1 SETs.
12381 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12382 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12384 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12385 that links with OpenSSL (well at least cause lots of warnings), but fear
12386 not: the conversion is trivial, and it eliminates loads of evil casts. A
12387 few STACKed things have been converted already. Feel free to convert more.
12388 In the fullness of time, I'll do away with the STACK type altogether.
12391 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12392 specified in <certfile> by updating the entry in the index.txt file.
12393 This way one no longer has to edit the index.txt file manually for
12394 revoking a certificate. The -revoke option does the gory details now.
12395 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12397 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12398 `-text' option at all and this way the `-noout -text' combination was
12399 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12400 [Ralf S. Engelschall]
12402 *) Make sure a corresponding plain text error message exists for the
12403 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12404 verify callback function determined that a certificate was revoked.
12405 [Ralf S. Engelschall]
12407 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12408 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12409 all available ciphers including rc5, which was forgotten until now.
12410 In order to let the testing shell script know which algorithms
12411 are available, a new (up to now undocumented) command
12412 "openssl list-cipher-commands" is used.
12415 *) Bugfix: s_client occasionally would sleep in select() when
12416 it should have checked SSL_pending() first.
12419 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12420 the raw DSA values prior to ASN.1 encoding.
12423 *) Tweaks to Configure
12424 [Niels Poppe <niels@netbox.org>]
12426 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12430 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12433 *) New config option to avoid instructions that are illegal on the 80386.
12434 The default code is faster, but requires at least a 486.
12437 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12438 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12439 same as SSL2_VERSION anyway.
12442 *) New "-showcerts" option for s_client.
12445 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12446 application. Various cleanups and fixes.
12449 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12450 modify error routines to work internally. Add error codes and PBE init
12451 to library startup routines.
12454 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12455 packing functions to asn1 and evp. Changed function names and error
12456 codes along the way.
12459 *) PKCS12 integration: and so it begins... First of several patches to
12460 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12461 objects to objects.h
12464 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12465 and display support for Thawte strong extranet extension.
12468 *) Add LinuxPPC support.
12469 [Jeff Dubrule <igor@pobox.org>]
12471 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12472 bn_div_words in alpha.s.
12473 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12475 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12476 OAEP isn't supported when OpenSSL is built with RSAref.
12477 [Ulf Moeller <ulf@fitug.de>]
12479 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12480 so they no longer are missing under -DNOPROTO.
12481 [Soren S. Jorvang <soren@t.dk>]
12484 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12486 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12487 doesn't work when the session is reused. Coming soon!
12490 *) Fix a security hole, that allows sessions to be reused in the wrong
12491 context thus bypassing client cert protection! All software that uses
12492 client certs and session caches in multiple contexts NEEDS PATCHING to
12493 allow session reuse! A fuller solution is in the works.
12494 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12496 *) Some more source tree cleanups (removed obsolete files
12497 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12498 permission on "config" script to be executable) and a fix for the INSTALL
12500 [Ulf Moeller <ulf@fitug.de>]
12502 *) Remove some legacy and erroneous uses of malloc, free instead of
12504 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12506 *) Make rsa_oaep_test return non-zero on error.
12507 [Ulf Moeller <ulf@fitug.de>]
12509 *) Add support for native Solaris shared libraries. Configure
12510 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12511 if someone would make that last step automatic.
12512 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12514 *) ctx_size was not built with the right compiler during "make links". Fixed.
12517 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12518 except NULL ciphers". This means the default cipher list will no longer
12519 enable NULL ciphers. They need to be specifically enabled e.g. with
12520 the string "DEFAULT:eNULL".
12523 *) Fix to RSA private encryption routines: if p < q then it would
12524 occasionally produce an invalid result. This will only happen with
12525 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12528 *) Be less restrictive and allow also `perl util/perlpath.pl
12529 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12530 because this way one can also use an interpreter named `perl5' (which is
12531 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12532 installed as `perl').
12533 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12535 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12536 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12538 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12539 advapi32.lib to Win32 build and change the pem test comparison
12540 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12541 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12542 and crypto/des/ede_cbcm_enc.c.
12545 *) DES quad checksum was broken on big-endian architectures. Fixed.
12548 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12549 Win32 test batch file so it (might) work again. The Win32 test batch file
12550 is horrible: I feel ill....
12553 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12554 in e_os.h. Audit of header files to check ANSI and non ANSI
12555 sections: 10 functions were absent from non ANSI section and not exported
12556 from Windows DLLs. Fixed up libeay.num for new functions.
12559 *) Make `openssl version' output lines consistent.
12560 [Ralf S. Engelschall]
12562 *) Fix Win32 symbol export lists for BIO functions: Added
12563 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12564 to ms/libeay{16,32}.def.
12565 [Ralf S. Engelschall]
12567 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12568 fine under Unix and passes some trivial tests I've now added. But the
12569 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12570 added to make sure no one expects that this stuff really works in the
12571 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12572 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12574 [Ralf S. Engelschall]
12576 *) Fix the generation of two part addresses in perl.
12577 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12579 *) Add config entry for Linux on MIPS.
12580 [John Tobey <jtobey@channel1.com>]
12582 *) Make links whenever Configure is run, unless we are on Windoze.
12585 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12586 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12590 *) Add a useful kludge to allow package maintainers to specify compiler and
12591 other platforms details on the command line without having to patch the
12592 Configure script everytime: One now can use ``perl Configure
12593 <id>:<details>'', i.e. platform ids are allowed to have details appended
12594 to them (separated by colons). This is treated as there would be a static
12595 pre-configured entry in Configure's %table under key <id> with value
12596 <details> and ``perl Configure <id>'' is called. So, when you want to
12597 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12598 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12599 now, which overrides the FreeBSD-elf entry on-the-fly.
12600 [Ralf S. Engelschall]
12602 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12605 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12606 on the `perl Configure ...' command line. This way one can compile
12607 OpenSSL libraries with Position Independent Code (PIC) which is needed
12608 for linking it into DSOs.
12609 [Ralf S. Engelschall]
12611 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12615 *) Cleaned up the LICENSE document: The official contact for any license
12616 questions now is the OpenSSL core team under openssl-core@openssl.org.
12617 And add a paragraph about the dual-license situation to make sure people
12618 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12619 to the OpenSSL toolkit.
12620 [Ralf S. Engelschall]
12622 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12623 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12624 Additionally cleaned up the `make links' target: Remove unnecessary
12625 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12626 to speed processing and no longer clutter the display with confusing
12627 stuff. Instead only the actually done links are displayed.
12628 [Ralf S. Engelschall]
12630 *) Permit null encryption ciphersuites, used for authentication only. It used
12631 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12632 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12636 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12637 signed attributes when verifying signatures (this would break them),
12638 the detached data encoding was wrong and public keys obtained using
12639 X509_get_pubkey() weren't freed.
12642 *) Add text documentation for the BUFFER functions. Also added a work around
12643 to a Win95 console bug. This was triggered by the password read stuff: the
12644 last character typed gets carried over to the next fread(). If you were
12645 generating a new cert request using 'req' for example then the last
12646 character of the passphrase would be CR which would then enter the first
12650 *) Added the new `Includes OpenSSL Cryptography Software' button as
12651 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12652 button and can be used by applications based on OpenSSL to show the
12653 relationship to the OpenSSL project.
12654 [Ralf S. Engelschall]
12656 *) Remove confusing variables in function signatures in files
12657 ssl/ssl_lib.c and ssl/ssl.h.
12658 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12660 *) Don't install bss_file.c under PREFIX/include/
12661 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12663 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12664 functions that return function pointers and has support for NT specific
12665 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12666 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12667 unsigned to signed types: this was killing the Win32 compile.
12670 *) Add new certificate file to stack functions,
12671 SSL_add_dir_cert_subjects_to_stack() and
12672 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12673 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12674 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12675 This means that Apache-SSL and similar packages don't have to mess around
12676 to add as many CAs as they want to the preferred list.
12679 *) Experiment with doxygen documentation. Currently only partially applied to
12681 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12682 openssl.doxy as the configuration file.
12685 *) Get rid of remaining C++-style comments which strict C compilers hate.
12686 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12688 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12689 compiled in by default: it has problems with large keys.
12692 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12693 DH private keys and/or callback functions which directly correspond to
12694 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12695 is needed for applications which have to configure certificates on a
12696 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12698 For the RSA certificate situation is makes no difference, but
12699 for the DSA certificate situation this fixes the "no shared cipher"
12700 problem where the OpenSSL cipher selection procedure failed because the
12701 temporary keys were not overtaken from the context and the API provided
12702 no way to reconfigure them.
12703 The new functions now let applications reconfigure the stuff and they
12704 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12705 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12706 non-public-API function ssl_cert_instantiate() is used as a helper
12707 function and also to reduce code redundancy inside ssl_rsa.c.
12708 [Ralf S. Engelschall]
12710 *) Move s_server -dcert and -dkey options out of the undocumented feature
12711 area because they are useful for the DSA situation and should be
12712 recognized by the users.
12713 [Ralf S. Engelschall]
12715 *) Fix the cipher decision scheme for export ciphers: the export bits are
12716 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12717 SSL_EXP_MASK. So, the original variable has to be used instead of the
12718 already masked variable.
12719 [Richard Levitte <levitte@stacken.kth.se>]
12721 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12722 [Richard Levitte <levitte@stacken.kth.se>]
12724 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12725 from `int' to `unsigned int' because it's a length and initialized by
12726 EVP_DigestFinal() which expects an `unsigned int *'.
12727 [Richard Levitte <levitte@stacken.kth.se>]
12729 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12730 script. Instead use the usual Shell->Perl transition trick.
12731 [Ralf S. Engelschall]
12733 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12734 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12735 -noout -modulus' as it's already the case for `openssl rsa -noout
12736 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12737 currently the public key is printed (a decision which was already done by
12738 `openssl dsa -modulus' in the past) which serves a similar purpose.
12739 Additionally the NO_RSA no longer completely removes the whole -modulus
12740 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12742 [Ralf S. Engelschall]
12744 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12745 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12746 [Arne Ansper <arne@ats.cyber.ee>]
12748 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12749 to be added. Now both 'req' and 'ca' can use new objects defined in the
12753 *) Add cool BIO that does syslog (or event log on NT).
12754 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12756 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12757 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12758 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12759 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12762 *) Add preliminary config info for new extension code.
12765 *) Make RSA_NO_PADDING really use no padding.
12766 [Ulf Moeller <ulf@fitug.de>]
12768 *) Generate errors when private/public key check is done.
12771 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12772 for some CRL extensions and new objects added.
12775 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12776 key usage extension and fuller support for authority key id.
12779 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12780 padding method for RSA, which is recommended for new applications in PKCS
12781 #1 v2.0 (RFC 2437, October 1998).
12782 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12783 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12784 against Bleichbacher's attack on RSA.
12785 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12788 *) Updates to the new SSL compression code
12789 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12791 *) Fix so that the version number in the master secret, when passed
12792 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12793 (because the server will not accept higher), that the version number
12794 is 0x03,0x01, not 0x03,0x00
12795 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12797 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12798 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12799 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12802 *) Support for RAW extensions where an arbitrary extension can be
12803 created by including its DER encoding. See apps/openssl.cnf for
12807 *) Make sure latest Perl versions don't interpret some generated C array
12808 code as Perl array code in the crypto/err/err_genc.pl script.
12809 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12811 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12812 not many people have the assembler. Various Win32 compilation fixes and
12813 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12814 build instructions.
12817 *) Modify configure script 'Configure' to automatically create crypto/date.h
12818 file under Win32 and also build pem.h from pem.org. New script
12819 util/mkfiles.pl to create the MINFO file on environments that can't do a
12820 'make files': perl util/mkfiles.pl >MINFO should work.
12823 *) Major rework of DES function declarations, in the pursuit of correctness
12824 and purity. As a result, many evil casts evaporated, and some weirdness,
12825 too. You may find this causes warnings in your code. Zapping your evil
12826 casts will probably fix them. Mostly.
12829 *) Fix for a typo in asn1.h. Bug fix to object creation script
12830 obj_dat.pl. It considered a zero in an object definition to mean
12831 "end of object": none of the objects in objects.h have any zeros
12832 so it wasn't spotted.
12833 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12835 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12836 Masking (CBCM). In the absence of test vectors, the best I have been able
12837 to do is check that the decrypt undoes the encrypt, so far. Send me test
12838 vectors if you have them.
12841 *) Correct calculation of key length for export ciphers (too much space was
12842 allocated for null ciphers). This has not been tested!
12845 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12846 message is now correct (it understands "crypto" and "ssl" on its
12847 command line). There is also now an "update" option. This will update
12848 the util/ssleay.num and util/libeay.num files with any new functions.
12850 perl util/mkdef.pl crypto ssl update
12851 it will update them.
12854 *) Overhauled the Perl interface (perl/*):
12855 - ported BN stuff to OpenSSL's different BN library
12856 - made the perl/ source tree CVS-aware
12857 - renamed the package from SSLeay to OpenSSL (the files still contain
12858 their history because I've copied them in the repository)
12859 - removed obsolete files (the test scripts will be replaced
12860 by better Test::Harness variants in the future)
12861 [Ralf S. Engelschall]
12863 *) First cut for a very conservative source tree cleanup:
12864 1. merge various obsolete readme texts into doc/ssleay.txt
12865 where we collect the old documents and readme texts.
12866 2. remove the first part of files where I'm already sure that we no
12867 longer need them because of three reasons: either they are just temporary
12868 files which were left by Eric or they are preserved original files where
12869 I've verified that the diff is also available in the CVS via "cvs diff
12870 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12871 the crypto/md/ stuff).
12872 [Ralf S. Engelschall]
12874 *) More extension code. Incomplete support for subject and issuer alt
12875 name, issuer and authority key id. Change the i2v function parameters
12876 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12877 what that's for :-) Fix to ASN1 macro which messed up
12878 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12881 *) Preliminary support for ENUMERATED type. This is largely copied from the
12885 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12886 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12888 *) Make sure `make rehash' target really finds the `openssl' program.
12889 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12891 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12892 like to hear about it if this slows down other processors.
12895 *) Add CygWin32 platform information to Configure script.
12896 [Alan Batie <batie@aahz.jf.intel.com>]
12898 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12899 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12901 *) New program nseq to manipulate netscape certificate sequences
12904 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12908 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12909 but the BN code had some problems that would cause failures when
12910 doing certificate verification and some other functions.
12911 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12913 *) Add ASN1 and PEM code to support netscape certificate sequences.
12916 *) Add ASN1 and PEM code to support netscape certificate sequences.
12919 *) Add several PKIX and private extended key usage OIDs.
12922 *) Modify the 'ca' program to handle the new extension code. Modify
12923 openssl.cnf for new extension format, add comments.
12926 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12927 and add a sample to openssl.cnf so req -x509 now adds appropriate
12931 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12932 error code, add initial support to X509_print() and x509 application.
12935 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12936 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12937 stuff is currently isolated and isn't even compiled yet.
12940 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12941 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12942 Removed the versions check from X509 routines when loading extensions:
12943 this allows certain broken certificates that don't set the version
12944 properly to be processed.
12947 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12948 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12949 can still be regenerated with "make depend".
12952 *) Spelling mistake in C version of CAST-128.
12953 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12955 *) Changes to the error generation code. The perl script err-code.pl
12956 now reads in the old error codes and retains the old numbers, only
12957 adding new ones if necessary. It also only changes the .err files if new
12958 codes are added. The makefiles have been modified to only insert errors
12959 when needed (to avoid needlessly modifying header files). This is done
12960 by only inserting errors if the .err file is newer than the auto generated
12961 C file. To rebuild all the error codes from scratch (the old behaviour)
12962 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12963 or delete all the .err files.
12966 *) CAST-128 was incorrectly implemented for short keys. The C version has
12967 been fixed, but is untested. The assembler versions are also fixed, but
12968 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12969 to regenerate it if needed.
12970 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12971 Hagino <itojun@kame.net>]
12973 *) File was opened incorrectly in randfile.c.
12974 [Ulf Möller <ulf@fitug.de>]
12976 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12977 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12978 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12979 al: it's just almost always a UTCTime. Note this patch adds new error
12980 codes so do a "make errors" if there are problems.
12983 *) Correct Linux 1 recognition in config.
12984 [Ulf Möller <ulf@fitug.de>]
12986 *) Remove pointless MD5 hash when using DSA keys in ca.
12987 [Anonymous <nobody@replay.com>]
12989 *) Generate an error if given an empty string as a cert directory. Also
12990 generate an error if handed NULL (previously returned 0 to indicate an
12991 error, but didn't set one).
12992 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12994 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12997 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12998 parameters. This was causing a warning which killed off the Win32 compile.
13001 *) Remove C++ style comments from crypto/bn/bn_local.h.
13002 [Neil Costigan <neil.costigan@celocom.com>]
13004 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13005 based on a text string, looking up short and long names and finally
13006 "dot" format. The "dot" format stuff didn't work. Added new function
13007 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13008 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13009 OID is not part of the table.
13012 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13013 X509_LOOKUP_by_alias().
13016 *) Sort openssl functions by name.
13019 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13020 encryption from sample DSA keys (in case anyone is interested the password
13024 *) Make _all_ *_free functions accept a NULL pointer.
13025 [Frans Heymans <fheymans@isaserver.be>]
13027 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13029 [Anonymous <nobody@replay.com>]
13031 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13032 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13034 *) Don't blow it for numeric -newkey arguments to apps/req.
13035 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13037 *) Temp key "for export" tests were wrong in s3_srvr.c.
13038 [Anonymous <nobody@replay.com>]
13040 *) Add prototype for temp key callback functions
13041 SSL_CTX_set_tmp_{rsa,dh}_callback().
13044 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13045 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13048 *) X509_name_add_entry() freed the wrong thing after an error.
13049 [Arne Ansper <arne@ats.cyber.ee>]
13051 *) rsa_eay.c would attempt to free a NULL context.
13052 [Arne Ansper <arne@ats.cyber.ee>]
13054 *) BIO_s_socket() had a broken should_retry() on Windoze.
13055 [Arne Ansper <arne@ats.cyber.ee>]
13057 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13058 [Arne Ansper <arne@ats.cyber.ee>]
13060 *) Make sure the already existing X509_STORE->depth variable is initialized
13061 in X509_STORE_new(), but document the fact that this variable is still
13062 unused in the certificate verification process.
13063 [Ralf S. Engelschall]
13065 *) Fix the various library and apps files to free up pkeys obtained from
13066 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13069 *) Fix reference counting in X509_PUBKEY_get(). This makes
13070 demos/maurice/example2.c work, amongst others, probably.
13071 [Steve Henson and Ben Laurie]
13073 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13074 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13075 are no longer created. This way we have a single and consistent command
13076 line interface `openssl <command>', similar to `cvs <command>'.
13077 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13079 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13080 BIT STRING wrapper always have zero unused bits.
13083 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13086 *) Make the top-level INSTALL documentation easier to understand.
13089 *) Makefiles updated to exit if an error occurs in a sub-directory
13090 make (including if user presses ^C) [Paul Sutton]
13092 *) Make Montgomery context stuff explicit in RSA data structure.
13095 *) Fix build order of pem and err to allow for generated pem.h.
13098 *) Fix renumbering bug in X509_NAME_delete_entry().
13101 *) Enhanced the err-ins.pl script so it makes the error library number
13102 global and can add a library name. This is needed for external ASN1 and
13103 other error libraries.
13106 *) Fixed sk_insert which never worked properly.
13109 *) Fix ASN1 macros so they can handle indefinite length constructed
13110 EXPLICIT tags. Some non standard certificates use these: they can now
13114 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13115 into a single doc/ssleay.txt bundle. This way the information is still
13116 preserved but no longer messes up this directory. Now it's new room for
13117 the new set of documentation files.
13118 [Ralf S. Engelschall]
13120 *) SETs were incorrectly DER encoded. This was a major pain, because they
13121 shared code with SEQUENCEs, which aren't coded the same. This means that
13122 almost everything to do with SETs or SEQUENCEs has either changed name or
13123 number of arguments.
13124 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13126 *) Fix test data to work with the above.
13129 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13130 was already fixed by Eric for 0.9.1 it seems.
13131 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13133 *) Autodetect FreeBSD3.
13136 *) Fix various bugs in Configure. This affects the following platforms:
13140 unixware-2.0-pentium
13144 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13145 before they are needed.
13148 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13152 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13154 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13155 changed SSLeay to OpenSSL in version strings.
13156 [Ralf S. Engelschall]
13158 *) Some fixups to the top-level documents.
13161 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13162 because the symlink to include/ was missing.
13163 [Ralf S. Engelschall]
13165 *) Incorporated the popular no-RSA/DSA-only patches
13166 which allow to compile a RSA-free SSLeay.
13167 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13169 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13170 when "ssleay" is still not found.
13171 [Ralf S. Engelschall]
13173 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13174 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13176 *) Updated the README file.
13177 [Ralf S. Engelschall]
13179 *) Added various .cvsignore files in the CVS repository subdirs
13180 to make a "cvs update" really silent.
13181 [Ralf S. Engelschall]
13183 *) Recompiled the error-definition header files and added
13184 missing symbols to the Win32 linker tables.
13185 [Ralf S. Engelschall]
13187 *) Cleaned up the top-level documents;
13188 o new files: CHANGES and LICENSE
13189 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13190 o merged COPYRIGHT into LICENSE
13191 o removed obsolete TODO file
13192 o renamed MICROSOFT to INSTALL.W32
13193 [Ralf S. Engelschall]
13195 *) Removed dummy files from the 0.9.1b source tree:
13196 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13197 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13198 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13199 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13200 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13201 [Ralf S. Engelschall]
13203 *) Added various platform portability fixes.
13206 *) The Genesis of the OpenSSL rpject:
13207 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13208 Young and Tim J. Hudson created while they were working for C2Net until
13210 [The OpenSSL Project]
13213 Changes between 0.9.0b and 0.9.1b [not released]
13215 *) Updated a few CA certificates under certs/
13218 *) Changed some BIGNUM api stuff.
13221 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13222 DGUX x86, Linux Alpha, etc.
13225 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13226 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13230 *) Add -strparse option to asn1pars program which parses nested
13232 [Dr Stephen Henson <shenson@bigfoot.com>]
13234 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13237 *) DSA fix for "ca" program.
13240 *) Added "-genkey" option to "dsaparam" program.
13243 *) Added RIPE MD160 (rmd160) message digest.
13246 *) Added -a (all) option to "ssleay version" command.
13249 *) Added PLATFORM define which is the id given to Configure.
13252 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13255 *) Extended the ASN.1 parser routines.
13258 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13261 *) Added a BN_CTX to the BN library.
13264 *) Fixed the weak key values in DES library
13267 *) Changed API in EVP library for cipher aliases.
13270 *) Added support for RC2/64bit cipher.
13273 *) Converted the lhash library to the crypto/mem.c functions.
13276 *) Added more recognized ASN.1 object ids.
13279 *) Added more RSA padding checks for SSL/TLS.
13282 *) Added BIO proxy/filter functionality.
13285 *) Added extra_certs to SSL_CTX which can be used
13286 send extra CA certificates to the client in the CA cert chain sending
13287 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13290 *) Now Fortezza is denied in the authentication phase because
13291 this is key exchange mechanism is not supported by SSLeay at all.
13294 *) Additional PKCS1 checks.
13297 *) Support the string "TLSv1" for all TLS v1 ciphers.
13300 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13301 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13304 *) Fixed a few memory leaks.
13307 *) Fixed various code and comment typos.
13310 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13311 bytes sent in the client random.
13312 [Edward Bishop <ebishop@spyglass.com>]