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 1.1.1a [xx XXX xxxx]
14 Changes between 1.1.1 and 1.1.1a [xx XXX xxxx]
16 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
17 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
18 of two gigabytes and the error handling improved.
20 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
21 categorized as a normal bug, not a security issue, because the DRBG reseeds
22 automatically and is fully functional even without additional randomness
23 provided by the application.
25 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
27 *) Add a new ClientHello callback. Provides a callback interface that gives
28 the application the ability to adjust the nascent SSL object at the
29 earliest stage of ClientHello processing, immediately after extensions have
30 been collected but before they have been processed. In particular, this
31 callback can adjust the supported TLS versions in response to the contents
35 *) Add SM2 base algorithm support.
38 *) s390x assembly pack: add (improved) hardware-support for the following
39 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
40 aes-cfb/cfb8, aes-ecb.
43 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
44 parameter is no longer accepted, as it leads to a corrupt table. NULL
45 pem_str is reserved for alias entries only.
48 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
49 step for prime curves. The new implementation is based on formulae from
50 differential addition-and-doubling in homogeneous projective coordinates
51 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
52 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
53 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
54 to work in projective coordinates.
55 [Billy Bob Brumley, Nicola Tuveri]
57 *) Change generating and checking of primes so that the error rate of not
58 being prime depends on the intended use based on the size of the input.
59 For larger primes this will result in more rounds of Miller-Rabin.
60 The maximal error rate for primes with more than 1080 bits is lowered
62 [Kurt Roeckx, Annie Yousar]
64 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
67 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
68 moving between systems, and to avoid confusion when a Windows build is
69 done with mingw vs with MSVC. For POSIX installs, there's still a
70 symlink or copy named 'tsget' to avoid that confusion as well.
73 *) Revert blinding in ECDSA sign and instead make problematic addition
74 length-invariant. Switch even to fixed-length Montgomery multiplication.
77 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
78 step for binary curves. The new implementation is based on formulae from
79 differential addition-and-doubling in mixed Lopez-Dahab projective
80 coordinates, modified to independently blind the operands.
81 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
83 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
84 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
85 EC_METHODs to implement their own specialized "ladder step", to take
86 advantage of more favorable coordinate systems or more efficient
87 differential addition-and-doubling algorithms.
88 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
90 *) Modified the random device based seed sources to keep the relevant
91 file descriptors open rather than reopening them on each access.
92 This allows such sources to operate in a chroot() jail without
93 the associated device nodes being available. This behaviour can be
94 controlled using RAND_keep_random_devices_open().
97 *) Numerous side-channel attack mitigations have been applied. This may have
98 performance impacts for some algorithms for the benefit of improved
99 security. Specific changes are noted in this change log by their respective
103 *) AIX shared library support overhaul. Switch to AIX "natural" way of
104 handling shared libraries, which means collecting shared objects of
105 different versions and bitnesses in one common archive. This allows to
106 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
107 doesn't affect the way 3rd party applications are linked, only how
108 multi-version installation is managed.
111 *) Make ec_group_do_inverse_ord() more robust and available to other
112 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
113 mitigations are applied to the fallback BN_mod_inverse().
114 When using this function rather than BN_mod_inverse() directly, new
115 EC cryptosystem implementations are then safer-by-default.
118 *) Add coordinate blinding for EC_POINT and implement projective
119 coordinate blinding for generic prime curves as a countermeasure to
120 chosen point SCA attacks.
121 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
123 *) Add blinding to ECDSA and DSA signatures to protect against side channel
124 attacks discovered by Keegan Ryan (NCC Group).
127 *) Enforce checking in the pkeyutl command line app to ensure that the input
128 length does not exceed the maximum supported digest length when performing
129 a sign, verify or verifyrecover operation.
132 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
133 I/O in combination with something like select() or poll() will hang. This
134 can be turned off again using SSL_CTX_clear_mode().
135 Many applications do not properly handle non-application data records, and
136 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
137 around the problems in those applications, but can also break some.
138 It's recommended to read the manpages about SSL_read(), SSL_write(),
139 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
140 SSL_CTX_set_read_ahead() again.
143 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
144 now allow empty (zero character) pass phrases.
147 *) Apply blinding to binary field modular inversion and remove patent
148 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
151 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
152 binary and prime elliptic curves.
155 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
156 constant time fixed point multiplication.
159 *) Revise elliptic curve scalar multiplication with timing attack
160 defenses: ec_wNAF_mul redirects to a constant time implementation
161 when computing fixed point and variable point multiplication (which
162 in OpenSSL are mostly used with secret scalars in keygen, sign,
163 ECDH derive operations).
164 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
167 *) Updated CONTRIBUTING
170 *) Updated DRBG / RAND to request nonce and additional low entropy
171 randomness from the system.
172 [Matthias St. Pierre]
174 *) Updated 'openssl rehash' to use OpenSSL consistent default.
177 *) Moved the load of the ssl_conf module to libcrypto, which helps
178 loading engines that libssl uses before libssl is initialised.
181 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
184 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
185 [Ingo Schwarze, Rich Salz]
187 *) Added output of accepting IP address and port for 'openssl s_server'
190 *) Added a new API for TLSv1.3 ciphersuites:
191 SSL_CTX_set_ciphersuites()
192 SSL_set_ciphersuites()
195 *) Memory allocation failures consistenly add an error to the error
199 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
200 in libcrypto when run as setuid/setgid.
203 *) Load any config file by default when libssl is used.
206 *) Added new public header file <openssl/rand_drbg.h> and documentation
207 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
208 [Matthias St. Pierre]
210 *) QNX support removed (cannot find contributors to get their approval
211 for the license change).
214 *) TLSv1.3 replay protection for early data has been implemented. See the
215 SSL_read_early_data() man page for further details.
218 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
219 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
220 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
221 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
222 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
223 configuration has been separated out. See the ciphers man page or the
224 SSL_CTX_set_ciphersuites() man page for more information.
227 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
228 in responder mode now supports the new "-multi" option, which
229 spawns the specified number of child processes to handle OCSP
230 requests. The "-timeout" option now also limits the OCSP
231 responder's patience to wait to receive the full client request
232 on a newly accepted connection. Child processes are respawned
233 as needed, and the CA index file is automatically reloaded
234 when changed. This makes it possible to run the "ocsp" responder
235 as a long-running service, making the OpenSSL CA somewhat more
236 feature-complete. In this mode, most diagnostic messages logged
237 after entering the event loop are logged via syslog(3) rather than
241 *) Added support for X448 and Ed448. Heavily based on original work by
245 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
246 objects loaded. This adds the functions OSSL_STORE_expect() and
247 OSSL_STORE_find() as well as needed tools to construct searches and
248 get the search data out of them.
251 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
252 version of OpenSSL should review their configuration settings to ensure
253 that they are still appropriate for TLSv1.3. For further information see:
254 https://wiki.openssl.org/index.php/TLS1.3
257 *) Grand redesign of the OpenSSL random generator
259 The default RAND method now utilizes an AES-CTR DRBG according to
260 NIST standard SP 800-90Ar1. The new random generator is essentially
261 a port of the default random generator from the OpenSSL FIPS 2.0
262 object module. It is a hybrid deterministic random bit generator
263 using an AES-CTR bit stream and which seeds and reseeds itself
264 automatically using trusted system entropy sources.
266 Some of its new features are:
267 o Support for multiple DRBG instances with seed chaining.
268 o The default RAND method makes use of a DRBG.
269 o There is a public and private DRBG instance.
270 o The DRBG instances are fork-safe.
271 o Keep all global DRBG instances on the secure heap if it is enabled.
272 o The public and private DRBG instance are per thread for lock free
274 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
276 *) Changed Configure so it only says what it does and doesn't dump
277 so much data. Instead, ./configdata.pm should be used as a script
278 to display all sorts of configuration data.
281 *) Added processing of "make variables" to Configure.
284 *) Added SHA512/224 and SHA512/256 algorithm support.
287 *) The last traces of Netware support, first removed in 1.1.0, have
291 *) Get rid of Makefile.shared, and in the process, make the processing
292 of certain files (rc.obj, or the .def/.map/.opt files produced from
293 the ordinal files) more visible and hopefully easier to trace and
294 debug (or make silent).
297 *) Make it possible to have environment variable assignments as
298 arguments to config / Configure.
301 *) Add multi-prime RSA (RFC 8017) support.
304 *) Add SM3 implemented according to GB/T 32905-2016
305 [ Jack Lloyd <jack.lloyd@ribose.com>,
306 Ronald Tse <ronald.tse@ribose.com>,
307 Erick Borsboom <erick.borsboom@ribose.com> ]
309 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
310 as documented in RFC6066.
311 Based on a patch from Tomasz Moń
312 [Filipe Raimundo da Silva]
314 *) Add SM4 implemented according to GB/T 32907-2016.
315 [ Jack Lloyd <jack.lloyd@ribose.com>,
316 Ronald Tse <ronald.tse@ribose.com>,
317 Erick Borsboom <erick.borsboom@ribose.com> ]
319 *) Reimplement -newreq-nodes and ERR_error_string_n; the
320 original author does not agree with the license change.
323 *) Add ARIA AEAD TLS support.
326 *) Some macro definitions to support VS6 have been removed. Visual
327 Studio 6 has not worked since 1.1.0
330 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
331 without clearing the errors.
334 *) Add "atfork" functions. If building on a system that without
335 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
336 requirements. The RAND facility now uses/requires this.
342 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
343 not possible to disable entirely. However, it's still possible to
344 disable the console reading UI method, UI_OpenSSL() (use UI_null()
347 To disable, configure with 'no-ui-console'. 'no-ui' is still
348 possible to use as an alias. Check at compile time with the
349 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
350 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
353 *) Add a STORE module, which implements a uniform and URI based reader of
354 stores that can contain keys, certificates, CRLs and numerous other
355 objects. The main API is loosely based on a few stdio functions,
356 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
357 OSSL_STORE_error and OSSL_STORE_close.
358 The implementation uses backends called "loaders" to implement arbitrary
359 URI schemes. There is one built in "loader" for the 'file' scheme.
362 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
363 then adjusted to work on FreeBSD 8.4 as well.
364 Enable by configuring with 'enable-devcryptoeng'. This is done by default
365 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
368 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
369 util/mkerr.pl, which is adapted to allow those prefixes, leading to
370 error code calls like this:
372 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
374 With this change, we claim the namespaces OSSL and OPENSSL in a manner
375 that can be encoded in C. For the foreseeable future, this will only
377 [Richard Levitte and Tim Hudson]
379 *) Removed BSD cryptodev engine.
382 *) Add a build target 'build_all_generated', to build all generated files
383 and only that. This can be used to prepare everything that requires
384 things like perl for a system that lacks perl and then move everything
385 to that system and do the rest of the build there.
388 *) In the UI interface, make it possible to duplicate the user data. This
389 can be used by engines that need to retain the data for a longer time
390 than just the call where this user data is passed.
393 *) Ignore the '-named_curve auto' value for compatibility of applications
395 [Tomas Mraz <tmraz@fedoraproject.org>]
397 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
398 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
399 alerts across multiple records (some of which could be empty). In practice
400 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
401 prohibts this altogether and other libraries (BoringSSL, NSS) do not
402 support this at all. Supporting it adds significant complexity to the
403 record layer, and its removal is unlikely to cause inter-operability
407 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
408 with Z. These are meant to replace LONG and ZLONG and to be size safe.
409 The use of LONG and ZLONG is discouraged and scheduled for deprecation
413 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
414 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
415 [Richard Levitte, Andy Polyakov]
417 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
421 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
422 platform rather than 'mingw'.
425 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
426 success if they are asked to add an object which already exists
427 in the store. This change cascades to other functions which load
428 certificates and CRLs.
431 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
432 facilitate stack unwinding even from assembly subroutines.
435 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
436 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
439 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
440 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
441 which is the minimum version we support.
444 *) Certificate time validation (X509_cmp_time) enforces stricter
445 compliance with RFC 5280. Fractional seconds and timezone offsets
446 are no longer allowed.
449 *) Add support for ARIA
452 *) s_client will now send the Server Name Indication (SNI) extension by
453 default unless the new "-noservername" option is used. The server name is
454 based on the host provided to the "-connect" option unless overridden by
458 *) Add support for SipHash
461 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
462 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
463 prevent issues where no progress is being made and the peer continually
464 sends unrecognised record types, using up resources processing them.
467 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
468 using the algorithm defined in
469 https://www.akkadia.org/drepper/SHA-crypt.txt
472 *) Heartbeat support has been removed; the ABI is changed for now.
473 [Richard Levitte, Rich Salz]
475 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
478 *) The RSA "null" method, which was partially supported to avoid patent
479 issues, has been replaced to always returns NULL.
483 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
485 *) Client DoS due to large DH parameter
487 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
488 malicious server can send a very large prime value to the client. This will
489 cause the client to spend an unreasonably long period of time generating a
490 key for this prime resulting in a hang until the client has finished. This
491 could be exploited in a Denial Of Service attack.
493 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
497 *) Cache timing vulnerability in RSA Key Generation
499 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
500 a cache timing side channel attack. An attacker with sufficient access to
501 mount cache timing attacks during the RSA key generation process could
502 recover the private key.
504 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
505 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
509 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
510 parameter is no longer accepted, as it leads to a corrupt table. NULL
511 pem_str is reserved for alias entries only.
514 *) Revert blinding in ECDSA sign and instead make problematic addition
515 length-invariant. Switch even to fixed-length Montgomery multiplication.
518 *) Change generating and checking of primes so that the error rate of not
519 being prime depends on the intended use based on the size of the input.
520 For larger primes this will result in more rounds of Miller-Rabin.
521 The maximal error rate for primes with more than 1080 bits is lowered
523 [Kurt Roeckx, Annie Yousar]
525 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
528 *) Add blinding to ECDSA and DSA signatures to protect against side channel
529 attacks discovered by Keegan Ryan (NCC Group).
532 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
533 now allow empty (zero character) pass phrases.
536 *) Certificate time validation (X509_cmp_time) enforces stricter
537 compliance with RFC 5280. Fractional seconds and timezone offsets
538 are no longer allowed.
541 *) Fixed a text canonicalisation bug in CMS
543 Where a CMS detached signature is used with text content the text goes
544 through a canonicalisation process first prior to signing or verifying a
545 signature. This process strips trailing space at the end of lines, converts
546 line terminators to CRLF and removes additional trailing line terminators
547 at the end of a file. A bug in the canonicalisation process meant that
548 some characters, such as form-feed, were incorrectly treated as whitespace
549 and removed. This is contrary to the specification (RFC5485). This fix
550 could mean that detached text data signed with an earlier version of
551 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
552 signed with a fixed OpenSSL may fail to verify with an earlier version of
553 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
554 and use the "-binary" flag (for the "cms" command line application) or set
555 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
558 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
560 *) Constructed ASN.1 types with a recursive definition could exceed the stack
562 Constructed ASN.1 types with a recursive definition (such as can be found
563 in PKCS7) could eventually exceed the stack given malicious input with
564 excessive recursion. This could result in a Denial Of Service attack. There
565 are no such structures used within SSL/TLS that come from untrusted sources
566 so this is considered safe.
568 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
573 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
575 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
576 effectively reduced to only comparing the least significant bit of each
577 byte. This allows an attacker to forge messages that would be considered as
578 authenticated in an amount of tries lower than that guaranteed by the
579 security claims of the scheme. The module can only be compiled by the
580 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
582 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
587 *) Add a build target 'build_all_generated', to build all generated files
588 and only that. This can be used to prepare everything that requires
589 things like perl for a system that lacks perl and then move everything
590 to that system and do the rest of the build there.
593 *) Backport SSL_OP_NO_RENGOTIATION
595 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
596 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
597 changes this is no longer possible in 1.1.0. Therefore the new
598 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
599 1.1.0 to provide equivalent functionality.
601 Note that if an application built against 1.1.0h headers (or above) is run
602 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
603 accepted but nothing will happen, i.e. renegotiation will not be prevented.
606 *) Removed the OS390-Unix config target. It relied on a script that doesn't
610 *) rsaz_1024_mul_avx2 overflow bug on x86_64
612 There is an overflow bug in the AVX2 Montgomery multiplication procedure
613 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
614 Analysis suggests that attacks against RSA and DSA as a result of this
615 defect would be very difficult to perform and are not believed likely.
616 Attacks against DH1024 are considered just feasible, because most of the
617 work necessary to deduce information about a private key may be performed
618 offline. The amount of resources required for such an attack would be
619 significant. However, for an attack on TLS to be meaningful, the server
620 would have to share the DH1024 private key among multiple clients, which is
621 no longer an option since CVE-2016-0701.
623 This only affects processors that support the AVX2 but not ADX extensions
624 like Intel Haswell (4th generation).
626 This issue was reported to OpenSSL by David Benjamin (Google). The issue
627 was originally found via the OSS-Fuzz project.
631 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
633 *) bn_sqrx8x_internal carry bug on x86_64
635 There is a carry propagating bug in the x86_64 Montgomery squaring
636 procedure. No EC algorithms are affected. Analysis suggests that attacks
637 against RSA and DSA as a result of this defect would be very difficult to
638 perform and are not believed likely. Attacks against DH are considered just
639 feasible (although very difficult) because most of the work necessary to
640 deduce information about a private key may be performed offline. The amount
641 of resources required for such an attack would be very significant and
642 likely only accessible to a limited number of attackers. An attacker would
643 additionally need online access to an unpatched system using the target
644 private key in a scenario with persistent DH parameters and a private
645 key that is shared between multiple clients.
647 This only affects processors that support the BMI1, BMI2 and ADX extensions
648 like Intel Broadwell (5th generation) and later or AMD Ryzen.
650 This issue was reported to OpenSSL by the OSS-Fuzz project.
654 *) Malformed X.509 IPAddressFamily could cause OOB read
656 If an X.509 certificate has a malformed IPAddressFamily extension,
657 OpenSSL could do a one-byte buffer overread. The most likely result
658 would be an erroneous display of the certificate in text format.
660 This issue was reported to OpenSSL by the OSS-Fuzz project.
664 Changes between 1.1.0e and 1.1.0f [25 May 2017]
666 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
667 platform rather than 'mingw'.
670 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
671 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
672 which is the minimum version we support.
675 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
677 *) Encrypt-Then-Mac renegotiation crash
679 During a renegotiation handshake if the Encrypt-Then-Mac extension is
680 negotiated where it was not in the original handshake (or vice-versa) then
681 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
682 and servers are affected.
684 This issue was reported to OpenSSL by Joe Orton (Red Hat).
688 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
690 *) Truncated packet could crash via OOB read
692 If one side of an SSL/TLS path is running on a 32-bit host and a specific
693 cipher is being used, then a truncated packet can cause that host to
694 perform an out-of-bounds read, usually resulting in a crash.
696 This issue was reported to OpenSSL by Robert Święcki of Google.
700 *) Bad (EC)DHE parameters cause a client crash
702 If a malicious server supplies bad parameters for a DHE or ECDHE key
703 exchange then this can result in the client attempting to dereference a
704 NULL pointer leading to a client crash. This could be exploited in a Denial
707 This issue was reported to OpenSSL by Guido Vranken.
711 *) BN_mod_exp may produce incorrect results on x86_64
713 There is a carry propagating bug in the x86_64 Montgomery squaring
714 procedure. No EC algorithms are affected. Analysis suggests that attacks
715 against RSA and DSA as a result of this defect would be very difficult to
716 perform and are not believed likely. Attacks against DH are considered just
717 feasible (although very difficult) because most of the work necessary to
718 deduce information about a private key may be performed offline. The amount
719 of resources required for such an attack would be very significant and
720 likely only accessible to a limited number of attackers. An attacker would
721 additionally need online access to an unpatched system using the target
722 private key in a scenario with persistent DH parameters and a private
723 key that is shared between multiple clients. For example this can occur by
724 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
725 similar to CVE-2015-3193 but must be treated as a separate problem.
727 This issue was reported to OpenSSL by the OSS-Fuzz project.
731 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
733 *) ChaCha20/Poly1305 heap-buffer-overflow
735 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
736 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
737 crash. This issue is not considered to be exploitable beyond a DoS.
739 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
743 *) CMS Null dereference
745 Applications parsing invalid CMS structures can crash with a NULL pointer
746 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
747 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
748 structure callback if an attempt is made to free certain invalid encodings.
749 Only CHOICE structures using a callback which do not handle NULL value are
752 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
756 *) Montgomery multiplication may produce incorrect results
758 There is a carry propagating bug in the Broadwell-specific Montgomery
759 multiplication procedure that handles input lengths divisible by, but
760 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
761 and DH private keys are impossible. This is because the subroutine in
762 question is not used in operations with the private key itself and an input
763 of the attacker's direct choice. Otherwise the bug can manifest itself as
764 transient authentication and key negotiation failures or reproducible
765 erroneous outcome of public-key operations with specially crafted input.
766 Among EC algorithms only Brainpool P-512 curves are affected and one
767 presumably can attack ECDH key negotiation. Impact was not analyzed in
768 detail, because pre-requisites for attack are considered unlikely. Namely
769 multiple clients have to choose the curve in question and the server has to
770 share the private key among them, neither of which is default behaviour.
771 Even then only clients that chose the curve will be affected.
773 This issue was publicly reported as transient failures and was not
774 initially recognized as a security issue. Thanks to Richard Morgan for
775 providing reproducible case.
779 *) Removed automatic addition of RPATH in shared libraries and executables,
780 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
783 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
785 *) Fix Use After Free for large message sizes
787 The patch applied to address CVE-2016-6307 resulted in an issue where if a
788 message larger than approx 16k is received then the underlying buffer to
789 store the incoming message is reallocated and moved. Unfortunately a
790 dangling pointer to the old location is left which results in an attempt to
791 write to the previously freed location. This is likely to result in a
792 crash, however it could potentially lead to execution of arbitrary code.
794 This issue only affects OpenSSL 1.1.0a.
796 This issue was reported to OpenSSL by Robert Święcki.
800 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
802 *) OCSP Status Request extension unbounded memory growth
804 A malicious client can send an excessively large OCSP Status Request
805 extension. If that client continually requests renegotiation, sending a
806 large OCSP Status Request extension each time, then there will be unbounded
807 memory growth on the server. This will eventually lead to a Denial Of
808 Service attack through memory exhaustion. Servers with a default
809 configuration are vulnerable even if they do not support OCSP. Builds using
810 the "no-ocsp" build time option are not affected.
812 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
816 *) SSL_peek() hang on empty record
818 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
819 sends an empty record. This could be exploited by a malicious peer in a
820 Denial Of Service attack.
822 This issue was reported to OpenSSL by Alex Gaynor.
826 *) Excessive allocation of memory in tls_get_message_header() and
827 dtls1_preprocess_fragment()
829 A (D)TLS message includes 3 bytes for its length in the header for the
830 message. This would allow for messages up to 16Mb in length. Messages of
831 this length are excessive and OpenSSL includes a check to ensure that a
832 peer is sending reasonably sized messages in order to avoid too much memory
833 being consumed to service a connection. A flaw in the logic of version
834 1.1.0 means that memory for the message is allocated too early, prior to
835 the excessive message length check. Due to way memory is allocated in
836 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
837 to service a connection. This could lead to a Denial of Service through
838 memory exhaustion. However, the excessive message length check still takes
839 place, and this would cause the connection to immediately fail. Assuming
840 that the application calls SSL_free() on the failed connection in a timely
841 manner then the 21Mb of allocated memory will then be immediately freed
842 again. Therefore the excessive memory allocation will be transitory in
843 nature. This then means that there is only a security impact if:
845 1) The application does not call SSL_free() in a timely manner in the event
846 that the connection fails
848 2) The application is working in a constrained environment where there is
849 very little free memory
851 3) The attacker initiates multiple connection attempts such that there are
852 multiple connections in a state where memory has been allocated for the
853 connection; SSL_free() has not yet been called; and there is insufficient
854 memory to service the multiple requests.
856 Except in the instance of (1) above any Denial Of Service is likely to be
857 transitory because as soon as the connection fails the memory is
858 subsequently freed again in the SSL_free() call. However there is an
859 increased risk during this period of application crashes due to the lack of
860 memory - which would then mean a more serious Denial of Service.
862 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
863 (CVE-2016-6307 and CVE-2016-6308)
866 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
867 had to be removed. Primary reason is that vendor assembler can't
868 assemble our modules with -KPIC flag. As result it, assembly
869 support, was not even available as option. But its lack means
870 lack of side-channel resistant code, which is incompatible with
871 security by todays standards. Fortunately gcc is readily available
872 prepackaged option, which we firmly point at...
875 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
877 *) Windows command-line tool supports UTF-8 opt-in option for arguments
878 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
879 (to any value) allows Windows user to access PKCS#12 file generated
880 with Windows CryptoAPI and protected with non-ASCII password, as well
881 as files generated under UTF-8 locale on Linux also protected with
885 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
886 have been disabled by default and removed from DEFAULT, just like RC4.
887 See the RC4 item below to re-enable both.
890 *) The method for finding the storage location for the Windows RAND seed file
891 has changed. First we check %RANDFILE%. If that is not set then we check
892 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
893 all else fails we fall back to C:\.
896 *) The EVP_EncryptUpdate() function has had its return type changed from void
897 to int. A return of 0 indicates and error while a return of 1 indicates
901 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
902 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
903 off the constant time implementation for RSA, DSA and DH have been made
904 no-ops and deprecated.
907 *) Windows RAND implementation was simplified to only get entropy by
908 calling CryptGenRandom(). Various other RAND-related tickets
910 [Joseph Wylie Yandle, Rich Salz]
912 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
913 and OPENSSL_LH_, respectively. The old names are available
914 with API compatibility. They new names are now completely documented.
917 *) Unify TYPE_up_ref(obj) methods signature.
918 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
919 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
920 int (instead of void) like all others TYPE_up_ref() methods.
921 So now these methods also check the return value of CRYPTO_atomic_add(),
922 and the validity of object reference counter.
923 [fdasilvayy@gmail.com]
925 *) With Windows Visual Studio builds, the .pdb files are installed
926 alongside the installed libraries and executables. For a static
927 library installation, ossl_static.pdb is the associate compiler
928 generated .pdb file to be used when linking programs.
931 *) Remove openssl.spec. Packaging files belong with the packagers.
934 *) Automatic Darwin/OSX configuration has had a refresh, it will now
935 recognise x86_64 architectures automatically. You can still decide
936 to build for a different bitness with the environment variable
937 KERNEL_BITS (can be 32 or 64), for example:
939 KERNEL_BITS=32 ./config
943 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
944 256 bit AES and HMAC with SHA256.
947 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
950 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
953 *) To enable users to have their own config files and build file templates,
954 Configure looks in the directory indicated by the environment variable
955 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
956 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
957 name and is used as is.
960 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
961 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
962 X509_CERT_FILE_CTX was removed.
965 *) "shared" builds are now the default. To create only static libraries use
966 the "no-shared" Configure option.
969 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
970 All of these option have not worked for some while and are fundamental
974 *) Make various cleanup routines no-ops and mark them as deprecated. Most
975 global cleanup functions are no longer required because they are handled
976 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
977 Explicitly de-initing can cause problems (e.g. where a library that uses
978 OpenSSL de-inits, but an application is still using it). The affected
979 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
980 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
981 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
985 *) --strict-warnings no longer enables runtime debugging options
986 such as REF_DEBUG. Instead, debug options are automatically
987 enabled with '--debug' builds.
988 [Andy Polyakov, Emilia Käsper]
990 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
991 have been moved out of the public header files. New functions for managing
992 these have been added.
995 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
996 objects have been moved out of the public header files. New
997 functions for managing these have been added.
1000 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1001 have been moved out of the public header files. New functions for managing
1002 these have been added.
1005 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1006 moved out of the public header files. New functions for managing these
1010 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1013 *) Removed the mk1mf build scripts.
1016 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1017 it is always safe to #include a header now.
1020 *) Removed the aged BC-32 config and all its supporting scripts
1023 *) Removed support for Ultrix, Netware, and OS/2.
1026 *) Add support for HKDF.
1027 [Alessandro Ghedini]
1029 *) Add support for blake2b and blake2s
1032 *) Added support for "pipelining". Ciphers that have the
1033 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1034 encryptions/decryptions simultaneously. There are currently no built-in
1035 ciphers with this property but the expectation is that engines will be able
1036 to offer it to significantly improve throughput. Support has been extended
1037 into libssl so that multiple records for a single connection can be
1038 processed in one go (for >=TLS 1.1).
1041 *) Added the AFALG engine. This is an async capable engine which is able to
1042 offload work to the Linux kernel. In this initial version it only supports
1043 AES128-CBC. The kernel must be version 4.1.0 or greater.
1046 *) OpenSSL now uses a new threading API. It is no longer necessary to
1047 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1048 are two supported threading models: pthreads and windows threads. It is
1049 also possible to configure OpenSSL at compile time for "no-threads". The
1050 old threading API should no longer be used. The functions have been
1051 replaced with "no-op" compatibility macros.
1052 [Alessandro Ghedini, Matt Caswell]
1054 *) Modify behavior of ALPN to invoke callback after SNI/servername
1055 callback, such that updates to the SSL_CTX affect ALPN.
1058 *) Add SSL_CIPHER queries for authentication and key-exchange.
1061 *) Changes to the DEFAULT cipherlist:
1062 - Prefer (EC)DHE handshakes over plain RSA.
1063 - Prefer AEAD ciphers over legacy ciphers.
1064 - Prefer ECDSA over RSA when both certificates are available.
1065 - Prefer TLSv1.2 ciphers/PRF.
1066 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1070 *) Change the ECC default curve list to be this, in order: x25519,
1071 secp256r1, secp521r1, secp384r1.
1074 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1075 disabled by default. They can be re-enabled using the
1076 enable-weak-ssl-ciphers option to Configure.
1079 *) If the server has ALPN configured, but supports no protocols that the
1080 client advertises, send a fatal "no_application_protocol" alert.
1081 This behaviour is SHALL in RFC 7301, though it isn't universally
1082 implemented by other servers.
1085 *) Add X25519 support.
1086 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1087 for public and private key encoding using the format documented in
1088 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1089 key generation and key derivation.
1091 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1095 *) Deprecate SRP_VBASE_get_by_user.
1096 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1097 In order to fix an unavoidable memory leak (CVE-2016-0798),
1098 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1099 seed, even if the seed is configured.
1101 Users should use SRP_VBASE_get1_by_user instead. Note that in
1102 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1103 also that even though configuring the SRP seed attempts to hide
1104 invalid usernames by continuing the handshake with fake
1105 credentials, this behaviour is not constant time and no strong
1106 guarantees are made that the handshake is indistinguishable from
1107 that of a valid user.
1110 *) Configuration change; it's now possible to build dynamic engines
1111 without having to build shared libraries and vice versa. This
1112 only applies to the engines in engines/, those in crypto/engine/
1113 will always be built into libcrypto (i.e. "static").
1115 Building dynamic engines is enabled by default; to disable, use
1116 the configuration option "disable-dynamic-engine".
1118 The only requirements for building dynamic engines are the
1119 presence of the DSO module and building with position independent
1120 code, so they will also automatically be disabled if configuring
1121 with "disable-dso" or "disable-pic".
1123 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1124 are also taken away from openssl/opensslconf.h, as they are
1128 *) Configuration change; if there is a known flag to compile
1129 position independent code, it will always be applied on the
1130 libcrypto and libssl object files, and never on the application
1131 object files. This means other libraries that use routines from
1132 libcrypto / libssl can be made into shared libraries regardless
1133 of how OpenSSL was configured.
1135 If this isn't desirable, the configuration options "disable-pic"
1136 or "no-pic" can be used to disable the use of PIC. This will
1137 also disable building shared libraries and dynamic engines.
1140 *) Removed JPAKE code. It was experimental and has no wide use.
1143 *) The INSTALL_PREFIX Makefile variable has been renamed to
1144 DESTDIR. That makes for less confusion on what this variable
1145 is for. Also, the configuration option --install_prefix is
1149 *) Heartbeat for TLS has been removed and is disabled by default
1150 for DTLS; configure with enable-heartbeats. Code that uses the
1151 old #define's might need to be updated.
1152 [Emilia Käsper, Rich Salz]
1154 *) Rename REF_CHECK to REF_DEBUG.
1157 *) New "unified" build system
1159 The "unified" build system is aimed to be a common system for all
1160 platforms we support. With it comes new support for VMS.
1162 This system builds supports building in a different directory tree
1163 than the source tree. It produces one Makefile (for unix family
1164 or lookalikes), or one descrip.mms (for VMS).
1166 The source of information to make the Makefile / descrip.mms is
1167 small files called 'build.info', holding the necessary
1168 information for each directory with source to compile, and a
1169 template in Configurations, like unix-Makefile.tmpl or
1172 With this change, the library names were also renamed on Windows
1173 and on VMS. They now have names that are closer to the standard
1174 on Unix, and include the major version number, and in certain
1175 cases, the architecture they are built for. See "Notes on shared
1176 libraries" in INSTALL.
1178 We rely heavily on the perl module Text::Template.
1181 *) Added support for auto-initialisation and de-initialisation of the library.
1182 OpenSSL no longer requires explicit init or deinit routines to be called,
1183 except in certain circumstances. See the OPENSSL_init_crypto() and
1184 OPENSSL_init_ssl() man pages for further information.
1187 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1188 "peer" argument is now expected to be a BIO_ADDR object.
1190 *) Rewrite of BIO networking library. The BIO library lacked consistent
1191 support of IPv6, and adding it required some more extensive
1192 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1193 which hold all types of addresses and chains of address information.
1194 It also introduces a new API, with functions like BIO_socket,
1195 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1196 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1197 have been adapted accordingly.
1200 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1204 *) CRIME protection: disable compression by default, even if OpenSSL is
1205 compiled with zlib enabled. Applications can still enable compression
1206 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1207 using the SSL_CONF library to configure compression.
1210 *) The signature of the session callback configured with
1211 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1212 was explicitly marked as 'const unsigned char*' instead of
1216 *) Always DPURIFY. Remove the use of uninitialized memory in the
1217 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1220 *) Removed many obsolete configuration items, including
1221 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1222 MD2_CHAR, MD2_INT, MD2_LONG
1224 IDEA_SHORT, IDEA_LONG
1225 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1226 [Rich Salz, with advice from Andy Polyakov]
1228 *) Many BN internals have been moved to an internal header file.
1229 [Rich Salz with help from Andy Polyakov]
1231 *) Configuration and writing out the results from it has changed.
1232 Files such as Makefile include/openssl/opensslconf.h and are now
1233 produced through general templates, such as Makefile.in and
1234 crypto/opensslconf.h.in and some help from the perl module
1237 Also, the center of configuration information is no longer
1238 Makefile. Instead, Configure produces a perl module in
1239 configdata.pm which holds most of the config data (in the hash
1240 table %config), the target data that comes from the target
1241 configuration in one of the Configurations/*.conf files (in
1245 *) To clarify their intended purposes, the Configure options
1246 --prefix and --openssldir change their semantics, and become more
1247 straightforward and less interdependent.
1249 --prefix shall be used exclusively to give the location INSTALLTOP
1250 where programs, scripts, libraries, include files and manuals are
1251 going to be installed. The default is now /usr/local.
1253 --openssldir shall be used exclusively to give the default
1254 location OPENSSLDIR where certificates, private keys, CRLs are
1255 managed. This is also where the default openssl.cnf gets
1257 If the directory given with this option is a relative path, the
1258 values of both the --prefix value and the --openssldir value will
1259 be combined to become OPENSSLDIR.
1260 The default for --openssldir is INSTALLTOP/ssl.
1262 Anyone who uses --openssldir to specify where OpenSSL is to be
1263 installed MUST change to use --prefix instead.
1266 *) The GOST engine was out of date and therefore it has been removed. An up
1267 to date GOST engine is now being maintained in an external repository.
1268 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1269 support for GOST ciphersuites (these are only activated if a GOST engine
1273 *) EGD is no longer supported by default; use enable-egd when
1275 [Ben Kaduk and Rich Salz]
1277 *) The distribution now has Makefile.in files, which are used to
1278 create Makefile's when Configure is run. *Configure must be run
1279 before trying to build now.*
1282 *) The return value for SSL_CIPHER_description() for error conditions
1286 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1288 Obtaining and performing DNSSEC validation of TLSA records is
1289 the application's responsibility. The application provides
1290 the TLSA records of its choice to OpenSSL, and these are then
1291 used to authenticate the peer.
1293 The TLSA records need not even come from DNS. They can, for
1294 example, be used to implement local end-entity certificate or
1295 trust-anchor "pinning", where the "pin" data takes the form
1296 of TLSA records, which can augment or replace verification
1297 based on the usual WebPKI public certification authorities.
1300 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1301 continues to support deprecated interfaces in default builds.
1302 However, applications are strongly advised to compile their
1303 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1304 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1305 or the 1.1.0 releases.
1307 In environments in which all applications have been ported to
1308 not use any deprecated interfaces OpenSSL's Configure script
1309 should be used with the --api=1.1.0 option to entirely remove
1310 support for the deprecated features from the library and
1311 unconditionally disable them in the installed headers.
1312 Essentially the same effect can be achieved with the "no-deprecated"
1313 argument to Configure, except that this will always restrict
1314 the build to just the latest API, rather than a fixed API
1317 As applications are ported to future revisions of the API,
1318 they should update their compile-time OPENSSL_API_COMPAT define
1319 accordingly, but in most cases should be able to continue to
1320 compile with later releases.
1322 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1323 0x10000000L and 0x00908000L, respectively. However those
1324 versions did not support the OPENSSL_API_COMPAT feature, and
1325 so applications are not typically tested for explicit support
1326 of just the undeprecated features of either release.
1329 *) Add support for setting the minimum and maximum supported protocol.
1330 It can bet set via the SSL_set_min_proto_version() and
1331 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1332 MaxProtocol. It's recommended to use the new APIs to disable
1333 protocols instead of disabling individual protocols using
1334 SSL_set_options() or SSL_CONF's Protocol. This change also
1335 removes support for disabling TLS 1.2 in the OpenSSL TLS
1336 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1339 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1342 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1343 and integrates ECDSA and ECDH functionality into EC. Implementations can
1344 now redirect key generation and no longer need to convert to or from
1347 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1348 include the ec.h header file instead.
1351 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1352 ciphers who are no longer supported and drops support the ephemeral RSA key
1353 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1356 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1357 opaque. For HMAC_CTX, the following constructors and destructors
1360 HMAC_CTX *HMAC_CTX_new(void);
1361 void HMAC_CTX_free(HMAC_CTX *ctx);
1363 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1364 destroy such methods has been added. See EVP_MD_meth_new(3) and
1365 EVP_CIPHER_meth_new(3) for documentation.
1368 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1369 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1370 EVP_MD_CTX_reset() should be called instead to reinitialise
1371 an already created structure.
1372 2) For consistency with the majority of our object creators and
1373 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1374 EVP_MD_CTX_(new|free). The old names are retained as macros
1375 for deprecated builds.
1378 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1379 cryptographic operations to be performed asynchronously as long as an
1380 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1381 further details. Libssl has also had this capability integrated with the
1382 introduction of the new mode SSL_MODE_ASYNC and associated error
1383 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1384 pages. This work was developed in partnership with Intel Corp.
1387 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1388 always enabled now. If you want to disable the support you should
1389 exclude it using the list of supported ciphers. This also means that the
1390 "-no_ecdhe" option has been removed from s_server.
1393 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1394 SSL_{CTX_}set1_curves() which can set a list.
1397 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1398 curve you want to support using SSL_{CTX_}set1_curves().
1401 *) State machine rewrite. The state machine code has been significantly
1402 refactored in order to remove much duplication of code and solve issues
1403 with the old code (see ssl/statem/README for further details). This change
1404 does have some associated API changes. Notably the SSL_state() function
1405 has been removed and replaced by SSL_get_state which now returns an
1406 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1407 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1411 *) All instances of the string "ssleay" in the public API were replaced
1412 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1413 Some error codes related to internal RSA_eay API's were renamed.
1416 *) The demo files in crypto/threads were moved to demo/threads.
1419 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1421 [Matt Caswell, Rich Salz]
1423 *) New ASN.1 embed macro.
1425 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1426 structure is not allocated: it is part of the parent. That is instead of
1434 This reduces memory fragmentation and make it impossible to accidentally
1435 set a mandatory field to NULL.
1437 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1438 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1439 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1443 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1446 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1447 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1448 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1449 DES and RC4 ciphersuites.
1452 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1453 This changes the decoding behaviour for some invalid messages,
1454 though the change is mostly in the more lenient direction, and
1455 legacy behaviour is preserved as much as possible.
1458 *) Fix no-stdio build.
1459 [ David Woodhouse <David.Woodhouse@intel.com> and also
1460 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1462 *) New testing framework
1463 The testing framework has been largely rewritten and is now using
1464 perl and the perl modules Test::Harness and an extended variant of
1465 Test::More called OpenSSL::Test to do its work. All test scripts in
1466 test/ have been rewritten into test recipes, and all direct calls to
1467 executables in test/Makefile have become individual recipes using the
1468 simplified testing OpenSSL::Test::Simple.
1470 For documentation on our testing modules, do:
1472 perldoc test/testlib/OpenSSL/Test/Simple.pm
1473 perldoc test/testlib/OpenSSL/Test.pm
1477 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1478 are used; the latter aborts on memory leaks (usually checked on exit).
1479 Some undocumented "set malloc, etc., hooks" functions were removed
1480 and others were changed. All are now documented.
1483 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1485 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1487 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1488 from RFC4279, RFC4785, RFC5487, RFC5489.
1490 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1491 original RSA_PSK patch.
1494 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1495 era flag was never set throughout the codebase (only read). Also removed
1496 SSL3_FLAGS_POP_BUFFER which was only used if
1497 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1500 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1501 to be "oneline" instead of "compat".
1504 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1505 not aware of clients that still exhibit this bug, and the workaround
1506 hasn't been working properly for a while.
1509 *) The return type of BIO_number_read() and BIO_number_written() as well as
1510 the corresponding num_read and num_write members in the BIO structure has
1511 changed from unsigned long to uint64_t. On platforms where an unsigned
1512 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1516 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1517 OpenSSL without support for them. It also means that maintaining
1518 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1519 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1522 *) Removed support for the two export grade static DH ciphersuites
1523 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1524 were newly added (along with a number of other static DH ciphersuites) to
1525 1.0.2. However the two export ones have *never* worked since they were
1526 introduced. It seems strange in any case to be adding new export
1527 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1530 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1531 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1532 and turned into macros which simply call the new preferred function names
1533 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1534 should use the new names instead. Also as part of this change the ssl23.h
1535 header file has been removed.
1538 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1539 code and the associated standard is no longer considered fit-for-purpose.
1542 *) RT2547 was closed. When generating a private key, try to make the
1543 output file readable only by the owner. This behavior change might
1544 be noticeable when interacting with other software.
1546 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1550 *) Added HTTP GET support to the ocsp command.
1553 *) Changed default digest for the dgst and enc commands from MD5 to
1557 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1560 *) Added support for TLS extended master secret from
1561 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1562 initial patch which was a great help during development.
1565 *) All libssl internal structures have been removed from the public header
1566 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1567 now redundant). Users should not attempt to access internal structures
1568 directly. Instead they should use the provided API functions.
1571 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1572 Access to deprecated functions can be re-enabled by running config with
1573 "enable-deprecated". In addition applications wishing to use deprecated
1574 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1575 will, by default, disable some transitive includes that previously existed
1576 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1579 *) Added support for OCB mode. OpenSSL has been granted a patent license
1580 compatible with the OpenSSL license for use of OCB. Details are available
1581 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1582 for OCB can be removed by calling config with no-ocb.
1585 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1586 compatible client hello.
1589 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1590 done while fixing the error code for the key-too-small case.
1591 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1593 *) CA.sh has been removed; use CA.pl instead.
1596 *) Removed old DES API.
1599 *) Remove various unsupported platforms:
1605 Sinix/ReliantUNIX RM400
1610 16-bit platforms such as WIN16
1613 *) Clean up OPENSSL_NO_xxx #define's
1614 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1615 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1616 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1617 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1618 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1619 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1620 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1621 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1622 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1623 Remove MS_STATIC; it's a relic from platforms <32 bits.
1626 *) Cleaned up dead code
1627 Remove all but one '#ifdef undef' which is to be looked at.
1630 *) Clean up calling of xxx_free routines.
1631 Just like free(), fix most of the xxx_free routines to accept
1632 NULL. Remove the non-null checks from callers. Save much code.
1635 *) Add secure heap for storage of private keys (when possible).
1636 Add BIO_s_secmem(), CBIGNUM, etc.
1637 Contributed by Akamai Technologies under our Corporate CLA.
1640 *) Experimental support for a new, fast, unbiased prime candidate generator,
1641 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1642 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1644 *) New output format NSS in the sess_id command line tool. This allows
1645 exporting the session id and the master key in NSS keylog format.
1646 [Martin Kaiser <martin@kaiser.cx>]
1648 *) Harmonize version and its documentation. -f flag is used to display
1650 [mancha <mancha1@zoho.com>]
1652 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1653 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1654 [mancha <mancha1@zoho.com>]
1656 *) Fix some double frees. These are not thought to be exploitable.
1657 [mancha <mancha1@zoho.com>]
1659 *) A missing bounds check in the handling of the TLS heartbeat extension
1660 can be used to reveal up to 64k of memory to a connected client or
1663 Thanks for Neel Mehta of Google Security for discovering this bug and to
1664 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1665 preparing the fix (CVE-2014-0160)
1666 [Adam Langley, Bodo Moeller]
1668 *) Fix for the attack described in the paper "Recovering OpenSSL
1669 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1670 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1671 http://eprint.iacr.org/2014/140
1673 Thanks to Yuval Yarom and Naomi Benger for discovering this
1674 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1675 [Yuval Yarom and Naomi Benger]
1677 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1678 this fixes a limitation in previous versions of OpenSSL.
1681 *) Experimental encrypt-then-mac support.
1683 Experimental support for encrypt then mac from
1684 draft-gutmann-tls-encrypt-then-mac-02.txt
1686 To enable it set the appropriate extension number (0x42 for the test
1687 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1689 For non-compliant peers (i.e. just about everything) this should have no
1692 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1696 *) Add EVP support for key wrapping algorithms, to avoid problems with
1697 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1698 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1699 algorithms and include tests cases.
1702 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1706 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1707 MGF1 digest and OAEP label.
1710 *) Make openssl verify return errors.
1711 [Chris Palmer <palmer@google.com> and Ben Laurie]
1713 *) New function ASN1_TIME_diff to calculate the difference between two
1714 ASN1_TIME structures or one structure and the current time.
1717 *) Update fips_test_suite to support multiple command line options. New
1718 test to induce all self test errors in sequence and check expected
1722 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1723 sign or verify all in one operation.
1726 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1727 test programs and fips_test_suite. Includes functionality to parse
1728 the minimal script output of fipsalgest.pl directly.
1731 *) Add authorisation parameter to FIPS_module_mode_set().
1734 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1737 *) Use separate DRBG fields for internal and external flags. New function
1738 FIPS_drbg_health_check() to perform on demand health checking. Add
1739 generation tests to fips_test_suite with reduced health check interval to
1740 demonstrate periodic health checking. Add "nodh" option to
1741 fips_test_suite to skip very slow DH test.
1744 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1748 *) More extensive health check for DRBG checking many more failure modes.
1749 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1750 combination: call this in fips_test_suite.
1753 *) Add support for canonical generation of DSA parameter 'g'. See
1756 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1757 POST to handle HMAC cases.
1760 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1761 to return numerical and string versions of the FIPS module number.
1764 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1765 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1766 outside the validated module in the FIPS capable OpenSSL.
1769 *) Minor change to DRBG entropy callback semantics. In some cases
1770 there is no multiple of the block length between min_len and
1771 max_len. Allow the callback to return more than max_len bytes
1772 of entropy but discard any extra: it is the callback's responsibility
1773 to ensure that the extra data discarded does not impact the
1774 requested amount of entropy.
1777 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1778 information in FIPS186-3, SP800-57 and SP800-131A.
1781 *) CCM support via EVP. Interface is very similar to GCM case except we
1782 must supply all data in one chunk (i.e. no update, final) and the
1783 message length must be supplied if AAD is used. Add algorithm test
1787 *) Initial version of POST overhaul. Add POST callback to allow the status
1788 of POST to be monitored and/or failures induced. Modify fips_test_suite
1789 to use callback. Always run all selftests even if one fails.
1792 *) XTS support including algorithm test driver in the fips_gcmtest program.
1793 Note: this does increase the maximum key length from 32 to 64 bytes but
1794 there should be no binary compatibility issues as existing applications
1795 will never use XTS mode.
1798 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1799 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1800 performs algorithm blocking for unapproved PRNG types. Also do not
1801 set PRNG type in FIPS_mode_set(): leave this to the application.
1802 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1803 the standard OpenSSL PRNG: set additional data to a date time vector.
1806 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1807 This shouldn't present any incompatibility problems because applications
1808 shouldn't be using these directly and any that are will need to rethink
1809 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1812 *) Extensive self tests and health checking required by SP800-90 DRBG.
1813 Remove strength parameter from FIPS_drbg_instantiate and always
1814 instantiate at maximum supported strength.
1817 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1820 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1823 *) New function DH_compute_key_padded() to compute a DH key and pad with
1824 leading zeroes if needed: this complies with SP800-56A et al.
1827 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1828 anything, incomplete, subject to change and largely untested at present.
1831 *) Modify fipscanisteronly build option to only build the necessary object
1832 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1835 *) Add experimental option FIPSSYMS to give all symbols in
1836 fipscanister.o and FIPS or fips prefix. This will avoid
1837 conflicts with future versions of OpenSSL. Add perl script
1838 util/fipsas.pl to preprocess assembly language source files
1839 and rename any affected symbols.
1842 *) Add selftest checks and algorithm block of non-fips algorithms in
1843 FIPS mode. Remove DES2 from selftests.
1846 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1847 return internal method without any ENGINE dependencies. Add new
1848 tiny fips sign and verify functions.
1851 *) New build option no-ec2m to disable characteristic 2 code.
1854 *) New build option "fipscanisteronly". This only builds fipscanister.o
1855 and (currently) associated fips utilities. Uses the file Makefile.fips
1856 instead of Makefile.org as the prototype.
1859 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1860 Update fips_gcmtest to use IV generator.
1863 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1864 setting output buffer to NULL. The *Final function must be
1865 called although it will not retrieve any additional data. The tag
1866 can be set or retrieved with a ctrl. The IV length is by default 12
1867 bytes (96 bits) but can be set to an alternative value. If the IV
1868 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1872 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1873 underlying do_cipher function handles all cipher semantics itself
1874 including padding and finalisation. This is useful if (for example)
1875 an ENGINE cipher handles block padding itself. The behaviour of
1876 do_cipher is subtly changed if this flag is set: the return value
1877 is the number of characters written to the output buffer (zero is
1878 no longer an error code) or a negative error code. Also if the
1879 input buffer is NULL and length 0 finalisation should be performed.
1882 *) If a candidate issuer certificate is already part of the constructed
1883 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1886 *) Improve forward-security support: add functions
1888 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1889 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1891 for use by SSL/TLS servers; the callback function will be called whenever a
1892 new session is created, and gets to decide whether the session may be
1893 cached to make it resumable (return 0) or not (return 1). (As by the
1894 SSL/TLS protocol specifications, the session_id sent by the server will be
1895 empty to indicate that the session is not resumable; also, the server will
1896 not generate RFC 4507 (RFC 5077) session tickets.)
1898 A simple reasonable callback implementation is to return is_forward_secure.
1899 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1900 by the SSL/TLS server library, indicating whether it can provide forward
1902 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1904 *) New -verify_name option in command line utilities to set verification
1908 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1909 Add CMAC pkey methods.
1912 *) Experimental renegotiation in s_server -www mode. If the client
1913 browses /reneg connection is renegotiated. If /renegcert it is
1914 renegotiated requesting a certificate.
1917 *) Add an "external" session cache for debugging purposes to s_server. This
1918 should help trace issues which normally are only apparent in deployed
1919 multi-process servers.
1922 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1923 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1924 BIO_set_cipher() and some obscure PEM functions were changed so they
1925 can now return an error. The RAND changes required a change to the
1926 RAND_METHOD structure.
1929 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1930 a gcc attribute to warn if the result of a function is ignored. This
1931 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1932 whose return value is often ignored.
1935 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1936 These allow SCTs (signed certificate timestamps) to be requested and
1937 validated when establishing a connection.
1938 [Rob Percival <robpercival@google.com>]
1940 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1942 *) Prevent padding oracle in AES-NI CBC MAC check
1944 A MITM attacker can use a padding oracle attack to decrypt traffic
1945 when the connection uses an AES CBC cipher and the server support
1948 This issue was introduced as part of the fix for Lucky 13 padding
1949 attack (CVE-2013-0169). The padding check was rewritten to be in
1950 constant time by making sure that always the same bytes are read and
1951 compared against either the MAC or padding bytes. But it no longer
1952 checked that there was enough data to have both the MAC and padding
1955 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1959 *) Fix EVP_EncodeUpdate overflow
1961 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1962 Base64 encoding of binary data. If an attacker is able to supply very large
1963 amounts of input data then a length check can overflow resulting in a heap
1966 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1967 the PEM_write_bio* family of functions. These are mainly used within the
1968 OpenSSL command line applications, so any application which processes data
1969 from an untrusted source and outputs it as a PEM file should be considered
1970 vulnerable to this issue. User applications that call these APIs directly
1971 with large amounts of untrusted data may also be vulnerable.
1973 This issue was reported by Guido Vranken.
1977 *) Fix EVP_EncryptUpdate overflow
1979 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1980 is able to supply very large amounts of input data after a previous call to
1981 EVP_EncryptUpdate() with a partial block then a length check can overflow
1982 resulting in a heap corruption. Following an analysis of all OpenSSL
1983 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1984 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1985 the first called function after an EVP_EncryptInit(), and therefore that
1986 specific call must be safe. The second form is where the length passed to
1987 EVP_EncryptUpdate() can be seen from the code to be some small value and
1988 therefore there is no possibility of an overflow. Since all instances are
1989 one of these two forms, it is believed that there can be no overflows in
1990 internal code due to this problem. It should be noted that
1991 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1992 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1993 of these calls have also been analysed too and it is believed there are no
1994 instances in internal usage where an overflow could occur.
1996 This issue was reported by Guido Vranken.
2000 *) Prevent ASN.1 BIO excessive memory allocation
2002 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2003 a short invalid encoding can cause allocation of large amounts of memory
2004 potentially consuming excessive resources or exhausting memory.
2006 Any application parsing untrusted data through d2i BIO functions is
2007 affected. The memory based functions such as d2i_X509() are *not* affected.
2008 Since the memory based functions are used by the TLS library, TLS
2009 applications are not affected.
2011 This issue was reported by Brian Carpenter.
2017 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2018 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2019 in arbitrary stack data being returned in the buffer.
2021 This issue was reported by Guido Vranken.
2025 *) Modify behavior of ALPN to invoke callback after SNI/servername
2026 callback, such that updates to the SSL_CTX affect ALPN.
2029 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2033 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2034 methods are enabled and ssl2 is disabled the methods return NULL.
2037 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2039 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2040 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2041 provide any "EXPORT" or "LOW" strength ciphers.
2044 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2045 is by default disabled at build-time. Builds that are not configured with
2046 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2047 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2048 will need to explicitly call either of:
2050 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2052 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2054 as appropriate. Even if either of those is used, or the application
2055 explicitly uses the version-specific SSLv2_method() or its client and
2056 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2057 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2058 ciphers, and SSLv2 56-bit DES are no longer available.
2062 *) Fix a double-free in DSA code
2064 A double free bug was discovered when OpenSSL parses malformed DSA private
2065 keys and could lead to a DoS attack or memory corruption for applications
2066 that receive DSA private keys from untrusted sources. This scenario is
2069 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2074 *) Disable SRP fake user seed to address a server memory leak.
2076 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2078 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2079 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2080 was changed to ignore the "fake user" SRP seed, even if the seed
2083 Users should use SRP_VBASE_get1_by_user instead. Note that in
2084 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2085 also that even though configuring the SRP seed attempts to hide
2086 invalid usernames by continuing the handshake with fake
2087 credentials, this behaviour is not constant time and no strong
2088 guarantees are made that the handshake is indistinguishable from
2089 that of a valid user.
2093 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2095 In the BN_hex2bn function the number of hex digits is calculated using an
2096 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2097 large values of |i| this can result in |bn_expand| not allocating any
2098 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2099 field as NULL leading to a subsequent NULL ptr deref. For very large values
2100 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2101 In this case memory is allocated to the internal BIGNUM data field, but it
2102 is insufficiently sized leading to heap corruption. A similar issue exists
2103 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2104 is ever called by user applications with very large untrusted hex/dec data.
2105 This is anticipated to be a rare occurrence.
2107 All OpenSSL internal usage of these functions use data that is not expected
2108 to be untrusted, e.g. config file data or application command line
2109 arguments. If user developed applications generate config file data based
2110 on untrusted data then it is possible that this could also lead to security
2111 consequences. This is also anticipated to be rare.
2113 This issue was reported to OpenSSL by Guido Vranken.
2117 *) Fix memory issues in BIO_*printf functions
2119 The internal |fmtstr| function used in processing a "%s" format string in
2120 the BIO_*printf functions could overflow while calculating the length of a
2121 string and cause an OOB read when printing very long strings.
2123 Additionally the internal |doapr_outch| function can attempt to write to an
2124 OOB memory location (at an offset from the NULL pointer) in the event of a
2125 memory allocation failure. In 1.0.2 and below this could be caused where
2126 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2127 could be in processing a very long "%s" format string. Memory leaks can
2130 The first issue may mask the second issue dependent on compiler behaviour.
2131 These problems could enable attacks where large amounts of untrusted data
2132 is passed to the BIO_*printf functions. If applications use these functions
2133 in this way then they could be vulnerable. OpenSSL itself uses these
2134 functions when printing out human-readable dumps of ASN.1 data. Therefore
2135 applications that print this data could be vulnerable if the data is from
2136 untrusted sources. OpenSSL command line applications could also be
2137 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2138 as command line arguments.
2140 Libssl is not considered directly vulnerable. Additionally certificates etc
2141 received via remote connections via libssl are also unlikely to be able to
2142 trigger these issues because of message size limits enforced within libssl.
2144 This issue was reported to OpenSSL Guido Vranken.
2148 *) Side channel attack on modular exponentiation
2150 A side-channel attack was found which makes use of cache-bank conflicts on
2151 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2152 of RSA keys. The ability to exploit this issue is limited as it relies on
2153 an attacker who has control of code in a thread running on the same
2154 hyper-threaded core as the victim thread which is performing decryptions.
2156 This issue was reported to OpenSSL by Yuval Yarom, The University of
2157 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2158 Nadia Heninger, University of Pennsylvania with more information at
2159 http://cachebleed.info.
2163 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2164 if no keysize is specified with default_bits. This fixes an
2165 omission in an earlier change that changed all RSA/DSA key generation
2166 apps to use 2048 bits by default.
2169 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2170 *) DH small subgroups
2172 Historically OpenSSL only ever generated DH parameters based on "safe"
2173 primes. More recently (in version 1.0.2) support was provided for
2174 generating X9.42 style parameter files such as those required for RFC 5114
2175 support. The primes used in such files may not be "safe". Where an
2176 application is using DH configured with parameters based on primes that are
2177 not "safe" then an attacker could use this fact to find a peer's private
2178 DH exponent. This attack requires that the attacker complete multiple
2179 handshakes in which the peer uses the same private DH exponent. For example
2180 this could be used to discover a TLS server's private DH exponent if it's
2181 reusing the private DH exponent or it's using a static DH ciphersuite.
2183 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2184 TLS. It is not on by default. If the option is not set then the server
2185 reuses the same private DH exponent for the life of the server process and
2186 would be vulnerable to this attack. It is believed that many popular
2187 applications do set this option and would therefore not be at risk.
2189 The fix for this issue adds an additional check where a "q" parameter is
2190 available (as is the case in X9.42 based parameters). This detects the
2191 only known attack, and is the only possible defense for static DH
2192 ciphersuites. This could have some performance impact.
2194 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2195 default and cannot be disabled. This could have some performance impact.
2197 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2201 *) SSLv2 doesn't block disabled ciphers
2203 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2204 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2205 been disabled, provided that the SSLv2 protocol was not also disabled via
2208 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2209 and Sebastian Schinzel.
2213 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2215 *) BN_mod_exp may produce incorrect results on x86_64
2217 There is a carry propagating bug in the x86_64 Montgomery squaring
2218 procedure. No EC algorithms are affected. Analysis suggests that attacks
2219 against RSA and DSA as a result of this defect would be very difficult to
2220 perform and are not believed likely. Attacks against DH are considered just
2221 feasible (although very difficult) because most of the work necessary to
2222 deduce information about a private key may be performed offline. The amount
2223 of resources required for such an attack would be very significant and
2224 likely only accessible to a limited number of attackers. An attacker would
2225 additionally need online access to an unpatched system using the target
2226 private key in a scenario with persistent DH parameters and a private
2227 key that is shared between multiple clients. For example this can occur by
2228 default in OpenSSL DHE based SSL/TLS ciphersuites.
2230 This issue was reported to OpenSSL by Hanno Böck.
2234 *) Certificate verify crash with missing PSS parameter
2236 The signature verification routines will crash with a NULL pointer
2237 dereference if presented with an ASN.1 signature using the RSA PSS
2238 algorithm and absent mask generation function parameter. Since these
2239 routines are used to verify certificate signature algorithms this can be
2240 used to crash any certificate verification operation and exploited in a
2241 DoS attack. Any application which performs certificate verification is
2242 vulnerable including OpenSSL clients and servers which enable client
2245 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2249 *) X509_ATTRIBUTE memory leak
2251 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2252 memory. This structure is used by the PKCS#7 and CMS routines so any
2253 application which reads PKCS#7 or CMS data from untrusted sources is
2254 affected. SSL/TLS is not affected.
2256 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2261 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2262 This changes the decoding behaviour for some invalid messages,
2263 though the change is mostly in the more lenient direction, and
2264 legacy behaviour is preserved as much as possible.
2267 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2269 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2271 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2273 *) Alternate chains certificate forgery
2275 During certificate verification, OpenSSL will attempt to find an
2276 alternative certificate chain if the first attempt to build such a chain
2277 fails. An error in the implementation of this logic can mean that an
2278 attacker could cause certain checks on untrusted certificates to be
2279 bypassed, such as the CA flag, enabling them to use a valid leaf
2280 certificate to act as a CA and "issue" an invalid certificate.
2282 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2286 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2288 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2289 incompatibility in the handling of HMAC. The previous ABI has now been
2293 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2295 *) Malformed ECParameters causes infinite loop
2297 When processing an ECParameters structure OpenSSL enters an infinite loop
2298 if the curve specified is over a specially malformed binary polynomial
2301 This can be used to perform denial of service against any
2302 system which processes public keys, certificate requests or
2303 certificates. This includes TLS clients and TLS servers with
2304 client authentication enabled.
2306 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2310 *) Exploitable out-of-bounds read in X509_cmp_time
2312 X509_cmp_time does not properly check the length of the ASN1_TIME
2313 string and can read a few bytes out of bounds. In addition,
2314 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2317 An attacker can use this to craft malformed certificates and CRLs of
2318 various sizes and potentially cause a segmentation fault, resulting in
2319 a DoS on applications that verify certificates or CRLs. TLS clients
2320 that verify CRLs are affected. TLS clients and servers with client
2321 authentication enabled may be affected if they use custom verification
2324 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2325 independently by Hanno Böck.
2329 *) PKCS7 crash with missing EnvelopedContent
2331 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2332 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2333 with missing content and trigger a NULL pointer dereference on parsing.
2335 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2336 structures from untrusted sources are affected. OpenSSL clients and
2337 servers are not affected.
2339 This issue was reported to OpenSSL by Michal Zalewski (Google).
2343 *) CMS verify infinite loop with unknown hash function
2345 When verifying a signedData message the CMS code can enter an infinite loop
2346 if presented with an unknown hash function OID. This can be used to perform
2347 denial of service against any system which verifies signedData messages using
2349 This issue was reported to OpenSSL by Johannes Bauer.
2353 *) Race condition handling NewSessionTicket
2355 If a NewSessionTicket is received by a multi-threaded client when attempting to
2356 reuse a previous ticket then a race condition can occur potentially leading to
2357 a double free of the ticket data.
2361 *) Only support 256-bit or stronger elliptic curves with the
2362 'ecdh_auto' setting (server) or by default (client). Of supported
2363 curves, prefer P-256 (both).
2366 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2368 *) ClientHello sigalgs DoS fix
2370 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2371 invalid signature algorithms extension a NULL pointer dereference will
2372 occur. This can be exploited in a DoS attack against the server.
2374 This issue was was reported to OpenSSL by David Ramos of Stanford
2377 [Stephen Henson and Matt Caswell]
2379 *) Multiblock corrupted pointer fix
2381 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2382 feature only applies on 64 bit x86 architecture platforms that support AES
2383 NI instructions. A defect in the implementation of "multiblock" can cause
2384 OpenSSL's internal write buffer to become incorrectly set to NULL when
2385 using non-blocking IO. Typically, when the user application is using a
2386 socket BIO for writing, this will only result in a failed connection.
2387 However if some other BIO is used then it is likely that a segmentation
2388 fault will be triggered, thus enabling a potential DoS attack.
2390 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2394 *) Segmentation fault in DTLSv1_listen fix
2396 The DTLSv1_listen function is intended to be stateless and processes the
2397 initial ClientHello from many peers. It is common for user code to loop
2398 over the call to DTLSv1_listen until a valid ClientHello is received with
2399 an associated cookie. A defect in the implementation of DTLSv1_listen means
2400 that state is preserved in the SSL object from one invocation to the next
2401 that can lead to a segmentation fault. Errors processing the initial
2402 ClientHello can trigger this scenario. An example of such an error could be
2403 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2406 This issue was reported to OpenSSL by Per Allansson.
2410 *) Segmentation fault in ASN1_TYPE_cmp fix
2412 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2413 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2414 certificate signature algorithm consistency this can be used to crash any
2415 certificate verification operation and exploited in a DoS attack. Any
2416 application which performs certificate verification is vulnerable including
2417 OpenSSL clients and servers which enable client authentication.
2421 *) Segmentation fault for invalid PSS parameters fix
2423 The signature verification routines will crash with a NULL pointer
2424 dereference if presented with an ASN.1 signature using the RSA PSS
2425 algorithm and invalid parameters. Since these routines are used to verify
2426 certificate signature algorithms this can be used to crash any
2427 certificate verification operation and exploited in a DoS attack. Any
2428 application which performs certificate verification is vulnerable including
2429 OpenSSL clients and servers which enable client authentication.
2431 This issue was was reported to OpenSSL by Brian Carpenter.
2435 *) ASN.1 structure reuse memory corruption fix
2437 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2438 memory corruption via an invalid write. Such reuse is and has been
2439 strongly discouraged and is believed to be rare.
2441 Applications that parse structures containing CHOICE or ANY DEFINED BY
2442 components may be affected. Certificate parsing (d2i_X509 and related
2443 functions) are however not affected. OpenSSL clients and servers are
2448 *) PKCS7 NULL pointer dereferences fix
2450 The PKCS#7 parsing code does not handle missing outer ContentInfo
2451 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2452 missing content and trigger a NULL pointer dereference on parsing.
2454 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2455 otherwise parse PKCS#7 structures from untrusted sources are
2456 affected. OpenSSL clients and servers are not affected.
2458 This issue was reported to OpenSSL by Michal Zalewski (Google).
2462 *) DoS via reachable assert in SSLv2 servers fix
2464 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2465 servers that both support SSLv2 and enable export cipher suites by sending
2466 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2468 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2469 (OpenSSL development team).
2473 *) Empty CKE with client auth and DHE fix
2475 If client auth is used then a server can seg fault in the event of a DHE
2476 ciphersuite being selected and a zero length ClientKeyExchange message
2477 being sent by the client. This could be exploited in a DoS attack.
2481 *) Handshake with unseeded PRNG fix
2483 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2484 with an unseeded PRNG. The conditions are:
2485 - The client is on a platform where the PRNG has not been seeded
2486 automatically, and the user has not seeded manually
2487 - A protocol specific client method version has been used (i.e. not
2488 SSL_client_methodv23)
2489 - A ciphersuite is used that does not require additional random data from
2490 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2492 If the handshake succeeds then the client random that has been used will
2493 have been generated from a PRNG with insufficient entropy and therefore the
2494 output may be predictable.
2496 For example using the following command with an unseeded openssl will
2497 succeed on an unpatched platform:
2499 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2503 *) Use After Free following d2i_ECPrivatekey error fix
2505 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2506 could cause a use after free condition. This, in turn, could cause a double
2507 free in several private key parsing functions (such as d2i_PrivateKey
2508 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2509 for applications that receive EC private keys from untrusted
2510 sources. This scenario is considered rare.
2512 This issue was discovered by the BoringSSL project and fixed in their
2517 *) X509_to_X509_REQ NULL pointer deref fix
2519 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2520 the certificate key is invalid. This function is rarely used in practice.
2522 This issue was discovered by Brian Carpenter.
2526 *) Removed the export ciphers from the DEFAULT ciphers
2529 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2531 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2532 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2533 So far those who have to target multiple platforms would compromise
2534 and argue that binary targeting say ARMv5 would still execute on
2535 ARMv8. "Universal" build resolves this compromise by providing
2536 near-optimal performance even on newer platforms.
2539 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2540 (other platforms pending).
2541 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2543 *) Add support for the SignedCertificateTimestampList certificate and
2544 OCSP response extensions from RFC6962.
2547 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2548 for corner cases. (Certain input points at infinity could lead to
2549 bogus results, with non-infinity inputs mapped to infinity too.)
2552 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2553 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2554 common cases are optimized and there still is room for further
2555 improvements. Vector Permutation AES for Altivec is also added.
2558 *) Add support for little-endian ppc64 Linux target.
2559 [Marcelo Cerri (IBM)]
2561 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2562 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2563 are optimized and there still is room for further improvements.
2564 Both 32- and 64-bit modes are supported.
2565 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2567 *) Improved ARMv7 NEON support.
2570 *) Support for SPARC Architecture 2011 crypto extensions, first
2571 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2572 SHA256/512, MD5, GHASH and modular exponentiation.
2573 [Andy Polyakov, David Miller]
2575 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2577 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2579 *) Support for new and upcoming Intel processors, including AVX2,
2580 BMI and SHA ISA extensions. This includes additional "stitched"
2581 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2584 This work was sponsored by Intel Corp.
2587 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2588 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2589 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2592 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2593 this fixes a limitation in previous versions of OpenSSL.
2596 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2597 MGF1 digest and OAEP label.
2600 *) Add EVP support for key wrapping algorithms, to avoid problems with
2601 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2602 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2603 algorithms and include tests cases.
2606 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2608 [Douglas E. Engert, Steve Henson]
2610 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2611 difference in days and seconds between two tm or ASN1_TIME structures.
2614 *) Add -rev test option to s_server to just reverse order of characters
2615 received by client and send back to server. Also prints an abbreviated
2616 summary of the connection parameters.
2619 *) New option -brief for s_client and s_server to print out a brief summary
2620 of connection parameters.
2623 *) Add callbacks for arbitrary TLS extensions.
2624 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2626 *) New option -crl_download in several openssl utilities to download CRLs
2627 from CRLDP extension in certificates.
2630 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2633 *) New function X509_CRL_diff to generate a delta CRL from the difference
2634 of two full CRLs. Add support to "crl" utility.
2637 *) New functions to set lookup_crls function and to retrieve
2638 X509_STORE from X509_STORE_CTX.
2641 *) Print out deprecated issuer and subject unique ID fields in
2645 *) Extend OCSP I/O functions so they can be used for simple general purpose
2646 HTTP as well as OCSP. New wrapper function which can be used to download
2647 CRLs using the OCSP API.
2650 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2653 *) SSL_CONF* functions. These provide a common framework for application
2654 configuration using configuration files or command lines.
2657 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2658 message callback and prints the results. Needs compile time option
2659 "enable-ssl-trace". New options to s_client and s_server to enable
2663 *) New ctrl and macro to retrieve supported points extensions.
2664 Print out extension in s_server and s_client.
2667 *) New functions to retrieve certificate signature and signature
2671 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2675 *) New Suite B modes for TLS code. These use and enforce the requirements
2676 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2677 only use Suite B curves. The Suite B modes can be set by using the
2678 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2681 *) New chain verification flags for Suite B levels of security. Check
2682 algorithms are acceptable when flags are set in X509_verify_cert.
2685 *) Make tls1_check_chain return a set of flags indicating checks passed
2686 by a certificate chain. Add additional tests to handle client
2687 certificates: checks for matching certificate type and issuer name
2691 *) If an attempt is made to use a signature algorithm not in the peer
2692 preference list abort the handshake. If client has no suitable
2693 signature algorithms in response to a certificate request do not
2694 use the certificate.
2697 *) If server EC tmp key is not in client preference list abort handshake.
2700 *) Add support for certificate stores in CERT structure. This makes it
2701 possible to have different stores per SSL structure or one store in
2702 the parent SSL_CTX. Include distinct stores for certificate chain
2703 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2704 to build and store a certificate chain in CERT structure: returning
2705 an error if the chain cannot be built: this will allow applications
2706 to test if a chain is correctly configured.
2708 Note: if the CERT based stores are not set then the parent SSL_CTX
2709 store is used to retain compatibility with existing behaviour.
2713 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2714 mask based on the current session, check mask when sending client
2715 hello and checking the requested ciphersuite.
2718 *) New ctrls to retrieve and set certificate types in a certificate
2719 request message. Print out received values in s_client. If certificate
2720 types is not set with custom values set sensible values based on
2721 supported signature algorithms.
2724 *) Support for distinct client and server supported signature algorithms.
2727 *) Add certificate callback. If set this is called whenever a certificate
2728 is required by client or server. An application can decide which
2729 certificate chain to present based on arbitrary criteria: for example
2730 supported signature algorithms. Add very simple example to s_server.
2731 This fixes many of the problems and restrictions of the existing client
2732 certificate callback: for example you can now clear an existing
2733 certificate and specify the whole chain.
2736 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2737 the certificate can be used for (if anything). Set valid_flags field
2738 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2739 to have similar checks in it.
2741 Add new "cert_flags" field to CERT structure and include a "strict mode".
2742 This enforces some TLS certificate requirements (such as only permitting
2743 certificate signature algorithms contained in the supported algorithms
2744 extension) which some implementations ignore: this option should be used
2745 with caution as it could cause interoperability issues.
2748 *) Update and tidy signature algorithm extension processing. Work out
2749 shared signature algorithms based on preferences and peer algorithms
2750 and print them out in s_client and s_server. Abort handshake if no
2751 shared signature algorithms.
2754 *) Add new functions to allow customised supported signature algorithms
2755 for SSL and SSL_CTX structures. Add options to s_client and s_server
2759 *) New function SSL_certs_clear() to delete all references to certificates
2760 from an SSL structure. Before this once a certificate had been added
2761 it couldn't be removed.
2764 *) Integrate hostname, email address and IP address checking with certificate
2765 verification. New verify options supporting checking in openssl utility.
2768 *) Fixes and wildcard matching support to hostname and email checking
2769 functions. Add manual page.
2770 [Florian Weimer (Red Hat Product Security Team)]
2772 *) New functions to check a hostname email or IP address against a
2773 certificate. Add options x509 utility to print results of checks against
2777 *) Fix OCSP checking.
2778 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2780 *) Initial experimental support for explicitly trusted non-root CAs.
2781 OpenSSL still tries to build a complete chain to a root but if an
2782 intermediate CA has a trust setting included that is used. The first
2783 setting is used: whether to trust (e.g., -addtrust option to the x509
2787 *) Add -trusted_first option which attempts to find certificates in the
2788 trusted store even if an untrusted chain is also supplied.
2791 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2792 platform support for Linux and Android.
2795 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2798 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2799 When in FIPS mode the approved implementations are used as normal,
2800 when not in FIPS mode the internal unapproved versions are used instead.
2801 This means that the FIPS capable OpenSSL isn't forced to use the
2802 (often lower performance) FIPS implementations outside FIPS mode.
2805 *) Transparently support X9.42 DH parameters when calling
2806 PEM_read_bio_DHparameters. This means existing applications can handle
2807 the new parameter format automatically.
2810 *) Initial experimental support for X9.42 DH parameter format: mainly
2811 to support use of 'q' parameter for RFC5114 parameters.
2814 *) Add DH parameters from RFC5114 including test data to dhtest.
2817 *) Support for automatic EC temporary key parameter selection. If enabled
2818 the most preferred EC parameters are automatically used instead of
2819 hardcoded fixed parameters. Now a server just has to call:
2820 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2821 support ECDH and use the most appropriate parameters.
2824 *) Enhance and tidy EC curve and point format TLS extension code. Use
2825 static structures instead of allocation if default values are used.
2826 New ctrls to set curves we wish to support and to retrieve shared curves.
2827 Print out shared curves in s_server. New options to s_server and s_client
2828 to set list of supported curves.
2831 *) New ctrls to retrieve supported signature algorithms and
2832 supported curve values as an array of NIDs. Extend openssl utility
2833 to print out received values.
2836 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2837 between NIDs and the more common NIST names such as "P-256". Enhance
2838 ecparam utility and ECC method to recognise the NIST names for curves.
2841 *) Enhance SSL/TLS certificate chain handling to support different
2842 chains for each certificate instead of one chain in the parent SSL_CTX.
2845 *) Support for fixed DH ciphersuite client authentication: where both
2846 server and client use DH certificates with common parameters.
2849 *) Support for fixed DH ciphersuites: those requiring DH server
2853 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2855 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2856 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2857 X509_CINF_get_signature were reverted post internal team review.
2859 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2861 *) Build fixes for the Windows and OpenVMS platforms
2862 [Matt Caswell and Richard Levitte]
2864 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2866 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2867 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2868 dereference. This could lead to a Denial Of Service attack. Thanks to
2869 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2873 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2874 dtls1_buffer_record function under certain conditions. In particular this
2875 could occur if an attacker sent repeated DTLS records with the same
2876 sequence number but for the next epoch. The memory leak could be exploited
2877 by an attacker in a Denial of Service attack through memory exhaustion.
2878 Thanks to Chris Mueller for reporting this issue.
2882 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2883 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2884 method would be set to NULL which could later result in a NULL pointer
2885 dereference. Thanks to Frank Schmirler for reporting this issue.
2889 *) Abort handshake if server key exchange message is omitted for ephemeral
2892 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2893 reporting this issue.
2897 *) Remove non-export ephemeral RSA code on client and server. This code
2898 violated the TLS standard by allowing the use of temporary RSA keys in
2899 non-export ciphersuites and could be used by a server to effectively
2900 downgrade the RSA key length used to a value smaller than the server
2901 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2902 INRIA or reporting this issue.
2906 *) Fixed issue where DH client certificates are accepted without verification.
2907 An OpenSSL server will accept a DH certificate for client authentication
2908 without the certificate verify message. This effectively allows a client to
2909 authenticate without the use of a private key. This only affects servers
2910 which trust a client certificate authority which issues certificates
2911 containing DH keys: these are extremely rare and hardly ever encountered.
2912 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2917 *) Ensure that the session ID context of an SSL is updated when its
2918 SSL_CTX is updated via SSL_set_SSL_CTX.
2920 The session ID context is typically set from the parent SSL_CTX,
2921 and can vary with the CTX.
2924 *) Fix various certificate fingerprint issues.
2926 By using non-DER or invalid encodings outside the signed portion of a
2927 certificate the fingerprint can be changed without breaking the signature.
2928 Although no details of the signed portion of the certificate can be changed
2929 this can cause problems with some applications: e.g. those using the
2930 certificate fingerprint for blacklists.
2932 1. Reject signatures with non zero unused bits.
2934 If the BIT STRING containing the signature has non zero unused bits reject
2935 the signature. All current signature algorithms require zero unused bits.
2937 2. Check certificate algorithm consistency.
2939 Check the AlgorithmIdentifier inside TBS matches the one in the
2940 certificate signature. NB: this will result in signature failure
2941 errors for some broken certificates.
2943 Thanks to Konrad Kraszewski from Google for reporting this issue.
2945 3. Check DSA/ECDSA signatures use DER.
2947 Re-encode DSA/ECDSA signatures and compare with the original received
2948 signature. Return an error if there is a mismatch.
2950 This will reject various cases including garbage after signature
2951 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2952 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2953 (negative or with leading zeroes).
2955 Further analysis was conducted and fixes were developed by Stephen Henson
2956 of the OpenSSL core team.
2961 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2962 results on some platforms, including x86_64. This bug occurs at random
2963 with a very low probability, and is not known to be exploitable in any
2964 way, though its exact impact is difficult to determine. Thanks to Pieter
2965 Wuille (Blockstream) who reported this issue and also suggested an initial
2966 fix. Further analysis was conducted by the OpenSSL development team and
2967 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2968 the OpenSSL core team.
2972 *) Do not resume sessions on the server if the negotiated protocol
2973 version does not match the session's version. Resuming with a different
2974 version, while not strictly forbidden by the RFC, is of questionable
2975 sanity and breaks all known clients.
2976 [David Benjamin, Emilia Käsper]
2978 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2979 early CCS messages during renegotiation. (Note that because
2980 renegotiation is encrypted, this early CCS was not exploitable.)
2983 *) Tighten client-side session ticket handling during renegotiation:
2984 ensure that the client only accepts a session ticket if the server sends
2985 the extension anew in the ServerHello. Previously, a TLS client would
2986 reuse the old extension state and thus accept a session ticket if one was
2987 announced in the initial ServerHello.
2989 Similarly, ensure that the client requires a session ticket if one
2990 was advertised in the ServerHello. Previously, a TLS client would
2991 ignore a missing NewSessionTicket message.
2994 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2996 *) SRTP Memory Leak.
2998 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2999 sends a carefully crafted handshake message, to cause OpenSSL to fail
3000 to free up to 64k of memory causing a memory leak. This could be
3001 exploited in a Denial Of Service attack. This issue affects OpenSSL
3002 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3003 whether SRTP is used or configured. Implementations of OpenSSL that
3004 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3006 The fix was developed by the OpenSSL team.
3010 *) Session Ticket Memory Leak.
3012 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3013 integrity of that ticket is first verified. In the event of a session
3014 ticket integrity check failing, OpenSSL will fail to free memory
3015 causing a memory leak. By sending a large number of invalid session
3016 tickets an attacker could exploit this issue in a Denial Of Service
3021 *) Build option no-ssl3 is incomplete.
3023 When OpenSSL is configured with "no-ssl3" as a build option, servers
3024 could accept and complete a SSL 3.0 handshake, and clients could be
3025 configured to send them.
3027 [Akamai and the OpenSSL team]
3029 *) Add support for TLS_FALLBACK_SCSV.
3030 Client applications doing fallback retries should call
3031 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3033 [Adam Langley, Bodo Moeller]
3035 *) Add additional DigestInfo checks.
3037 Re-encode DigestInto in DER and check against the original when
3038 verifying RSA signature: this will reject any improperly encoded
3039 DigestInfo structures.
3041 Note: this is a precautionary measure and no attacks are currently known.
3045 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3047 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3048 SRP code can be overrun an internal buffer. Add sanity check that
3049 g, A, B < N to SRP code.
3051 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3052 Group for discovering this issue.
3056 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3057 TLS 1.0 instead of higher protocol versions when the ClientHello message
3058 is badly fragmented. This allows a man-in-the-middle attacker to force a
3059 downgrade to TLS 1.0 even if both the server and the client support a
3060 higher protocol version, by modifying the client's TLS records.
3062 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3063 researching this issue.
3067 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3068 to a denial of service attack. A malicious server can crash the client
3069 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3070 ciphersuite and sending carefully crafted handshake messages.
3072 Thanks to Felix Gröbert (Google) for discovering and researching this
3077 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3078 to leak memory. This can be exploited through a Denial of Service attack.
3079 Thanks to Adam Langley for discovering and researching this issue.
3083 *) An attacker can force openssl to consume large amounts of memory whilst
3084 processing DTLS handshake messages. This can be exploited through a
3085 Denial of Service attack.
3086 Thanks to Adam Langley for discovering and researching this issue.
3090 *) An attacker can force an error condition which causes openssl to crash
3091 whilst processing DTLS packets due to memory being freed twice. This
3092 can be exploited through a Denial of Service attack.
3093 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3098 *) If a multithreaded client connects to a malicious server using a resumed
3099 session and the server sends an ec point format extension it could write
3100 up to 255 bytes to freed memory.
3102 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3107 *) A malicious server can crash an OpenSSL client with a null pointer
3108 dereference (read) by specifying an SRP ciphersuite even though it was not
3109 properly negotiated with the client. This can be exploited through a
3110 Denial of Service attack.
3112 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3113 discovering and researching this issue.
3117 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3118 X509_name_oneline, X509_name_print_ex et al. to leak some information
3119 from the stack. Applications may be affected if they echo pretty printing
3120 output to the attacker.
3122 Thanks to Ivan Fratric (Google) for discovering this issue.
3124 [Emilia Käsper, and Steve Henson]
3126 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3127 for corner cases. (Certain input points at infinity could lead to
3128 bogus results, with non-infinity inputs mapped to infinity too.)
3131 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3133 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3134 handshake can force the use of weak keying material in OpenSSL
3135 SSL/TLS clients and servers.
3137 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3138 researching this issue. (CVE-2014-0224)
3139 [KIKUCHI Masashi, Steve Henson]
3141 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3142 OpenSSL DTLS client the code can be made to recurse eventually crashing
3145 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3147 [Imre Rad, Steve Henson]
3149 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3150 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3151 client or server. This is potentially exploitable to run arbitrary
3152 code on a vulnerable client or server.
3154 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3155 [Jüri Aedla, Steve Henson]
3157 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3158 are subject to a denial of service attack.
3160 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3161 this issue. (CVE-2014-3470)
3162 [Felix Gröbert, Ivan Fratric, Steve Henson]
3164 *) Harmonize version and its documentation. -f flag is used to display
3166 [mancha <mancha1@zoho.com>]
3168 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3169 in i2d_ECPrivateKey.
3170 [mancha <mancha1@zoho.com>]
3172 *) Fix some double frees. These are not thought to be exploitable.
3173 [mancha <mancha1@zoho.com>]
3175 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3177 *) A missing bounds check in the handling of the TLS heartbeat extension
3178 can be used to reveal up to 64k of memory to a connected client or
3181 Thanks for Neel Mehta of Google Security for discovering this bug and to
3182 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3183 preparing the fix (CVE-2014-0160)
3184 [Adam Langley, Bodo Moeller]
3186 *) Fix for the attack described in the paper "Recovering OpenSSL
3187 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3188 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3189 http://eprint.iacr.org/2014/140
3191 Thanks to Yuval Yarom and Naomi Benger for discovering this
3192 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3193 [Yuval Yarom and Naomi Benger]
3195 *) TLS pad extension: draft-agl-tls-padding-03
3197 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3198 TLS client Hello record length value would otherwise be > 255 and
3199 less that 512 pad with a dummy extension containing zeroes so it
3200 is at least 512 bytes long.
3202 [Adam Langley, Steve Henson]
3204 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3206 *) Fix for TLS record tampering bug. A carefully crafted invalid
3207 handshake could crash OpenSSL with a NULL pointer exception.
3208 Thanks to Anton Johansson for reporting this issues.
3211 *) Keep original DTLS digest and encryption contexts in retransmission
3212 structures so we can use the previous session parameters if they need
3213 to be resent. (CVE-2013-6450)
3216 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3217 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3218 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3219 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3220 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3221 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3222 [Rob Stradling, Adam Langley]
3224 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3226 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3227 supporting platforms or when small records were transferred.
3228 [Andy Polyakov, Steve Henson]
3230 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3232 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3234 This addresses the flaw in CBC record processing discovered by
3235 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3236 at: http://www.isg.rhul.ac.uk/tls/
3238 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3239 Security Group at Royal Holloway, University of London
3240 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3241 Emilia Käsper for the initial patch.
3243 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3245 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3246 ciphersuites which can be exploited in a denial of service attack.
3247 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3248 and detecting this bug and to Wolfgang Ettlinger
3249 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3253 *) Return an error when checking OCSP signatures when key is NULL.
3254 This fixes a DoS attack. (CVE-2013-0166)
3257 *) Make openssl verify return errors.
3258 [Chris Palmer <palmer@google.com> and Ben Laurie]
3260 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3261 the right response is stapled. Also change SSL_get_certificate()
3262 so it returns the certificate actually sent.
3263 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3264 [Rob Stradling <rob.stradling@comodo.com>]
3266 *) Fix possible deadlock when decoding public keys.
3269 *) Don't use TLS 1.0 record version number in initial client hello
3273 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3275 *) Sanity check record length before skipping explicit IV in TLS
3276 1.2, 1.1 and DTLS to fix DoS attack.
3278 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3279 fuzzing as a service testing platform.
3283 *) Initialise tkeylen properly when encrypting CMS messages.
3284 Thanks to Solar Designer of Openwall for reporting this issue.
3287 *) In FIPS mode don't try to use composite ciphers as they are not
3291 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3293 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3294 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3295 mean any application compiled against OpenSSL 1.0.0 headers setting
3296 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3297 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3298 0x10000000L Any application which was previously compiled against
3299 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3300 will need to be recompiled as a result. Letting be results in
3301 inability to disable specifically TLS 1.1 and in client context,
3302 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3305 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3306 disable just protocol X, but all protocols above X *if* there are
3307 protocols *below* X still enabled. In more practical terms it means
3308 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3309 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3310 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3314 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3316 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3317 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3318 in CRYPTO_realloc_clean.
3320 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3321 issue and to Adam Langley <agl@chromium.org> for fixing it.
3323 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3325 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3328 *) Workarounds for some broken servers that "hang" if a client hello
3329 record length exceeds 255 bytes.
3331 1. Do not use record version number > TLS 1.0 in initial client
3332 hello: some (but not all) hanging servers will now work.
3333 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3334 the number of ciphers sent in the client hello. This should be
3335 set to an even number, such as 50, for example by passing:
3336 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3337 Most broken servers should now work.
3338 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3339 TLS 1.2 client support entirely.
3342 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3345 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3347 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3348 STRING form instead of a DigestInfo.
3351 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3352 and the RSA_sign/RSA_verify functions. This was made more apparent when
3353 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3354 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3355 the correct format in RSA_verify so both forms transparently work.
3358 *) Some servers which support TLS 1.0 can choke if we initially indicate
3359 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3360 encrypted premaster secret. As a workaround use the maximum permitted
3361 client version in client hello, this should keep such servers happy
3362 and still work with previous versions of OpenSSL.
3365 *) Add support for TLS/DTLS heartbeats.
3366 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3368 *) Add support for SCTP.
3369 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3371 *) Improved PRNG seeding for VOS.
3372 [Paul Green <Paul.Green@stratus.com>]
3374 *) Extensive assembler packs updates, most notably:
3376 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3377 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3378 - x86_64: bit-sliced AES implementation;
3379 - ARM: NEON support, contemporary platforms optimizations;
3380 - s390x: z196 support;
3381 - *: GHASH and GF(2^m) multiplication implementations;
3385 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3386 (removal of unnecessary code)
3387 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3389 *) Add TLS key material exporter from RFC 5705.
3392 *) Add DTLS-SRTP negotiation from RFC 5764.
3395 *) Add Next Protocol Negotiation,
3396 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3397 disabled with a no-npn flag to config or Configure. Code donated
3399 [Adam Langley <agl@google.com> and Ben Laurie]
3401 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3402 NIST-P256, NIST-P521, with constant-time single point multiplication on
3403 typical inputs. Compiler support for the nonstandard type __uint128_t is
3404 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3405 Code made available under Apache License version 2.0.
3407 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3408 line to include this in your build of OpenSSL, and run "make depend" (or
3409 "make update"). This enables the following EC_METHODs:
3411 EC_GFp_nistp224_method()
3412 EC_GFp_nistp256_method()
3413 EC_GFp_nistp521_method()
3415 EC_GROUP_new_by_curve_name() will automatically use these (while
3416 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3418 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3420 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3421 all platforms. Move ssize_t definition from e_os.h to the public
3422 header file e_os2.h as it now appears in public header file cms.h
3425 *) New -sigopt option to the ca, req and x509 utilities. Additional
3426 signature parameters can be passed using this option and in
3430 *) Add RSA PSS signing function. This will generate and set the
3431 appropriate AlgorithmIdentifiers for PSS based on those in the
3432 corresponding EVP_MD_CTX structure. No application support yet.
3435 *) Support for companion algorithm specific ASN1 signing routines.
3436 New function ASN1_item_sign_ctx() signs a pre-initialised
3437 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3438 the appropriate parameters.
3441 *) Add new algorithm specific ASN1 verification initialisation function
3442 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3443 handling will be the same no matter what EVP_PKEY_METHOD is used.
3444 Add a PSS handler to support verification of PSS signatures: checked
3445 against a number of sample certificates.
3448 *) Add signature printing for PSS. Add PSS OIDs.
3449 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3451 *) Add algorithm specific signature printing. An individual ASN1 method
3452 can now print out signatures instead of the standard hex dump.
3454 More complex signatures (e.g. PSS) can print out more meaningful
3455 information. Include DSA version that prints out the signature
3459 *) Password based recipient info support for CMS library: implementing
3463 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3464 neatly separates the code into cipher and PBE sections and is required
3465 for some algorithms that split PBES2 into separate pieces (such as
3466 password based CMS).
3469 *) Session-handling fixes:
3470 - Fix handling of connections that are resuming with a session ID,
3471 but also support Session Tickets.
3472 - Fix a bug that suppressed issuing of a new ticket if the client
3473 presented a ticket with an expired session.
3474 - Try to set the ticket lifetime hint to something reasonable.
3475 - Make tickets shorter by excluding irrelevant information.
3476 - On the client side, don't ignore renewed tickets.
3477 [Adam Langley, Bodo Moeller (Google)]
3479 *) Fix PSK session representation.
3482 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3484 This work was sponsored by Intel.
3487 *) Add GCM support to TLS library. Some custom code is needed to split
3488 the IV between the fixed (from PRF) and explicit (from TLS record)
3489 portions. This adds all GCM ciphersuites supported by RFC5288 and
3490 RFC5289. Generalise some AES* cipherstrings to include GCM and
3491 add a special AESGCM string for GCM only.
3494 *) Expand range of ctrls for AES GCM. Permit setting invocation
3495 field on decrypt and retrieval of invocation field only on encrypt.
3498 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3499 As required by RFC5289 these ciphersuites cannot be used if for
3500 versions of TLS earlier than 1.2.
3503 *) For FIPS capable OpenSSL interpret a NULL default public key method
3504 as unset and return the appropriate default but do *not* set the default.
3505 This means we can return the appropriate method in applications that
3506 switch between FIPS and non-FIPS modes.
3509 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3510 ENGINE is used then we cannot handle that in the FIPS module so we
3511 keep original code iff non-FIPS operations are allowed.
3514 *) Add -attime option to openssl utilities.
3515 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3517 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3520 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3521 FIPS EC methods unconditionally for now.
3524 *) New build option no-ec2m to disable characteristic 2 code.
3527 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3528 all cases can be covered as some introduce binary incompatibilities.
3531 *) Redirect RSA operations to FIPS module including keygen,
3532 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3535 *) Add similar low level API blocking to ciphers.
3538 *) Low level digest APIs are not approved in FIPS mode: any attempt
3539 to use these will cause a fatal error. Applications that *really* want
3540 to use them can use the private_* version instead.
3543 *) Redirect cipher operations to FIPS module for FIPS builds.
3546 *) Redirect digest operations to FIPS module for FIPS builds.
3549 *) Update build system to add "fips" flag which will link in fipscanister.o
3550 for static and shared library builds embedding a signature if needed.
3553 *) Output TLS supported curves in preference order instead of numerical
3554 order. This is currently hardcoded for the highest order curves first.
3555 This should be configurable so applications can judge speed vs strength.
3558 *) Add TLS v1.2 server support for client authentication.
3561 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3565 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3566 FIPS modules versions.
3569 *) Add TLS v1.2 client side support for client authentication. Keep cache
3570 of handshake records longer as we don't know the hash algorithm to use
3571 until after the certificate request message is received.
3574 *) Initial TLS v1.2 client support. Add a default signature algorithms
3575 extension including all the algorithms we support. Parse new signature
3576 format in client key exchange. Relax some ECC signing restrictions for
3577 TLS v1.2 as indicated in RFC5246.
3580 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3581 to new signature format when needed using client digest preference.
3582 All server ciphersuites should now work correctly in TLS v1.2. No client
3583 support yet and no support for client certificates.
3586 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3587 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3588 ciphersuites. At present only RSA key exchange ciphersuites work with
3589 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3590 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3591 and version checking.
3594 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3595 with this defined it will not be affected by any changes to ssl internal
3596 structures. Add several utility functions to allow openssl application
3597 to work with OPENSSL_NO_SSL_INTERN defined.
3600 *) A long standing patch to add support for SRP from EdelWeb (Peter
3601 Sylvester and Christophe Renou) was integrated.
3602 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3603 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3606 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3609 *) Permit abbreviated handshakes when renegotiating using the function
3610 SSL_renegotiate_abbreviated().
3611 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3613 *) Add call to ENGINE_register_all_complete() to
3614 ENGINE_load_builtin_engines(), so some implementations get used
3615 automatically instead of needing explicit application support.
3618 *) Add support for TLS key exporter as described in RFC5705.
3619 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3621 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3622 a few changes are required:
3624 Add SSL_OP_NO_TLSv1_1 flag.
3625 Add TLSv1_1 methods.
3626 Update version checking logic to handle version 1.1.
3627 Add explicit IV handling (ported from DTLS code).
3628 Add command line options to s_client/s_server.
3631 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3633 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3634 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3635 content decryption and always return the same error. Note: this attack
3636 needs on average 2^20 messages so it only affects automated senders. The
3637 old behaviour can be re-enabled in the CMS code by setting the
3638 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3639 an MMA defence is not necessary.
3640 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3641 this issue. (CVE-2012-0884)
3644 *) Fix CVE-2011-4619: make sure we really are receiving a
3645 client hello before rejecting multiple SGC restarts. Thanks to
3646 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3649 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3651 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3652 Thanks to Antonio Martin, Enterprise Secure Access Research and
3653 Development, Cisco Systems, Inc. for discovering this bug and
3654 preparing a fix. (CVE-2012-0050)
3657 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3659 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3660 of the Vaudenay padding oracle attack on CBC mode encryption
3661 which enables an efficient plaintext recovery attack against
3662 the OpenSSL implementation of DTLS. Their attack exploits timing
3663 differences arising during decryption processing. A research
3664 paper describing this attack can be found at:
3665 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3666 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3667 Security Group at Royal Holloway, University of London
3668 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3669 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3670 for preparing the fix. (CVE-2011-4108)
3671 [Robin Seggelmann, Michael Tuexen]
3673 *) Clear bytes used for block padding of SSL 3.0 records.
3675 [Adam Langley (Google)]
3677 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3678 Kadianakis <desnacked@gmail.com> for discovering this issue and
3679 Adam Langley for preparing the fix. (CVE-2011-4619)
3680 [Adam Langley (Google)]
3682 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3683 [Andrey Kulikov <amdeich@gmail.com>]
3685 *) Prevent malformed RFC3779 data triggering an assertion failure.
3686 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3687 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3688 [Rob Austein <sra@hactrn.net>]
3690 *) Improved PRNG seeding for VOS.
3691 [Paul Green <Paul.Green@stratus.com>]
3693 *) Fix ssl_ciph.c set-up race.
3694 [Adam Langley (Google)]
3696 *) Fix spurious failures in ecdsatest.c.
3697 [Emilia Käsper (Google)]
3699 *) Fix the BIO_f_buffer() implementation (which was mixing different
3700 interpretations of the '..._len' fields).
3701 [Adam Langley (Google)]
3703 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3704 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3705 threads won't reuse the same blinding coefficients.
3707 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3708 lock to call BN_BLINDING_invert_ex, and avoids one use of
3709 BN_BLINDING_update for each BN_BLINDING structure (previously,
3710 the last update always remained unused).
3711 [Emilia Käsper (Google)]
3713 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3714 [Bob Buckholz (Google)]
3716 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3718 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3719 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3720 [Kaspar Brand <ossl@velox.ch>]
3722 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3723 for multi-threaded use of ECDH. (CVE-2011-3210)
3724 [Adam Langley (Google)]
3726 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3729 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3730 signature public key algorithm by using OID xref utilities instead.
3731 Before this you could only use some ECC ciphersuites with SHA1 only.
3734 *) Add protection against ECDSA timing attacks as mentioned in the paper
3735 by Billy Bob Brumley and Nicola Tuveri, see:
3737 http://eprint.iacr.org/2011/232.pdf
3739 [Billy Bob Brumley and Nicola Tuveri]
3741 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3743 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3744 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3746 *) Fix bug in string printing code: if *any* escaping is enabled we must
3747 escape the escape character (backslash) or the resulting string is
3751 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3753 *) Disable code workaround for ancient and obsolete Netscape browsers
3754 and servers: an attacker can use it in a ciphersuite downgrade attack.
3755 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3758 *) Fixed J-PAKE implementation error, originally discovered by
3759 Sebastien Martini, further info and confirmation from Stefan
3760 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3763 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3765 *) Fix extension code to avoid race conditions which can result in a buffer
3766 overrun vulnerability: resumed sessions must not be modified as they can
3767 be shared by multiple threads. CVE-2010-3864
3770 *) Fix WIN32 build system to correctly link an ENGINE directory into
3774 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3776 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3778 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3780 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3782 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3783 context. The operation can be customised via the ctrl mechanism in
3784 case ENGINEs want to include additional functionality.
3787 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3790 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3791 output hashes compatible with older versions of OpenSSL.
3792 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3794 *) Fix compression algorithm handling: if resuming a session use the
3795 compression algorithm of the resumed session instead of determining
3796 it from client hello again. Don't allow server to change algorithm.
3799 *) Add load_crls() function to apps tidying load_certs() too. Add option
3800 to verify utility to allow additional CRLs to be included.
3803 *) Update OCSP request code to permit adding custom headers to the request:
3804 some responders need this.
3807 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3809 [Julia Lawall <julia@diku.dk>]
3811 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3812 needlessly dereferenced structures, used obsolete functions and
3813 didn't handle all updated verify codes correctly.
3816 *) Disable MD2 in the default configuration.
3819 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3820 indicate the initial BIO being pushed or popped. This makes it possible
3821 to determine whether the BIO is the one explicitly called or as a result
3822 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3823 it handles reference counts correctly and doesn't zero out the I/O bio
3824 when it is not being explicitly popped. WARNING: applications which
3825 included workarounds for the old buggy behaviour will need to be modified
3826 or they could free up already freed BIOs.
3829 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3830 renaming to all platforms (within the 0.9.8 branch, this was
3831 done conditionally on Netware platforms to avoid a name clash).
3832 [Guenter <lists@gknw.net>]
3834 *) Add ECDHE and PSK support to DTLS.
3835 [Michael Tuexen <tuexen@fh-muenster.de>]
3837 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3841 *) Add "missing" function EVP_MD_flags() (without this the only way to
3842 retrieve a digest flags is by accessing the structure directly. Update
3843 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3844 or cipher is registered as in the "from" argument. Print out all
3845 registered digests in the dgst usage message instead of manually
3846 attempting to work them out.
3849 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3850 this allows the use of compression and extensions. Change default cipher
3851 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3852 by default unless an application cipher string requests it.
3855 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3856 key ids to find matching certificates and keys but some PKCS#12 files
3857 don't follow the (somewhat unwritten) rules and this strategy fails.
3858 Now just gather all certificates together and the first private key
3859 then look for the first certificate that matches the key.
3862 *) Support use of registered digest and cipher names for dgst and cipher
3863 commands instead of having to add each one as a special case. So now
3870 openssl dgst -sha256 foo
3872 and this works for ENGINE based algorithms too.
3876 *) Update Gost ENGINE to support parameter files.
3877 [Victor B. Wagner <vitus@cryptocom.ru>]
3879 *) Support GeneralizedTime in ca utility.
3880 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3882 *) Enhance the hash format used for certificate directory links. The new
3883 form uses the canonical encoding (meaning equivalent names will work
3884 even if they aren't identical) and uses SHA1 instead of MD5. This form
3885 is incompatible with the older format and as a result c_rehash should
3886 be used to rebuild symbolic links.
3889 *) Make PKCS#8 the default write format for private keys, replacing the
3890 traditional format. This form is standardised, more secure and doesn't
3891 include an implicit MD5 dependency.
3894 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3895 committed to OpenSSL should pass this lot as a minimum.
3898 *) Add session ticket override functionality for use by EAP-FAST.
3899 [Jouni Malinen <j@w1.fi>]
3901 *) Modify HMAC functions to return a value. Since these can be implemented
3902 in an ENGINE errors can occur.
3905 *) Type-checked OBJ_bsearch_ex.
3908 *) Type-checked OBJ_bsearch. Also some constification necessitated
3909 by type-checking. Still to come: TXT_DB, bsearch(?),
3910 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3914 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3915 seconds to a tm structure directly, instead of going through OS
3916 specific date routines. This avoids any issues with OS routines such
3917 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3918 and X509_time_adj_ex() to cover the extended range. The existing
3919 X509_time_adj() is still usable and will no longer have any date issues.
3922 *) Delta CRL support. New use deltas option which will attempt to locate
3923 and search any appropriate delta CRLs available.
3925 This work was sponsored by Google.
3928 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3929 code and add additional score elements. Validate alternate CRL paths
3930 as part of the CRL checking and indicate a new error "CRL path validation
3931 error" in this case. Applications wanting additional details can use
3932 the verify callback and check the new "parent" field. If this is not
3933 NULL CRL path validation is taking place. Existing applications won't
3934 see this because it requires extended CRL support which is off by
3937 This work was sponsored by Google.
3940 *) Support for freshest CRL extension.
3942 This work was sponsored by Google.
3945 *) Initial indirect CRL support. Currently only supported in the CRLs
3946 passed directly and not via lookup. Process certificate issuer
3947 CRL entry extension and lookup CRL entries by bother issuer name
3948 and serial number. Check and process CRL issuer entry in IDP extension.
3950 This work was sponsored by Google.
3953 *) Add support for distinct certificate and CRL paths. The CRL issuer
3954 certificate is validated separately in this case. Only enabled if
3955 an extended CRL support flag is set: this flag will enable additional
3956 CRL functionality in future.
3958 This work was sponsored by Google.
3961 *) Add support for policy mappings extension.
3963 This work was sponsored by Google.
3966 *) Fixes to pathlength constraint, self issued certificate handling,
3967 policy processing to align with RFC3280 and PKITS tests.
3969 This work was sponsored by Google.
3972 *) Support for name constraints certificate extension. DN, email, DNS
3973 and URI types are currently supported.
3975 This work was sponsored by Google.
3978 *) To cater for systems that provide a pointer-based thread ID rather
3979 than numeric, deprecate the current numeric thread ID mechanism and
3980 replace it with a structure and associated callback type. This
3981 mechanism allows a numeric "hash" to be extracted from a thread ID in
3982 either case, and on platforms where pointers are larger than 'long',
3983 mixing is done to help ensure the numeric 'hash' is usable even if it
3984 can't be guaranteed unique. The default mechanism is to use "&errno"
3985 as a pointer-based thread ID to distinguish between threads.
3987 Applications that want to provide their own thread IDs should now use
3988 CRYPTO_THREADID_set_callback() to register a callback that will call
3989 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3991 Note that ERR_remove_state() is now deprecated, because it is tied
3992 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3993 to free the current thread's error state should be replaced by
3994 ERR_remove_thread_state(NULL).
3996 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3997 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3998 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3999 application was previously providing a numeric thread callback that
4000 was inappropriate for distinguishing threads, then uniqueness might
4001 have been obtained with &errno that happened immediately in the
4002 intermediate development versions of OpenSSL; this is no longer the
4003 case, the numeric thread callback will now override the automatic use
4005 [Geoff Thorpe, with help from Bodo Moeller]
4007 *) Initial support for different CRL issuing certificates. This covers a
4008 simple case where the self issued certificates in the chain exist and
4009 the real CRL issuer is higher in the existing chain.
4011 This work was sponsored by Google.
4014 *) Removed effectively defunct crypto/store from the build.
4017 *) Revamp of STACK to provide stronger type-checking. Still to come:
4018 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4019 ASN1_STRING, CONF_VALUE.
4022 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4023 RAM on SSL connections. This option can save about 34k per idle SSL.
4026 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4027 STACK, TXT_DB, bsearch, qsort.
4030 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4031 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4032 support for data, signedData, compressedData, digestedData and
4033 encryptedData, envelopedData types included. Scripts to check against
4034 RFC4134 examples draft and interop and consistency checks of many
4035 content types and variants.
4038 *) Add options to enc utility to support use of zlib compression BIO.
4041 *) Extend mk1mf to support importing of options and assembly language
4042 files from Configure script, currently only included in VC-WIN32.
4043 The assembly language rules can now optionally generate the source
4044 files from the associated perl scripts.
4047 *) Implement remaining functionality needed to support GOST ciphersuites.
4048 Interop testing has been performed using CryptoPro implementations.
4049 [Victor B. Wagner <vitus@cryptocom.ru>]
4051 *) s390x assembler pack.
4054 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4058 *) Implement Opaque PRF Input TLS extension as specified in
4059 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4060 official specification yet and no extension type assignment by
4061 IANA exists, this extension (for now) will have to be explicitly
4062 enabled when building OpenSSL by providing the extension number
4063 to use. For example, specify an option
4065 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4067 to the "config" or "Configure" script to enable the extension,
4068 assuming extension number 0x9527 (which is a completely arbitrary
4069 and unofficial assignment based on the MD5 hash of the Internet
4070 Draft). Note that by doing so, you potentially lose
4071 interoperability with other TLS implementations since these might
4072 be using the same extension number for other purposes.
4074 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4075 opaque PRF input value to use in the handshake. This will create
4076 an internal copy of the length-'len' string at 'src', and will
4077 return non-zero for success.
4079 To get more control and flexibility, provide a callback function
4082 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4083 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4087 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4090 Callback function 'cb' will be called in handshakes, and is
4091 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4092 Argument 'arg' is for application purposes (the value as given to
4093 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4094 be provided to the callback function). The callback function
4095 has to return non-zero to report success: usually 1 to use opaque
4096 PRF input just if possible, or 2 to enforce use of the opaque PRF
4097 input. In the latter case, the library will abort the handshake
4098 if opaque PRF input is not successfully negotiated.
4100 Arguments 'peerinput' and 'len' given to the callback function
4101 will always be NULL and 0 in the case of a client. A server will
4102 see the client's opaque PRF input through these variables if
4103 available (NULL and 0 otherwise). Note that if the server
4104 provides an opaque PRF input, the length must be the same as the
4105 length of the client's opaque PRF input.
4107 Note that the callback function will only be called when creating
4108 a new session (session resumption can resume whatever was
4109 previously negotiated), and will not be called in SSL 2.0
4110 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4111 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4112 for applications that need to enforce opaque PRF input.
4116 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4119 [Victor B. Wagner <vitus@cryptocom.ru>]
4121 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4122 RFC4507bis. The encrypted ticket format is an encrypted encoded
4123 SSL_SESSION structure, that way new session features are automatically
4126 If a client application caches session in an SSL_SESSION structure
4127 support is transparent because tickets are now stored in the encoded
4130 The SSL_CTX structure automatically generates keys for ticket
4131 protection in servers so again support should be possible
4132 with no application modification.
4134 If a client or server wishes to disable RFC4507 support then the option
4135 SSL_OP_NO_TICKET can be set.
4137 Add a TLS extension debugging callback to allow the contents of any client
4138 or server extensions to be examined.
4140 This work was sponsored by Google.
4143 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4144 OpenSSL should now compile cleanly on gcc 4.2
4145 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4147 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4148 support including streaming MAC support: this is required for GOST
4149 ciphersuite support.
4150 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4152 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4153 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4154 to output in BER and PEM format.
4157 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4158 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4159 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4160 ENGINE support for HMAC keys which are unextractable. New -mac and
4161 -macopt options to dgst utility.
4164 *) New option -sigopt to dgst utility. Update dgst to use
4165 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4166 alternative signing parameters such as X9.31 or PSS in the dgst
4170 *) Change ssl_cipher_apply_rule(), the internal function that does
4171 the work each time a ciphersuite string requests enabling
4172 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4173 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4174 the order of disabled ciphersuites such that those ciphersuites
4175 that most recently went from enabled to disabled not only stay
4176 in order with respect to each other, but also have higher priority
4177 than other disabled ciphersuites the next time ciphersuites are
4180 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4181 the same ciphersuites as with "HIGH" alone, but in a specific
4182 order where the PSK ciphersuites come first (since they are the
4183 most recently disabled ciphersuites when "HIGH" is parsed).
4185 Also, change ssl_create_cipher_list() (using this new
4186 functionality) such that between otherwise identical
4187 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4191 *) Change ssl_create_cipher_list() so that it automatically
4192 arranges the ciphersuites in reasonable order before starting
4193 to process the rule string. Thus, the definition for "DEFAULT"
4194 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4195 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4196 This makes it much easier to arrive at a reasonable default order
4197 in applications for which anonymous ciphers are OK (meaning
4198 that you can't actually use DEFAULT).
4199 [Bodo Moeller; suggested by Victor Duchovni]
4201 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4202 processing) into multiple integers instead of setting
4203 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4204 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4205 (These masks as well as the individual bit definitions are hidden
4206 away into the non-exported interface ssl/ssl_locl.h, so this
4207 change to the definition of the SSL_CIPHER structure shouldn't
4208 affect applications.) This give us more bits for each of these
4209 categories, so there is no longer a need to coagulate AES128 and
4210 AES256 into a single algorithm bit, and to coagulate Camellia128
4211 and Camellia256 into a single algorithm bit, which has led to all
4214 Thus, among other things, the kludge introduced in 0.9.7m and
4215 0.9.8e for masking out AES256 independently of AES128 or masking
4216 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4218 With the change, we also introduce new ciphersuite aliases that
4219 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4223 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4224 Use the leftmost N bytes of the signature input if the input is
4225 larger than the prime q (with N being the size in bytes of q).
4228 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4229 it yet and it is largely untested.
4232 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4235 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4236 some compilers (gcc 4.2 and later) reject their use. Safestack is
4237 reimplemented. Update ASN1 to avoid use of legacy functions.
4240 *) Win32/64 targets are linked with Winsock2.
4243 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4244 to external functions. This can be used to increase CRL handling
4245 efficiency especially when CRLs are very large by (for example) storing
4246 the CRL revoked certificates in a database.
4249 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4250 new CRLs added to a directory can be used. New command line option
4251 -verify_return_error to s_client and s_server. This causes real errors
4252 to be returned by the verify callback instead of carrying on no matter
4253 what. This reflects the way a "real world" verify callback would behave.
4256 *) GOST engine, supporting several GOST algorithms and public key formats.
4257 Kindly donated by Cryptocom.
4260 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4261 partitioned by DP are handled but no indirect CRL or reason partitioning
4262 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4263 selected via a scoring technique which handles IDP and AKID in CRLs.
4266 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4267 will ultimately be used for all verify operations: this will remove the
4268 X509_STORE dependency on certificate verification and allow alternative
4269 lookup methods. X509_STORE based implementations of these two callbacks.
4272 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4273 Modify get_crl() to find a valid (unexpired) CRL if possible.
4276 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4277 this would be called X509_CRL_cmp() but that name is already used by
4278 a function that just compares CRL issuer names. Cache several CRL
4279 extensions in X509_CRL structure and cache CRLDP in X509.
4282 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4283 this maps equivalent X509_NAME structures into a consistent structure.
4284 Name comparison can then be performed rapidly using memcmp().
4287 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4291 *) Allow digests to supply their own micalg string for S/MIME type using
4292 the ctrl EVP_MD_CTRL_MICALG.
4295 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4296 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4297 ctrl. It can then customise the structure before and/or after signing
4301 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4302 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4303 to free up any added signature OIDs.
4306 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4307 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4308 digest and cipher tables. New options added to openssl utility:
4309 list-message-digest-algorithms and list-cipher-algorithms.
4312 *) Change the array representation of binary polynomials: the list
4313 of degrees of non-zero coefficients is now terminated with -1.
4314 Previously it was terminated with 0, which was also part of the
4315 value; thus, the array representation was not applicable to
4316 polynomials where t^0 has coefficient zero. This change makes
4317 the array representation useful in a more general context.
4320 *) Various modifications and fixes to SSL/TLS cipher string
4321 handling. For ECC, the code now distinguishes between fixed ECDH
4322 with RSA certificates on the one hand and with ECDSA certificates
4323 on the other hand, since these are separate ciphersuites. The
4324 unused code for Fortezza ciphersuites has been removed.
4326 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4327 (not "ECDHE"). For consistency with the code for DH
4328 certificates, use of ECDH certificates is now considered ECDH
4329 authentication, not RSA or ECDSA authentication (the latter is
4330 merely the CA's signing algorithm and not actively used in the
4333 The temporary ciphersuite alias "ECCdraft" is no longer
4334 available, and ECC ciphersuites are no longer excluded from "ALL"
4335 and "DEFAULT". The following aliases now exist for RFC 4492
4336 ciphersuites, most of these by analogy with the DH case:
4338 kECDHr - ECDH cert, signed with RSA
4339 kECDHe - ECDH cert, signed with ECDSA
4340 kECDH - ECDH cert (signed with either RSA or ECDSA)
4341 kEECDH - ephemeral ECDH
4342 ECDH - ECDH cert or ephemeral ECDH
4348 AECDH - anonymous ECDH
4349 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4353 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4354 Use correct micalg parameters depending on digest(s) in signed message.
4357 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4358 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4361 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4362 an engine to register a method. Add ENGINE lookups for methods and
4363 functional reference processing.
4366 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4367 EVP_{Sign,Verify}* which allow an application to customise the signature
4371 *) New -resign option to smime utility. This adds one or more signers
4372 to an existing PKCS#7 signedData structure. Also -md option to use an
4373 alternative message digest algorithm for signing.
4376 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4377 create PKCS7 structures containing multiple signers. Update smime
4378 application to support multiple signers.
4381 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4385 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4386 Reorganize PBE internals to lookup from a static table using NIDs,
4387 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4388 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4389 PRF which will be automatically used with PBES2.
4392 *) Replace the algorithm specific calls to generate keys in "req" with the
4396 *) Update PKCS#7 enveloped data routines to use new API. This is now
4397 supported by any public key method supporting the encrypt operation. A
4398 ctrl is added to allow the public key algorithm to examine or modify
4399 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4403 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4404 a default digest type to use. In most cases this will be SHA1 but some
4405 algorithms (such as GOST) need to specify an alternative digest. The
4406 return value indicates how strong the preference is 1 means optional and
4407 2 is mandatory (that is it is the only supported type). Modify
4408 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4409 use the default md. Update openssl utilities to use the default digest
4410 type for signing if it is not explicitly indicated.
4413 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4414 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4415 signing method from the key type. This effectively removes the link
4416 between digests and public key types.
4419 *) Add an OID cross reference table and utility functions. Its purpose is to
4420 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4421 rsaEncryption. This will allow some of the algorithm specific hackery
4422 needed to use the correct OID to be removed.
4425 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4426 structures for PKCS7_sign(). They are now set up by the relevant public
4430 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4433 *) Add support for key derivation (agreement) in the API, DH method and
4437 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4438 public and private key formats. As a side effect these add additional
4439 command line functionality not previously available: DSA signatures can be
4440 generated and verified using pkeyutl and DH key support and generation in
4445 [Oliver Tappe <zooey@hirschkaefer.de>]
4447 *) New make target "install_html_docs" installs HTML renditions of the
4449 [Oliver Tappe <zooey@hirschkaefer.de>]
4451 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4452 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4453 support key and parameter generation and add initial key generation
4454 functionality for RSA.
4457 *) Add functions for main EVP_PKEY_method operations. The undocumented
4458 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4459 EVP_PKEY_{encrypt,decrypt}_old.
4462 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4463 key API, doesn't do much yet.
4466 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4467 public key algorithms. New option to openssl utility:
4468 "list-public-key-algorithms" to print out info.
4471 *) Implement the Supported Elliptic Curves Extension for
4472 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4475 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4476 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4479 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4480 utilities such as rsa, dsa, dsaparam etc except they process any key
4484 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4485 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4486 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4490 *) Initial support for pluggable public key ASN1.
4491 De-spaghettify the public key ASN1 handling. Move public and private
4492 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4493 algorithm specific handling to a single module within the relevant
4494 algorithm directory. Add functions to allow (near) opaque processing
4495 of public and private key structures.
4498 *) Implement the Supported Point Formats Extension for
4499 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4502 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4503 for the psk identity [hint] and the psk callback functions to the
4504 SSL_SESSION, SSL and SSL_CTX structure.
4507 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4511 SSL_CTX_use_psk_identity_hint
4512 SSL_get_psk_identity_hint
4513 SSL_get_psk_identity
4514 SSL_use_psk_identity_hint
4516 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4518 *) Add RFC 3161 compliant time stamp request creation, response generation
4519 and response verification functionality.
4520 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4522 *) Add initial support for TLS extensions, specifically for the server_name
4523 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4524 have new members for a host name. The SSL data structure has an
4525 additional member SSL_CTX *initial_ctx so that new sessions can be
4526 stored in that context to allow for session resumption, even after the
4527 SSL has been switched to a new SSL_CTX in reaction to a client's
4528 server_name extension.
4530 New functions (subject to change):
4532 SSL_get_servername()
4533 SSL_get_servername_type()
4536 New CTRL codes and macros (subject to change):
4538 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4539 - SSL_CTX_set_tlsext_servername_callback()
4540 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4541 - SSL_CTX_set_tlsext_servername_arg()
4542 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4544 openssl s_client has a new '-servername ...' option.
4546 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4547 '-key2 ...', '-servername_fatal' (subject to change). This allows
4548 testing the HostName extension for a specific single host name ('-cert'
4549 and '-key' remain fallbacks for handshakes without HostName
4550 negotiation). If the unrecognized_name alert has to be sent, this by
4551 default is a warning; it becomes fatal with the '-servername_fatal'
4554 [Peter Sylvester, Remy Allais, Christophe Renou]
4556 *) Whirlpool hash implementation is added.
4559 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4560 bn(64,32). Because of instruction set limitations it doesn't have
4561 any negative impact on performance. This was done mostly in order
4562 to make it possible to share assembler modules, such as bn_mul_mont
4563 implementations, between 32- and 64-bit builds without hassle.
4566 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4567 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4571 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4572 dedicated Montgomery multiplication procedure, is introduced.
4573 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4574 "64-bit" performance on certain 32-bit targets.
4577 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4578 in SSL structures. New SSL ctrl to set maximum send fragment size.
4579 Save memory by setting the I/O buffer sizes dynamically instead of
4580 using the maximum available value.
4583 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4584 in addition to the text details.
4587 *) Very, very preliminary EXPERIMENTAL support for printing of general
4588 ASN1 structures. This currently produces rather ugly output and doesn't
4589 handle several customised structures at all.
4592 *) Integrated support for PVK file format and some related formats such
4593 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4594 these in the 'rsa' and 'dsa' utilities.
4597 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4600 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4601 place for the (very old) "NETSCAPE" format certificates which are now
4602 handled using new ASN1 code equivalents.
4605 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4606 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4607 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4610 *) Modify CRL distribution points extension code to print out previously
4611 unsupported fields. Enhance extension setting code to allow setting of
4615 *) Add print and set support for Issuing Distribution Point CRL extension.
4618 *) Change 'Configure' script to enable Camellia by default.
4621 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4623 *) When rejecting SSL/TLS records due to an incorrect version number, never
4624 update s->server with a new major version number. As of
4625 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4626 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4627 the previous behavior could result in a read attempt at NULL when
4628 receiving specific incorrect SSL/TLS records once record payload
4629 protection is active. (CVE-2010-0740)
4630 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4632 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4633 could be crashed if the relevant tables were not present (e.g. chrooted).
4634 [Tomas Hoger <thoger@redhat.com>]
4636 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4638 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4639 [Martin Olsson, Neel Mehta]
4641 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4642 accommodate for stack sorting, always a write lock!).
4645 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4646 excessive delays in the RAND_poll(): over a minute. As a workaround
4647 include a time check in the inner Heap32Next loop too.
4650 *) The code that handled flushing of data in SSL/TLS originally used the
4651 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4652 the problem outlined in PR#1949. The fix suggested there however can
4653 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4654 of Apache). So instead simplify the code to flush unconditionally.
4655 This should be fine since flushing with no data to flush is a no op.
4658 *) Handle TLS versions 2.0 and later properly and correctly use the
4659 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4660 off ancient servers have a habit of sticking around for a while...
4663 *) Modify compression code so it frees up structures without using the
4664 ex_data callbacks. This works around a problem where some applications
4665 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4666 restarting) then use compression (e.g. SSL with compression) later.
4667 This results in significant per-connection memory leaks and
4668 has caused some security issues including CVE-2008-1678 and
4672 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4673 change when encrypting or decrypting.
4676 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4677 connect and renegotiate with servers which do not support RI.
4678 Until RI is more widely deployed this option is enabled by default.
4681 *) Add "missing" ssl ctrls to clear options and mode.
4684 *) If client attempts to renegotiate and doesn't support RI respond with
4685 a no_renegotiation alert as required by RFC5746. Some renegotiating
4686 TLS clients will continue a connection gracefully when they receive
4687 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4688 waiting for a server hello which it will never receive. Now we treat a
4689 received no_renegotiation alert as a fatal error. This is because
4690 applications requesting a renegotiation might well expect it to succeed
4691 and would have no code in place to handle the server denying it so the
4692 only safe thing to do is to terminate the connection.
4695 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4696 peer supports secure renegotiation and 0 otherwise. Print out peer
4697 renegotiation support in s_client/s_server.
4700 *) Replace the highly broken and deprecated SPKAC certification method with
4701 the updated NID creation version. This should correctly handle UTF8.
4704 *) Implement RFC5746. Re-enable renegotiation but require the extension
4705 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4706 turns out to be a bad idea. It has been replaced by
4707 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4708 SSL_CTX_set_options(). This is really not recommended unless you
4709 know what you are doing.
4710 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4712 *) Fixes to stateless session resumption handling. Use initial_ctx when
4713 issuing and attempting to decrypt tickets in case it has changed during
4714 servername handling. Use a non-zero length session ID when attempting
4715 stateless session resumption: this makes it possible to determine if
4716 a resumption has occurred immediately after receiving server hello
4717 (several places in OpenSSL subtly assume this) instead of later in
4721 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4722 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4723 fixes for a few places where the return code is not checked
4725 [Julia Lawall <julia@diku.dk>]
4727 *) Add --strict-warnings option to Configure script to include devteam
4728 warnings in other configurations.
4731 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4732 makes it possible to install openssl libraries in locations which
4733 have names other than "lib", for example "/usr/lib64" which some
4735 [Steve Henson, based on patch from Jeremy Utley]
4737 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4738 X690 8.9.12 and can produce some misleading textual output of OIDs.
4739 [Steve Henson, reported by Dan Kaminsky]
4741 *) Delete MD2 from algorithm tables. This follows the recommendation in
4742 several standards that it is not used in new applications due to
4743 several cryptographic weaknesses. For binary compatibility reasons
4744 the MD2 API is still compiled in by default.
4747 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4751 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4752 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4754 [Guenter <lists@gknw.net>]
4756 *) Fix the server certificate chain building code to use X509_verify_cert(),
4757 it used to have an ad-hoc builder which was unable to cope with anything
4758 other than a simple chain.
4759 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4761 *) Don't check self signed certificate signatures in X509_verify_cert()
4762 by default (a flag can override this): it just wastes time without
4763 adding any security. As a useful side effect self signed root CAs
4764 with non-FIPS digests are now usable in FIPS mode.
4767 *) In dtls1_process_out_of_seq_message() the check if the current message
4768 is already buffered was missing. For every new message was memory
4769 allocated, allowing an attacker to perform an denial of service attack
4770 with sending out of seq handshake messages until there is no memory
4771 left. Additionally every future message was buffered, even if the
4772 sequence number made no sense and would be part of another handshake.
4773 So only messages with sequence numbers less than 10 in advance will be
4774 buffered. (CVE-2009-1378)
4775 [Robin Seggelmann, discovered by Daniel Mentz]
4777 *) Records are buffered if they arrive with a future epoch to be
4778 processed after finishing the corresponding handshake. There is
4779 currently no limitation to this buffer allowing an attacker to perform
4780 a DOS attack with sending records with future epochs until there is no
4781 memory left. This patch adds the pqueue_size() function to determine
4782 the size of a buffer and limits the record buffer to 100 entries.
4784 [Robin Seggelmann, discovered by Daniel Mentz]
4786 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4787 parent structure is freed. (CVE-2009-1379)
4790 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4791 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4794 [Ilya O. <vrghost@gmail.com>]
4796 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4798 *) Disable renegotiation completely - this fixes a severe security
4799 problem (CVE-2009-3555) at the cost of breaking all
4800 renegotiation. Renegotiation can be re-enabled by setting
4801 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4802 run-time. This is really not recommended unless you know what
4806 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4808 *) Don't set val to NULL when freeing up structures, it is freed up by
4809 underlying code. If sizeof(void *) > sizeof(long) this can result in
4810 zeroing past the valid field. (CVE-2009-0789)
4811 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4813 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4814 checked correctly. This would allow some invalid signed attributes to
4815 appear to verify correctly. (CVE-2009-0591)
4816 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4818 *) Reject UniversalString and BMPString types with invalid lengths. This
4819 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4820 a legal length. (CVE-2009-0590)
4823 *) Set S/MIME signing as the default purpose rather than setting it
4824 unconditionally. This allows applications to override it at the store
4828 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4829 to handle some structures.
4832 *) Improve efficiency of mem_gets: don't search whole buffer each time
4834 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4836 *) New -hex option for openssl rand.
4839 *) Print out UTF8String and NumericString when parsing ASN1.
4842 *) Support NumericString type for name components.
4845 *) Allow CC in the environment to override the automatically chosen
4846 compiler. Note that nothing is done to ensure flags work with the
4850 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4852 *) Properly check EVP_VerifyFinal() and similar return values
4854 [Ben Laurie, Bodo Moeller, Google Security Team]
4856 *) Enable TLS extensions by default.
4859 *) Allow the CHIL engine to be loaded, whether the application is
4860 multithreaded or not. (This does not release the developer from the
4861 obligation to set up the dynamic locking callbacks.)
4862 [Sander Temme <sander@temme.net>]
4864 *) Use correct exit code if there is an error in dgst command.
4865 [Steve Henson; problem pointed out by Roland Dirlewanger]
4867 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4868 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4871 *) Add experimental JPAKE support, including demo authentication in
4872 s_client and s_server.
4875 *) Set the comparison function in v3_addr_canonize().
4876 [Rob Austein <sra@hactrn.net>]
4878 *) Add support for XMPP STARTTLS in s_client.
4879 [Philip Paeps <philip@freebsd.org>]
4881 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4882 to ensure that even with this option, only ciphersuites in the
4883 server's preference list will be accepted. (Note that the option
4884 applies only when resuming a session, so the earlier behavior was
4885 just about the algorithm choice for symmetric cryptography.)
4888 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4890 *) Fix NULL pointer dereference if a DTLS server received
4891 ChangeCipherSpec as first record (CVE-2009-1386).
4894 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4895 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4898 *) The fix in 0.9.8c that supposedly got rid of unsafe
4899 double-checked locking was incomplete for RSA blinding,
4900 addressing just one layer of what turns out to have been
4901 doubly unsafe triple-checked locking.
4903 So now fix this for real by retiring the MONT_HELPER macro
4904 in crypto/rsa/rsa_eay.c.
4906 [Bodo Moeller; problem pointed out by Marius Schilder]
4908 *) Various precautionary measures:
4910 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4912 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4913 (NB: This would require knowledge of the secret session ticket key
4914 to exploit, in which case you'd be SOL either way.)
4916 - Change bn_nist.c so that it will properly handle input BIGNUMs
4917 outside the expected range.
4919 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4922 [Neel Mehta, Bodo Moeller]
4924 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4925 the load fails. Useful for distros.
4926 [Ben Laurie and the FreeBSD team]
4928 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4931 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4934 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4936 This work was sponsored by Logica.
4939 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4940 keystores. Support for SSL/TLS client authentication too.
4941 Not compiled unless enable-capieng specified to Configure.
4943 This work was sponsored by Logica.
4946 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4947 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4948 attribute creation routines such as certificate requests and PKCS#12
4952 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4954 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4955 handshake which could lead to a client crash as found using the
4956 Codenomicon TLS test suite (CVE-2008-1672)
4957 [Steve Henson, Mark Cox]
4959 *) Fix double free in TLS server name extensions which could lead to
4960 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4963 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4965 Clear the error queue to ensure that error entries left from
4966 older function calls do not interfere with the correct operation.
4967 [Lutz Jaenicke, Erik de Castro Lopo]
4969 *) Remove root CA certificates of commercial CAs:
4971 The OpenSSL project does not recommend any specific CA and does not
4972 have any policy with respect to including or excluding any CA.
4973 Therefore it does not make any sense to ship an arbitrary selection
4974 of root CA certificates with the OpenSSL software.
4977 *) RSA OAEP patches to fix two separate invalid memory reads.
4978 The first one involves inputs when 'lzero' is greater than
4979 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4980 before the beginning of from). The second one involves inputs where
4981 the 'db' section contains nothing but zeroes (there is a one-byte
4982 invalid read after the end of 'db').
4983 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4985 *) Partial backport from 0.9.9-dev:
4987 Introduce bn_mul_mont (dedicated Montgomery multiplication
4988 procedure) as a candidate for BIGNUM assembler implementation.
4989 While 0.9.9-dev uses assembler for various architectures, only
4990 x86_64 is available by default here in the 0.9.8 branch, and
4991 32-bit x86 is available through a compile-time setting.
4993 To try the 32-bit x86 assembler implementation, use Configure
4994 option "enable-montasm" (which exists only for this backport).
4996 As "enable-montasm" for 32-bit x86 disclaims code stability
4997 anyway, in this constellation we activate additional code
4998 backported from 0.9.9-dev for further performance improvements,
4999 namely BN_from_montgomery_word. (To enable this otherwise,
5000 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5002 [Andy Polyakov (backport partially by Bodo Moeller)]
5004 *) Add TLS session ticket callback. This allows an application to set
5005 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5006 values. This is useful for key rollover for example where several key
5007 sets may exist with different names.
5010 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5011 This was broken until now in 0.9.8 releases, such that the only way
5012 a registered ENGINE could be used (assuming it initialises
5013 successfully on the host) was to explicitly set it as the default
5014 for the relevant algorithms. This is in contradiction with 0.9.7
5015 behaviour and the documentation. With this fix, when an ENGINE is
5016 registered into a given algorithm's table of implementations, the
5017 'uptodate' flag is reset so that auto-discovery will be used next
5018 time a new context for that algorithm attempts to select an
5020 [Ian Lister (tweaked by Geoff Thorpe)]
5022 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5023 implementation in the following ways:
5025 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5028 Lack of BER streaming support means one pass streaming processing is
5029 only supported if data is detached: setting the streaming flag is
5030 ignored for embedded content.
5032 CMS support is disabled by default and must be explicitly enabled
5033 with the enable-cms configuration option.
5036 *) Update the GMP engine glue to do direct copies between BIGNUM and
5037 mpz_t when openssl and GMP use the same limb size. Otherwise the
5038 existing "conversion via a text string export" trick is still used.
5039 [Paul Sheer <paulsheer@gmail.com>]
5041 *) Zlib compression BIO. This is a filter BIO which compressed and
5042 uncompresses any data passed through it.
5045 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5046 RFC3394 compatible AES key wrapping.
5049 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5050 sets string data without copying. X509_ALGOR_set0() and
5051 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5052 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5053 from an X509_ATTRIBUTE structure optionally checking it occurs only
5054 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5058 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5059 to get the expected BN_FLG_CONSTTIME behavior.
5060 [Bodo Moeller (Google)]
5064 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5065 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5066 - added some more tests to do_tests.pl
5067 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5068 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5069 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5070 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5071 - various changes to netware.pl to enable gcc-cross builds on Win32
5073 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5074 - various changes to fix missing prototype warnings
5075 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5076 - added AES, WHIRLPOOL and CPUID assembler code to build files
5077 - added missing AES assembler make rules to mk1mf.pl
5078 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5079 [Guenter Knauf <eflash@gmx.net>]
5081 *) Implement certificate status request TLS extension defined in RFC3546.
5082 A client can set the appropriate parameters and receive the encoded
5083 OCSP response via a callback. A server can query the supplied parameters
5084 and set the encoded OCSP response in the callback. Add simplified examples
5085 to s_client and s_server.
5088 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5090 *) Fix various bugs:
5091 + Binary incompatibility of ssl_ctx_st structure
5092 + DTLS interoperation with non-compliant servers
5093 + Don't call get_session_cb() without proposed session
5094 + Fix ia64 assembler code
5095 [Andy Polyakov, Steve Henson]
5097 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5099 *) DTLS Handshake overhaul. There were longstanding issues with
5100 OpenSSL DTLS implementation, which were making it impossible for
5101 RFC 4347 compliant client to communicate with OpenSSL server.
5102 Unfortunately just fixing these incompatibilities would "cut off"
5103 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5104 server keeps tolerating non RFC compliant syntax. The opposite is
5105 not true, 0.9.8f client can not communicate with earlier server.
5106 This update even addresses CVE-2007-4995.
5109 *) Changes to avoid need for function casts in OpenSSL: some compilers
5110 (gcc 4.2 and later) reject their use.
5111 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5114 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5115 RFC4507bis. The encrypted ticket format is an encrypted encoded
5116 SSL_SESSION structure, that way new session features are automatically
5119 If a client application caches session in an SSL_SESSION structure
5120 support is transparent because tickets are now stored in the encoded
5123 The SSL_CTX structure automatically generates keys for ticket
5124 protection in servers so again support should be possible
5125 with no application modification.
5127 If a client or server wishes to disable RFC4507 support then the option
5128 SSL_OP_NO_TICKET can be set.
5130 Add a TLS extension debugging callback to allow the contents of any client
5131 or server extensions to be examined.
5133 This work was sponsored by Google.
5136 *) Add initial support for TLS extensions, specifically for the server_name
5137 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5138 have new members for a host name. The SSL data structure has an
5139 additional member SSL_CTX *initial_ctx so that new sessions can be
5140 stored in that context to allow for session resumption, even after the
5141 SSL has been switched to a new SSL_CTX in reaction to a client's
5142 server_name extension.
5144 New functions (subject to change):
5146 SSL_get_servername()
5147 SSL_get_servername_type()
5150 New CTRL codes and macros (subject to change):
5152 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5153 - SSL_CTX_set_tlsext_servername_callback()
5154 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5155 - SSL_CTX_set_tlsext_servername_arg()
5156 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5158 openssl s_client has a new '-servername ...' option.
5160 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5161 '-key2 ...', '-servername_fatal' (subject to change). This allows
5162 testing the HostName extension for a specific single host name ('-cert'
5163 and '-key' remain fallbacks for handshakes without HostName
5164 negotiation). If the unrecognized_name alert has to be sent, this by
5165 default is a warning; it becomes fatal with the '-servername_fatal'
5168 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5170 *) Add AES and SSE2 assembly language support to VC++ build.
5173 *) Mitigate attack on final subtraction in Montgomery reduction.
5176 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5177 (which previously caused an internal error).
5180 *) Squeeze another 10% out of IGE mode when in != out.
5183 *) AES IGE mode speedup.
5184 [Dean Gaudet (Google)]
5186 *) Add the Korean symmetric 128-bit cipher SEED (see
5187 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5188 add SEED ciphersuites from RFC 4162:
5190 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5191 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5192 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5193 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5195 To minimize changes between patchlevels in the OpenSSL 0.9.8
5196 series, SEED remains excluded from compilation unless OpenSSL
5197 is configured with 'enable-seed'.
5198 [KISA, Bodo Moeller]
5200 *) Mitigate branch prediction attacks, which can be practical if a
5201 single processor is shared, allowing a spy process to extract
5202 information. For detailed background information, see
5203 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5204 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5205 and Necessary Software Countermeasures"). The core of the change
5206 are new versions BN_div_no_branch() and
5207 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5208 respectively, which are slower, but avoid the security-relevant
5209 conditional branches. These are automatically called by BN_div()
5210 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5211 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5212 remove a conditional branch.
5214 BN_FLG_CONSTTIME is the new name for the previous
5215 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5216 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5217 in the exponent causes BN_mod_exp_mont() to use the alternative
5218 implementation in BN_mod_exp_mont_consttime().) The old name
5219 remains as a deprecated alias.
5221 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5222 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5223 constant-time implementations for more than just exponentiation.
5224 Here too the old name is kept as a deprecated alias.
5226 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5227 the BN_BLINDING structure gets an independent copy of the
5228 modulus. This means that the previous "BIGNUM *m" argument to
5229 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5230 essentially becomes "const BIGNUM *m", although we can't actually
5231 change this in the header file before 0.9.9. It allows
5232 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5233 enable BN_FLG_CONSTTIME.
5235 [Matthew D Wood (Intel Corp)]
5237 *) In the SSL/TLS server implementation, be strict about session ID
5238 context matching (which matters if an application uses a single
5239 external cache for different purposes). Previously,
5240 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5241 set. This did ensure strict client verification, but meant that,
5242 with applications using a single external cache for quite
5243 different requirements, clients could circumvent ciphersuite
5244 restrictions for a given session ID context by starting a session
5245 in a different context.
5248 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5249 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5250 authentication-only ciphersuites.
5253 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5254 not complete and could lead to a possible single byte overflow
5255 (CVE-2007-5135) [Ben Laurie]
5257 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5259 *) Since AES128 and AES256 (and similarly Camellia128 and
5260 Camellia256) share a single mask bit in the logic of
5261 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5262 kludge to work properly if AES128 is available and AES256 isn't
5263 (or if Camellia128 is available and Camellia256 isn't).
5266 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5267 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5268 When a point or a seed is encoded in a BIT STRING, we need to
5269 prevent the removal of trailing zero bits to get the proper DER
5270 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5271 of a NamedBitList, for which trailing 0 bits need to be removed.)
5274 *) Have SSL/TLS server implementation tolerate "mismatched" record
5275 protocol version while receiving ClientHello even if the
5276 ClientHello is fragmented. (The server can't insist on the
5277 particular protocol version it has chosen before the ServerHello
5278 message has informed the client about his choice.)
5281 *) Add RFC 3779 support.
5282 [Rob Austein for ARIN, Ben Laurie]
5284 *) Load error codes if they are not already present instead of using a
5285 static variable. This allows them to be cleanly unloaded and reloaded.
5286 Improve header file function name parsing.
5289 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5290 or CAPABILITY handshake as required by RFCs.
5293 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5295 *) Introduce limits to prevent malicious keys being able to
5296 cause a denial of service. (CVE-2006-2940)
5297 [Steve Henson, Bodo Moeller]
5299 *) Fix ASN.1 parsing of certain invalid structures that can result
5300 in a denial of service. (CVE-2006-2937) [Steve Henson]
5302 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5303 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5305 *) Fix SSL client code which could crash if connecting to a
5306 malicious SSLv2 server. (CVE-2006-4343)
5307 [Tavis Ormandy and Will Drewry, Google Security Team]
5309 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5310 match only those. Before that, "AES256-SHA" would be interpreted
5311 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5312 the same strength classification in 0.9.7h) as we currently only
5313 have a single AES bit in the ciphersuite description bitmap.
5314 That change, however, also applied to ciphersuite strings such as
5315 "RC4-MD5" that intentionally matched multiple ciphersuites --
5316 namely, SSL 2.0 ciphersuites in addition to the more common ones
5317 from SSL 3.0/TLS 1.0.
5319 So we change the selection algorithm again: Naming an explicit
5320 ciphersuite selects this one ciphersuite, and any other similar
5321 ciphersuite (same bitmap) from *other* protocol versions.
5322 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5323 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5325 Since SSL 2.0 does not have any ciphersuites for which the
5326 128/256 bit distinction would be relevant, this works for now.
5327 The proper fix will be to use different bits for AES128 and
5328 AES256, which would have avoided the problems from the beginning;
5329 however, bits are scarce, so we can only do this in a new release
5330 (not just a patchlevel) when we can change the SSL_CIPHER
5331 definition to split the single 'unsigned long mask' bitmap into
5332 multiple values to extend the available space.
5336 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5338 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5339 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5341 *) Add AES IGE and biIGE modes.
5344 *) Change the Unix randomness entropy gathering to use poll() when
5345 possible instead of select(), since the latter has some
5346 undesirable limitations.
5347 [Darryl Miles via Richard Levitte and Bodo Moeller]
5349 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5350 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5351 cannot be implicitly activated as part of, e.g., the "AES" alias.
5352 However, please upgrade to OpenSSL 0.9.9[-dev] for
5353 non-experimental use of the ECC ciphersuites to get TLS extension
5354 support, which is required for curve and point format negotiation
5355 to avoid potential handshake problems.
5358 *) Disable rogue ciphersuites:
5360 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5361 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5362 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5364 The latter two were purportedly from
5365 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5368 Also deactivate the remaining ciphersuites from
5369 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5370 unofficial, and the ID has long expired.
5373 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5374 dual-core machines) and other potential thread-safety issues.
5377 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5378 versions), which is now available for royalty-free use
5379 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5380 Also, add Camellia TLS ciphersuites from RFC 4132.
5382 To minimize changes between patchlevels in the OpenSSL 0.9.8
5383 series, Camellia remains excluded from compilation unless OpenSSL
5384 is configured with 'enable-camellia'.
5387 *) Disable the padding bug check when compression is in use. The padding
5388 bug check assumes the first packet is of even length, this is not
5389 necessarily true if compression is enabled and can result in false
5390 positives causing handshake failure. The actual bug test is ancient
5391 code so it is hoped that implementations will either have fixed it by
5392 now or any which still have the bug do not support compression.
5395 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5397 *) When applying a cipher rule check to see if string match is an explicit
5398 cipher suite and only match that one cipher suite if it is.
5401 *) Link in manifests for VC++ if needed.
5402 [Austin Ziegler <halostatue@gmail.com>]
5404 *) Update support for ECC-based TLS ciphersuites according to
5405 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5406 TLS extensions, which are supported starting with the 0.9.9
5407 branch, not in the OpenSSL 0.9.8 branch).
5410 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5411 opaque EVP_CIPHER_CTX handling.
5414 *) Fixes and enhancements to zlib compression code. We now only use
5415 "zlib1.dll" and use the default __cdecl calling convention on Win32
5416 to conform with the standards mentioned here:
5417 http://www.zlib.net/DLL_FAQ.txt
5418 Static zlib linking now works on Windows and the new --with-zlib-include
5419 --with-zlib-lib options to Configure can be used to supply the location
5420 of the headers and library. Gracefully handle case where zlib library
5424 *) Several fixes and enhancements to the OID generation code. The old code
5425 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5426 handle numbers larger than ULONG_MAX, truncated printing and had a
5427 non standard OBJ_obj2txt() behaviour.
5430 *) Add support for building of engines under engine/ as shared libraries
5431 under VC++ build system.
5434 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5435 Hopefully, we will not see any false combination of paths any more.
5438 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5440 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5441 (part of SSL_OP_ALL). This option used to disable the
5442 countermeasure against man-in-the-middle protocol-version
5443 rollback in the SSL 2.0 server implementation, which is a bad
5444 idea. (CVE-2005-2969)
5446 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5447 for Information Security, National Institute of Advanced Industrial
5448 Science and Technology [AIST], Japan)]
5450 *) Add two function to clear and return the verify parameter flags.
5453 *) Keep cipherlists sorted in the source instead of sorting them at
5454 runtime, thus removing the need for a lock.
5457 *) Avoid some small subgroup attacks in Diffie-Hellman.
5458 [Nick Mathewson and Ben Laurie]
5460 *) Add functions for well-known primes.
5463 *) Extended Windows CE support.
5464 [Satoshi Nakamura and Andy Polyakov]
5466 *) Initialize SSL_METHOD structures at compile time instead of during
5467 runtime, thus removing the need for a lock.
5470 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5471 attempting to decrypt each encrypted key in turn. Add support to
5475 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5477 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5480 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5483 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5484 key into the same file any more.
5487 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5490 *) Add -utf8 command line and config file option to 'ca'.
5491 [Stefan <stf@udoma.org]
5493 *) Removed the macro des_crypt(), as it seems to conflict with some
5494 libraries. Use DES_crypt().
5497 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5498 involves renaming the source and generated shared-libs for
5499 both. The engines will accept the corrected or legacy ids
5500 ('ncipher' and '4758_cca' respectively) when binding. NB,
5501 this only applies when building 'shared'.
5502 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5504 *) Add attribute functions to EVP_PKEY structure. Modify
5505 PKCS12_create() to recognize a CSP name attribute and
5506 use it. Make -CSP option work again in pkcs12 utility.
5509 *) Add new functionality to the bn blinding code:
5510 - automatic re-creation of the BN_BLINDING parameters after
5511 a fixed number of uses (currently 32)
5512 - add new function for parameter creation
5513 - introduce flags to control the update behaviour of the
5514 BN_BLINDING parameters
5515 - hide BN_BLINDING structure
5516 Add a second BN_BLINDING slot to the RSA structure to improve
5517 performance when a single RSA object is shared among several
5521 *) Add support for DTLS.
5522 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5524 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5525 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5528 *) Remove buggy and incomplete DH cert support from
5529 ssl/ssl_rsa.c and ssl/s3_both.c
5532 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5533 the apps/openssl applications.
5536 *) Compile clean with "-Wall -Wmissing-prototypes
5537 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5538 DEBUG_SAFESTACK must also be set.
5541 *) Change ./Configure so that certain algorithms can be disabled by default.
5542 The new counterpiece to "no-xxx" is "enable-xxx".
5544 The patented RC5 and MDC2 algorithms will now be disabled unless
5545 "enable-rc5" and "enable-mdc2", respectively, are specified.
5547 (IDEA remains enabled despite being patented. This is because IDEA
5548 is frequently required for interoperability, and there is no license
5549 fee for non-commercial use. As before, "no-idea" can be used to
5550 avoid this algorithm.)
5554 *) Add processing of proxy certificates (see RFC 3820). This work was
5555 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5556 EGEE (Enabling Grids for E-science in Europe).
5559 *) RC4 performance overhaul on modern architectures/implementations, such
5560 as Intel P4, IA-64 and AMD64.
5563 *) New utility extract-section.pl. This can be used specify an alternative
5564 section number in a pod file instead of having to treat each file as
5565 a separate case in Makefile. This can be done by adding two lines to the
5568 =for comment openssl_section:XXX
5570 The blank line is mandatory.
5574 *) New arguments -certform, -keyform and -pass for s_client and s_server
5575 to allow alternative format key and certificate files and passphrase
5579 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5580 update associated structures and add various utility functions.
5582 Add new policy related verify parameters, include policy checking in
5583 standard verify code. Enhance 'smime' application with extra parameters
5584 to support policy checking and print out.
5587 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5588 Nehemiah processors. These extensions support AES encryption in hardware
5589 as well as RNG (though RNG support is currently disabled).
5590 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5592 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5595 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5596 [Andy Polyakov and a number of other people]
5598 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5599 implementation contributed by IBM.
5600 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5602 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5603 exponent rather than 'unsigned long'. There is a corresponding change to
5604 the new 'rsa_keygen' element of the RSA_METHOD structure.
5605 [Jelte Jansen, Geoff Thorpe]
5607 *) Functionality for creating the initial serial number file is now
5608 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5610 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5611 number file to 1, which is bound to cause problems. To avoid
5612 the problems while respecting compatibility between different 0.9.7
5613 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5614 CA.pl for serial number initialization. With the new release 0.9.8,
5615 we can fix the problem directly in the 'ca' utility.)
5618 *) Reduced header interdependencies by declaring more opaque objects in
5619 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5620 give fewer recursive includes, which could break lazy source code - so
5621 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5622 developers should define this symbol when building and using openssl to
5623 ensure they track the recommended behaviour, interfaces, [etc], but
5624 backwards-compatible behaviour prevails when this isn't defined.
5627 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5630 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5631 This will generate a random key of the appropriate length based on the
5632 cipher context. The EVP_CIPHER can provide its own random key generation
5633 routine to support keys of a specific form. This is used in the des and
5634 3des routines to generate a key of the correct parity. Update S/MIME
5635 code to use new functions and hence generate correct parity DES keys.
5636 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5637 valid (weak or incorrect parity).
5640 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5641 as looking them up. This is useful when the verified structure may contain
5642 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5643 present unless the new PKCS7_NO_CRL flag is asserted.
5646 *) Extend ASN1 oid configuration module. It now additionally accepts the
5649 shortName = some long name, 1.2.3.4
5652 *) Reimplemented the BN_CTX implementation. There is now no more static
5653 limitation on the number of variables it can handle nor the depth of the
5654 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5655 information can now expand as required, and rather than having a single
5656 static array of bignums, BN_CTX now uses a linked-list of such arrays
5657 allowing it to expand on demand whilst maintaining the usefulness of
5658 BN_CTX's "bundling".
5661 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5662 to allow all RSA operations to function using a single BN_CTX.
5665 *) Preliminary support for certificate policy evaluation and checking. This
5666 is initially intended to pass the tests outlined in "Conformance Testing
5667 of Relying Party Client Certificate Path Processing Logic" v1.07.
5670 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5671 remained unused and not that useful. A variety of other little bignum
5672 tweaks and fixes have also been made continuing on from the audit (see
5676 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5677 associated ASN1, EVP and SSL functions and old ASN1 macros.
5680 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5681 and this should never fail. So the return value from the use of
5682 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5683 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5686 *) BN_CTX_get() should return zero-valued bignums, providing the same
5687 initialised value as BN_new().
5688 [Geoff Thorpe, suggested by Ulf Möller]
5690 *) Support for inhibitAnyPolicy certificate extension.
5693 *) An audit of the BIGNUM code is underway, for which debugging code is
5694 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5695 is considered valid when processing BIGNUMs, and causes execution to
5696 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5697 further steps are taken to deliberately pollute unused data in BIGNUM
5698 structures to try and expose faulty code further on. For now, openssl will
5699 (in its default mode of operation) continue to tolerate the inconsistent
5700 forms that it has tolerated in the past, but authors and packagers should
5701 consider trying openssl and their own applications when compiled with
5702 these debugging symbols defined. It will help highlight potential bugs in
5703 their own code, and will improve the test coverage for OpenSSL itself. At
5704 some point, these tighter rules will become openssl's default to improve
5705 maintainability, though the assert()s and other overheads will remain only
5706 in debugging configurations. See bn.h for more details.
5707 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5709 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5710 that can only be obtained through BN_CTX_new() (which implicitly
5711 initialises it). The presence of this function only made it possible
5712 to overwrite an existing structure (and cause memory leaks).
5715 *) Because of the callback-based approach for implementing LHASH as a
5716 template type, lh_insert() adds opaque objects to hash-tables and
5717 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5718 to clean up those corresponding objects before destroying the hash table
5719 (and losing the object pointers). So some over-zealous constifications in
5720 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5721 objects as "const" and the lh_doall[_arg] callback wrappers are not
5722 prototyped to have "const" restrictions on the object pointers they are
5723 given (and so aren't required to cast them away any more).
5726 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5727 (speed) prefers to use its own implementation. The two implementations
5728 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5729 its object type properly exposed (MS_TM) instead of casting to/from "char
5730 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5731 aren't necessarily the greatest nomenclatures - but this is what was used
5732 internally to the implementation so I've used that for now.
5735 *) Ensure that deprecated functions do not get compiled when
5736 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5737 the self-tests were still using deprecated key-generation functions so
5738 these have been updated also.
5741 *) Reorganise PKCS#7 code to separate the digest location functionality
5742 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5743 New function PKCS7_set_digest() to set the digest type for PKCS#7
5744 digestedData type. Add additional code to correctly generate the
5745 digestedData type and add support for this type in PKCS7 initialization
5749 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5750 structure of type "other".
5753 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5754 sure the loop does correctly stop and breaking ("division by zero")
5755 modulus operations are not performed. The (pre-generated) prime
5756 table crypto/bn/bn_prime.h was already correct, but it could not be
5757 re-generated on some platforms because of the "division by zero"
5758 situation in the script.
5759 [Ralf S. Engelschall]
5761 *) Update support for ECC-based TLS ciphersuites according to
5762 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5763 SHA-1 now is only used for "small" curves (where the
5764 representation of a field element takes up to 24 bytes); for
5765 larger curves, the field element resulting from ECDH is directly
5766 used as premaster secret.
5767 [Douglas Stebila (Sun Microsystems Laboratories)]
5769 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5770 curve secp160r1 to the tests.
5771 [Douglas Stebila (Sun Microsystems Laboratories)]
5773 *) Add the possibility to load symbols globally with DSO.
5774 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5776 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5777 control of the error stack.
5780 *) Add support for STORE in ENGINE.
5783 *) Add the STORE type. The intention is to provide a common interface
5784 to certificate and key stores, be they simple file-based stores, or
5785 HSM-type store, or LDAP stores, or...
5786 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5789 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5790 pass a list of arguments to any function as well as provide a way
5791 for a function to pass data back to the caller.
5794 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5795 works like BUF_strdup() but can be used to duplicate a portion of
5796 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5800 *) Add the function sk_find_ex() which works like sk_find(), but will
5801 return an index to an element even if an exact match couldn't be
5802 found. The index is guaranteed to point at the element where the
5803 searched-for key would be inserted to preserve sorting order.
5806 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5807 takes an extra flags argument for optional functionality. Currently,
5808 the following flags are defined:
5810 OBJ_BSEARCH_VALUE_ON_NOMATCH
5811 This one gets OBJ_bsearch_ex() to return a pointer to the first
5812 element where the comparing function returns a negative or zero
5815 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5816 This one gets OBJ_bsearch_ex() to return a pointer to the first
5817 element where the comparing function returns zero. This is useful
5818 if there are more than one element where the comparing function
5822 *) Make it possible to create self-signed certificates with 'openssl ca'
5823 in such a way that the self-signed certificate becomes part of the
5824 CA database and uses the same mechanisms for serial number generation
5825 as all other certificate signing. The new flag '-selfsign' enables
5826 this functionality. Adapt CA.sh and CA.pl.in.
5829 *) Add functionality to check the public key of a certificate request
5830 against a given private. This is useful to check that a certificate
5831 request can be signed by that key (self-signing).
5834 *) Make it possible to have multiple active certificates with the same
5835 subject in the CA index file. This is done only if the keyword
5836 'unique_subject' is set to 'no' in the main CA section (default
5837 if 'CA_default') of the configuration file. The value is saved
5838 with the database itself in a separate index attribute file,
5839 named like the index file with '.attr' appended to the name.
5842 *) Generate multi-valued AVAs using '+' notation in config files for
5846 *) Support for nameConstraints certificate extension.
5849 *) Support for policyConstraints certificate extension.
5852 *) Support for policyMappings certificate extension.
5855 *) Make sure the default DSA_METHOD implementation only uses its
5856 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5857 and change its own handlers to be NULL so as to remove unnecessary
5858 indirection. This lets alternative implementations fallback to the
5859 default implementation more easily.
5862 *) Support for directoryName in GeneralName related extensions
5866 *) Make it possible to link applications using Makefile.shared.
5867 Make that possible even when linking against static libraries!
5870 *) Support for single pass processing for S/MIME signing. This now
5871 means that S/MIME signing can be done from a pipe, in addition
5872 cleartext signing (multipart/signed type) is effectively streaming
5873 and the signed data does not need to be all held in memory.
5875 This is done with a new flag PKCS7_STREAM. When this flag is set
5876 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5877 is done after the data is output (and digests calculated) in
5878 SMIME_write_PKCS7().
5881 *) Add full support for -rpath/-R, both in shared libraries and
5882 applications, at least on the platforms where it's known how
5886 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5887 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5888 will now compute a table of multiples of the generator that
5889 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5890 faster (notably in the case of a single point multiplication,
5891 scalar * generator).
5892 [Nils Larsch, Bodo Moeller]
5894 *) IPv6 support for certificate extensions. The various extensions
5895 which use the IP:a.b.c.d can now take IPv6 addresses using the
5896 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5900 *) Added an ENGINE that implements RSA by performing private key
5901 exponentiations with the GMP library. The conversions to and from
5902 GMP's mpz_t format aren't optimised nor are any montgomery forms
5903 cached, and on x86 it appears OpenSSL's own performance has caught up.
5904 However there are likely to be other architectures where GMP could
5905 provide a boost. This ENGINE is not built in by default, but it can be
5906 specified at Configure time and should be accompanied by the necessary
5907 linker additions, eg;
5908 ./config -DOPENSSL_USE_GMP -lgmp
5911 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5912 testing availability of engines with "-t" - the old behaviour is
5913 produced by increasing the feature's verbosity with "-tt".
5916 *) ECDSA routines: under certain error conditions uninitialized BN objects
5917 could be freed. Solution: make sure initialization is performed early
5918 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5922 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5923 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5924 software implementations. For DSA and DH, parameter generation can
5925 also be overridden by providing the appropriate method callbacks.
5928 *) Change the "progress" mechanism used in key-generation and
5929 primality testing to functions that take a new BN_GENCB pointer in
5930 place of callback/argument pairs. The new API functions have "_ex"
5931 postfixes and the older functions are reimplemented as wrappers for
5932 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5933 declarations of the old functions to help (graceful) attempts to
5934 migrate to the new functions. Also, the new key-generation API
5935 functions operate on a caller-supplied key-structure and return
5936 success/failure rather than returning a key or NULL - this is to
5937 help make "keygen" another member function of RSA_METHOD etc.
5939 Example for using the new callback interface:
5941 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5945 BN_GENCB_set(&my_cb, my_callback, my_arg);
5947 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5948 /* For the meaning of a, b in calls to my_callback(), see the
5949 * documentation of the function that calls the callback.
5950 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5951 * my_callback should return 1 if it wants BN_is_prime_ex()
5952 * to continue, or 0 to stop.
5957 *) Change the ZLIB compression method to be stateful, and make it
5958 available to TLS with the number defined in
5959 draft-ietf-tls-compression-04.txt.
5962 *) Add the ASN.1 structures and functions for CertificatePair, which
5963 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5965 CertificatePair ::= SEQUENCE {
5966 forward [0] Certificate OPTIONAL,
5967 reverse [1] Certificate OPTIONAL,
5968 -- at least one of the pair shall be present -- }
5970 Also implement the PEM functions to read and write certificate
5971 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5973 This needed to be defined, mostly for the sake of the LDAP
5974 attribute crossCertificatePair, but may prove useful elsewhere as
5978 *) Make it possible to inhibit symlinking of shared libraries in
5979 Makefile.shared, for Cygwin's sake.
5982 *) Extend the BIGNUM API by creating a function
5983 void BN_set_negative(BIGNUM *a, int neg);
5984 and a macro that behave like
5985 int BN_is_negative(const BIGNUM *a);
5987 to avoid the need to access 'a->neg' directly in applications.
5990 *) Implement fast modular reduction for pseudo-Mersenne primes
5991 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5992 EC_GROUP_new_curve_GFp() will now automatically use this
5994 [Nils Larsch <nla@trustcenter.de>]
5996 *) Add new lock type (CRYPTO_LOCK_BN).
5999 *) Change the ENGINE framework to automatically load engines
6000 dynamically from specific directories unless they could be
6001 found to already be built in or loaded. Move all the
6002 current engines except for the cryptodev one to a new
6004 The engines in engines/ are built as shared libraries if
6005 the "shared" options was given to ./Configure or ./config.
6006 Otherwise, they are inserted in libcrypto.a.
6007 /usr/local/ssl/engines is the default directory for dynamic
6008 engines, but that can be overridden at configure time through
6009 the usual use of --prefix and/or --openssldir, and at run
6010 time with the environment variable OPENSSL_ENGINES.
6011 [Geoff Thorpe and Richard Levitte]
6013 *) Add Makefile.shared, a helper makefile to build shared
6014 libraries. Adapt Makefile.org.
6017 *) Add version info to Win32 DLLs.
6018 [Peter 'Luna' Runestig" <peter@runestig.com>]
6020 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6021 can be added using this API to created arbitrary PKCS#12
6022 files while avoiding the low level API.
6024 New options to PKCS12_create(), key or cert can be NULL and
6025 will then be omitted from the output file. The encryption
6026 algorithm NIDs can be set to -1 for no encryption, the mac
6027 iteration count can be set to 0 to omit the mac.
6029 Enhance pkcs12 utility by making the -nokeys and -nocerts
6030 options work when creating a PKCS#12 file. New option -nomac
6031 to omit the mac, NONE can be set for an encryption algorithm.
6032 New code is modified to use the enhanced PKCS12_create()
6033 instead of the low level API.
6036 *) Extend ASN1 encoder to support indefinite length constructed
6037 encoding. This can output sequences tags and octet strings in
6038 this form. Modify pk7_asn1.c to support indefinite length
6039 encoding. This is experimental and needs additional code to
6040 be useful, such as an ASN1 bio and some enhanced streaming
6043 Extend template encode functionality so that tagging is passed
6044 down to the template encoder.
6047 *) Let 'openssl req' fail if an argument to '-newkey' is not
6048 recognized instead of using RSA as a default.
6051 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6052 As these are not official, they are not included in "ALL";
6053 the "ECCdraft" ciphersuite group alias can be used to select them.
6054 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6056 *) Add ECDH engine support.
6057 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6059 *) Add ECDH in new directory crypto/ecdh/.
6060 [Douglas Stebila (Sun Microsystems Laboratories)]
6062 *) Let BN_rand_range() abort with an error after 100 iterations
6063 without success (which indicates a broken PRNG).
6066 *) Change BN_mod_sqrt() so that it verifies that the input value
6067 is really the square of the return value. (Previously,
6068 BN_mod_sqrt would show GIGO behaviour.)
6071 *) Add named elliptic curves over binary fields from X9.62, SECG,
6072 and WAP/WTLS; add OIDs that were still missing.
6074 [Sheueling Chang Shantz and Douglas Stebila
6075 (Sun Microsystems Laboratories)]
6077 *) Extend the EC library for elliptic curves over binary fields
6078 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6081 EC_GF2m_simple_method
6085 EC_GROUP_new_curve_GF2m
6086 EC_GROUP_set_curve_GF2m
6087 EC_GROUP_get_curve_GF2m
6088 EC_POINT_set_affine_coordinates_GF2m
6089 EC_POINT_get_affine_coordinates_GF2m
6090 EC_POINT_set_compressed_coordinates_GF2m
6092 Point compression for binary fields is disabled by default for
6093 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6096 As binary polynomials are represented as BIGNUMs, various members
6097 of the EC_GROUP and EC_POINT data structures can be shared
6098 between the implementations for prime fields and binary fields;
6099 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6100 are essentially identical to their ..._GFp counterparts.
6101 (For simplicity, the '..._GFp' prefix has been dropped from
6102 various internal method names.)
6104 An internal 'field_div' method (similar to 'field_mul' and
6105 'field_sqr') has been added; this is used only for binary fields.
6107 [Sheueling Chang Shantz and Douglas Stebila
6108 (Sun Microsystems Laboratories)]
6110 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6111 through methods ('mul', 'precompute_mult').
6113 The generic implementations (now internally called 'ec_wNAF_mul'
6114 and 'ec_wNAF_precomputed_mult') remain the default if these
6115 methods are undefined.
6117 [Sheueling Chang Shantz and Douglas Stebila
6118 (Sun Microsystems Laboratories)]
6120 *) New function EC_GROUP_get_degree, which is defined through
6121 EC_METHOD. For curves over prime fields, this returns the bit
6122 length of the modulus.
6124 [Sheueling Chang Shantz and Douglas Stebila
6125 (Sun Microsystems Laboratories)]
6127 *) New functions EC_GROUP_dup, EC_POINT_dup.
6128 (These simply call ..._new and ..._copy).
6130 [Sheueling Chang Shantz and Douglas Stebila
6131 (Sun Microsystems Laboratories)]
6133 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6134 Polynomials are represented as BIGNUMs (where the sign bit is not
6135 used) in the following functions [macros]:
6138 BN_GF2m_sub [= BN_GF2m_add]
6139 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6140 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6141 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6143 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6144 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6145 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6146 BN_GF2m_cmp [= BN_ucmp]
6148 (Note that only the 'mod' functions are actually for fields GF(2^m).
6149 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6151 For some functions, an the irreducible polynomial defining a
6152 field can be given as an 'unsigned int[]' with strictly
6153 decreasing elements giving the indices of those bits that are set;
6154 i.e., p[] represents the polynomial
6155 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6157 p[0] > p[1] > ... > p[k] = 0.
6158 This applies to the following functions:
6163 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6164 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6166 BN_GF2m_mod_sqrt_arr
6167 BN_GF2m_mod_solve_quad_arr
6171 Conversion can be performed by the following functions:
6176 bntest.c has additional tests for binary polynomial arithmetic.
6178 Two implementations for BN_GF2m_mod_div() are available.
6179 The default algorithm simply uses BN_GF2m_mod_inv() and
6180 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6181 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6182 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6184 [Sheueling Chang Shantz and Douglas Stebila
6185 (Sun Microsystems Laboratories)]
6187 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6188 functionality is disabled at compile-time.
6189 [Douglas Stebila <douglas.stebila@sun.com>]
6191 *) Change default behaviour of 'openssl asn1parse' so that more
6192 information is visible when viewing, e.g., a certificate:
6194 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6195 mode the content of non-printable OCTET STRINGs is output in a
6196 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6197 avoid the appearance of a printable string.
6198 [Nils Larsch <nla@trustcenter.de>]
6200 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6202 EC_GROUP_set_asn1_flag()
6203 EC_GROUP_get_asn1_flag()
6204 EC_GROUP_set_point_conversion_form()
6205 EC_GROUP_get_point_conversion_form()
6206 These control ASN1 encoding details:
6207 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6208 has been set to OPENSSL_EC_NAMED_CURVE.
6209 - Points are encoded in uncompressed form by default; options for
6210 asn1_for are as for point2oct, namely
6211 POINT_CONVERSION_COMPRESSED
6212 POINT_CONVERSION_UNCOMPRESSED
6213 POINT_CONVERSION_HYBRID
6215 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6218 EC_GROUP_get0_seed()
6219 EC_GROUP_get_seed_len()
6220 This is used only for ASN1 purposes (so far).
6221 [Nils Larsch <nla@trustcenter.de>]
6223 *) Add 'field_type' member to EC_METHOD, which holds the NID
6224 of the appropriate field type OID. The new function
6225 EC_METHOD_get_field_type() returns this value.
6226 [Nils Larsch <nla@trustcenter.de>]
6231 EC_POINT_point2hex()
6232 EC_POINT_hex2point()
6233 providing useful interfaces to EC_POINT_point2oct() and
6234 EC_POINT_oct2point().
6235 [Nils Larsch <nla@trustcenter.de>]
6237 *) Change internals of the EC library so that the functions
6238 EC_GROUP_set_generator()
6239 EC_GROUP_get_generator()
6240 EC_GROUP_get_order()
6241 EC_GROUP_get_cofactor()
6242 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6243 to methods, which would lead to unnecessary code duplication when
6244 adding different types of curves.
6245 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6247 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6248 arithmetic, and such that modified wNAFs are generated
6249 (which avoid length expansion in many cases).
6252 *) Add a function EC_GROUP_check_discriminant() (defined via
6253 EC_METHOD) that verifies that the curve discriminant is non-zero.
6255 Add a function EC_GROUP_check() that makes some sanity tests
6256 on a EC_GROUP, its generator and order. This includes
6257 EC_GROUP_check_discriminant().
6258 [Nils Larsch <nla@trustcenter.de>]
6260 *) Add ECDSA in new directory crypto/ecdsa/.
6262 Add applications 'openssl ecparam' and 'openssl ecdsa'
6263 (these are based on 'openssl dsaparam' and 'openssl dsa').
6265 ECDSA support is also included in various other files across the
6266 library. Most notably,
6267 - 'openssl req' now has a '-newkey ecdsa:file' option;
6268 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6269 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6270 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6271 them suitable for ECDSA where domain parameters must be
6272 extracted before the specific public key;
6273 - ECDSA engine support has been added.
6274 [Nils Larsch <nla@trustcenter.de>]
6276 *) Include some named elliptic curves, and add OIDs from X9.62,
6277 SECG, and WAP/WTLS. Each curve can be obtained from the new
6279 EC_GROUP_new_by_curve_name(),
6280 and the list of available named curves can be obtained with
6281 EC_get_builtin_curves().
6282 Also add a 'curve_name' member to EC_GROUP objects, which can be
6284 EC_GROUP_set_curve_name()
6285 EC_GROUP_get_curve_name()
6286 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6288 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6289 was actually never needed) and in BN_mul(). The removal in BN_mul()
6290 required a small change in bn_mul_part_recursive() and the addition
6291 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6292 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6293 bn_sub_words() and bn_add_words() except they take arrays with
6297 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6299 *) Cleanse PEM buffers before freeing them since they may contain
6301 [Benjamin Bennett <ben@psc.edu>]
6303 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6304 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6305 authentication-only ciphersuites.
6308 *) Since AES128 and AES256 share a single mask bit in the logic of
6309 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6310 kludge to work properly if AES128 is available and AES256 isn't.
6313 *) Expand security boundary to match 1.1.1 module.
6316 *) Remove redundant features: hash file source, editing of test vectors
6317 modify fipsld to use external fips_premain.c signature.
6320 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6321 run algorithm test programs.
6324 *) Make algorithm test programs more tolerant of whitespace.
6327 *) Have SSL/TLS server implementation tolerate "mismatched" record
6328 protocol version while receiving ClientHello even if the
6329 ClientHello is fragmented. (The server can't insist on the
6330 particular protocol version it has chosen before the ServerHello
6331 message has informed the client about his choice.)
6334 *) Load error codes if they are not already present instead of using a
6335 static variable. This allows them to be cleanly unloaded and reloaded.
6338 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6340 *) Introduce limits to prevent malicious keys being able to
6341 cause a denial of service. (CVE-2006-2940)
6342 [Steve Henson, Bodo Moeller]
6344 *) Fix ASN.1 parsing of certain invalid structures that can result
6345 in a denial of service. (CVE-2006-2937) [Steve Henson]
6347 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6348 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6350 *) Fix SSL client code which could crash if connecting to a
6351 malicious SSLv2 server. (CVE-2006-4343)
6352 [Tavis Ormandy and Will Drewry, Google Security Team]
6354 *) Change ciphersuite string processing so that an explicit
6355 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6356 will no longer include "AES128-SHA"), and any other similar
6357 ciphersuite (same bitmap) from *other* protocol versions (so that
6358 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6359 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6360 changes from 0.9.8b and 0.9.8d.
6363 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6365 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6366 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6368 *) Change the Unix randomness entropy gathering to use poll() when
6369 possible instead of select(), since the latter has some
6370 undesirable limitations.
6371 [Darryl Miles via Richard Levitte and Bodo Moeller]
6373 *) Disable rogue ciphersuites:
6375 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6376 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6377 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6379 The latter two were purportedly from
6380 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6383 Also deactivate the remaining ciphersuites from
6384 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6385 unofficial, and the ID has long expired.
6388 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6389 dual-core machines) and other potential thread-safety issues.
6392 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6394 *) Adapt fipsld and the build system to link against the validated FIPS
6395 module in FIPS mode.
6398 *) Fixes for VC++ 2005 build under Windows.
6401 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6402 from a Windows bash shell such as MSYS. It is autodetected from the
6403 "config" script when run from a VC++ environment. Modify standard VC++
6404 build to use fipscanister.o from the GNU make build.
6407 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6409 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6410 The value now differs depending on if you build for FIPS or not.
6411 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6412 safely run with a non-FIPSed libcrypto, as it may crash because of
6413 the difference induced by this change.
6416 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6418 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6419 (part of SSL_OP_ALL). This option used to disable the
6420 countermeasure against man-in-the-middle protocol-version
6421 rollback in the SSL 2.0 server implementation, which is a bad
6422 idea. (CVE-2005-2969)
6424 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6425 for Information Security, National Institute of Advanced Industrial
6426 Science and Technology [AIST], Japan)]
6428 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6429 mainly for FIPS compliance and not fully integrated at this stage.
6432 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6433 the exponentiation using a fixed-length exponent. (Otherwise,
6434 the information leaked through timing could expose the secret key
6435 after many signatures; cf. Bleichenbacher's attack on DSA with
6439 *) Make a new fixed-window mod_exp implementation the default for
6440 RSA, DSA, and DH private-key operations so that the sequence of
6441 squares and multiplies and the memory access pattern are
6442 independent of the particular secret key. This will mitigate
6443 cache-timing and potential related attacks.
6445 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6446 and this is automatically used by BN_mod_exp_mont() if the new flag
6447 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6448 will use this BN flag for private exponents unless the flag
6449 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6450 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6452 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6454 *) Change the client implementation for SSLv23_method() and
6455 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6456 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6457 (Previously, the SSL 2.0 backwards compatible Client Hello
6458 message format would be used even with SSL_OP_NO_SSLv2.)
6461 *) Add support for smime-type MIME parameter in S/MIME messages which some
6465 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6466 a threadsafe manner. Modify rsa code to use new function and add calls
6467 to dsa and dh code (which had race conditions before).
6470 *) Include the fixed error library code in the C error file definitions
6471 instead of fixing them up at runtime. This keeps the error code
6472 structures constant.
6475 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6477 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6480 *) Fixes for newer kerberos headers. NB: the casts are needed because
6481 the 'length' field is signed on one version and unsigned on another
6482 with no (?) obvious way to tell the difference, without these VC++
6483 complains. Also the "definition" of FAR (blank) is no longer included
6484 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6485 some needed definitions.
6488 *) Undo Cygwin change.
6491 *) Added support for proxy certificates according to RFC 3820.
6492 Because they may be a security thread to unaware applications,
6493 they must be explicitly allowed in run-time. See
6494 docs/HOWTO/proxy_certificates.txt for further information.
6497 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6499 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6500 server and client random values. Previously
6501 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6502 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6504 This change has negligible security impact because:
6506 1. Server and client random values still have 24 bytes of pseudo random
6509 2. Server and client random values are sent in the clear in the initial
6512 3. The master secret is derived using the premaster secret (48 bytes in
6513 size for static RSA ciphersuites) as well as client server and random
6516 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6519 [Stephen Henson, reported by UK NISCC]
6521 *) Use Windows randomness collection on Cygwin.
6524 *) Fix hang in EGD/PRNGD query when communication socket is closed
6525 prematurely by EGD/PRNGD.
6526 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6528 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6531 *) Back-port of selected performance improvements from development
6532 branch, as well as improved support for PowerPC platforms.
6535 *) Add lots of checks for memory allocation failure, error codes to indicate
6536 failure and freeing up memory if a failure occurs.
6537 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6539 *) Add new -passin argument to dgst.
6542 *) Perform some character comparisons of different types in X509_NAME_cmp:
6543 this is needed for some certificates that re-encode DNs into UTF8Strings
6544 (in violation of RFC3280) and can't or won't issue name rollover
6548 *) Make an explicit check during certificate validation to see that
6549 the CA setting in each certificate on the chain is correct. As a
6550 side effect always do the following basic checks on extensions,
6551 not just when there's an associated purpose to the check:
6553 - if there is an unhandled critical extension (unless the user
6554 has chosen to ignore this fault)
6555 - if the path length has been exceeded (if one is set at all)
6556 - that certain extensions fit the associated purpose (if one has
6560 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6562 *) Avoid a race condition when CRLs are checked in a multi threaded
6563 environment. This would happen due to the reordering of the revoked
6564 entries during signature checking and serial number lookup. Now the
6565 encoding is cached and the serial number sort performed under a lock.
6566 Add new STACK function sk_is_sorted().
6569 *) Add Delta CRL to the extension code.
6572 *) Various fixes to s3_pkt.c so alerts are sent properly.
6573 [David Holmes <d.holmes@f5.com>]
6575 *) Reduce the chances of duplicate issuer name and serial numbers (in
6576 violation of RFC3280) using the OpenSSL certificate creation utilities.
6577 This is done by creating a random 64 bit value for the initial serial
6578 number when a serial number file is created or when a self signed
6579 certificate is created using 'openssl req -x509'. The initial serial
6580 number file is created using 'openssl x509 -next_serial' in CA.pl
6581 rather than being initialized to 1.
6584 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6586 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6587 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6588 [Joe Orton, Steve Henson]
6590 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6592 [Joe Orton, Steve Henson]
6594 *) Make it possible to have multiple active certificates with the same
6595 subject in the CA index file. This is done only if the keyword
6596 'unique_subject' is set to 'no' in the main CA section (default
6597 if 'CA_default') of the configuration file. The value is saved
6598 with the database itself in a separate index attribute file,
6599 named like the index file with '.attr' appended to the name.
6602 *) X509 verify fixes. Disable broken certificate workarounds when
6603 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6604 keyUsage extension present. Don't accept CRLs with unhandled critical
6605 extensions: since verify currently doesn't process CRL extensions this
6606 rejects a CRL with *any* critical extensions. Add new verify error codes
6610 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6611 A clarification of RFC2560 will require the use of OCTET STRINGs and
6612 some implementations cannot handle the current raw format. Since OpenSSL
6613 copies and compares OCSP nonces as opaque blobs without any attempt at
6614 parsing them this should not create any compatibility issues.
6617 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6618 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6619 this HMAC (and other) operations are several times slower than OpenSSL
6623 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6624 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6626 *) Use the correct content when signing type "other".
6629 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6631 *) Fix various bugs revealed by running the NISCC test suite:
6633 Stop out of bounds reads in the ASN1 code when presented with
6634 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6636 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6638 If verify callback ignores invalid public key errors don't try to check
6639 certificate signature with the NULL public key.
6643 *) New -ignore_err option in ocsp application to stop the server
6644 exiting on the first error in a request.
6647 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6648 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6652 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6653 extra data after the compression methods not only for TLS 1.0
6654 but also for SSL 3.0 (as required by the specification).
6655 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6657 *) Change X509_certificate_type() to mark the key as exported/exportable
6658 when it's 512 *bits* long, not 512 bytes.
6661 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6662 blocks during encryption.
6665 *) Various fixes to base64 BIO and non blocking I/O. On write
6666 flushes were not handled properly if the BIO retried. On read
6667 data was not being buffered properly and had various logic bugs.
6668 This also affects blocking I/O when the data being decoded is a
6672 *) Various S/MIME bugfixes and compatibility changes:
6673 output correct application/pkcs7 MIME type if
6674 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6675 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6676 of files as .eml work). Correctly handle very long lines in MIME
6680 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6682 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6683 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6684 a protocol version number mismatch like a decryption error
6685 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6688 *) Turn on RSA blinding by default in the default implementation
6689 to avoid a timing attack. Applications that don't want it can call
6690 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6691 They would be ill-advised to do so in most cases.
6692 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6694 *) Change RSA blinding code so that it works when the PRNG is not
6695 seeded (in this case, the secret RSA exponent is abused as
6696 an unpredictable seed -- if it is not unpredictable, there
6697 is no point in blinding anyway). Make RSA blinding thread-safe
6698 by remembering the creator's thread ID in rsa->blinding and
6699 having all other threads use local one-time blinding factors
6700 (this requires more computation than sharing rsa->blinding, but
6701 avoids excessive locking; and if an RSA object is not shared
6702 between threads, blinding will still be very fast).
6705 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6706 ENGINE as defaults for all supported algorithms irrespective of
6707 the 'flags' parameter. 'flags' is now honoured, so applications
6708 should make sure they are passing it correctly.
6711 *) Target "mingw" now allows native Windows code to be generated in
6712 the Cygwin environment as well as with the MinGW compiler.
6715 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6717 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6718 via timing by performing a MAC computation even if incorrect
6719 block cipher padding has been found. This is a countermeasure
6720 against active attacks where the attacker has to distinguish
6721 between bad padding and a MAC verification error. (CVE-2003-0078)
6723 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6724 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6725 Martin Vuagnoux (EPFL, Ilion)]
6727 *) Make the no-err option work as intended. The intention with no-err
6728 is not to have the whole error stack handling routines removed from
6729 libcrypto, it's only intended to remove all the function name and
6730 reason texts, thereby removing some of the footprint that may not
6731 be interesting if those errors aren't displayed anyway.
6733 NOTE: it's still possible for any application or module to have it's
6734 own set of error texts inserted. The routines are there, just not
6735 used by default when no-err is given.
6738 *) Add support for FreeBSD on IA64.
6739 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6741 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6742 Kerberos function mit_des_cbc_cksum(). Before this change,
6743 the value returned by DES_cbc_cksum() was like the one from
6744 mit_des_cbc_cksum(), except the bytes were swapped.
6745 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6747 *) Allow an application to disable the automatic SSL chain building.
6748 Before this a rather primitive chain build was always performed in
6749 ssl3_output_cert_chain(): an application had no way to send the
6750 correct chain if the automatic operation produced an incorrect result.
6752 Now the chain builder is disabled if either:
6754 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6756 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6758 The reasoning behind this is that an application would not want the
6759 auto chain building to take place if extra chain certificates are
6760 present and it might also want a means of sending no additional
6761 certificates (for example the chain has two certificates and the
6765 *) Add the possibility to build without the ENGINE framework.
6766 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6768 *) Under Win32 gmtime() can return NULL: check return value in
6769 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6772 *) DSA routines: under certain error conditions uninitialized BN objects
6773 could be freed. Solution: make sure initialization is performed early
6774 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6775 Nils Larsch <nla@trustcenter.de> via PR#459)
6778 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6779 checked on reconnect on the client side, therefore session resumption
6780 could still fail with a "ssl session id is different" error. This
6781 behaviour is masked when SSL_OP_ALL is used due to
6782 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6783 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6784 followup to PR #377.
6787 *) IA-32 assembler support enhancements: unified ELF targets, support
6788 for SCO/Caldera platforms, fix for Cygwin shared build.
6791 *) Add support for FreeBSD on sparc64. As a consequence, support for
6792 FreeBSD on non-x86 processors is separate from x86 processors on
6793 the config script, much like the NetBSD support.
6794 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6796 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6798 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6801 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6802 code (06) was taken as the first octet of the session ID and the last
6803 octet was ignored consequently. As a result SSLv2 client side session
6804 caching could not have worked due to the session ID mismatch between
6806 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6810 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6811 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6815 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6816 seems that in spite of existing for more than a year, many application
6817 author have done nothing to provide the necessary callbacks, which
6818 means that this particular engine will not work properly anywhere.
6819 This is a very unfortunate situation which forces us, in the name
6820 of usability, to give the hw_ncipher.c a static lock, which is part
6822 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6823 appear in 0.9.8 or later. We EXPECT application authors to have
6824 dealt properly with this when 0.9.8 is released (unless we actually
6825 make such changes in the libcrypto locking code that changes will
6826 have to be made anyway).
6829 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6830 octets have been read, EOF or an error occurs. Without this change
6831 some truncated ASN1 structures will not produce an error.
6834 *) Disable Heimdal support, since it hasn't been fully implemented.
6835 Still give the possibility to force the use of Heimdal, but with
6836 warnings and a request that patches get sent to openssl-dev.
6839 *) Add the VC-CE target, introduce the WINCE sysname, and add
6840 INSTALL.WCE and appropriate conditionals to make it build.
6841 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6843 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6844 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6845 edit numbers of the version.
6846 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6848 *) Introduce safe string copy and catenation functions
6849 (BUF_strlcpy() and BUF_strlcat()).
6850 [Ben Laurie (CHATS) and Richard Levitte]
6852 *) Avoid using fixed-size buffers for one-line DNs.
6853 [Ben Laurie (CHATS)]
6855 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6856 resizing buffers containing secrets, and use where appropriate.
6857 [Ben Laurie (CHATS)]
6859 *) Avoid using fixed size buffers for configuration file location.
6860 [Ben Laurie (CHATS)]
6862 *) Avoid filename truncation for various CA files.
6863 [Ben Laurie (CHATS)]
6865 *) Use sizeof in preference to magic numbers.
6866 [Ben Laurie (CHATS)]
6868 *) Avoid filename truncation in cert requests.
6869 [Ben Laurie (CHATS)]
6871 *) Add assertions to check for (supposedly impossible) buffer
6873 [Ben Laurie (CHATS)]
6875 *) Don't cache truncated DNS entries in the local cache (this could
6876 potentially lead to a spoofing attack).
6877 [Ben Laurie (CHATS)]
6879 *) Fix various buffers to be large enough for hex/decimal
6880 representations in a platform independent manner.
6881 [Ben Laurie (CHATS)]
6883 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6884 resizing buffers containing secrets, and use where appropriate.
6885 [Ben Laurie (CHATS)]
6887 *) Add BIO_indent() to avoid much slightly worrying code to do
6889 [Ben Laurie (CHATS)]
6891 *) Convert sprintf()/BIO_puts() to BIO_printf().
6892 [Ben Laurie (CHATS)]
6894 *) buffer_gets() could terminate with the buffer only half
6896 [Ben Laurie (CHATS)]
6898 *) Add assertions to prevent user-supplied crypto functions from
6899 overflowing internal buffers by having large block sizes, etc.
6900 [Ben Laurie (CHATS)]
6902 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6904 [Ben Laurie (CHATS)]
6906 *) Eliminate unused copy of key in RC4.
6907 [Ben Laurie (CHATS)]
6909 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6910 [Ben Laurie (CHATS)]
6912 *) Fix off-by-one error in EGD path.
6913 [Ben Laurie (CHATS)]
6915 *) If RANDFILE path is too long, ignore instead of truncating.
6916 [Ben Laurie (CHATS)]
6918 *) Eliminate unused and incorrectly sized X.509 structure
6920 [Ben Laurie (CHATS)]
6922 *) Eliminate unused and dangerous function knumber().
6923 [Ben Laurie (CHATS)]
6925 *) Eliminate unused and dangerous structure, KSSL_ERR.
6926 [Ben Laurie (CHATS)]
6928 *) Protect against overlong session ID context length in an encoded
6929 session object. Since these are local, this does not appear to be
6931 [Ben Laurie (CHATS)]
6933 *) Change from security patch (see 0.9.6e below) that did not affect
6934 the 0.9.6 release series:
6936 Remote buffer overflow in SSL3 protocol - an attacker could
6937 supply an oversized master key in Kerberos-enabled versions.
6939 [Ben Laurie (CHATS)]
6941 *) Change the SSL kerb5 codes to match RFC 2712.
6944 *) Make -nameopt work fully for req and add -reqopt switch.
6945 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6947 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6948 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6950 *) Make sure tests can be performed even if the corresponding algorithms
6951 have been removed entirely. This was also the last step to make
6952 OpenSSL compilable with DJGPP under all reasonable conditions.
6953 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6955 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6956 to allow version independent disabling of normally unselected ciphers,
6957 which may be activated as a side-effect of selecting a single cipher.
6959 (E.g., cipher list string "RSA" enables ciphersuites that are left
6960 out of "ALL" because they do not provide symmetric encryption.
6961 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6962 [Lutz Jaenicke, Bodo Moeller]
6964 *) Add appropriate support for separate platform-dependent build
6965 directories. The recommended way to make a platform-dependent
6966 build directory is the following (tested on Linux), maybe with
6969 # Place yourself outside of the OpenSSL source tree. In
6970 # this example, the environment variable OPENSSL_SOURCE
6971 # is assumed to contain the absolute OpenSSL source directory.
6972 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6973 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6974 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6975 mkdir -p `dirname $F`
6976 ln -s $OPENSSL_SOURCE/$F $F
6979 To be absolutely sure not to disturb the source tree, a "make clean"
6980 is a good thing. If it isn't successful, don't worry about it,
6981 it probably means the source directory is very clean.
6984 *) Make sure any ENGINE control commands make local copies of string
6985 pointers passed to them whenever necessary. Otherwise it is possible
6986 the caller may have overwritten (or deallocated) the original string
6987 data when a later ENGINE operation tries to use the stored values.
6988 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6990 *) Improve diagnostics in file reading and command-line digests.
6991 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6993 *) Add AES modes CFB and OFB to the object database. Correct an
6994 error in AES-CFB decryption.
6997 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6998 allows existing EVP_CIPHER_CTX structures to be reused after
6999 calling EVP_*Final(). This behaviour is used by encryption
7000 BIOs and some applications. This has the side effect that
7001 applications must explicitly clean up cipher contexts with
7002 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7005 *) Check the values of dna and dnb in bn_mul_recursive before calling
7006 bn_mul_comba (a non zero value means the a or b arrays do not contain
7007 n2 elements) and fallback to bn_mul_normal if either is not zero.
7010 *) Fix escaping of non-ASCII characters when using the -subj option
7011 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7014 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7015 form for "surname", serialNumber has no short form.
7016 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7017 therefore remove "mail" short name for "internet 7".
7018 The OID for unique identifiers in X509 certificates is
7019 x500UniqueIdentifier, not uniqueIdentifier.
7020 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7023 *) Add an "init" command to the ENGINE config module and auto initialize
7024 ENGINEs. Without any "init" command the ENGINE will be initialized
7025 after all ctrl commands have been executed on it. If init=1 the
7026 ENGINE is initialized at that point (ctrls before that point are run
7027 on the uninitialized ENGINE and after on the initialized one). If
7028 init=0 then the ENGINE will not be initialized at all.
7031 *) Fix the 'app_verify_callback' interface so that the user-defined
7032 argument is actually passed to the callback: In the
7033 SSL_CTX_set_cert_verify_callback() prototype, the callback
7034 declaration has been changed from
7037 int (*cb)(X509_STORE_CTX *,void *);
7038 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7039 i=s->ctx->app_verify_callback(&ctx)
7040 has been changed into
7041 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7043 To update applications using SSL_CTX_set_cert_verify_callback(),
7044 a dummy argument can be added to their callback functions.
7045 [D. K. Smetters <smetters@parc.xerox.com>]
7047 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7048 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7050 *) Add and OPENSSL_LOAD_CONF define which will cause
7051 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7052 This allows older applications to transparently support certain
7053 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7054 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7055 load the config file and OPENSSL_add_all_algorithms_conf() which will
7056 always load it have also been added.
7059 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7060 Adjust NIDs and EVP layer.
7061 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7063 *) Config modules support in openssl utility.
7065 Most commands now load modules from the config file,
7066 though in a few (such as version) this isn't done
7067 because it couldn't be used for anything.
7069 In the case of ca and req the config file used is
7070 the same as the utility itself: that is the -config
7071 command line option can be used to specify an
7075 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7076 use "openssl_conf" if filename is NULL use default openssl config file.
7079 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7080 config section name. Add a new flag to tolerate a missing config file
7081 and move code to CONF_modules_load_file().
7084 *) Support for crypto accelerator cards from Accelerated Encryption
7085 Processing, www.aep.ie. (Use engine 'aep')
7086 The support was copied from 0.9.6c [engine] and adapted/corrected
7087 to work with the new engine framework.
7088 [AEP Inc. and Richard Levitte]
7090 *) Support for SureWare crypto accelerator cards from Baltimore
7091 Technologies. (Use engine 'sureware')
7092 The support was copied from 0.9.6c [engine] and adapted
7093 to work with the new engine framework.
7096 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7097 make the newer ENGINE framework commands for the CHIL engine work.
7098 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7100 *) Make it possible to produce shared libraries on ReliantUNIX.
7101 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7103 *) Add the configuration target debug-linux-ppro.
7104 Make 'openssl rsa' use the general key loading routines
7105 implemented in apps.c, and make those routines able to
7106 handle the key format FORMAT_NETSCAPE and the variant
7108 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7110 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7111 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7113 *) Add -keyform to rsautl, and document -engine.
7114 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7116 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7117 BIO_R_NO_SUCH_FILE error code rather than the generic
7118 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7121 *) Add new functions
7123 ERR_peek_last_error_line
7124 ERR_peek_last_error_line_data.
7125 These are similar to
7128 ERR_peek_error_line_data,
7129 but report on the latest error recorded rather than the first one
7130 still in the error queue.
7131 [Ben Laurie, Bodo Moeller]
7133 *) default_algorithms option in ENGINE config module. This allows things
7135 default_algorithms = ALL
7136 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7139 *) Preliminary ENGINE config module.
7142 *) New experimental application configuration code.
7145 *) Change the AES code to follow the same name structure as all other
7146 symmetric ciphers, and behave the same way. Move everything to
7147 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7148 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7150 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7151 [Ben Laurie and Theo de Raadt]
7153 *) Add option to output public keys in req command.
7154 [Massimiliano Pala madwolf@openca.org]
7156 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7157 (up to about 10% better than before for P-192 and P-224).
7160 *) New functions/macros
7162 SSL_CTX_set_msg_callback(ctx, cb)
7163 SSL_CTX_set_msg_callback_arg(ctx, arg)
7164 SSL_set_msg_callback(ssl, cb)
7165 SSL_set_msg_callback_arg(ssl, arg)
7167 to request calling a callback function
7169 void cb(int write_p, int version, int content_type,
7170 const void *buf, size_t len, SSL *ssl, void *arg)
7172 whenever a protocol message has been completely received
7173 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7174 protocol version according to which the SSL library interprets
7175 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7176 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7177 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7178 specification (change_cipher_spec(20), alert(21), handshake(22)).
7179 'buf' and 'len' point to the actual message, 'ssl' to the
7180 SSL object, and 'arg' is the application-defined value set by
7181 SSL[_CTX]_set_msg_callback_arg().
7183 'openssl s_client' and 'openssl s_server' have new '-msg' options
7184 to enable a callback that displays all protocol messages.
7187 *) Change the shared library support so shared libraries are built as
7188 soon as the corresponding static library is finished, and thereby get
7189 openssl and the test programs linked against the shared library.
7190 This still only happens when the keyword "shard" has been given to
7191 the configuration scripts.
7193 NOTE: shared library support is still an experimental thing, and
7194 backward binary compatibility is still not guaranteed.
7195 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7197 *) Add support for Subject Information Access extension.
7198 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7200 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7201 additional bytes when new memory had to be allocated, not just
7202 when reusing an existing buffer.
7205 *) New command line and configuration option 'utf8' for the req command.
7206 This allows field values to be specified as UTF8 strings.
7209 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7210 runs for the former and machine-readable output for the latter.
7213 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7214 of the e-mail address in the DN (i.e., it will go into a certificate
7215 extension only). The new configuration file option 'email_in_dn = no'
7216 has the same effect.
7217 [Massimiliano Pala madwolf@openca.org]
7219 *) Change all functions with names starting with des_ to be starting
7220 with DES_ instead. Add wrappers that are compatible with libdes,
7221 but are named _ossl_old_des_*. Finally, add macros that map the
7222 des_* symbols to the corresponding _ossl_old_des_* if libdes
7223 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7224 desired, the des_* symbols will be mapped to DES_*, with one
7227 Since we provide two compatibility mappings, the user needs to
7228 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7229 compatibility is desired. The default (i.e., when that macro
7230 isn't defined) is OpenSSL 0.9.6c compatibility.
7232 There are also macros that enable and disable the support of old
7233 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7234 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7235 are defined, the default will apply: to support the old des routines.
7237 In either case, one must include openssl/des.h to get the correct
7238 definitions. Do not try to just include openssl/des_old.h, that
7241 NOTE: This is a major break of an old API into a new one. Software
7242 authors are encouraged to switch to the DES_ style functions. Some
7243 time in the future, des_old.h and the libdes compatibility functions
7244 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7245 default), and then completely removed.
7248 *) Test for certificates which contain unsupported critical extensions.
7249 If such a certificate is found during a verify operation it is
7250 rejected by default: this behaviour can be overridden by either
7251 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7252 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7253 X509_supported_extension() has also been added which returns 1 if a
7254 particular extension is supported.
7257 *) Modify the behaviour of EVP cipher functions in similar way to digests
7258 to retain compatibility with existing code.
7261 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7262 compatibility with existing code. In particular the 'ctx' parameter does
7263 not have to be to be initialized before the call to EVP_DigestInit() and
7264 it is tidied up after a call to EVP_DigestFinal(). New function
7265 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7266 EVP_MD_CTX_copy() changed to not require the destination to be
7267 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7268 requires the destination to be valid.
7270 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7271 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7274 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7275 so that complete 'Handshake' protocol structures are kept in memory
7276 instead of overwriting 'msg_type' and 'length' with 'body' data.
7279 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7280 [Massimo Santin via Richard Levitte]
7282 *) Major restructuring to the underlying ENGINE code. This includes
7283 reduction of linker bloat, separation of pure "ENGINE" manipulation
7284 (initialisation, etc) from functionality dealing with implementations
7285 of specific crypto interfaces. This change also introduces integrated
7286 support for symmetric ciphers and digest implementations - so ENGINEs
7287 can now accelerate these by providing EVP_CIPHER and EVP_MD
7288 implementations of their own. This is detailed in crypto/engine/README
7289 as it couldn't be adequately described here. However, there are a few
7290 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7291 were changed in the original introduction of ENGINE code have now
7292 reverted back - the hooking from this code to ENGINE is now a good
7293 deal more passive and at run-time, operations deal directly with
7294 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7295 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7296 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7297 they were not being used by the framework as there is no concept of a
7298 BIGNUM_METHOD and they could not be generalised to the new
7299 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7300 ENGINE_cpy() has been removed as it cannot be consistently defined in
7304 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7307 *) Change mkdef.pl to sort symbols that get the same entry number,
7308 and make sure the automatically generated functions ERR_load_*
7309 become part of libeay.num as well.
7312 *) New function SSL_renegotiate_pending(). This returns true once
7313 renegotiation has been requested (either SSL_renegotiate() call
7314 or HelloRequest/ClientHello received from the peer) and becomes
7315 false once a handshake has been completed.
7316 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7317 sends a HelloRequest, but does not ensure that a handshake takes
7318 place. SSL_renegotiate_pending() is useful for checking if the
7319 client has followed the request.)
7322 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7323 By default, clients may request session resumption even during
7324 renegotiation (if session ID contexts permit); with this option,
7325 session resumption is possible only in the first handshake.
7327 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7328 more bits available for options that should not be part of
7329 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7332 *) Add some demos for certificate and certificate request creation.
7335 *) Make maximum certificate chain size accepted from the peer application
7336 settable (SSL*_get/set_max_cert_list()), as proposed by
7337 "Douglas E. Engert" <deengert@anl.gov>.
7340 *) Add support for shared libraries for Unixware-7
7341 (Boyd Lynn Gerber <gerberb@zenez.com>).
7344 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7345 be done prior to destruction. Use this to unload error strings from
7346 ENGINEs that load their own error strings. NB: This adds two new API
7347 functions to "get" and "set" this destroy handler in an ENGINE.
7350 *) Alter all existing ENGINE implementations (except "openssl" and
7351 "openbsd") to dynamically instantiate their own error strings. This
7352 makes them more flexible to be built both as statically-linked ENGINEs
7353 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7354 Also, add stub code to each that makes building them as self-contained
7355 shared-libraries easier (see README.ENGINE).
7358 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7359 implementations into applications that are completely implemented in
7360 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7361 commands that can be used to configure what shared-library to load and
7362 to control aspects of the way it is handled. Also, made an update to
7363 the README.ENGINE file that brings its information up-to-date and
7364 provides some information and instructions on the "dynamic" ENGINE
7365 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7368 *) Make it possible to unload ranges of ERR strings with a new
7369 "ERR_unload_strings" function.
7372 *) Add a copy() function to EVP_MD.
7375 *) Make EVP_MD routines take a context pointer instead of just the
7376 md_data void pointer.
7379 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7380 that the digest can only process a single chunk of data
7381 (typically because it is provided by a piece of
7382 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7383 is only going to provide a single chunk of data, and hence the
7384 framework needn't accumulate the data for oneshot drivers.
7387 *) As with "ERR", make it possible to replace the underlying "ex_data"
7388 functions. This change also alters the storage and management of global
7389 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7390 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7391 index counters. The API functions that use this state have been changed
7392 to take a "class_index" rather than pointers to the class's local STACK
7393 and counter, and there is now an API function to dynamically create new
7394 classes. This centralisation allows us to (a) plug a lot of the
7395 thread-safety problems that existed, and (b) makes it possible to clean
7396 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7397 such data would previously have always leaked in application code and
7398 workarounds were in place to make the memory debugging turn a blind eye
7399 to it. Application code that doesn't use this new function will still
7400 leak as before, but their memory debugging output will announce it now
7401 rather than letting it slide.
7403 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7404 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7405 has a return value to indicate success or failure.
7408 *) Make it possible to replace the underlying "ERR" functions such that the
7409 global state (2 LHASH tables and 2 locks) is only used by the "default"
7410 implementation. This change also adds two functions to "get" and "set"
7411 the implementation prior to it being automatically set the first time
7412 any other ERR function takes place. Ie. an application can call "get",
7413 pass the return value to a module it has just loaded, and that module
7414 can call its own "set" function using that value. This means the
7415 module's "ERR" operations will use (and modify) the error state in the
7416 application and not in its own statically linked copy of OpenSSL code.
7419 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7420 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7421 the operation, and provides a more encapsulated way for external code
7422 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7423 to use these functions rather than manually incrementing the counts.
7425 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7428 *) Add EVP test program.
7431 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7434 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7435 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7436 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7437 These allow a CRL to be built without having to access X509_CRL fields
7438 directly. Modify 'ca' application to use new functions.
7441 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7442 bug workarounds. Rollback attack detection is a security feature.
7443 The problem will only arise on OpenSSL servers when TLSv1 is not
7444 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7445 Software authors not wanting to support TLSv1 will have special reasons
7446 for their choice and can explicitly enable this option.
7447 [Bodo Moeller, Lutz Jaenicke]
7449 *) Rationalise EVP so it can be extended: don't include a union of
7450 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7451 (similar to those existing for EVP_CIPHER_CTX).
7456 EVP_MD_CTX_init(&md); /* new function call */
7457 EVP_DigestInit(&md, EVP_sha1());
7458 EVP_DigestUpdate(&md, in, len);
7459 EVP_DigestFinal(&md, out, NULL);
7460 EVP_MD_CTX_cleanup(&md); /* new function call */
7464 *) Make DES key schedule conform to the usual scheme, as well as
7465 correcting its structure. This means that calls to DES functions
7466 now have to pass a pointer to a des_key_schedule instead of a
7467 plain des_key_schedule (which was actually always a pointer
7470 des_key_schedule ks;
7472 des_set_key_checked(..., &ks);
7473 des_ncbc_encrypt(..., &ks, ...);
7475 (Note that a later change renames 'des_...' into 'DES_...'.)
7478 *) Initial reduction of linker bloat: the use of some functions, such as
7479 PEM causes large amounts of unused functions to be linked in due to
7480 poor organisation. For example pem_all.c contains every PEM function
7481 which has a knock on effect of linking in large amounts of (unused)
7482 ASN1 code. Grouping together similar functions and splitting unrelated
7483 functions prevents this.
7486 *) Cleanup of EVP macros.
7489 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7490 correct _ecb suffix.
7493 *) Add initial OCSP responder support to ocsp application. The
7494 revocation information is handled using the text based index
7495 use by the ca application. The responder can either handle
7496 requests generated internally, supplied in files (for example
7497 via a CGI script) or using an internal minimal server.
7500 *) Add configuration choices to get zlib compression for TLS.
7503 *) Changes to Kerberos SSL for RFC 2712 compliance:
7504 1. Implemented real KerberosWrapper, instead of just using
7505 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7506 2. Implemented optional authenticator field of KerberosWrapper.
7508 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7509 and authenticator structs; see crypto/krb5/.
7511 Generalized Kerberos calls to support multiple Kerberos libraries.
7512 [Vern Staats <staatsvr@asc.hpc.mil>,
7513 Jeffrey Altman <jaltman@columbia.edu>
7514 via Richard Levitte]
7516 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7517 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7518 values for each of the key sizes rather than having just
7519 parameters (and 'speed' generating keys each time).
7522 *) Speed up EVP routines.
7525 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7526 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7527 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7528 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7530 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7531 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7532 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7535 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7537 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7540 *) Added the OS2-EMX target.
7541 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7543 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7544 to support NCONF routines in extension code. New function CONF_set_nconf()
7545 to allow functions which take an NCONF to also handle the old LHASH
7546 structure: this means that the old CONF compatible routines can be
7547 retained (in particular wrt extensions) without having to duplicate the
7548 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7551 *) Enhance the general user interface with mechanisms for inner control
7552 and with possibilities to have yes/no kind of prompts.
7555 *) Change all calls to low level digest routines in the library and
7556 applications to use EVP. Add missing calls to HMAC_cleanup() and
7557 don't assume HMAC_CTX can be copied using memcpy().
7558 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7560 *) Add the possibility to control engines through control names but with
7561 arbitrary arguments instead of just a string.
7562 Change the key loaders to take a UI_METHOD instead of a callback
7563 function pointer. NOTE: this breaks binary compatibility with earlier
7564 versions of OpenSSL [engine].
7565 Adapt the nCipher code for these new conditions and add a card insertion
7569 *) Enhance the general user interface with mechanisms to better support
7570 dialog box interfaces, application-defined prompts, the possibility
7571 to use defaults (for example default passwords from somewhere else)
7572 and interrupts/cancellations.
7575 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7576 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7579 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7580 tidy up some unnecessarily weird code in 'sk_new()').
7581 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7583 *) Change the key loading routines for ENGINEs to use the same kind
7584 callback (pem_password_cb) as all other routines that need this
7588 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7589 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7590 than this minimum value is recommended.
7593 *) New random seeder for OpenVMS, using the system process statistics
7594 that are easily reachable.
7597 *) Windows apparently can't transparently handle global
7598 variables defined in DLLs. Initialisations such as:
7600 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7602 won't compile. This is used by the any applications that need to
7603 declare their own ASN1 modules. This was fixed by adding the option
7604 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7605 needed for static libraries under Win32.
7608 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7609 setting of purpose and trust fields. New X509_STORE trust and
7610 purpose functions and tidy up setting in other SSL functions.
7613 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7614 structure. These are inherited by X509_STORE_CTX when it is
7615 initialised. This allows various defaults to be set in the
7616 X509_STORE structure (such as flags for CRL checking and custom
7617 purpose or trust settings) for functions which only use X509_STORE_CTX
7618 internally such as S/MIME.
7620 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7621 trust settings if they are not set in X509_STORE. This allows X509_STORE
7622 purposes and trust (in S/MIME for example) to override any set by default.
7624 Add command line options for CRL checking to smime, s_client and s_server
7628 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7629 are set then the CRL is looked up in the X509_STORE structure and
7630 its validity and signature checked, then if the certificate is found
7631 in the CRL the verify fails with a revoked error.
7633 Various new CRL related callbacks added to X509_STORE_CTX structure.
7635 Command line options added to 'verify' application to support this.
7637 This needs some additional work, such as being able to handle multiple
7638 CRLs with different times, extension based lookup (rather than just
7639 by subject name) and ultimately more complete V2 CRL extension
7643 *) Add a general user interface API (crypto/ui/). This is designed
7644 to replace things like des_read_password and friends (backward
7645 compatibility functions using this new API are provided).
7646 The purpose is to remove prompting functions from the DES code
7647 section as well as provide for prompting through dialog boxes in
7648 a window system and the like.
7651 *) Add "ex_data" support to ENGINE so implementations can add state at a
7652 per-structure level rather than having to store it globally.
7655 *) Make it possible for ENGINE structures to be copied when retrieved by
7656 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7657 This causes the "original" ENGINE structure to act like a template,
7658 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7659 operational state can be localised to each ENGINE structure, despite the
7660 fact they all share the same "methods". New ENGINE structures returned in
7661 this case have no functional references and the return value is the single
7662 structural reference. This matches the single structural reference returned
7663 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7667 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7668 needs to match any other type at all we need to manually clear the
7672 *) Changes to the "openssl engine" utility to include;
7673 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7674 about an ENGINE's available control commands.
7675 - executing control commands from command line arguments using the
7676 '-pre' and '-post' switches. '-post' is only used if '-t' is
7677 specified and the ENGINE is successfully initialised. The syntax for
7678 the individual commands are colon-separated, for example;
7679 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7682 *) New dynamic control command support for ENGINEs. ENGINEs can now
7683 declare their own commands (numbers), names (strings), descriptions,
7684 and input types for run-time discovery by calling applications. A
7685 subset of these commands are implicitly classed as "executable"
7686 depending on their input type, and only these can be invoked through
7687 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7688 can be based on user input, config files, etc). The distinction is
7689 that "executable" commands cannot return anything other than a boolean
7690 result and can only support numeric or string input, whereas some
7691 discoverable commands may only be for direct use through
7692 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7693 pointers, or other custom uses. The "executable" commands are to
7694 support parameterisations of ENGINE behaviour that can be
7695 unambiguously defined by ENGINEs and used consistently across any
7696 OpenSSL-based application. Commands have been added to all the
7697 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7698 control over shared-library paths without source code alterations.
7701 *) Changed all ENGINE implementations to dynamically allocate their
7702 ENGINEs rather than declaring them statically. Apart from this being
7703 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7704 this also allows the implementations to compile without using the
7705 internal engine_int.h header.
7708 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7709 'const' value. Any code that should be able to modify a RAND_METHOD
7710 should already have non-const pointers to it (ie. they should only
7711 modify their own ones).
7714 *) Made a variety of little tweaks to the ENGINE code.
7715 - "atalla" and "ubsec" string definitions were moved from header files
7716 to C code. "nuron" string definitions were placed in variables
7717 rather than hard-coded - allowing parameterisation of these values
7718 later on via ctrl() commands.
7719 - Removed unused "#if 0"'d code.
7720 - Fixed engine list iteration code so it uses ENGINE_free() to release
7721 structural references.
7722 - Constified the RAND_METHOD element of ENGINE structures.
7723 - Constified various get/set functions as appropriate and added
7724 missing functions (including a catch-all ENGINE_cpy that duplicates
7725 all ENGINE values onto a new ENGINE except reference counts/state).
7726 - Removed NULL parameter checks in get/set functions. Setting a method
7727 or function to NULL is a way of cancelling out a previously set
7728 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7729 and doesn't justify the extra error symbols and code.
7730 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7731 flags from engine_int.h to engine.h.
7732 - Changed prototypes for ENGINE handler functions (init(), finish(),
7733 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7736 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7737 to the algorithm using long division. The binary algorithm can be
7738 used only if the modulus is odd. On 32-bit systems, it is faster
7739 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7740 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7741 up to 450 bits. In 64-bit environments, the binary algorithm
7742 appears to be advantageous for much longer moduli; here we use it
7743 for moduli up to 2048 bits.
7746 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7747 could not support the combine flag in choice fields.
7750 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7751 extensions from a certificate request to the certificate.
7754 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7755 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7756 file: this allows the display of the certificate about to be
7757 signed to be customised, to allow certain fields to be included
7758 or excluded and extension details. The old system didn't display
7759 multicharacter strings properly, omitted fields not in the policy
7760 and couldn't display additional details such as extensions.
7763 *) Function EC_POINTs_mul for multiple scalar multiplication
7764 of an arbitrary number of elliptic curve points
7765 \sum scalars[i]*points[i],
7766 optionally including the generator defined for the EC_GROUP:
7767 scalar*generator + \sum scalars[i]*points[i].
7769 EC_POINT_mul is a simple wrapper function for the typical case
7770 that the point list has just one item (besides the optional
7774 *) First EC_METHODs for curves over GF(p):
7776 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7777 operations and provides various method functions that can also
7778 operate with faster implementations of modular arithmetic.
7780 EC_GFp_mont_method() reuses most functions that are part of
7781 EC_GFp_simple_method, but uses Montgomery arithmetic.
7783 [Bodo Moeller; point addition and point doubling
7784 implementation directly derived from source code provided by
7785 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7787 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7788 crypto/ec/ec_lib.c):
7790 Curves are EC_GROUP objects (with an optional group generator)
7791 based on EC_METHODs that are built into the library.
7793 Points are EC_POINT objects based on EC_GROUP objects.
7795 Most of the framework would be able to handle curves over arbitrary
7796 finite fields, but as there are no obvious types for fields other
7797 than GF(p), some functions are limited to that for now.
7800 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7801 that the file contains a complete HTTP response.
7804 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7805 change the def and num file printf format specifier from "%-40sXXX"
7806 to "%-39s XXX". The latter will always guarantee a space after the
7807 field while the former will cause them to run together if the field
7808 is 40 of more characters long.
7811 *) Constify the cipher and digest 'method' functions and structures
7812 and modify related functions to take constant EVP_MD and EVP_CIPHER
7816 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7817 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7820 *) Modify EVP_Digest*() routines so they now return values. Although the
7821 internal software routines can never fail additional hardware versions
7825 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7827 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7828 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7831 ERR_R_NESTED_ASN1_ERROR
7833 ERR_R_MISSING_ASN1_EOS
7834 were 4 .. 9, conflicting with
7835 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7837 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7838 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7840 Add new error code 'ERR_R_INTERNAL_ERROR'.
7843 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7847 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7848 sets the subject name for a new request or supersedes the
7849 subject name in a given request. Formats that can be parsed are
7850 'CN=Some Name, OU=myOU, C=IT'
7852 'CN=Some Name/OU=myOU/C=IT'.
7854 Add options '-batch' and '-verbose' to 'openssl req'.
7855 [Massimiliano Pala <madwolf@hackmasters.net>]
7857 *) Introduce the possibility to access global variables through
7858 functions on platform were that's the best way to handle exporting
7859 global variables in shared libraries. To enable this functionality,
7860 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7861 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7862 is normally done by Configure or something similar).
7864 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7865 in the source file (foo.c) like this:
7867 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7868 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7870 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7871 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7873 OPENSSL_DECLARE_GLOBAL(int,foo);
7874 #define foo OPENSSL_GLOBAL_REF(foo)
7875 OPENSSL_DECLARE_GLOBAL(double,bar);
7876 #define bar OPENSSL_GLOBAL_REF(bar)
7878 The #defines are very important, and therefore so is including the
7879 header file everywhere where the defined globals are used.
7881 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7882 of ASN.1 items, but that structure is a bit different.
7884 The largest change is in util/mkdef.pl which has been enhanced with
7885 better and easier to understand logic to choose which symbols should
7886 go into the Windows .def files as well as a number of fixes and code
7887 cleanup (among others, algorithm keywords are now sorted
7888 lexicographically to avoid constant rewrites).
7891 *) In BN_div() keep a copy of the sign of 'num' before writing the
7892 result to 'rm' because if rm==num the value will be overwritten
7893 and produce the wrong result if 'num' is negative: this caused
7894 problems with BN_mod() and BN_nnmod().
7897 *) Function OCSP_request_verify(). This checks the signature on an
7898 OCSP request and verifies the signer certificate. The signer
7899 certificate is just checked for a generic purpose and OCSP request
7903 *) Add OCSP_check_validity() function to check the validity of OCSP
7904 responses. OCSP responses are prepared in real time and may only
7905 be a few seconds old. Simply checking that the current time lies
7906 between thisUpdate and nextUpdate max reject otherwise valid responses
7907 caused by either OCSP responder or client clock inaccuracy. Instead
7908 we allow thisUpdate and nextUpdate to fall within a certain period of
7909 the current time. The age of the response can also optionally be
7910 checked. Two new options -validity_period and -status_age added to
7914 *) If signature or public key algorithm is unrecognized print out its
7915 OID rather that just UNKNOWN.
7918 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7919 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7920 ID to be generated from the issuer certificate alone which can then be
7921 passed to OCSP_id_issuer_cmp().
7924 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7925 ASN1 modules to export functions returning ASN1_ITEM pointers
7926 instead of the ASN1_ITEM structures themselves. This adds several
7927 new macros which allow the underlying ASN1 function/structure to
7928 be accessed transparently. As a result code should not use ASN1_ITEM
7929 references directly (such as &X509_it) but instead use the relevant
7930 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7931 use of the new ASN1 code on platforms where exporting structures
7932 is problematical (for example in shared libraries) but exporting
7933 functions returning pointers to structures is not.
7936 *) Add support for overriding the generation of SSL/TLS session IDs.
7937 These callbacks can be registered either in an SSL_CTX or per SSL.
7938 The purpose of this is to allow applications to control, if they wish,
7939 the arbitrary values chosen for use as session IDs, particularly as it
7940 can be useful for session caching in multiple-server environments. A
7941 command-line switch for testing this (and any client code that wishes
7942 to use such a feature) has been added to "s_server".
7943 [Geoff Thorpe, Lutz Jaenicke]
7945 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7946 of the form '#if defined(...) || defined(...) || ...' and
7947 '#if !defined(...) && !defined(...) && ...'. This also avoids
7948 the growing number of special cases it was previously handling.
7951 *) Make all configuration macros available for application by making
7952 sure they are available in opensslconf.h, by giving them names starting
7953 with "OPENSSL_" to avoid conflicts with other packages and by making
7954 sure e_os2.h will cover all platform-specific cases together with
7956 Additionally, it is now possible to define configuration/platform-
7957 specific names (called "system identities"). In the C code, these
7958 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7959 macro with the name beginning with "OPENSSL_SYS_", which is determined
7960 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7964 *) New option -set_serial to 'req' and 'x509' this allows the serial
7965 number to use to be specified on the command line. Previously self
7966 signed certificates were hard coded with serial number 0 and the
7967 CA options of 'x509' had to use a serial number in a file which was
7971 *) New options to 'ca' utility to support V2 CRL entry extensions.
7972 Currently CRL reason, invalidity date and hold instruction are
7973 supported. Add new CRL extensions to V3 code and some new objects.
7976 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7977 disable standard block padding (aka PKCS#5 padding) in the EVP
7978 API, which was previously mandatory. This means that the data is
7979 not padded in any way and so the total length much be a multiple
7980 of the block size, otherwise an error occurs.
7983 *) Initial (incomplete) OCSP SSL support.
7986 *) New function OCSP_parse_url(). This splits up a URL into its host,
7987 port and path components: primarily to parse OCSP URLs. New -url
7988 option to ocsp utility.
7991 *) New nonce behavior. The return value of OCSP_check_nonce() now
7992 reflects the various checks performed. Applications can decide
7993 whether to tolerate certain situations such as an absent nonce
7994 in a response when one was present in a request: the ocsp application
7995 just prints out a warning. New function OCSP_add1_basic_nonce()
7996 this is to allow responders to include a nonce in a response even if
7997 the request is nonce-less.
8000 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8001 skipped when using openssl x509 multiple times on a single input file,
8002 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8005 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8006 set string type: to handle setting ASN1_TIME structures. Fix ca
8007 utility to correctly initialize revocation date of CRLs.
8010 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8011 the clients preferred ciphersuites and rather use its own preferences.
8012 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8013 Internet Explorer by ensuring unchanged hash method during stepup.
8014 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8017 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8018 to aes and add a new 'exist' option to print out symbols that don't
8022 *) Additional options to ocsp utility to allow flags to be set and
8023 additional certificates supplied.
8026 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8027 OCSP client a number of certificate to only verify the response
8031 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8032 handle the new API. Currently only ECB, CBC modes supported. Add new
8035 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8036 Encryption Standard (AES) Ciphersuites for Transport Layer
8037 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8038 not enabled by default and were not part of the "ALL" ciphersuite
8039 alias because they were not yet official; they could be
8040 explicitly requested by specifying the "AESdraft" ciphersuite
8041 group alias. In the final release of OpenSSL 0.9.7, the group
8042 alias is called "AES" and is part of "ALL".)
8043 [Ben Laurie, Steve Henson, Bodo Moeller]
8045 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8046 request to response.
8049 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8050 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8051 extract information from a certificate request. OCSP_response_create()
8052 creates a response and optionally adds a basic response structure.
8053 OCSP_basic_add1_status() adds a complete single response to a basic
8054 response and returns the OCSP_SINGLERESP structure just added (to allow
8055 extensions to be included for example). OCSP_basic_add1_cert() adds a
8056 certificate to a basic response and OCSP_basic_sign() signs a basic
8057 response with various flags. New helper functions ASN1_TIME_check()
8058 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8059 (converts ASN1_TIME to GeneralizedTime).
8062 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8063 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8064 structure from a certificate. X509_pubkey_digest() digests the public_key
8065 contents: this is used in various key identifiers.
8068 *) Make sk_sort() tolerate a NULL argument.
8069 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8071 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8072 passed by the function are trusted implicitly. If any of them signed the
8073 response then it is assumed to be valid and is not verified.
8076 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8077 to data. This was previously part of the PKCS7 ASN1 code. This
8078 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8079 [Steve Henson, reported by Kenneth R. Robinette
8080 <support@securenetterm.com>]
8082 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8083 routines: without these tracing memory leaks is very painful.
8084 Fix leaks in PKCS12 and PKCS7 routines.
8087 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8088 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8089 effectively meant GeneralizedTime would never be used. Now it
8090 is initialised to -1 but X509_time_adj() now has to check the value
8091 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8092 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8093 [Steve Henson, reported by Kenneth R. Robinette
8094 <support@securenetterm.com>]
8096 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8097 result in a zero length in the ASN1_INTEGER structure which was
8098 not consistent with the structure when d2i_ASN1_INTEGER() was used
8099 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8100 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8101 where it did not print out a minus for negative ASN1_INTEGER.
8104 *) Add summary printout to ocsp utility. The various functions which
8105 convert status values to strings have been renamed to:
8106 OCSP_response_status_str(), OCSP_cert_status_str() and
8107 OCSP_crl_reason_str() and are no longer static. New options
8108 to verify nonce values and to disable verification. OCSP response
8109 printout format cleaned up.
8112 *) Add additional OCSP certificate checks. These are those specified
8113 in RFC2560. This consists of two separate checks: the CA of the
8114 certificate being checked must either be the OCSP signer certificate
8115 or the issuer of the OCSP signer certificate. In the latter case the
8116 OCSP signer certificate must contain the OCSP signing extended key
8117 usage. This check is performed by attempting to match the OCSP
8118 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8119 in the OCSP_CERTID structures of the response.
8122 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8123 and related routines. This uses the standard OpenSSL certificate
8124 verify routines to perform initial checks (just CA validity) and
8125 to obtain the certificate chain. Then additional checks will be
8126 performed on the chain. Currently the root CA is checked to see
8127 if it is explicitly trusted for OCSP signing. This is used to set
8128 a root CA as a global signing root: that is any certificate that
8129 chains to that CA is an acceptable OCSP signing certificate.
8132 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8133 extensions from a separate configuration file.
8134 As when reading extensions from the main configuration file,
8135 the '-extensions ...' option may be used for specifying the
8137 [Massimiliano Pala <madwolf@comune.modena.it>]
8139 *) New OCSP utility. Allows OCSP requests to be generated or
8140 read. The request can be sent to a responder and the output
8141 parsed, outputed or printed in text form. Not complete yet:
8142 still needs to check the OCSP response validity.
8145 *) New subcommands for 'openssl ca':
8146 'openssl ca -status <serial>' prints the status of the cert with
8147 the given serial number (according to the index file).
8148 'openssl ca -updatedb' updates the expiry status of certificates
8150 [Massimiliano Pala <madwolf@comune.modena.it>]
8152 *) New '-newreq-nodes' command option to CA.pl. This is like
8153 '-newreq', but calls 'openssl req' with the '-nodes' option
8154 so that the resulting key is not encrypted.
8155 [Damien Miller <djm@mindrot.org>]
8157 *) New configuration for the GNU Hurd.
8158 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8160 *) Initial code to implement OCSP basic response verify. This
8161 is currently incomplete. Currently just finds the signer's
8162 certificate and verifies the signature on the response.
8165 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8166 value of OPENSSLDIR. This is available via the new '-d' option
8167 to 'openssl version', and is also included in 'openssl version -a'.
8170 *) Allowing defining memory allocation callbacks that will be given
8171 file name and line number information in additional arguments
8172 (a const char* and an int). The basic functionality remains, as
8173 well as the original possibility to just replace malloc(),
8174 realloc() and free() by functions that do not know about these
8175 additional arguments. To register and find out the current
8176 settings for extended allocation functions, the following
8177 functions are provided:
8179 CRYPTO_set_mem_ex_functions
8180 CRYPTO_set_locked_mem_ex_functions
8181 CRYPTO_get_mem_ex_functions
8182 CRYPTO_get_locked_mem_ex_functions
8184 These work the same way as CRYPTO_set_mem_functions and friends.
8185 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8186 extended allocation function is enabled.
8187 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8188 a conventional allocation function is enabled.
8189 [Richard Levitte, Bodo Moeller]
8191 *) Finish off removing the remaining LHASH function pointer casts.
8192 There should no longer be any prototype-casting required when using
8193 the LHASH abstraction, and any casts that remain are "bugs". See
8194 the callback types and macros at the head of lhash.h for details
8195 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8198 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8199 If /dev/[u]random devices are not available or do not return enough
8200 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8202 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8203 /etc/entropy will be queried once each in this sequence, querying stops
8204 when enough entropy was collected without querying more sockets.
8207 *) Change the Unix RAND_poll() variant to be able to poll several
8208 random devices, as specified by DEVRANDOM, until a sufficient amount
8209 of data has been collected. We spend at most 10 ms on each file
8210 (select timeout) and read in non-blocking mode. DEVRANDOM now
8211 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8212 (previously it was just the string "/dev/urandom"), so on typical
8213 platforms the 10 ms delay will never occur.
8214 Also separate out the Unix variant to its own file, rand_unix.c.
8215 For VMS, there's a currently-empty rand_vms.c.
8218 *) Move OCSP client related routines to ocsp_cl.c. These
8219 provide utility functions which an application needing
8220 to issue a request to an OCSP responder and analyse the
8221 response will typically need: as opposed to those which an
8222 OCSP responder itself would need which will be added later.
8224 OCSP_request_sign() signs an OCSP request with an API similar
8225 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8226 response. OCSP_response_get1_basic() extracts basic response
8227 from response. OCSP_resp_find_status(): finds and extracts status
8228 information from an OCSP_CERTID structure (which will be created
8229 when the request structure is built). These are built from lower
8230 level functions which work on OCSP_SINGLERESP structures but
8231 won't normally be used unless the application wishes to examine
8232 extensions in the OCSP response for example.
8234 Replace nonce routines with a pair of functions.
8235 OCSP_request_add1_nonce() adds a nonce value and optionally
8236 generates a random value. OCSP_check_nonce() checks the
8237 validity of the nonce in an OCSP response.
8240 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8241 This doesn't copy the supplied OCSP_CERTID and avoids the
8242 need to free up the newly created id. Change return type
8243 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8244 This can then be used to add extensions to the request.
8245 Deleted OCSP_request_new(), since most of its functionality
8246 is now in OCSP_REQUEST_new() (and the case insensitive name
8247 clash) apart from the ability to set the request name which
8248 will be added elsewhere.
8251 *) Update OCSP API. Remove obsolete extensions argument from
8252 various functions. Extensions are now handled using the new
8253 OCSP extension code. New simple OCSP HTTP function which
8254 can be used to send requests and parse the response.
8257 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8258 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8259 uses the special reorder version of SET OF to sort the attributes
8260 and reorder them to match the encoded order. This resolves a long
8261 standing problem: a verify on a PKCS7 structure just after signing
8262 it used to fail because the attribute order did not match the
8263 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8264 it uses the received order. This is necessary to tolerate some broken
8265 software that does not order SET OF. This is handled by encoding
8266 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8267 to produce the required SET OF.
8270 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8271 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8272 files to get correct declarations of the ASN.1 item variables.
8275 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8276 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8277 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8278 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8279 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8280 ASN1_ITEM and no wrapper functions.
8283 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8284 replace the old function pointer based I/O routines. Change most of
8285 the *_d2i_bio() and *_d2i_fp() functions to use these.
8288 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8289 lines, recognize more "algorithms" that can be deselected, and make
8290 it complain about algorithm deselection that isn't recognised.
8293 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8294 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8295 to use new functions. Add NO_ASN1_OLD which can be set to remove
8296 some old style ASN1 functions: this can be used to determine if old
8297 code will still work when these eventually go away.
8300 *) New extension functions for OCSP structures, these follow the
8301 same conventions as certificates and CRLs.
8304 *) New function X509V3_add1_i2d(). This automatically encodes and
8305 adds an extension. Its behaviour can be customised with various
8306 flags to append, replace or delete. Various wrappers added for
8307 certificates and CRLs.
8310 *) Fix to avoid calling the underlying ASN1 print routine when
8311 an extension cannot be parsed. Correct a typo in the
8312 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8315 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8316 entries for variables.
8319 *) Add functionality to apps/openssl.c for detecting locking
8320 problems: As the program is single-threaded, all we have
8321 to do is register a locking callback using an array for
8322 storing which locks are currently held by the program.
8325 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8326 SSL_get_ex_data_X509_STORE_idx(), which is used in
8327 ssl_verify_cert_chain() and thus can be called at any time
8328 during TLS/SSL handshakes so that thread-safety is essential.
8329 Unfortunately, the ex_data design is not at all suited
8330 for multi-threaded use, so it probably should be abolished.
8333 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8334 [Broadcom, tweaked and integrated by Geoff Thorpe]
8336 *) Move common extension printing code to new function
8337 X509V3_print_extensions(). Reorganise OCSP print routines and
8338 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8341 *) New function X509_signature_print() to remove duplication in some
8345 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8346 set (this was treated exactly the same as SET OF previously). This
8347 is used to reorder the STACK representing the structure to match the
8348 encoding. This will be used to get round a problem where a PKCS7
8349 structure which was signed could not be verified because the STACK
8350 order did not reflect the encoded order.
8353 *) Reimplement the OCSP ASN1 module using the new code.
8356 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8357 for its ASN1 operations. The old style function pointers still exist
8358 for now but they will eventually go away.
8361 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8362 completely replaces the old ASN1 functionality with a table driven
8363 encoder and decoder which interprets an ASN1_ITEM structure describing
8364 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8365 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8366 has also been converted to the new form.
8369 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8370 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8371 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8372 for negative moduli.
8375 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8376 of not touching the result's sign bit.
8379 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8383 *) Changed the LHASH code to use prototypes for callbacks, and created
8384 macros to declare and implement thin (optionally static) functions
8385 that provide type-safety and avoid function pointer casting for the
8386 type-specific callbacks.
8389 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8391 [Veers Staats <staatsvr@asc.hpc.mil>,
8392 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8394 *) Reformat the FAQ so the different questions and answers can be divided
8395 in sections depending on the subject.
8398 *) Have the zlib compression code load ZLIB.DLL dynamically under
8402 *) New function BN_mod_sqrt for computing square roots modulo a prime
8403 (using the probabilistic Tonelli-Shanks algorithm unless
8404 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8405 be handled deterministically).
8406 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8408 *) Make BN_mod_inverse faster by explicitly handling small quotients
8409 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8410 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8413 *) New function BN_kronecker.
8416 *) Fix BN_gcd so that it works on negative inputs; the result is
8417 positive unless both parameters are zero.
8418 Previously something reasonably close to an infinite loop was
8419 possible because numbers could be growing instead of shrinking
8420 in the implementation of Euclid's algorithm.
8423 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8424 sign of the number in question.
8426 Fix BN_is_word(a,w) to work correctly for w == 0.
8428 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8429 because its test if the absolute value of 'a' equals 'w'.
8430 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8431 it exists mostly for use in the implementations of BN_is_zero(),
8432 BN_is_one(), and BN_is_word().
8435 *) New function BN_swap.
8438 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8439 the exponentiation functions are more likely to produce reasonable
8440 results on negative inputs.
8443 *) Change BN_mod_mul so that the result is always non-negative.
8444 Previously, it could be negative if one of the factors was negative;
8445 I don't think anyone really wanted that behaviour.
8448 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8449 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8450 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8451 and add new functions:
8460 BN_mod_lshift1_quick
8464 These functions always generate non-negative results.
8466 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8467 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8469 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8470 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8471 be reduced modulo m.
8472 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8475 The following entry accidentally appeared in the CHANGES file
8476 distributed with OpenSSL 0.9.7. The modifications described in
8477 it do *not* apply to OpenSSL 0.9.7.
8479 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8480 was actually never needed) and in BN_mul(). The removal in BN_mul()
8481 required a small change in bn_mul_part_recursive() and the addition
8482 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8483 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8484 bn_sub_words() and bn_add_words() except they take arrays with
8489 *) In 'openssl passwd', verify passwords read from the terminal
8490 unless the '-salt' option is used (which usually means that
8491 verification would just waste user's time since the resulting
8492 hash is going to be compared with some given password hash)
8493 or the new '-noverify' option is used.
8495 This is an incompatible change, but it does not affect
8496 non-interactive use of 'openssl passwd' (passwords on the command
8497 line, '-stdin' option, '-in ...' option) and thus should not
8501 *) Remove all references to RSAref, since there's no more need for it.
8504 *) Make DSO load along a path given through an environment variable
8505 (SHLIB_PATH) with shl_load().
8508 *) Constify the ENGINE code as a result of BIGNUM constification.
8509 Also constify the RSA code and most things related to it. In a
8510 few places, most notable in the depth of the ASN.1 code, ugly
8511 casts back to non-const were required (to be solved at a later
8515 *) Make it so the openssl application has all engines loaded by default.
8518 *) Constify the BIGNUM routines a little more.
8521 *) Add the following functions:
8523 ENGINE_load_cswift()
8525 ENGINE_load_atalla()
8527 ENGINE_load_builtin_engines()
8529 That way, an application can itself choose if external engines that
8530 are built-in in OpenSSL shall ever be used or not. The benefit is
8531 that applications won't have to be linked with libdl or other dso
8532 libraries unless it's really needed.
8534 Changed 'openssl engine' to load all engines on demand.
8535 Changed the engine header files to avoid the duplication of some
8536 declarations (they differed!).
8539 *) 'openssl engine' can now list capabilities.
8542 *) Better error reporting in 'openssl engine'.
8545 *) Never call load_dh_param(NULL) in s_server.
8548 *) Add engine application. It can currently list engines by name and
8549 identity, and test if they are actually available.
8552 *) Improve RPM specification file by forcing symbolic linking and making
8553 sure the installed documentation is also owned by root.root.
8554 [Damien Miller <djm@mindrot.org>]
8556 *) Give the OpenSSL applications more possibilities to make use of
8557 keys (public as well as private) handled by engines.
8560 *) Add OCSP code that comes from CertCo.
8563 *) Add VMS support for the Rijndael code.
8566 *) Added untested support for Nuron crypto accelerator.
8569 *) Add support for external cryptographic devices. This code was
8570 previously distributed separately as the "engine" branch.
8571 [Geoff Thorpe, Richard Levitte]
8573 *) Rework the filename-translation in the DSO code. It is now possible to
8574 have far greater control over how a "name" is turned into a filename
8575 depending on the operating environment and any oddities about the
8576 different shared library filenames on each system.
8579 *) Support threads on FreeBSD-elf in Configure.
8582 *) Fix for SHA1 assembly problem with MASM: it produces
8583 warnings about corrupt line number information when assembling
8584 with debugging information. This is caused by the overlapping
8586 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8589 NCONF_get_number() has no error checking at all. As a replacement,
8590 NCONF_get_number_e() is defined (_e for "error checking") and is
8591 promoted strongly. The old NCONF_get_number is kept around for
8592 binary backward compatibility.
8593 Make it possible for methods to load from something other than a BIO,
8594 by providing a function pointer that is given a name instead of a BIO.
8595 For example, this could be used to load configuration data from an
8599 *) Fix for non blocking accept BIOs. Added new I/O special reason
8600 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8601 with non blocking I/O was not possible because no retry code was
8602 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8606 *) Added the beginnings of Rijndael support.
8609 *) Fix for bug in DirectoryString mask setting. Add support for
8610 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8611 to allow certificate printing to more controllable, additional
8612 'certopt' option to 'x509' to allow new printing options to be
8616 *) Clean old EAY MD5 hack from e_os.h.
8619 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8621 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8622 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8623 [Joe Orton, Steve Henson]
8625 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8627 *) Fix additional bug revealed by the NISCC test suite:
8629 Stop bug triggering large recursion when presented with
8630 certain ASN.1 tags (CVE-2003-0851)
8633 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8635 *) Fix various bugs revealed by running the NISCC test suite:
8637 Stop out of bounds reads in the ASN1 code when presented with
8638 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8640 If verify callback ignores invalid public key errors don't try to check
8641 certificate signature with the NULL public key.
8645 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8646 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8650 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8651 extra data after the compression methods not only for TLS 1.0
8652 but also for SSL 3.0 (as required by the specification).
8653 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8655 *) Change X509_certificate_type() to mark the key as exported/exportable
8656 when it's 512 *bits* long, not 512 bytes.
8659 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8661 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8662 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8663 a protocol version number mismatch like a decryption error
8664 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8667 *) Turn on RSA blinding by default in the default implementation
8668 to avoid a timing attack. Applications that don't want it can call
8669 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8670 They would be ill-advised to do so in most cases.
8671 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8673 *) Change RSA blinding code so that it works when the PRNG is not
8674 seeded (in this case, the secret RSA exponent is abused as
8675 an unpredictable seed -- if it is not unpredictable, there
8676 is no point in blinding anyway). Make RSA blinding thread-safe
8677 by remembering the creator's thread ID in rsa->blinding and
8678 having all other threads use local one-time blinding factors
8679 (this requires more computation than sharing rsa->blinding, but
8680 avoids excessive locking; and if an RSA object is not shared
8681 between threads, blinding will still be very fast).
8684 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8686 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8687 via timing by performing a MAC computation even if incorrect
8688 block cipher padding has been found. This is a countermeasure
8689 against active attacks where the attacker has to distinguish
8690 between bad padding and a MAC verification error. (CVE-2003-0078)
8692 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8693 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8694 Martin Vuagnoux (EPFL, Ilion)]
8696 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8698 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8699 memory from it's contents. This is done with a counter that will
8700 place alternating values in each byte. This can be used to solve
8701 two issues: 1) the removal of calls to memset() by highly optimizing
8702 compilers, and 2) cleansing with other values than 0, since those can
8703 be read through on certain media, for example a swap space on disk.
8706 *) Bugfix: client side session caching did not work with external caching,
8707 because the session->cipher setting was not restored when reloading
8708 from the external cache. This problem was masked, when
8709 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8710 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8713 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8714 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8715 [Zeev Lieber <zeev-l@yahoo.com>]
8717 *) Undo an undocumented change introduced in 0.9.6e which caused
8718 repeated calls to OpenSSL_add_all_ciphers() and
8719 OpenSSL_add_all_digests() to be ignored, even after calling
8723 *) Change the default configuration reader to deal with last line not
8724 being properly terminated.
8727 *) Change X509_NAME_cmp() so it applies the special rules on handling
8728 DN values that are of type PrintableString, as well as RDNs of type
8729 emailAddress where the value has the type ia5String.
8730 [stefank@valicert.com via Richard Levitte]
8732 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8733 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8734 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8735 the bitwise-OR of the two for use by the majority of applications
8736 wanting this behaviour, and update the docs. The documented
8737 behaviour and actual behaviour were inconsistent and had been
8738 changing anyway, so this is more a bug-fix than a behavioural
8740 [Geoff Thorpe, diagnosed by Nadav Har'El]
8742 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8743 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8746 *) Fix initialization code race conditions in
8747 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8748 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8749 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8750 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8751 ssl2_get_cipher_by_char(),
8752 ssl3_get_cipher_by_char().
8753 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8755 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8756 the cached sessions are flushed, as the remove_cb() might use ex_data
8757 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8758 (see [openssl.org #212]).
8759 [Geoff Thorpe, Lutz Jaenicke]
8761 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8762 length, instead of the encoding length to d2i_ASN1_OBJECT.
8765 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8767 *) [In 0.9.6g-engine release:]
8768 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8769 [Lynn Gazis <lgazis@rainbow.com>]
8771 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8773 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8774 and get fix the header length calculation.
8775 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8776 Alon Kantor <alonk@checkpoint.com> (and others),
8779 *) Use proper error handling instead of 'assertions' in buffer
8780 overflow checks added in 0.9.6e. This prevents DoS (the
8781 assertions could call abort()).
8782 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8784 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8786 *) Add various sanity checks to asn1_get_length() to reject
8787 the ASN1 length bytes if they exceed sizeof(long), will appear
8788 negative or the content length exceeds the length of the
8790 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8792 *) Fix cipher selection routines: ciphers without encryption had no flags
8793 for the cipher strength set and where therefore not handled correctly
8794 by the selection routines (PR #130).
8797 *) Fix EVP_dsa_sha macro.
8801 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8802 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8803 that was added in OpenSSL 0.9.6d.
8805 As the countermeasure turned out to be incompatible with some
8806 broken SSL implementations, the new option is part of SSL_OP_ALL.
8807 SSL_OP_ALL is usually employed when compatibility with weird SSL
8808 implementations is desired (e.g. '-bugs' option to 's_client' and
8809 's_server'), so the new option is automatically set in many
8813 *) Changes in security patch:
8815 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8816 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8817 Air Force Materiel Command, USAF, under agreement number
8820 *) Add various sanity checks to asn1_get_length() to reject
8821 the ASN1 length bytes if they exceed sizeof(long), will appear
8822 negative or the content length exceeds the length of the
8823 supplied buffer. (CVE-2002-0659)
8824 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8826 *) Assertions for various potential buffer overflows, not known to
8828 [Ben Laurie (CHATS)]
8830 *) Various temporary buffers to hold ASCII versions of integers were
8831 too small for 64 bit platforms. (CVE-2002-0655)
8832 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8834 *) Remote buffer overflow in SSL3 protocol - an attacker could
8835 supply an oversized session ID to a client. (CVE-2002-0656)
8836 [Ben Laurie (CHATS)]
8838 *) Remote buffer overflow in SSL2 protocol - an attacker could
8839 supply an oversized client master key. (CVE-2002-0656)
8840 [Ben Laurie (CHATS)]
8842 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8844 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8845 encoded as NULL) with id-dsa-with-sha1.
8846 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8848 *) Check various X509_...() return values in apps/req.c.
8849 [Nils Larsch <nla@trustcenter.de>]
8851 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8852 an end-of-file condition would erroneously be flagged, when the CRLF
8853 was just at the end of a processed block. The bug was discovered when
8854 processing data through a buffering memory BIO handing the data to a
8855 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8856 <ptsekov@syntrex.com> and Nedelcho Stanev.
8859 *) Implement a countermeasure against a vulnerability recently found
8860 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8861 before application data chunks to avoid the use of known IVs
8862 with data potentially chosen by the attacker.
8865 *) Fix length checks in ssl3_get_client_hello().
8868 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8869 to prevent ssl3_read_internal() from incorrectly assuming that
8870 ssl3_read_bytes() found application data while handshake
8871 processing was enabled when in fact s->s3->in_read_app_data was
8872 merely automatically cleared during the initial handshake.
8873 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8875 *) Fix object definitions for Private and Enterprise: they were not
8876 recognized in their shortname (=lowercase) representation. Extend
8877 obj_dat.pl to issue an error when using undefined keywords instead
8878 of silently ignoring the problem (Svenning Sorensen
8879 <sss@sss.dnsalias.net>).
8882 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8883 generators, i.e. generators other than 2 and 5. (Previously, the
8884 code did not properly initialise the 'add' and 'rem' values to
8885 BN_generate_prime().)
8887 In the new general case, we do not insist that 'generator' is
8888 actually a primitive root: This requirement is rather pointless;
8889 a generator of the order-q subgroup is just as good, if not
8893 *) Map new X509 verification errors to alerts. Discovered and submitted by
8894 Tom Wu <tom@arcot.com>.
8897 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8898 returning non-zero before the data has been completely received
8899 when using non-blocking I/O.
8900 [Bodo Moeller; problem pointed out by John Hughes]
8902 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8903 [Ben Laurie, Lutz Jaenicke]
8905 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8906 Yoram Zahavi <YoramZ@gilian.com>).
8909 *) Add information about CygWin 1.3 and on, and preserve proper
8910 configuration for the versions before that.
8911 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8913 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8914 check whether we deal with a copy of a session and do not delete from
8915 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8916 <izhar@checkpoint.com>.
8919 *) Do not store session data into the internal session cache, if it
8920 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8921 flag is set). Proposed by Aslam <aslam@funk.com>.
8924 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8928 *) [In 0.9.6d-engine release:]
8929 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8930 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8932 *) Add the configuration target linux-s390x.
8933 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8935 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8936 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8937 variable as an indication that a ClientHello message has been
8938 received. As the flag value will be lost between multiple
8939 invocations of ssl3_accept when using non-blocking I/O, the
8940 function may not be aware that a handshake has actually taken
8941 place, thus preventing a new session from being added to the
8944 To avoid this problem, we now set s->new_session to 2 instead of
8945 using a local variable.
8946 [Lutz Jaenicke, Bodo Moeller]
8948 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8949 if the SSL_R_LENGTH_MISMATCH error is detected.
8950 [Geoff Thorpe, Bodo Moeller]
8952 *) New 'shared_ldflag' column in Configure platform table.
8955 *) Fix EVP_CIPHER_mode macro.
8956 ["Dan S. Camper" <dan@bti.net>]
8958 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8959 type, we must throw them away by setting rr->length to 0.
8960 [D P Chang <dpc@qualys.com>]
8962 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8964 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8965 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8966 worked incorrectly for those cases where range = 10..._2 and
8967 3*range is two bits longer than range.)
8970 *) Only add signing time to PKCS7 structures if it is not already
8974 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8975 OBJ_ld_ce should be OBJ_id_ce.
8976 Also some ip-pda OIDs in crypto/objects/objects.txt were
8977 incorrect (cf. RFC 3039).
8978 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8980 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8981 returns early because it has nothing to do.
8982 [Andy Schneider <andy.schneider@bjss.co.uk>]
8984 *) [In 0.9.6c-engine release:]
8985 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8986 [Andy Schneider <andy.schneider@bjss.co.uk>]
8988 *) [In 0.9.6c-engine release:]
8989 Add support for Cryptographic Appliance's keyserver technology.
8990 (Use engine 'keyclient')
8991 [Cryptographic Appliances and Geoff Thorpe]
8993 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8994 is called via tools/c89.sh because arguments have to be
8995 rearranged (all '-L' options must appear before the first object
8997 [Richard Shapiro <rshapiro@abinitio.com>]
8999 *) [In 0.9.6c-engine release:]
9000 Add support for Broadcom crypto accelerator cards, backported
9002 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9004 *) [In 0.9.6c-engine release:]
9005 Add support for SureWare crypto accelerator cards from
9006 Baltimore Technologies. (Use engine 'sureware')
9007 [Baltimore Technologies and Mark Cox]
9009 *) [In 0.9.6c-engine release:]
9010 Add support for crypto accelerator cards from Accelerated
9011 Encryption Processing, www.aep.ie. (Use engine 'aep')
9012 [AEP Inc. and Mark Cox]
9014 *) Add a configuration entry for gcc on UnixWare.
9015 [Gary Benson <gbenson@redhat.com>]
9017 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9018 messages are stored in a single piece (fixed-length part and
9019 variable-length part combined) and fix various bugs found on the way.
9022 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9023 instead. BIO_gethostbyname() does not know what timeouts are
9024 appropriate, so entries would stay in cache even when they have
9026 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9028 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9029 faced with a pathologically small ClientHello fragment that does
9030 not contain client_version: Instead of aborting with an error,
9031 simply choose the highest available protocol version (i.e.,
9032 TLS 1.0 unless it is disabled). In practice, ClientHello
9033 messages are never sent like this, but this change gives us
9034 strictly correct behaviour at least for TLS.
9037 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9038 never resets s->method to s->ctx->method when called from within
9039 one of the SSL handshake functions.
9040 [Bodo Moeller; problem pointed out by Niko Baric]
9042 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9043 (sent using the client's version number) if client_version is
9044 smaller than the protocol version in use. Also change
9045 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9046 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9047 the client will at least see that alert.
9050 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9054 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9055 client receives HelloRequest while in a handshake.
9056 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9058 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9059 should end in 'break', not 'goto end' which circumvents various
9060 cleanups done in state SSL_ST_OK. But session related stuff
9061 must be disabled for SSL_ST_OK in the case that we just sent a
9064 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9065 before just sending a HelloRequest.
9066 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9068 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9069 reveal whether illegal block cipher padding was found or a MAC
9070 verification error occurred. (Neither SSLerr() codes nor alerts
9071 are directly visible to potential attackers, but the information
9072 may leak via logfiles.)
9074 Similar changes are not required for the SSL 2.0 implementation
9075 because the number of padding bytes is sent in clear for SSL 2.0,
9076 and the extra bytes are just ignored. However ssl/s2_pkt.c
9077 failed to verify that the purported number of padding bytes is in
9081 *) Add OpenUNIX-8 support including shared libraries
9082 (Boyd Lynn Gerber <gerberb@zenez.com>).
9085 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9086 'wristwatch attack' using huge encoding parameters (cf.
9087 James H. Manger's CRYPTO 2001 paper). Note that the
9088 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9089 encoding parameters and hence was not vulnerable.
9092 *) BN_sqr() bug fix.
9093 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9095 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9096 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9097 followed by modular reduction.
9098 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9100 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9101 equivalent based on BN_pseudo_rand() instead of BN_rand().
9104 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9105 This function was broken, as the check for a new client hello message
9106 to handle SGC did not allow these large messages.
9107 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9110 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9113 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9114 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9117 *) Rework the configuration and shared library support for Tru64 Unix.
9118 The configuration part makes use of modern compiler features and
9119 still retains old compiler behavior for those that run older versions
9120 of the OS. The shared library support part includes a variant that
9121 uses the RPATH feature, and is available through the special
9122 configuration target "alpha-cc-rpath", which will never be selected
9124 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9126 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9127 with the same message size as in ssl3_get_certificate_request().
9128 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9129 messages might inadvertently be reject as too long.
9130 [Petr Lampa <lampa@fee.vutbr.cz>]
9132 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9135 *) Modified SSL library such that the verify_callback that has been set
9136 specificly for an SSL object with SSL_set_verify() is actually being
9137 used. Before the change, a verify_callback set with this function was
9138 ignored and the verify_callback() set in the SSL_CTX at the time of
9139 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9140 to allow the necessary settings.
9143 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9144 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9145 done automatically (in contradiction to the requirements of the C
9146 standard). This made problems when used from OpenSSH.
9149 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9150 dh->length and always used
9152 BN_rand_range(priv_key, dh->p).
9154 BN_rand_range() is not necessary for Diffie-Hellman, and this
9155 specific range makes Diffie-Hellman unnecessarily inefficient if
9156 dh->length (recommended exponent length) is much smaller than the
9157 length of dh->p. We could use BN_rand_range() if the order of
9158 the subgroup was stored in the DH structure, but we only have
9163 BN_rand(priv_key, l, ...)
9165 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9171 RSA_eay_public_encrypt
9172 RSA_eay_private_decrypt
9173 RSA_eay_private_encrypt (signing)
9174 RSA_eay_public_decrypt (signature verification)
9176 (default implementations for RSA_public_encrypt,
9177 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9178 always reject numbers >= n.
9181 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9182 to synchronize access to 'locking_thread'. This is necessary on
9183 systems where access to 'locking_thread' (an 'unsigned long'
9184 variable) is not atomic.
9187 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9188 *before* setting the 'crypto_lock_rand' flag. The previous code had
9189 a race condition if 0 is a valid thread ID.
9190 [Travis Vitek <vitek@roguewave.com>]
9192 *) Add support for shared libraries under Irix.
9193 [Albert Chin-A-Young <china@thewrittenword.com>]
9195 *) Add configuration option to build on Linux on both big-endian and
9197 [Ralf Baechle <ralf@uni-koblenz.de>]
9199 *) Add the possibility to create shared libraries on HP-UX.
9202 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9204 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9205 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9206 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9207 PRNG state recovery was possible based on the output of
9208 one PRNG request appropriately sized to gain knowledge on
9209 'md' followed by enough consecutive 1-byte PRNG requests
9210 to traverse all of 'state'.
9212 1. When updating 'md_local' (the current thread's copy of 'md')
9213 during PRNG output generation, hash all of the previous
9214 'md_local' value, not just the half used for PRNG output.
9216 2. Make the number of bytes from 'state' included into the hash
9217 independent from the number of PRNG bytes requested.
9219 The first measure alone would be sufficient to avoid
9220 Markku-Juhani's attack. (Actually it had never occurred
9221 to me that the half of 'md_local' used for chaining was the
9222 half from which PRNG output bytes were taken -- I had always
9223 assumed that the secret half would be used.) The second
9224 measure makes sure that additional data from 'state' is never
9225 mixed into 'md_local' in small portions; this heuristically
9226 further strengthens the PRNG.
9229 *) Fix crypto/bn/asm/mips3.s.
9232 *) When only the key is given to "enc", the IV is undefined. Print out
9233 an error message in this case.
9236 *) Handle special case when X509_NAME is empty in X509 printing routines.
9239 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9240 positive and less than q.
9243 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9244 used: it isn't thread safe and the add_lock_callback should handle
9246 [Paul Rose <Paul.Rose@bridge.com>]
9248 *) Verify that incoming data obeys the block size in
9249 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9253 [Ulf Möller, Bodo Möller]
9255 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9256 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9257 when fixing the server behaviour for backwards-compatible 'client
9258 hello' messages. (Note that the attack is impractical against
9259 SSL 3.0 and TLS 1.0 anyway because length and version checking
9260 means that the probability of guessing a valid ciphertext is
9261 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9264 Before 0.9.5, the countermeasure (hide the error by generating a
9265 random 'decryption result') did not work properly because
9266 ERR_clear_error() was missing, meaning that SSL_get_error() would
9267 detect the supposedly ignored error.
9269 Both problems are now fixed.
9272 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9273 (previously it was 1024).
9276 *) Fix for compatibility mode trust settings: ignore trust settings
9277 unless some valid trust or reject settings are present.
9280 *) Fix for blowfish EVP: its a variable length cipher.
9283 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9284 parameters in DSA public key structures and return an error in the
9285 DSA routines if parameters are absent.
9288 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9289 in the current directory if neither $RANDFILE nor $HOME was set.
9290 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9291 caused some confusion to Windows users who haven't defined $HOME.
9292 Thus RAND_file_name() is changed again: e_os.h can define a
9293 DEFAULT_HOME, which will be used if $HOME is not set.
9294 For Windows, we use "C:"; on other platforms, we still require
9295 environment variables.
9297 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9298 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9299 having multiple threads call RAND_poll() concurrently.
9302 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9303 combination of a flag and a thread ID variable.
9304 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9305 flag), *other* threads can enter ssleay_add_bytes without obeying
9306 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9307 that they do not hold after the first thread unsets add_do_not_lock).
9310 *) Change bctest again: '-x' expressions are not available in all
9314 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9316 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9317 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9319 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9320 the default extension for executables, if any. Also, make the perl
9321 scripts that use symlink() to test if it really exists and use "cp"
9322 if it doesn't. All this made OpenSSL compilable and installable in
9326 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9327 If SEQUENCE is length is indefinite just set c->slen to the total
9328 amount of data available.
9329 [Steve Henson, reported by shige@FreeBSD.org]
9330 [This change does not apply to 0.9.7.]
9332 *) Change bctest to avoid here-documents inside command substitution
9333 (workaround for FreeBSD /bin/sh bug).
9334 For compatibility with Ultrix, avoid shell functions (introduced
9335 in the bctest version that searches along $PATH).
9338 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9339 with des_encrypt() defined on some operating systems, like Solaris
9343 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9344 On the Importance of Eliminating Errors in Cryptographic
9345 Computations, J. Cryptology 14 (2001) 2, 101-119,
9346 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9349 *) MIPS assembler BIGNUM division bug fix.
9352 *) Disabled incorrect Alpha assembler code.
9355 *) Fix PKCS#7 decode routines so they correctly update the length
9356 after reading an EOC for the EXPLICIT tag.
9358 [This change does not apply to 0.9.7.]
9360 *) Fix bug in PKCS#12 key generation routines. This was triggered
9361 if a 3DES key was generated with a 0 initial byte. Include
9362 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9363 (but broken) behaviour.
9366 *) Enhance bctest to search for a working bc along $PATH and print
9368 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9370 *) Fix memory leaks in err.c: free err_data string if necessary;
9371 don't write to the wrong index in ERR_set_error_data.
9374 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9378 *) Replace rdtsc with _emit statements for VC++ version 5.
9379 [Jeremy Cooper <jeremy@baymoo.org>]
9381 *) Make it possible to reuse SSLv2 sessions.
9384 *) In copy_email() check for >= 0 as a return value for
9385 X509_NAME_get_index_by_NID() since 0 is a valid index.
9386 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9388 *) Avoid coredump with unsupported or invalid public keys by checking if
9389 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9390 PKCS7_verify() fails with non detached data.
9393 *) Don't use getenv in library functions when run as setuid/setgid.
9394 New function OPENSSL_issetugid().
9397 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9398 due to incorrect handling of multi-threading:
9400 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9402 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9404 3. Count how many times MemCheck_off() has been called so that
9405 nested use can be treated correctly. This also avoids
9406 inband-signalling in the previous code (which relied on the
9407 assumption that thread ID 0 is impossible).
9410 *) Add "-rand" option also to s_client and s_server.
9413 *) Fix CPU detection on Irix 6.x.
9414 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9415 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9417 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9420 [This change does not apply to 0.9.7.]
9422 *) Use the cached encoding of an X509_NAME structure rather than
9423 copying it. This is apparently the reason for the libsafe "errors"
9424 but the code is actually correct.
9427 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9428 Bleichenbacher's DSA attack.
9429 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9430 to be set and top=0 forces the highest bit to be set; top=-1 is new
9431 and leaves the highest bit random.
9432 [Ulf Moeller, Bodo Moeller]
9434 *) In the NCONF_...-based implementations for CONF_... queries
9435 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9436 a temporary CONF structure with the data component set to NULL
9437 (which gives segmentation faults in lh_retrieve).
9438 Instead, use NULL for the CONF pointer in CONF_get_string and
9439 CONF_get_number (which may use environment variables) and directly
9440 return NULL from CONF_get_section.
9443 *) Fix potential buffer overrun for EBCDIC.
9446 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9447 keyUsage if basicConstraints absent for a CA.
9450 *) Make SMIME_write_PKCS7() write mail header values with a format that
9451 is more generally accepted (no spaces before the semicolon), since
9452 some programs can't parse those values properly otherwise. Also make
9453 sure BIO's that break lines after each write do not create invalid
9457 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9458 macros previously used would not encode an empty SEQUENCE OF
9459 and break the signature.
9461 [This change does not apply to 0.9.7.]
9463 *) Zero the premaster secret after deriving the master secret in
9467 *) Add some EVP_add_digest_alias registrations (as found in
9468 OpenSSL_add_all_digests()) to SSL_library_init()
9469 aka OpenSSL_add_ssl_algorithms(). This provides improved
9470 compatibility with peers using X.509 certificates
9471 with unconventional AlgorithmIdentifier OIDs.
9474 *) Fix for Irix with NO_ASM.
9475 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9477 *) ./config script fixes.
9478 [Ulf Moeller, Richard Levitte]
9480 *) Fix 'openssl passwd -1'.
9483 *) Change PKCS12_key_gen_asc() so it can cope with non null
9484 terminated strings whose length is passed in the passlen
9485 parameter, for example from PEM callbacks. This was done
9486 by adding an extra length parameter to asc2uni().
9487 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9489 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9490 call failed, free the DSA structure.
9493 *) Fix to uni2asc() to cope with zero length Unicode strings.
9494 These are present in some PKCS#12 files.
9497 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9498 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9499 when writing a 32767 byte record.
9500 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9502 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9503 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9505 (RSA objects have a reference count access to which is protected
9506 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9507 so they are meant to be shared between threads.)
9508 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9509 "Reddie, Steven" <Steven.Reddie@ca.com>]
9511 *) Fix a deadlock in CRYPTO_mem_leaks().
9514 *) Use better test patterns in bntest.
9517 *) rand_win.c fix for Borland C.
9520 *) BN_rshift bugfix for n == 0.
9523 *) Add a 'bctest' script that checks for some known 'bc' bugs
9524 so that 'make test' does not abort just because 'bc' is broken.
9527 *) Store verify_result within SSL_SESSION also for client side to
9528 avoid potential security hole. (Re-used sessions on the client side
9529 always resulted in verify_result==X509_V_OK, not using the original
9530 result of the server certificate verification.)
9533 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9534 SSL3_RT_APPLICATION_DATA, return 0.
9535 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9539 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9540 releases, have been re-implemented by renaming the previous
9541 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9542 and ssl3_read_internal, respectively, and adding 'peek' parameters
9543 to them. The new ssl[23]_{read,peek} functions are calls to
9544 ssl[23]_read_internal with the 'peek' flag set appropriately.
9545 A 'peek' parameter has also been added to ssl3_read_bytes, which
9546 does the actual work for ssl3_read_internal.
9549 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9550 the method-specific "init()" handler. Also clean up ex_data after
9551 calling the method-specific "finish()" handler. Previously, this was
9552 happening the other way round.
9555 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9556 The previous value, 12, was not always sufficient for BN_mod_exp().
9559 *) Make sure that shared libraries get the internal name engine with
9560 the full version number and not just 0. This should mark the
9561 shared libraries as not backward compatible. Of course, this should
9562 be changed again when we can guarantee backward binary compatibility.
9565 *) Fix typo in get_cert_by_subject() in by_dir.c
9566 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9568 *) Rework the system to generate shared libraries:
9570 - Make note of the expected extension for the shared libraries and
9571 if there is a need for symbolic links from for example libcrypto.so.0
9572 to libcrypto.so.0.9.7. There is extended info in Configure for
9575 - Make as few rebuilds of the shared libraries as possible.
9577 - Still avoid linking the OpenSSL programs with the shared libraries.
9579 - When installing, install the shared libraries separately from the
9583 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9585 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9586 and not in SSL_clear because the latter is also used by the
9587 accept/connect functions; previously, the settings made by
9588 SSL_set_read_ahead would be lost during the handshake.
9589 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9591 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9592 Previously, it would create entries for disabled algorithms no
9596 *) Added several new manual pages for SSL_* function.
9599 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9601 *) In ssl23_get_client_hello, generate an error message when faced
9602 with an initial SSL 3.0/TLS record that is too small to contain the
9603 first two bytes of the ClientHello message, i.e. client_version.
9604 (Note that this is a pathologic case that probably has never happened
9605 in real life.) The previous approach was to use the version number
9606 from the record header as a substitute; but our protocol choice
9607 should not depend on that one because it is not authenticated
9608 by the Finished messages.
9611 *) More robust randomness gathering functions for Windows.
9612 [Jeffrey Altman <jaltman@columbia.edu>]
9614 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9615 not set then we don't setup the error code for issuer check errors
9616 to avoid possibly overwriting other errors which the callback does
9617 handle. If an application does set the flag then we assume it knows
9618 what it is doing and can handle the new informational codes
9622 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9623 a general "ANY" type, as such it should be able to decode anything
9624 including tagged types. However it didn't check the class so it would
9625 wrongly interpret tagged types in the same way as their universal
9626 counterpart and unknown types were just rejected. Changed so that the
9627 tagged and unknown types are handled in the same way as a SEQUENCE:
9628 that is the encoding is stored intact. There is also a new type
9629 "V_ASN1_OTHER" which is used when the class is not universal, in this
9630 case we have no idea what the actual type is so we just lump them all
9634 *) On VMS, stdout may very well lead to a file that is written to
9635 in a record-oriented fashion. That means that every write() will
9636 write a separate record, which will be read separately by the
9637 programs trying to read from it. This can be very confusing.
9639 The solution is to put a BIO filter in the way that will buffer
9640 text until a linefeed is reached, and then write everything a
9641 line at a time, so every record written will be an actual line,
9642 not chunks of lines and not (usually doesn't happen, but I've
9643 seen it once) several lines in one record. BIO_f_linebuffer() is
9646 Currently, it's a VMS-only method, because that's where it has
9647 been tested well enough.
9650 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9651 it can return incorrect results.
9652 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9653 but it was in 0.9.6-beta[12].)
9656 *) Disable the check for content being present when verifying detached
9657 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9658 include zero length content when signing messages.
9661 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9662 BIO_ctrl (for BIO pairs).
9665 *) Add DSO method for VMS.
9668 *) Bug fix: Montgomery multiplication could produce results with the
9672 *) Add RPM specification openssl.spec and modify it to build three
9673 packages. The default package contains applications, application
9674 documentation and run-time libraries. The devel package contains
9675 include files, static libraries and function documentation. The
9676 doc package contains the contents of the doc directory. The original
9677 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9680 *) Add a large number of documentation files for many SSL routines.
9681 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9683 *) Add a configuration entry for Sony News 4.
9684 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9686 *) Don't set the two most significant bits to one when generating a
9687 random number < q in the DSA library.
9690 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9691 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9692 the underlying transport is blocking) if a handshake took place.
9693 (The default behaviour is needed by applications such as s_client
9694 and s_server that use select() to determine when to use SSL_read;
9695 but for applications that know in advance when to expect data, it
9696 just makes things more complicated.)
9699 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9703 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9704 work better on such systems.
9705 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9707 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9708 Update PKCS12_parse() so it copies the friendlyName and the
9709 keyid to the certificates aux info.
9712 *) Fix bug in PKCS7_verify() which caused an infinite loop
9713 if there was more than one signature.
9714 [Sven Uszpelkat <su@celocom.de>]
9716 *) Major change in util/mkdef.pl to include extra information
9717 about each symbol, as well as presenting variables as well
9718 as functions. This change means that there's n more need
9719 to rebuild the .num files when some algorithms are excluded.
9722 *) Allow the verify time to be set by an application,
9723 rather than always using the current time.
9726 *) Phase 2 verify code reorganisation. The certificate
9727 verify code now looks up an issuer certificate by a
9728 number of criteria: subject name, authority key id
9729 and key usage. It also verifies self signed certificates
9730 by the same criteria. The main comparison function is
9731 X509_check_issued() which performs these checks.
9733 Lot of changes were necessary in order to support this
9734 without completely rewriting the lookup code.
9736 Authority and subject key identifier are now cached.
9738 The LHASH 'certs' is X509_STORE has now been replaced
9739 by a STACK_OF(X509_OBJECT). This is mainly because an
9740 LHASH can't store or retrieve multiple objects with
9741 the same hash value.
9743 As a result various functions (which were all internal
9744 use only) have changed to handle the new X509_STORE
9745 structure. This will break anything that messed round
9746 with X509_STORE internally.
9748 The functions X509_STORE_add_cert() now checks for an
9749 exact match, rather than just subject name.
9751 The X509_STORE API doesn't directly support the retrieval
9752 of multiple certificates matching a given criteria, however
9753 this can be worked round by performing a lookup first
9754 (which will fill the cache with candidate certificates)
9755 and then examining the cache for matches. This is probably
9756 the best we can do without throwing out X509_LOOKUP
9757 entirely (maybe later...).
9759 The X509_VERIFY_CTX structure has been enhanced considerably.
9761 All certificate lookup operations now go via a get_issuer()
9762 callback. Although this currently uses an X509_STORE it
9763 can be replaced by custom lookups. This is a simple way
9764 to bypass the X509_STORE hackery necessary to make this
9765 work and makes it possible to use more efficient techniques
9766 in future. A very simple version which uses a simple
9767 STACK for its trusted certificate store is also provided
9768 using X509_STORE_CTX_trusted_stack().
9770 The verify_cb() and verify() callbacks now have equivalents
9771 in the X509_STORE_CTX structure.
9773 X509_STORE_CTX also has a 'flags' field which can be used
9774 to customise the verify behaviour.
9777 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9778 excludes S/MIME capabilities.
9781 *) When a certificate request is read in keep a copy of the
9782 original encoding of the signed data and use it when outputting
9783 again. Signatures then use the original encoding rather than
9784 a decoded, encoded version which may cause problems if the
9785 request is improperly encoded.
9788 *) For consistency with other BIO_puts implementations, call
9789 buffer_write(b, ...) directly in buffer_puts instead of calling
9792 In BIO_puts, increment b->num_write as in BIO_write.
9793 [Peter.Sylvester@EdelWeb.fr]
9795 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9796 BN_zero, we may not return a BIGNUM with an array consisting of
9800 *) Avoid calling abort() from within the library when problems are
9801 detected, except if preprocessor symbols have been defined
9802 (such as REF_CHECK, BN_DEBUG etc.).
9805 *) New openssl application 'rsautl'. This utility can be
9806 used for low level RSA operations. DER public key
9807 BIO/fp routines also added.
9810 *) New Configure entry and patches for compiling on QNX 4.
9811 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9813 *) A demo state-machine implementation was sponsored by
9814 Nuron (http://www.nuron.com/) and is now available in
9815 demos/state_machine.
9818 *) New options added to the 'dgst' utility for signature
9819 generation and verification.
9822 *) Unrecognized PKCS#7 content types are now handled via a
9823 catch all ASN1_TYPE structure. This allows unsupported
9824 types to be stored as a "blob" and an application can
9825 encode and decode it manually.
9828 *) Fix various signed/unsigned issues to make a_strex.c
9830 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9832 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9833 length if passed a buffer. ASN1_INTEGER_to_BN failed
9834 if passed a NULL BN and its argument was negative.
9835 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9837 *) Modification to PKCS#7 encoding routines to output definite
9838 length encoding. Since currently the whole structures are in
9839 memory there's not real point in using indefinite length
9840 constructed encoding. However if OpenSSL is compiled with
9841 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9844 *) Added BIO_vprintf() and BIO_vsnprintf().
9847 *) Added more prefixes to parse for in the strings written
9848 through a logging bio, to cover all the levels that are available
9849 through syslog. The prefixes are now:
9851 PANIC, EMERG, EMR => LOG_EMERG
9852 ALERT, ALR => LOG_ALERT
9853 CRIT, CRI => LOG_CRIT
9854 ERROR, ERR => LOG_ERR
9855 WARNING, WARN, WAR => LOG_WARNING
9856 NOTICE, NOTE, NOT => LOG_NOTICE
9857 INFO, INF => LOG_INFO
9858 DEBUG, DBG => LOG_DEBUG
9860 and as before, if none of those prefixes are present at the
9861 beginning of the string, LOG_ERR is chosen.
9863 On Win32, the LOG_* levels are mapped according to this:
9865 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9866 LOG_WARNING => EVENTLOG_WARNING_TYPE
9867 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9871 *) Made it possible to reconfigure with just the configuration
9872 argument "reconf" or "reconfigure". The command line arguments
9873 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9874 and are retrieved from there when reconfiguring.
9878 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9880 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9883 *) The obj_dat.pl script was messing up the sorting of object
9884 names. The reason was that it compared the quoted version
9885 of strings as a result "OCSP" > "OCSP Signing" because
9886 " > SPACE. Changed script to store unquoted versions of
9887 names and add quotes on output. It was also omitting some
9888 names from the lookup table if they were given a default
9889 value (that is if SN is missing it is given the same
9890 value as LN and vice versa), these are now added on the
9891 grounds that if an object has a name we should be able to
9892 look it up. Finally added warning output when duplicate
9893 short or long names are found.
9896 *) Changes needed for Tandem NSK.
9897 [Scott Uroff <scott@xypro.com>]
9899 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9900 RSA_padding_check_SSLv23(), special padding was never detected
9901 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9902 version rollback attacks was not effective.
9904 In s23_clnt.c, don't use special rollback-attack detection padding
9905 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9906 client; similarly, in s23_srvr.c, don't do the rollback check if
9907 SSL 2.0 is the only protocol enabled in the server.
9910 *) Make it possible to get hexdumps of unprintable data with 'openssl
9911 asn1parse'. By implication, the functions ASN1_parse_dump() and
9912 BIO_dump_indent() are added.
9915 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9916 these print out strings and name structures based on various
9917 flags including RFC2253 support and proper handling of
9918 multibyte characters. Added options to the 'x509' utility
9919 to allow the various flags to be set.
9922 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9923 Also change the functions X509_cmp_current_time() and
9924 X509_gmtime_adj() work with an ASN1_TIME structure,
9925 this will enable certificates using GeneralizedTime in validity
9926 dates to be checked.
9929 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9930 negative public key encodings) on by default,
9931 NO_NEG_PUBKEY_BUG can be set to disable it.
9934 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9935 content octets. An i2c_ASN1_OBJECT is unnecessary because
9936 the encoding can be trivially obtained from the structure.
9939 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9940 not read locks (CRYPTO_r_[un]lock).
9943 *) A first attempt at creating official support for shared
9944 libraries through configuration. I've kept it so the
9945 default is static libraries only, and the OpenSSL programs
9946 are always statically linked for now, but there are
9947 preparations for dynamic linking in place.
9948 This has been tested on Linux and Tru64.
9951 *) Randomness polling function for Win9x, as described in:
9952 Peter Gutmann, Software Generation of Practically Strong
9956 *) Fix so PRNG is seeded in req if using an already existing
9960 *) New options to smime application. -inform and -outform
9961 allow alternative formats for the S/MIME message including
9962 PEM and DER. The -content option allows the content to be
9963 specified separately. This should allow things like Netscape
9964 form signing output easier to verify.
9967 *) Fix the ASN1 encoding of tags using the 'long form'.
9970 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9971 STRING types. These convert content octets to and from the
9972 underlying type. The actual tag and length octets are
9973 already assumed to have been read in and checked. These
9974 are needed because all other string types have virtually
9975 identical handling apart from the tag. By having versions
9976 of the ASN1 functions that just operate on content octets
9977 IMPLICIT tagging can be handled properly. It also allows
9978 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9979 and ASN1_INTEGER are identical apart from the tag.
9982 *) Change the handling of OID objects as follows:
9984 - New object identifiers are inserted in objects.txt, following
9985 the syntax given in objects.README.
9986 - objects.pl is used to process obj_mac.num and create a new
9988 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9991 This is currently kind of a hack, and the perl code in objects.pl
9992 isn't very elegant, but it works as I intended. The simplest way
9993 to check that it worked correctly is to look in obj_dat.h and
9994 check the array nid_objs and make sure the objects haven't moved
9995 around (this is important!). Additions are OK, as well as
9996 consistent name changes.
9999 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10002 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10003 The given file adds to whatever has already been seeded into the
10004 random pool through the RANDFILE configuration file option or
10005 environment variable, or the default random state file.
10008 *) mkstack.pl now sorts each macro group into lexical order.
10009 Previously the output order depended on the order the files
10010 appeared in the directory, resulting in needless rewriting
10014 *) Patches to make OpenSSL compile under Win32 again. Mostly
10015 work arounds for the VC++ problem that it treats func() as
10016 func(void). Also stripped out the parts of mkdef.pl that
10017 added extra typesafe functions: these no longer exist.
10020 *) Reorganisation of the stack code. The macros are now all
10021 collected in safestack.h . Each macro is defined in terms of
10022 a "stack macro" of the form SKM_<name>(type, a, b). The
10023 DEBUG_SAFESTACK is now handled in terms of function casts,
10024 this has the advantage of retaining type safety without the
10025 use of additional functions. If DEBUG_SAFESTACK is not defined
10026 then the non typesafe macros are used instead. Also modified the
10027 mkstack.pl script to handle the new form. Needs testing to see
10028 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10029 the default if no major problems. Similar behaviour for ASN1_SET_OF
10030 and PKCS12_STACK_OF.
10033 *) When some versions of IIS use the 'NET' form of private key the
10034 key derivation algorithm is different. Normally MD5(password) is
10035 used as a 128 bit RC4 key. In the modified case
10036 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10037 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10038 as the old Netscape_RSA functions except they have an additional
10039 'sgckey' parameter which uses the modified algorithm. Also added
10040 an -sgckey command line option to the rsa utility. Thanks to
10041 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10042 algorithm to openssl-dev.
10045 *) The evp_local.h macros were using 'c.##kname' which resulted in
10046 invalid expansion on some systems (SCO 5.0.5 for example).
10047 Corrected to 'c.kname'.
10048 [Phillip Porch <root@theporch.com>]
10050 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10051 a STACK of email addresses from a certificate or request, these look
10052 in the subject name and the subject alternative name extensions and
10053 omit any duplicate addresses.
10056 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10057 This makes DSA verification about 2 % faster.
10060 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10061 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10062 plus overhead for 1024 bit moduli).
10063 This makes exponentiations about 0.5 % faster for 1024 bit
10064 exponents (as measured by "openssl speed rsa2048").
10067 *) Rename memory handling macros to avoid conflicts with other
10069 Malloc => OPENSSL_malloc
10070 Malloc_locked => OPENSSL_malloc_locked
10071 Realloc => OPENSSL_realloc
10072 Free => OPENSSL_free
10075 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10076 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10079 *) CygWin32 support.
10080 [John Jarvie <jjarvie@newsguy.com>]
10082 *) The type-safe stack code has been rejigged. It is now only compiled
10083 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10084 by default all type-specific stack functions are "#define"d back to
10085 standard stack functions. This results in more streamlined output
10086 but retains the type-safety checking possibilities of the original
10090 *) The STACK code has been cleaned up, and certain type declarations
10091 that didn't make a lot of sense have been brought in line. This has
10092 also involved a cleanup of sorts in safestack.h to more correctly
10093 map type-safe stack functions onto their plain stack counterparts.
10094 This work has also resulted in a variety of "const"ifications of
10095 lots of the code, especially "_cmp" operations which should normally
10096 be prototyped with "const" parameters anyway.
10099 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10100 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10101 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10102 where all of 'md' is used each time the PRNG is used, but 'state'
10103 is used only indexed by a cyclic counter. As entropy may not be
10104 well distributed from the beginning, 'md' is important as a
10105 chaining variable. However, the output function chains only half
10106 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10107 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10108 in all of 'state' being rewritten, with the new values depending
10109 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10112 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10113 the handshake is continued after ssl_verify_cert_chain();
10114 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10115 can lead to 'unexplainable' connection aborts later.
10116 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10118 *) Major EVP API cipher revision.
10119 Add hooks for extra EVP features. This allows various cipher
10120 parameters to be set in the EVP interface. Support added for variable
10121 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10122 setting of RC2 and RC5 parameters.
10124 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10127 Remove lots of duplicated code from the EVP library. For example *every*
10128 cipher init() function handles the 'iv' in the same way according to the
10129 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10130 for CFB and OFB modes they zero ctx->num.
10132 New functionality allows removal of S/MIME code RC2 hack.
10134 Most of the routines have the same form and so can be declared in terms
10137 By shifting this to the top level EVP_CipherInit() it can be removed from
10138 all individual ciphers. If the cipher wants to handle IVs or keys
10139 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10142 Change lots of functions like EVP_EncryptUpdate() to now return a
10143 value: although software versions of the algorithms cannot fail
10144 any installed hardware versions can.
10147 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10148 this option is set, tolerate broken clients that send the negotiated
10149 protocol version number instead of the requested protocol version
10153 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10154 i.e. non-zero for export ciphersuites, zero otherwise.
10155 Previous versions had this flag inverted, inconsistent with
10156 rsa_tmp_cb (..._TMP_RSA_CB).
10157 [Bodo Moeller; problem reported by Amit Chopra]
10159 *) Add missing DSA library text string. Work around for some IIS
10160 key files with invalid SEQUENCE encoding.
10163 *) Add a document (doc/standards.txt) that list all kinds of standards
10164 and so on that are implemented in OpenSSL.
10167 *) Enhance c_rehash script. Old version would mishandle certificates
10168 with the same subject name hash and wouldn't handle CRLs at all.
10169 Added -fingerprint option to crl utility, to support new c_rehash
10173 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10176 *) Fix for SSL server purpose checking. Server checking was
10177 rejecting certificates which had extended key usage present
10178 but no ssl client purpose.
10179 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10181 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10182 is a little unclear about how a blank password is handled.
10183 Since the password in encoded as a BMPString with terminating
10184 double NULL a zero length password would end up as just the
10185 double NULL. However no password at all is different and is
10186 handled differently in the PKCS#12 key generation code. NS
10187 treats a blank password as zero length. MSIE treats it as no
10188 password on export: but it will try both on import. We now do
10189 the same: PKCS12_parse() tries zero length and no password if
10190 the password is set to "" or NULL (NULL is now a valid password:
10191 it wasn't before) as does the pkcs12 application.
10194 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10195 perror when PEM_read_bio_X509_REQ fails, the error message must
10196 be obtained from the error queue.
10199 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10200 it in ERR_remove_state if appropriate, and change ERR_get_state
10201 accordingly to avoid race conditions (this is necessary because
10202 thread_hash is no longer constant once set).
10205 *) Bugfix for linux-elf makefile.one.
10208 *) RSA_get_default_method() will now cause a default
10209 RSA_METHOD to be chosen if one doesn't exist already.
10210 Previously this was only set during a call to RSA_new()
10211 or RSA_new_method(NULL) meaning it was possible for
10212 RSA_get_default_method() to return NULL.
10215 *) Added native name translation to the existing DSO code
10216 that will convert (if the flag to do so is set) filenames
10217 that are sufficiently small and have no path information
10218 into a canonical native form. Eg. "blah" converted to
10219 "libblah.so" or "blah.dll" etc.
10222 *) New function ERR_error_string_n(e, buf, len) which is like
10223 ERR_error_string(e, buf), but writes at most 'len' bytes
10224 including the 0 terminator. For ERR_error_string_n, 'buf'
10226 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10228 *) CONF library reworked to become more general. A new CONF
10229 configuration file reader "class" is implemented as well as a
10230 new functions (NCONF_*, for "New CONF") to handle it. The now
10231 old CONF_* functions are still there, but are reimplemented to
10232 work in terms of the new functions. Also, a set of functions
10233 to handle the internal storage of the configuration data is
10234 provided to make it easier to write new configuration file
10235 reader "classes" (I can definitely see something reading a
10236 configuration file in XML format, for example), called _CONF_*,
10237 or "the configuration storage API"...
10239 The new configuration file reading functions are:
10241 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10242 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10244 NCONF_default, NCONF_WIN32
10246 NCONF_dump_fp, NCONF_dump_bio
10248 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10249 NCONF_new creates a new CONF object. This works in the same way
10250 as other interfaces in OpenSSL, like the BIO interface.
10251 NCONF_dump_* dump the internal storage of the configuration file,
10252 which is useful for debugging. All other functions take the same
10253 arguments as the old CONF_* functions wth the exception of the
10254 first that must be a `CONF *' instead of a `LHASH *'.
10256 To make it easer to use the new classes with the old CONF_* functions,
10257 the function CONF_set_default_method is provided.
10260 *) Add '-tls1' option to 'openssl ciphers', which was already
10261 mentioned in the documentation but had not been implemented.
10262 (This option is not yet really useful because even the additional
10263 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10266 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10267 OpenSSL-based applications) load shared libraries and bind to
10268 them in a portable way.
10269 [Geoff Thorpe, with contributions from Richard Levitte]
10271 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10273 *) Make sure _lrotl and _lrotr are only used with MSVC.
10275 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10276 (the default implementation of RAND_status).
10278 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10279 to '-clrext' (= clear extensions), as intended and documented.
10280 [Bodo Moeller; inconsistency pointed out by Michael Attili
10281 <attili@amaxo.com>]
10283 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10284 was larger than the MD block size.
10285 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10287 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10288 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10289 using the passed key: if the passed key was a private key the result
10290 of X509_print(), for example, would be to print out all the private key
10294 *) des_quad_cksum() byte order bug fix.
10295 [Ulf Möller, using the problem description in krb4-0.9.7, where
10296 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10298 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10300 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10302 *) For easily testing in shell scripts whether some command
10303 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10304 returns with exit code 0 iff no command of the given name is available.
10305 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10306 the output goes to stdout and nothing is printed to stderr.
10307 Additional arguments are always ignored.
10309 Since for each cipher there is a command of the same name,
10310 the 'no-cipher' compilation switches can be tested this way.
10312 ('openssl no-XXX' is not able to detect pseudo-commands such
10313 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10316 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10319 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10320 is set; it will be thrown away anyway because each handshake creates
10322 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10323 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10324 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10325 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10328 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10329 'Q' and 'R' lose their special meanings (quit/renegotiate).
10330 This is part of what -quiet does; unlike -quiet, -ign_eof
10331 does not suppress any output.
10334 *) Add compatibility options to the purpose and trust code. The
10335 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10336 accepts a certificate or CA, this was the previous behaviour,
10337 with all the associated security issues.
10339 X509_TRUST_COMPAT is the old trust behaviour: only and
10340 automatically trust self signed roots in certificate store. A
10341 new trust setting X509_TRUST_DEFAULT is used to specify that
10342 a purpose has no associated trust setting and it should instead
10343 use the value in the default purpose.
10346 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10347 and fix a memory leak.
10350 *) In util/mkerr.pl (which implements 'make errors'), preserve
10351 reason strings from the previous version of the .c file, as
10352 the default to have only downcase letters (and digits) in
10353 automatically generated reasons codes is not always appropriate.
10356 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10357 using strerror. Previously, ERR_reason_error_string() returned
10358 library names as reason strings for SYSerr; but SYSerr is a special
10359 case where small numbers are errno values, not library numbers.
10362 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10363 converts DSA parameters into DH parameters. (When creating parameters,
10364 DSA_generate_parameters is used.)
10367 *) Include 'length' (recommended exponent length) in C code generated
10368 by 'openssl dhparam -C'.
10371 *) The second argument to set_label in perlasm was already being used
10372 so couldn't be used as a "file scope" flag. Moved to third argument
10376 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10377 instead of RAND_bytes for encryption IVs and salts.
10380 *) Include RAND_status() into RAND_METHOD instead of implementing
10381 it only for md_rand.c Otherwise replacing the PRNG by calling
10382 RAND_set_rand_method would be impossible.
10385 *) Don't let DSA_generate_key() enter an infinite loop if the random
10386 number generation fails.
10389 *) New 'rand' application for creating pseudo-random output.
10392 *) Added configuration support for Linux/IA64
10393 [Rolf Haberrecker <rolf@suse.de>]
10395 *) Assembler module support for Mingw32.
10398 *) Shared library support for HPUX (in shlib/).
10399 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10401 *) Shared library support for Solaris gcc.
10402 [Lutz Behnke <behnke@trustcenter.de>]
10404 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10406 *) PKCS7_encrypt() was adding text MIME headers twice because they
10407 were added manually and by SMIME_crlf_copy().
10410 *) In bntest.c don't call BN_rand with zero bits argument.
10411 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10413 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10414 case was implemented. This caused BN_div_recp() to fail occasionally.
10417 *) Add an optional second argument to the set_label() in the perl
10418 assembly language builder. If this argument exists and is set
10419 to 1 it signals that the assembler should use a symbol whose
10420 scope is the entire file, not just the current function. This
10421 is needed with MASM which uses the format label:: for this scope.
10422 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10424 *) Change the ASN1 types so they are typedefs by default. Before
10425 almost all types were #define'd to ASN1_STRING which was causing
10426 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10430 *) Change names of new functions to the new get1/get0 naming
10431 convention: After 'get1', the caller owns a reference count
10432 and has to call ..._free; 'get0' returns a pointer to some
10433 data structure without incrementing reference counters.
10434 (Some of the existing 'get' functions increment a reference
10435 counter, some don't.)
10436 Similarly, 'set1' and 'add1' functions increase reference
10437 counters or duplicate objects.
10440 *) Allow for the possibility of temp RSA key generation failure:
10441 the code used to assume it always worked and crashed on failure.
10444 *) Fix potential buffer overrun problem in BIO_printf().
10445 [Ulf Möller, using public domain code by Patrick Powell; problem
10446 pointed out by David Sacerdote <das33@cornell.edu>]
10448 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10449 RAND_egd() and RAND_status(). In the command line application,
10450 the EGD socket can be specified like a seed file using RANDFILE
10454 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10455 Some CAs (e.g. Verisign) distribute certificates in this form.
10458 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10459 list to exclude them. This means that no special compilation option
10460 is needed to use anonymous DH: it just needs to be included in the
10464 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10465 EVP_MD_type. The old functionality is available in a new macro called
10466 EVP_MD_md(). Change code that uses it and update docs.
10469 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10470 where the 'void *' argument is replaced by a function pointer argument.
10471 Previously 'void *' was abused to point to functions, which works on
10472 many platforms, but is not correct. As these functions are usually
10473 called by macros defined in OpenSSL header files, most source code
10474 should work without changes.
10477 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10478 sections with information on -D... compiler switches used for
10479 compiling the library so that applications can see them. To enable
10480 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10481 must be defined. E.g.,
10482 #define OPENSSL_ALGORITHM_DEFINES
10483 #include <openssl/opensslconf.h>
10484 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10485 [Richard Levitte, Ulf and Bodo Möller]
10487 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10491 *) Change the 'other' type in certificate aux info to a STACK_OF
10492 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10493 the required ASN1 format: arbitrary types determined by an OID.
10496 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10497 argument to 'req'. This is not because the function is newer or
10498 better than others it just uses the work 'NEW' in the certificate
10499 request header lines. Some software needs this.
10502 *) Reorganise password command line arguments: now passwords can be
10503 obtained from various sources. Delete the PEM_cb function and make
10504 it the default behaviour: i.e. if the callback is NULL and the
10505 usrdata argument is not NULL interpret it as a null terminated pass
10506 phrase. If usrdata and the callback are NULL then the pass phrase
10507 is prompted for as usual.
10510 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10511 the support is automatically enabled. The resulting binaries will
10512 autodetect the card and use it if present.
10513 [Ben Laurie and Compaq Inc.]
10515 *) Work around for Netscape hang bug. This sends certificate request
10516 and server done in one record. Since this is perfectly legal in the
10517 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10518 the bugs/SSLv3 entry for more info.
10521 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10524 *) Add -rand argument to smime and pkcs12 applications and read/write
10528 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10531 *) Add command line password options to the remaining applications.
10534 *) Bug fix for BN_div_recp() for numerators with an even number of
10538 *) More tests in bntest.c, and changed test_bn output.
10541 *) ./config recognizes MacOS X now.
10544 *) Bug fix for BN_div() when the first words of num and divisor are
10545 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10548 *) Add support for various broken PKCS#8 formats, and command line
10549 options to produce them.
10552 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10553 get temporary BIGNUMs from a BN_CTX.
10556 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10560 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10561 include a #define from the old name to the new. The original intent
10562 was that statically linked binaries could for example just call
10563 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10564 link with digests. This never worked because SSLeay_add_all_digests()
10565 and SSLeay_add_all_ciphers() were in the same source file so calling
10566 one would link with the other. They are now in separate source files.
10569 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10572 *) Use a less unusual form of the Miller-Rabin primality test (it used
10573 a binary algorithm for exponentiation integrated into the Miller-Rabin
10574 loop, our standard modexp algorithms are faster).
10577 *) Support for the EBCDIC character set completed.
10578 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10580 *) Source code cleanups: use const where appropriate, eliminate casts,
10581 use void * instead of char * in lhash.
10584 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10585 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10586 this the server could overwrite ephemeral keys that the client
10590 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10591 using 50 iterations of the Rabin-Miller test.
10593 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10594 iterations of the Rabin-Miller test as required by the appendix
10595 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10596 As BN_is_prime_fasttest includes trial division, DSA parameter
10597 generation becomes much faster.
10599 This implies a change for the callback functions in DSA_is_prime
10600 and DSA_generate_parameters: The callback function is called once
10601 for each positive witness in the Rabin-Miller test, not just
10602 occasionally in the inner loop; and the parameters to the
10603 callback function now provide an iteration count for the outer
10604 loop rather than for the current invocation of the inner loop.
10605 DSA_generate_parameters additionally can call the callback
10606 function with an 'iteration count' of -1, meaning that a
10607 candidate has passed the trial division test (when q is generated
10608 from an application-provided seed, trial division is skipped).
10611 *) New function BN_is_prime_fasttest that optionally does trial
10612 division before starting the Rabin-Miller test and has
10613 an additional BN_CTX * argument (whereas BN_is_prime always
10614 has to allocate at least one BN_CTX).
10615 'callback(1, -1, cb_arg)' is called when a number has passed the
10616 trial division stage.
10619 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10623 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10626 *) New function BN_pseudo_rand().
10629 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10630 bignum version of BN_from_montgomery() with the working code from
10631 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10635 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10636 made it impossible to use the same SSL_SESSION data structure in
10637 SSL2 clients in multiple threads.
10640 *) The return value of RAND_load_file() no longer counts bytes obtained
10641 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10642 to seed the PRNG (previously an explicit byte count was required).
10643 [Ulf Möller, Bodo Möller]
10645 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10646 used (char *) instead of (void *) and had casts all over the place.
10649 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10652 *) Retain source code compatibility for BN_prime_checks macro:
10653 BN_is_prime(..., BN_prime_checks, ...) now uses
10654 BN_prime_checks_for_size to determine the appropriate number of
10655 Rabin-Miller iterations.
10658 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10659 DH_CHECK_P_NOT_SAFE_PRIME.
10660 (Check if this is true? OpenPGP calls them "strong".)
10663 *) Merge the functionality of "dh" and "gendh" programs into a new program
10664 "dhparam". The old programs are retained for now but will handle DH keys
10665 (instead of parameters) in future.
10668 *) Make the ciphers, s_server and s_client programs check the return values
10669 when a new cipher list is set.
10672 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10673 ciphers. Before when the 56bit ciphers were enabled the sorting was
10676 The syntax for the cipher sorting has been extended to support sorting by
10677 cipher-strength (using the strength_bits hard coded in the tables).
10678 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10680 Fix a bug in the cipher-command parser: when supplying a cipher command
10681 string with an "undefined" symbol (neither command nor alphanumeric
10682 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10683 an error is flagged.
10685 Due to the strength-sorting extension, the code of the
10686 ssl_create_cipher_list() function was completely rearranged. I hope that
10687 the readability was also increased :-)
10688 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10690 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10691 for the first serial number and places 2 in the serial number file. This
10692 avoids problems when the root CA is created with serial number zero and
10693 the first user certificate has the same issuer name and serial number
10697 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10698 the new code. Add documentation for this stuff.
10701 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10702 X509_*() to X509at_*() on the grounds that they don't handle X509
10703 structures and behave in an analogous way to the X509v3 functions:
10704 they shouldn't be called directly but wrapper functions should be used
10707 So we also now have some wrapper functions that call the X509at functions
10708 when passed certificate requests. (TO DO: similar things can be done with
10709 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10710 things. Some of these need some d2i or i2d and print functionality
10711 because they handle more complex structures.)
10714 *) Add missing #ifndefs that caused missing symbols when building libssl
10715 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10716 NO_RSA in ssl/s2*.c.
10717 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10719 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10720 has a return value which indicates the quality of the random data
10721 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10722 error queue. New function RAND_pseudo_bytes() generates output that is
10723 guaranteed to be unique but not unpredictable. RAND_add is like
10724 RAND_seed, but takes an extra argument for an entropy estimate
10725 (RAND_seed always assumes full entropy).
10728 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10729 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10730 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10731 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10732 false-positive rate of at most 2^-80 for random input.
10735 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10738 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10739 in the 0.9.5 release), this returns the chain
10740 from an X509_CTX structure with a dup of the stack and all
10741 the X509 reference counts upped: so the stack will exist
10742 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10745 Also make SSL_SESSION_print() print out the verify return
10749 *) Add manpage for the pkcs12 command. Also change the default
10750 behaviour so MAC iteration counts are used unless the new
10751 -nomaciter option is used. This improves file security and
10752 only older versions of MSIE (4.0 for example) need it.
10755 *) Honor the no-xxx Configure options when creating .DEF files.
10758 *) Add PKCS#10 attributes to field table: challengePassword,
10759 unstructuredName and unstructuredAddress. These are taken from
10760 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10761 international characters are used.
10763 More changes to X509_ATTRIBUTE code: allow the setting of types
10764 based on strings. Remove the 'loc' parameter when adding
10765 attributes because these will be a SET OF encoding which is sorted
10769 *) Initial changes to the 'req' utility to allow request generation
10770 automation. This will allow an application to just generate a template
10771 file containing all the field values and have req construct the
10774 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10775 used all over the place including certificate requests and PKCS#7
10776 structures. They are currently handled manually where necessary with
10777 some primitive wrappers for PKCS#7. The new functions behave in a
10778 manner analogous to the X509 extension functions: they allow
10779 attributes to be looked up by NID and added.
10781 Later something similar to the X509V3 code would be desirable to
10782 automatically handle the encoding, decoding and printing of the
10783 more complex types. The string types like challengePassword can
10784 be handled by the string table functions.
10786 Also modified the multi byte string table handling. Now there is
10787 a 'global mask' which masks out certain types. The table itself
10788 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10789 is useful when for example there is only one permissible type
10790 (as in countryName) and using the mask might result in no valid
10794 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10795 SSL_get_peer_finished to allow applications to obtain the latest
10796 Finished messages sent to the peer or expected from the peer,
10797 respectively. (SSL_get_peer_finished is usually the Finished message
10798 actually received from the peer, otherwise the protocol will be aborted.)
10800 As the Finished message are message digests of the complete handshake
10801 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10802 be used for external authentication procedures when the authentication
10803 provided by SSL/TLS is not desired or is not enough.
10806 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10807 the host supports BWX extension and if Compaq C is present on the
10808 $PATH. Just exploiting of the BWX extension results in 20-30%
10809 performance kick for some algorithms, e.g. DES and RC4 to mention
10810 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10814 *) Add support for MS "fast SGC". This is arguably a violation of the
10815 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10816 weak crypto and after checking the certificate is SGC a second one
10817 with strong crypto. MS SGC stops the first handshake after receiving
10818 the server certificate message and sends a second client hello. Since
10819 a server will typically do all the time consuming operations before
10820 expecting any further messages from the client (server key exchange
10821 is the most expensive) there is little difference between the two.
10823 To get OpenSSL to support MS SGC we have to permit a second client
10824 hello message after we have sent server done. In addition we have to
10825 reset the MAC if we do get this second client hello.
10828 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10829 if a DER encoded private key is RSA or DSA traditional format. Changed
10830 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10831 format DER encoded private key. Newer code should use PKCS#8 format which
10832 has the key type encoded in the ASN1 structure. Added DER private key
10833 support to pkcs8 application.
10836 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10837 ciphersuites has been selected (as required by the SSL 3/TLS 1
10838 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10839 is set, we interpret this as a request to violate the specification
10840 (the worst that can happen is a handshake failure, and 'correct'
10841 behaviour would result in a handshake failure anyway).
10844 *) In SSL_CTX_add_session, take into account that there might be multiple
10845 SSL_SESSION structures with the same session ID (e.g. when two threads
10846 concurrently obtain them from an external cache).
10847 The internal cache can handle only one SSL_SESSION with a given ID,
10848 so if there's a conflict, we now throw out the old one to achieve
10852 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10853 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10854 some routines that use cipher OIDs: some ciphers do not have OIDs
10855 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10859 *) Simplify the trust setting structure and code. Now we just have
10860 two sequences of OIDs for trusted and rejected settings. These will
10861 typically have values the same as the extended key usage extension
10862 and any application specific purposes.
10864 The trust checking code now has a default behaviour: it will just
10865 check for an object with the same NID as the passed id. Functions can
10866 be provided to override either the default behaviour or the behaviour
10867 for a given id. SSL client, server and email already have functions
10868 in place for compatibility: they check the NID and also return "trusted"
10869 if the certificate is self signed.
10872 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10873 traditional format into an EVP_PKEY structure.
10876 *) Add a password callback function PEM_cb() which either prompts for
10877 a password if usr_data is NULL or otherwise assumes it is a null
10878 terminated password. Allow passwords to be passed on command line
10879 environment or config files in a few more utilities.
10882 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10883 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10884 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10885 Update documentation.
10888 *) Support for ASN1 "NULL" type. This could be handled before by using
10889 ASN1_TYPE but there wasn't any function that would try to read a NULL
10890 and produce an error if it couldn't. For compatibility we also have
10891 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10892 don't allocate anything because they don't need to.
10895 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10897 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10899 *) Rebuild of the memory allocation routines used by OpenSSL code and
10900 possibly others as well. The purpose is to make an interface that
10901 provide hooks so anyone can build a separate set of allocation and
10902 deallocation routines to be used by OpenSSL, for example memory
10903 pool implementations, or something else, which was previously hard
10904 since Malloc(), Realloc() and Free() were defined as macros having
10905 the values malloc, realloc and free, respectively (except for Win32
10906 compilations). The same is provided for memory debugging code.
10907 OpenSSL already comes with functionality to find memory leaks, but
10908 this gives people a chance to debug other memory problems.
10910 With these changes, a new set of functions and macros have appeared:
10912 CRYPTO_set_mem_debug_functions() [F]
10913 CRYPTO_get_mem_debug_functions() [F]
10914 CRYPTO_dbg_set_options() [F]
10915 CRYPTO_dbg_get_options() [F]
10916 CRYPTO_malloc_debug_init() [M]
10918 The memory debug functions are NULL by default, unless the library
10919 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10920 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10921 gives the standard debugging functions that come with OpenSSL) or
10922 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10923 provided by the library user) must be used. When the standard
10924 debugging functions are used, CRYPTO_dbg_set_options can be used to
10925 request additional information:
10926 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10927 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10929 Also, things like CRYPTO_set_mem_functions will always give the
10930 expected result (the new set of functions is used for allocation
10931 and deallocation) at all times, regardless of platform and compiler
10934 To finish it up, some functions that were never use in any other
10935 way than through macros have a new API and new semantic:
10937 CRYPTO_dbg_malloc()
10938 CRYPTO_dbg_realloc()
10941 All macros of value have retained their old syntax.
10942 [Richard Levitte and Bodo Moeller]
10944 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10945 ordering of SMIMECapabilities wasn't in "strength order" and there
10946 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10950 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10951 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10952 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10954 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10955 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10956 functionality to handle multipart/signed properly) and a utility
10957 called 'smime' to call all this stuff. This is based on code I
10958 originally wrote for Celo who have kindly allowed it to be
10959 included in OpenSSL.
10962 *) Add variants des_set_key_checked and des_set_key_unchecked of
10963 des_set_key (aka des_key_sched). Global variable des_check_key
10964 decides which of these is called by des_set_key; this way
10965 des_check_key behaves as it always did, but applications and
10966 the library itself, which was buggy for des_check_key == 1,
10967 have a cleaner way to pick the version they need.
10970 *) New function PKCS12_newpass() which changes the password of a
10974 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10975 dynamic mix. In both cases the ids can be used as an index into the
10976 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10977 functions so they accept a list of the field values and the
10978 application doesn't need to directly manipulate the X509_TRUST
10982 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10986 *) Modify the way the V3 extension code looks up extensions. This now
10987 works in a similar way to the object code: we have some "standard"
10988 extensions in a static table which is searched with OBJ_bsearch()
10989 and the application can add dynamic ones if needed. The file
10990 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10991 updated whenever a new extension is added to the core code and kept
10992 in ext_nid order. There is a simple program 'tabtest.c' which checks
10993 this. New extensions are not added too often so this file can readily
10994 be maintained manually.
10996 There are two big advantages in doing things this way. The extensions
10997 can be looked up immediately and no longer need to be "added" using
10998 X509V3_add_standard_extensions(): this function now does nothing.
10999 [Side note: I get *lots* of email saying the extension code doesn't
11000 work because people forget to call this function]
11001 Also no dynamic allocation is done unless new extensions are added:
11002 so if we don't add custom extensions there is no need to call
11003 X509V3_EXT_cleanup().
11006 *) Modify enc utility's salting as follows: make salting the default. Add a
11007 magic header, so unsalted files fail gracefully instead of just decrypting
11008 to garbage. This is because not salting is a big security hole, so people
11009 should be discouraged from doing it.
11012 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11013 digest to be passed on the command line but it only used this
11014 parameter when signing a certificate. Modified so all relevant
11015 operations are affected by the digest parameter including the
11016 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11017 DSA key was used because it didn't fix the digest.
11020 *) Initial certificate chain verify code. Currently tests the untrusted
11021 certificates for consistency with the verify purpose (which is set
11022 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11024 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11025 this is because it will reject chains with invalid extensions whereas
11026 every previous version of OpenSSL and SSLeay made no checks at all.
11028 Trust code: checks the root CA for the relevant trust settings. Trust
11029 settings have an initial value consistent with the verify purpose: e.g.
11030 if the verify purpose is for SSL client use it expects the CA to be
11031 trusted for SSL client use. However the default value can be changed to
11032 permit custom trust settings: one example of this would be to only trust
11033 certificates from a specific "secure" set of CAs.
11035 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11036 which should be used for version portability: especially since the
11037 verify structure is likely to change more often now.
11039 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11040 to set them. If not set then assume SSL clients will verify SSL servers
11043 Two new options to the verify program: -untrusted allows a set of
11044 untrusted certificates to be passed in and -purpose which sets the
11045 intended purpose of the certificate. If a purpose is set then the
11046 new chain verify code is used to check extension consistency.
11049 *) Support for the authority information access extension.
11052 *) Modify RSA and DSA PEM read routines to transparently handle
11053 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11054 public keys in a format compatible with certificate
11055 SubjectPublicKeyInfo structures. Unfortunately there were already
11056 functions called *_PublicKey_* which used various odd formats so
11057 these are retained for compatibility: however the DSA variants were
11058 never in a public release so they have been deleted. Changed dsa/rsa
11059 utilities to handle the new format: note no releases ever handled public
11060 keys so we should be OK.
11062 The primary motivation for this change is to avoid the same fiasco
11063 that dogs private keys: there are several incompatible private key
11064 formats some of which are standard and some OpenSSL specific and
11065 require various evil hacks to allow partial transparent handling and
11066 even then it doesn't work with DER formats. Given the option anything
11067 other than PKCS#8 should be dumped: but the other formats have to
11068 stay in the name of compatibility.
11070 With public keys and the benefit of hindsight one standard format
11071 is used which works with EVP_PKEY, RSA or DSA structures: though
11072 it clearly returns an error if you try to read the wrong kind of key.
11074 Added a -pubkey option to the 'x509' utility to output the public key.
11075 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11076 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11077 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11078 that do the same as the EVP_PKEY_assign_*() except they up the
11079 reference count of the added key (they don't "swallow" the
11083 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11084 CRLs would fail if the file contained no certificates or no CRLs:
11085 added a new function to read in both types and return the number
11086 read: this means that if none are read it will be an error. The
11087 DER versions of the certificate and CRL reader would always fail
11088 because it isn't possible to mix certificates and CRLs in DER format
11089 without choking one or the other routine. Changed this to just read
11090 a certificate: this is the best we can do. Also modified the code
11091 in apps/verify.c to take notice of return codes: it was previously
11092 attempting to read in certificates from NULL pointers and ignoring
11093 any errors: this is one reason why the cert and CRL reader seemed
11094 to work. It doesn't check return codes from the default certificate
11095 routines: these may well fail if the certificates aren't installed.
11098 *) Code to support otherName option in GeneralName.
11101 *) First update to verify code. Change the verify utility
11102 so it warns if it is passed a self signed certificate:
11103 for consistency with the normal behaviour. X509_verify
11104 has been modified to it will now verify a self signed
11105 certificate if *exactly* the same certificate appears
11106 in the store: it was previously impossible to trust a
11107 single self signed certificate. This means that:
11108 openssl verify ss.pem
11109 now gives a warning about a self signed certificate but
11110 openssl verify -CAfile ss.pem ss.pem
11114 *) For servers, store verify_result in SSL_SESSION data structure
11115 (and add it to external session representation).
11116 This is needed when client certificate verifications fails,
11117 but an application-provided verification callback (set by
11118 SSL_CTX_set_cert_verify_callback) allows accepting the session
11119 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11120 but returns 1): When the session is reused, we have to set
11121 ssl->verify_result to the appropriate error code to avoid
11123 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11125 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11126 case in PKCS7_dataInit() where the signed PKCS7 structure
11127 didn't contain any existing data because it was being created.
11128 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11130 *) Add a salt to the key derivation routines in enc.c. This
11131 forms the first 8 bytes of the encrypted file. Also add a
11132 -S option to allow a salt to be input on the command line.
11135 *) New function X509_cmp(). Oddly enough there wasn't a function
11136 to compare two certificates. We do this by working out the SHA1
11137 hash and comparing that. X509_cmp() will be needed by the trust
11141 *) SSL_get1_session() is like SSL_get_session(), but increments
11142 the reference count in the SSL_SESSION returned.
11143 [Geoff Thorpe <geoff@eu.c2.net>]
11145 *) Fix for 'req': it was adding a null to request attributes.
11146 Also change the X509_LOOKUP and X509_INFO code to handle
11147 certificate auxiliary information.
11150 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11154 *) Add the possibility to add extra information to the memory leak
11155 detecting output, to form tracebacks, showing from where each
11156 allocation was originated: CRYPTO_push_info("constant string") adds
11157 the string plus current file name and line number to a per-thread
11158 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11159 is like calling CYRPTO_pop_info() until the stack is empty.
11160 Also updated memory leak detection code to be multi-thread-safe.
11163 *) Add options -text and -noout to pkcs7 utility and delete the
11164 encryption options which never did anything. Update docs.
11167 *) Add options to some of the utilities to allow the pass phrase
11168 to be included on either the command line (not recommended on
11169 OSes like Unix) or read from the environment. Update the
11170 manpages and fix a few bugs.
11173 *) Add a few manpages for some of the openssl commands.
11176 *) Fix the -revoke option in ca. It was freeing up memory twice,
11177 leaking and not finding already revoked certificates.
11180 *) Extensive changes to support certificate auxiliary information.
11181 This involves the use of X509_CERT_AUX structure and X509_AUX
11182 functions. An X509_AUX function such as PEM_read_X509_AUX()
11183 can still read in a certificate file in the usual way but it
11184 will also read in any additional "auxiliary information". By
11185 doing things this way a fair degree of compatibility can be
11186 retained: existing certificates can have this information added
11187 using the new 'x509' options.
11189 Current auxiliary information includes an "alias" and some trust
11190 settings. The trust settings will ultimately be used in enhanced
11191 certificate chain verification routines: currently a certificate
11192 can only be trusted if it is self signed and then it is trusted
11196 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11197 The problem was that one of the replacement routines had not been working
11198 since SSLeay releases. For now the offending routine has been replaced
11199 with non-optimised assembler. Even so, this now gives around 95%
11200 performance improvement for 1024 bit RSA signs.
11203 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11204 handling. Most clients have the effective key size in bits equal to
11205 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11206 A few however don't do this and instead use the size of the decrypted key
11207 to determine the RC2 key length and the AlgorithmIdentifier to determine
11208 the effective key length. In this case the effective key length can still
11209 be 40 bits but the key length can be 168 bits for example. This is fixed
11210 by manually forcing an RC2 key into the EVP_PKEY structure because the
11211 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11212 the key length and effective key length are equal.
11215 *) Add a bunch of functions that should simplify the creation of
11216 X509_NAME structures. Now you should be able to do:
11217 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11218 and have it automatically work out the correct field type and fill in
11219 the structures. The more adventurous can try:
11220 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11221 and it will (hopefully) work out the correct multibyte encoding.
11224 *) Change the 'req' utility to use the new field handling and multibyte
11225 copy routines. Before the DN field creation was handled in an ad hoc
11226 way in req, ca, and x509 which was rather broken and didn't support
11227 BMPStrings or UTF8Strings. Since some software doesn't implement
11228 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11229 using the dirstring_type option. See the new comment in the default
11230 openssl.cnf for more info.
11233 *) Make crypto/rand/md_rand.c more robust:
11234 - Assure unique random numbers after fork().
11235 - Make sure that concurrent threads access the global counter and
11236 md serializably so that we never lose entropy in them
11237 or use exactly the same state in multiple threads.
11238 Access to the large state is not always serializable because
11239 the additional locking could be a performance killer, and
11240 md should be large enough anyway.
11243 *) New file apps/app_rand.c with commonly needed functionality
11244 for handling the random seed file.
11246 Use the random seed file in some applications that previously did not:
11248 dsaparam -genkey (which also ignored its '-rand' option),
11251 x509 (when signing).
11252 Except on systems with /dev/urandom, it is crucial to have a random
11253 seed file at least for key creation, DSA signing, and for DH exchanges;
11254 for RSA signatures we could do without one.
11256 gendh and gendsa (unlike genrsa) used to read only the first byte
11257 of each file listed in the '-rand' option. The function as previously
11258 found in genrsa is now in app_rand.c and is used by all programs
11259 that support '-rand'.
11262 *) In RAND_write_file, use mode 0600 for creating files;
11263 don't just chmod when it may be too late.
11266 *) Report an error from X509_STORE_load_locations
11267 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11270 *) New function ASN1_mbstring_copy() this copies a string in either
11271 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11272 into an ASN1_STRING type. A mask of permissible types is passed
11273 and it chooses the "minimal" type to use or an error if not type
11277 *) Add function equivalents to the various macros in asn1.h. The old
11278 macros are retained with an M_ prefix. Code inside the library can
11279 use the M_ macros. External code (including the openssl utility)
11280 should *NOT* in order to be "shared library friendly".
11283 *) Add various functions that can check a certificate's extensions
11284 to see if it usable for various purposes such as SSL client,
11285 server or S/MIME and CAs of these types. This is currently
11286 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11287 verification. Also added a -purpose flag to x509 utility to
11288 print out all the purposes.
11291 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11295 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11296 for, obtain and decode and extension and obtain its critical flag.
11297 This allows all the necessary extension code to be handled in a
11298 single function call.
11301 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11302 platforms. See crypto/rc4/rc4_enc.c for further details.
11305 *) New -noout option to asn1parse. This causes no output to be produced
11306 its main use is when combined with -strparse and -out to extract data
11307 from a file (which may not be in ASN.1 format).
11310 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11311 when producing the local key id.
11312 [Richard Levitte <levitte@stacken.kth.se>]
11314 *) New option -dhparam in s_server. This allows a DH parameter file to be
11315 stated explicitly. If it is not stated then it tries the first server
11316 certificate file. The previous behaviour hard coded the filename
11320 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11321 a public key to be input or output. For example:
11322 openssl rsa -in key.pem -pubout -out pubkey.pem
11323 Also added necessary DSA public key functions to handle this.
11326 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11327 in the message. This was handled by allowing
11328 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11329 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11331 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11332 to the end of the strings whereas this didn't. This would cause problems
11333 if strings read with d2i_ASN1_bytes() were later modified.
11334 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11336 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11337 data and it contains EOF it will end up returning an error. This is
11338 caused by input 46 bytes long. The cause is due to the way base64
11339 BIOs find the start of base64 encoded data. They do this by trying a
11340 trial decode on each line until they find one that works. When they
11341 do a flag is set and it starts again knowing it can pass all the
11342 data directly through the decoder. Unfortunately it doesn't reset
11343 the context it uses. This means that if EOF is reached an attempt
11344 is made to pass two EOFs through the context and this causes the
11345 resulting error. This can also cause other problems as well. As is
11346 usual with these problems it takes *ages* to find and the fix is
11347 trivial: move one line.
11348 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11350 *) Ugly workaround to get s_client and s_server working under Windows. The
11351 old code wouldn't work because it needed to select() on sockets and the
11352 tty (for keypresses and to see if data could be written). Win32 only
11353 supports select() on sockets so we select() with a 1s timeout on the
11354 sockets and then see if any characters are waiting to be read, if none
11355 are present then we retry, we also assume we can always write data to
11356 the tty. This isn't nice because the code then blocks until we've
11357 received a complete line of data and it is effectively polling the
11358 keyboard at 1s intervals: however it's quite a bit better than not
11359 working at all :-) A dedicated Windows application might handle this
11360 with an event loop for example.
11363 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11364 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11365 will be called when RSA_sign() and RSA_verify() are used. This is useful
11366 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11367 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11368 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11369 This necessitated the support of an extra signature type NID_md5_sha1
11370 for SSL signatures and modifications to the SSL library to use it instead
11371 of calling RSA_public_decrypt() and RSA_private_encrypt().
11374 *) Add new -verify -CAfile and -CApath options to the crl program, these
11375 will lookup a CRL issuers certificate and verify the signature in a
11376 similar way to the verify program. Tidy up the crl program so it
11377 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11378 less strict. It will now permit CRL extensions even if it is not
11379 a V2 CRL: this will allow it to tolerate some broken CRLs.
11382 *) Initialize all non-automatic variables each time one of the openssl
11383 sub-programs is started (this is necessary as they may be started
11384 multiple times from the "OpenSSL>" prompt).
11385 [Lennart Bang, Bodo Moeller]
11387 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11388 removing all other RSA functionality (this is what NO_RSA does). This
11389 is so (for example) those in the US can disable those operations covered
11390 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11394 *) Non-copying interface to BIO pairs.
11395 (still largely untested)
11398 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11399 ASCII string. This was handled independently in various places before.
11402 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11403 UTF8 strings a character at a time.
11406 *) Use client_version from client hello to select the protocol
11407 (s23_srvr.c) and for RSA client key exchange verification
11408 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11411 *) Add various utility functions to handle SPKACs, these were previously
11412 handled by poking round in the structure internals. Added new function
11413 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11414 print, verify and generate SPKACs. Based on an original idea from
11415 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11418 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11421 *) Allow the config file extension section to be overwritten on the
11422 command line. Based on an original idea from Massimiliano Pala
11423 <madwolf@comune.modena.it>. The new option is called -extensions
11424 and can be applied to ca, req and x509. Also -reqexts to override
11425 the request extensions in req and -crlexts to override the crl extensions
11429 *) Add new feature to the SPKAC handling in ca. Now you can include
11430 the same field multiple times by preceding it by "XXXX." for example:
11433 this is the same syntax as used in the req config file.
11436 *) Allow certificate extensions to be added to certificate requests. These
11437 are specified in a 'req_extensions' option of the req section of the
11438 config file. They can be printed out with the -text option to req but
11439 are otherwise ignored at present.
11442 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11443 data read consists of only the final block it would not decrypted because
11444 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11445 A misplaced 'break' also meant the decrypted final block might not be
11446 copied until the next read.
11449 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11450 a few extra parameters to the DH structure: these will be useful if
11451 for example we want the value of 'q' or implement X9.42 DH.
11454 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11455 provides hooks that allow the default DSA functions or functions on a
11456 "per key" basis to be replaced. This allows hardware acceleration and
11457 hardware key storage to be handled without major modification to the
11458 library. Also added low level modexp hooks and CRYPTO_EX structure and
11459 associated functions.
11462 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11463 as "read only": it can't be written to and the buffer it points to will
11464 not be freed. Reading from a read only BIO is much more efficient than
11465 a normal memory BIO. This was added because there are several times when
11466 an area of memory needs to be read from a BIO. The previous method was
11467 to create a memory BIO and write the data to it, this results in two
11468 copies of the data and an O(n^2) reading algorithm. There is a new
11469 function BIO_new_mem_buf() which creates a read only memory BIO from
11470 an area of memory. Also modified the PKCS#7 routines to use read only
11474 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11475 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11476 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11477 but a retry condition occurred while trying to read the rest.
11480 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11481 NID_pkcs7_encrypted by default: this was wrong since this should almost
11482 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11483 the encrypted data type: this is a more sensible place to put it and it
11484 allows the PKCS#12 code to be tidied up that duplicated this
11488 *) Changed obj_dat.pl script so it takes its input and output files on
11489 the command line. This should avoid shell escape redirection problems
11493 *) Initial support for certificate extension requests, these are included
11494 in things like Xenroll certificate requests. Included functions to allow
11495 extensions to be obtained and added.
11498 *) -crlf option to s_client and s_server for sending newlines as
11499 CRLF (as required by many protocols).
11502 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11504 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11505 [Ralf S. Engelschall]
11507 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11508 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11510 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11514 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11515 DH parameters/keys (q is lost during that conversion, but the resulting
11516 DH parameters contain its length).
11518 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11519 much faster than DH_generate_parameters (which creates parameters
11520 where p = 2*q + 1), and also the smaller q makes DH computations
11521 much more efficient (160-bit exponentiation instead of 1024-bit
11522 exponentiation); so this provides a convenient way to support DHE
11523 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11524 utter importance to use
11525 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11527 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11528 when such DH parameters are used, because otherwise small subgroup
11529 attacks may become possible!
11532 *) Avoid memory leak in i2d_DHparams.
11535 *) Allow the -k option to be used more than once in the enc program:
11536 this allows the same encrypted message to be read by multiple recipients.
11539 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11540 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11541 it will always use the numerical form of the OID, even if it has a short
11545 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11546 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11547 otherwise bn_mod_exp was called. In the case of hardware keys for example
11548 no private key components need be present and it might store extra data
11549 in the RSA structure, which cannot be accessed from bn_mod_exp.
11550 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11551 private key operations.
11554 *) Added support for SPARC Linux.
11557 *) pem_password_cb function type incompatibly changed from
11558 typedef int pem_password_cb(char *buf, int size, int rwflag);
11560 ....(char *buf, int size, int rwflag, void *userdata);
11561 so that applications can pass data to their callbacks:
11562 The PEM[_ASN1]_{read,write}... functions and macros now take an
11563 additional void * argument, which is just handed through whenever
11564 the password callback is called.
11565 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11567 New function SSL_CTX_set_default_passwd_cb_userdata.
11569 Compatibility note: As many C implementations push function arguments
11570 onto the stack in reverse order, the new library version is likely to
11571 interoperate with programs that have been compiled with the old
11572 pem_password_cb definition (PEM_whatever takes some data that
11573 happens to be on the stack as its last argument, and the callback
11574 just ignores this garbage); but there is no guarantee whatsoever that
11577 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11578 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11579 problems not only on Windows, but also on some Unix platforms.
11580 To avoid problematic command lines, these definitions are now in an
11581 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11582 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11585 *) MIPS III/IV assembler module is reimplemented.
11588 *) More DES library cleanups: remove references to srand/rand and
11589 delete an unused file.
11592 *) Add support for the free Netwide assembler (NASM) under Win32,
11593 since not many people have MASM (ml) and it can be hard to obtain.
11594 This is currently experimental but it seems to work OK and pass all
11595 the tests. Check out INSTALL.W32 for info.
11598 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11599 without temporary keys kept an extra copy of the server key,
11600 and connections with temporary keys did not free everything in case
11604 *) New function RSA_check_key and new openssl rsa option -check
11605 for verifying the consistency of RSA keys.
11606 [Ulf Moeller, Bodo Moeller]
11608 *) Various changes to make Win32 compile work:
11609 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11610 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11611 comparison" warnings.
11612 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11615 *) Add a debugging option to PKCS#5 v2 key generation function: when
11616 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11617 derived keys are printed to stderr.
11620 *) Copy the flags in ASN1_STRING_dup().
11621 [Roman E. Pavlov <pre@mo.msk.ru>]
11623 *) The x509 application mishandled signing requests containing DSA
11624 keys when the signing key was also DSA and the parameters didn't match.
11626 It was supposed to omit the parameters when they matched the signing key:
11627 the verifying software was then supposed to automatically use the CA's
11628 parameters if they were absent from the end user certificate.
11630 Omitting parameters is no longer recommended. The test was also
11631 the wrong way round! This was probably due to unusual behaviour in
11632 EVP_cmp_parameters() which returns 1 if the parameters match.
11633 This meant that parameters were omitted when they *didn't* match and
11634 the certificate was useless. Certificates signed with 'ca' didn't have
11636 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11638 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11639 The interface is as follows:
11640 Applications can use
11641 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11642 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11643 "off" is now the default.
11644 The library internally uses
11645 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11646 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11647 to disable memory-checking temporarily.
11649 Some inconsistent states that previously were possible (and were
11650 even the default) are now avoided.
11652 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11653 with each memory chunk allocated; this is occasionally more helpful
11654 than just having a counter.
11656 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11658 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11662 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11663 which largely parallels "options", but is for changing API behaviour,
11664 whereas "options" are about protocol behaviour.
11665 Initial "mode" flags are:
11667 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11668 a single record has been written.
11669 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11670 retries use the same buffer location.
11671 (But all of the contents must be
11675 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11678 *) Fix problems with no-hmac etc.
11679 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11681 *) New functions RSA_get_default_method(), RSA_set_method() and
11682 RSA_get_method(). These allows replacement of RSA_METHODs without having
11683 to mess around with the internals of an RSA structure.
11686 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11687 Also really enable memory leak checks in openssl.c and in some
11689 [Chad C. Mulligan, Bodo Moeller]
11691 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11692 up the length of negative integers. This has now been simplified to just
11693 store the length when it is first determined and use it later, rather
11694 than trying to keep track of where data is copied and updating it to
11696 [Steve Henson, reported by Brien Wheeler
11697 <bwheeler@authentica-security.com>]
11699 *) Add a new function PKCS7_signatureVerify. This allows the verification
11700 of a PKCS#7 signature but with the signing certificate passed to the
11701 function itself. This contrasts with PKCS7_dataVerify which assumes the
11702 certificate is present in the PKCS#7 structure. This isn't always the
11703 case: certificates can be omitted from a PKCS#7 structure and be
11704 distributed by "out of band" means (such as a certificate database).
11707 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11708 function prototypes in pem.h, also change util/mkdef.pl to add the
11709 necessary function names.
11712 *) mk1mf.pl (used by Windows builds) did not properly read the
11713 options set by Configure in the top level Makefile, and Configure
11714 was not even able to write more than one option correctly.
11715 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11718 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11719 file to be loaded from a BIO or FILE pointer. The BIO version will
11720 for example allow memory BIOs to contain config info.
11723 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11724 Whoever hopes to achieve shared-library compatibility across versions
11725 must use this, not the compile-time macro.
11726 (Exercise 0.9.4: Which is the minimum library version required by
11728 Note: All this applies only to multi-threaded programs, others don't
11732 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11733 through a BIO pair triggered the default case, i.e.
11734 SSLerr(...,SSL_R_UNKNOWN_STATE).
11737 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11738 can use the SSL library even if none of the specific BIOs is
11742 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11743 for the encoded length.
11744 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11746 *) Add initial documentation of the X509V3 functions.
11749 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11750 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11751 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11752 secure PKCS#8 private key format with a high iteration count.
11755 *) Fix determination of Perl interpreter: A perl or perl5
11756 _directory_ in $PATH was also accepted as the interpreter.
11757 [Ralf S. Engelschall]
11759 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11760 wrong with it but it was very old and did things like calling
11761 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11762 unusual formatting.
11765 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11766 to use the new extension code.
11769 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11770 with macros. This should make it easier to change their form, add extra
11771 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11775 *) Add to configuration table a new entry that can specify an alternative
11776 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11777 according to Mark Crispin <MRC@Panda.COM>.
11781 *) DES CBC did not update the IV. Weird.
11784 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11785 Changing the behaviour of the former might break existing programs --
11786 where IV updating is needed, des_ncbc_encrypt can be used.
11789 *) When bntest is run from "make test" it drives bc to check its
11790 calculations, as well as internally checking them. If an internal check
11791 fails, it needs to cause bc to give a non-zero result or make test carries
11792 on without noticing the failure. Fixed.
11795 *) DES library cleanups.
11798 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11799 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11800 ciphers. NOTE: although the key derivation function has been verified
11801 against some published test vectors it has not been extensively tested
11802 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11806 *) Instead of "mkdir -p", which is not fully portable, use new
11807 Perl script "util/mkdir-p.pl".
11810 *) Rewrite the way password based encryption (PBE) is handled. It used to
11811 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11812 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11813 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11814 the 'parameter' field of the AlgorithmIdentifier is passed to the
11815 underlying key generation function so it must do its own ASN1 parsing.
11816 This has also changed the EVP_PBE_CipherInit() function which now has a
11817 'parameter' argument instead of literal salt and iteration count values
11818 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11821 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11822 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11823 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11824 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11825 value was just used as a "magic string" and not used directly its
11826 value doesn't matter.
11829 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11833 *) "linux-sparc64" configuration (ultrapenguin).
11834 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11835 "linux-sparc" configuration.
11836 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11838 *) config now generates no-xxx options for missing ciphers.
11841 *) Support the EBCDIC character set (work in progress).
11842 File ebcdic.c not yet included because it has a different license.
11843 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11845 *) Support BS2000/OSD-POSIX.
11846 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11848 *) Make callbacks for key generation use void * instead of char *.
11851 *) Make S/MIME samples compile (not yet tested).
11854 *) Additional typesafe stacks.
11857 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11861 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11863 *) New configuration variant "sco5-gcc".
11865 *) Updated some demos.
11866 [Sean O Riordain, Wade Scholine]
11868 *) Add missing BIO_free at exit of pkcs12 application.
11871 *) Fix memory leak in conf.c.
11874 *) Updates for Win32 to assembler version of MD5.
11877 *) Set #! path to perl in apps/der_chop to where we found it
11878 instead of using a fixed path.
11881 *) SHA library changes for irix64-mips4-cc.
11884 *) Improvements for VMS support.
11888 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11890 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11891 This also avoids the problems with SC4.2 and unpatched SC5.
11892 [Andy Polyakov <appro@fy.chalmers.se>]
11894 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11895 These are required because of the typesafe stack would otherwise break
11896 existing code. If old code used a structure member which used to be STACK
11897 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11898 sk_num or sk_value it would produce an error because the num, data members
11899 are not present in STACK_OF. Now it just produces a warning. sk_set
11900 replaces the old method of assigning a value to sk_value
11901 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11902 that does this will no longer work (and should use sk_set instead) but
11903 this could be regarded as a "questionable" behaviour anyway.
11906 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11907 correctly handle encrypted S/MIME data.
11910 *) Change type of various DES function arguments from des_cblock
11911 (which means, in function argument declarations, pointer to char)
11912 to des_cblock * (meaning pointer to array with 8 char elements),
11913 which allows the compiler to do more typechecking; it was like
11914 that back in SSLeay, but with lots of ugly casts.
11916 Introduce new type const_des_cblock.
11919 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11920 problems: find RecipientInfo structure that matches recipient certificate
11921 and initialise the ASN1 structures properly based on passed cipher.
11924 *) Belatedly make the BN tests actually check the results.
11927 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11928 to and from BNs: it was completely broken. New compilation option
11929 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11930 key elements as negative integers.
11933 *) Reorganize and speed up MD5.
11934 [Andy Polyakov <appro@fy.chalmers.se>]
11937 [Richard Levitte <richard@levitte.org>]
11939 *) New option -out to asn1parse to allow the parsed structure to be
11940 output to a file. This is most useful when combined with the -strparse
11941 option to examine the output of things like OCTET STRINGS.
11944 *) Make SSL library a little more fool-proof by not requiring any longer
11945 that SSL_set_{accept,connect}_state be called before
11946 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11947 in many applications because usually everything *appeared* to work as
11948 intended anyway -- now it really works as intended).
11951 *) Move openssl.cnf out of lib/.
11954 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11955 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11956 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11957 [Ralf S. Engelschall]
11959 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11960 handle PKCS#7 enveloped data properly.
11961 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11963 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11964 copying pointers. The cert_st handling is changed by this in
11965 various ways (and thus what used to be known as ctx->default_cert
11966 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11967 any longer when s->cert does not give us what we need).
11968 ssl_cert_instantiate becomes obsolete by this change.
11969 As soon as we've got the new code right (possibly it already is?),
11970 we have solved a couple of bugs of the earlier code where s->cert
11971 was used as if it could not have been shared with other SSL structures.
11973 Note that using the SSL API in certain dirty ways now will result
11974 in different behaviour than observed with earlier library versions:
11975 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11976 does not influence s as it used to.
11978 In order to clean up things more thoroughly, inside SSL_SESSION
11979 we don't use CERT any longer, but a new structure SESS_CERT
11980 that holds per-session data (if available); currently, this is
11981 the peer's certificate chain and, for clients, the server's certificate
11982 and temporary key. CERT holds only those values that can have
11983 meaningful defaults in an SSL_CTX.
11986 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11987 from the internal representation. Various PKCS#7 fixes: remove some
11988 evil casts and set the enc_dig_alg field properly based on the signing
11992 *) Allow PKCS#12 password to be set from the command line or the
11993 environment. Let 'ca' get its config file name from the environment
11994 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11998 *) Allow certificate policies extension to use an IA5STRING for the
11999 organization field. This is contrary to the PKIX definition but
12000 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12004 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12005 without disallowing inline assembler and the like for non-pedantic builds.
12008 *) Support Borland C++ builder.
12009 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12011 *) Support Mingw32.
12014 *) SHA-1 cleanups and performance enhancements.
12015 [Andy Polyakov <appro@fy.chalmers.se>]
12017 *) Sparc v8plus assembler for the bignum library.
12018 [Andy Polyakov <appro@fy.chalmers.se>]
12020 *) Accept any -xxx and +xxx compiler options in Configure.
12023 *) Update HPUX configuration.
12026 *) Add missing sk_<type>_unshift() function to safestack.h
12027 [Ralf S. Engelschall]
12029 *) New function SSL_CTX_use_certificate_chain_file that sets the
12030 "extra_cert"s in addition to the certificate. (This makes sense
12031 only for "PEM" format files, as chains as a whole are not
12035 *) Support verify_depth from the SSL API.
12036 x509_vfy.c had what can be considered an off-by-one-error:
12037 Its depth (which was not part of the external interface)
12038 was actually counting the number of certificates in a chain;
12039 now it really counts the depth.
12042 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12043 instead of X509err, which often resulted in confusing error
12044 messages since the error codes are not globally unique
12045 (e.g. an alleged error in ssl3_accept when a certificate
12046 didn't match the private key).
12048 *) New function SSL_CTX_set_session_id_context that allows to set a default
12049 value (so that you don't need SSL_set_session_id_context for each
12050 connection using the SSL_CTX).
12053 *) OAEP decoding bug fix.
12056 *) Support INSTALL_PREFIX for package builders, as proposed by
12060 *) New Configure options "threads" and "no-threads". For systems
12061 where the proper compiler options are known (currently Solaris
12062 and Linux), "threads" is the default.
12065 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12068 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12069 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12070 such as /usr/local/bin.
12073 *) "make linux-shared" to build shared libraries.
12074 [Niels Poppe <niels@netbox.org>]
12076 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12079 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12080 extension adding in x509 utility.
12083 *) Remove NOPROTO sections and error code comments.
12086 *) Partial rewrite of the DEF file generator to now parse the ANSI
12090 *) New Configure options --prefix=DIR and --openssldir=DIR.
12093 *) Complete rewrite of the error code script(s). It is all now handled
12094 by one script at the top level which handles error code gathering,
12095 header rewriting and C source file generation. It should be much better
12096 than the old method: it now uses a modified version of Ulf's parser to
12097 read the ANSI prototypes in all header files (thus the old K&R definitions
12098 aren't needed for error creation any more) and do a better job of
12099 translating function codes into names. The old 'ASN1 error code imbedded
12100 in a comment' is no longer necessary and it doesn't use .err files which
12101 have now been deleted. Also the error code call doesn't have to appear all
12102 on one line (which resulted in some large lines...).
12105 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12108 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12109 0 (which usually indicates a closed connection), but continue reading.
12112 *) Fix some race conditions.
12115 *) Add support for CRL distribution points extension. Add Certificate
12116 Policies and CRL distribution points documentation.
12119 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12122 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12123 8 of keying material. Merlin has also confirmed interop with this fix
12124 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12125 [Merlin Hughes <merlin@baltimore.ie>]
12127 *) Fix lots of warnings.
12128 [Richard Levitte <levitte@stacken.kth.se>]
12130 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12131 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12132 [Richard Levitte <levitte@stacken.kth.se>]
12134 *) Fix problems with sizeof(long) == 8.
12135 [Andy Polyakov <appro@fy.chalmers.se>]
12137 *) Change functions to ANSI C.
12140 *) Fix typos in error codes.
12141 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12143 *) Remove defunct assembler files from Configure.
12146 *) SPARC v8 assembler BIGNUM implementation.
12147 [Andy Polyakov <appro@fy.chalmers.se>]
12149 *) Support for Certificate Policies extension: both print and set.
12150 Various additions to support the r2i method this uses.
12153 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12154 return a const string when you are expecting an allocated buffer.
12157 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12158 types DirectoryString and DisplayText.
12161 *) Add code to allow r2i extensions to access the configuration database,
12162 add an LHASH database driver and add several ctx helper functions.
12165 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12166 fail when they extended the size of a BIGNUM.
12169 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12170 support typesafe stack.
12173 *) Fix typo in SSL_[gs]et_options().
12174 [Nils Frostberg <nils@medcom.se>]
12176 *) Delete various functions and files that belonged to the (now obsolete)
12177 old X509V3 handling code.
12180 *) New Configure option "rsaref".
12183 *) Don't auto-generate pem.h.
12186 *) Introduce type-safe ASN.1 SETs.
12189 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12190 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12192 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12193 that links with OpenSSL (well at least cause lots of warnings), but fear
12194 not: the conversion is trivial, and it eliminates loads of evil casts. A
12195 few STACKed things have been converted already. Feel free to convert more.
12196 In the fullness of time, I'll do away with the STACK type altogether.
12199 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12200 specified in <certfile> by updating the entry in the index.txt file.
12201 This way one no longer has to edit the index.txt file manually for
12202 revoking a certificate. The -revoke option does the gory details now.
12203 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12205 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12206 `-text' option at all and this way the `-noout -text' combination was
12207 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12208 [Ralf S. Engelschall]
12210 *) Make sure a corresponding plain text error message exists for the
12211 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12212 verify callback function determined that a certificate was revoked.
12213 [Ralf S. Engelschall]
12215 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12216 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12217 all available ciphers including rc5, which was forgotten until now.
12218 In order to let the testing shell script know which algorithms
12219 are available, a new (up to now undocumented) command
12220 "openssl list-cipher-commands" is used.
12223 *) Bugfix: s_client occasionally would sleep in select() when
12224 it should have checked SSL_pending() first.
12227 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12228 the raw DSA values prior to ASN.1 encoding.
12231 *) Tweaks to Configure
12232 [Niels Poppe <niels@netbox.org>]
12234 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12238 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12241 *) New config option to avoid instructions that are illegal on the 80386.
12242 The default code is faster, but requires at least a 486.
12245 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12246 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12247 same as SSL2_VERSION anyway.
12250 *) New "-showcerts" option for s_client.
12253 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12254 application. Various cleanups and fixes.
12257 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12258 modify error routines to work internally. Add error codes and PBE init
12259 to library startup routines.
12262 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12263 packing functions to asn1 and evp. Changed function names and error
12264 codes along the way.
12267 *) PKCS12 integration: and so it begins... First of several patches to
12268 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12269 objects to objects.h
12272 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12273 and display support for Thawte strong extranet extension.
12276 *) Add LinuxPPC support.
12277 [Jeff Dubrule <igor@pobox.org>]
12279 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12280 bn_div_words in alpha.s.
12281 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12283 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12284 OAEP isn't supported when OpenSSL is built with RSAref.
12285 [Ulf Moeller <ulf@fitug.de>]
12287 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12288 so they no longer are missing under -DNOPROTO.
12289 [Soren S. Jorvang <soren@t.dk>]
12292 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12294 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12295 doesn't work when the session is reused. Coming soon!
12298 *) Fix a security hole, that allows sessions to be reused in the wrong
12299 context thus bypassing client cert protection! All software that uses
12300 client certs and session caches in multiple contexts NEEDS PATCHING to
12301 allow session reuse! A fuller solution is in the works.
12302 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12304 *) Some more source tree cleanups (removed obsolete files
12305 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12306 permission on "config" script to be executable) and a fix for the INSTALL
12308 [Ulf Moeller <ulf@fitug.de>]
12310 *) Remove some legacy and erroneous uses of malloc, free instead of
12312 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12314 *) Make rsa_oaep_test return non-zero on error.
12315 [Ulf Moeller <ulf@fitug.de>]
12317 *) Add support for native Solaris shared libraries. Configure
12318 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12319 if someone would make that last step automatic.
12320 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12322 *) ctx_size was not built with the right compiler during "make links". Fixed.
12325 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12326 except NULL ciphers". This means the default cipher list will no longer
12327 enable NULL ciphers. They need to be specifically enabled e.g. with
12328 the string "DEFAULT:eNULL".
12331 *) Fix to RSA private encryption routines: if p < q then it would
12332 occasionally produce an invalid result. This will only happen with
12333 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12336 *) Be less restrictive and allow also `perl util/perlpath.pl
12337 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12338 because this way one can also use an interpreter named `perl5' (which is
12339 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12340 installed as `perl').
12341 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12343 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12344 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12346 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12347 advapi32.lib to Win32 build and change the pem test comparison
12348 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12349 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12350 and crypto/des/ede_cbcm_enc.c.
12353 *) DES quad checksum was broken on big-endian architectures. Fixed.
12356 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12357 Win32 test batch file so it (might) work again. The Win32 test batch file
12358 is horrible: I feel ill....
12361 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12362 in e_os.h. Audit of header files to check ANSI and non ANSI
12363 sections: 10 functions were absent from non ANSI section and not exported
12364 from Windows DLLs. Fixed up libeay.num for new functions.
12367 *) Make `openssl version' output lines consistent.
12368 [Ralf S. Engelschall]
12370 *) Fix Win32 symbol export lists for BIO functions: Added
12371 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12372 to ms/libeay{16,32}.def.
12373 [Ralf S. Engelschall]
12375 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12376 fine under Unix and passes some trivial tests I've now added. But the
12377 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12378 added to make sure no one expects that this stuff really works in the
12379 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12380 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12382 [Ralf S. Engelschall]
12384 *) Fix the generation of two part addresses in perl.
12385 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12387 *) Add config entry for Linux on MIPS.
12388 [John Tobey <jtobey@channel1.com>]
12390 *) Make links whenever Configure is run, unless we are on Windoze.
12393 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12394 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12398 *) Add a useful kludge to allow package maintainers to specify compiler and
12399 other platforms details on the command line without having to patch the
12400 Configure script everytime: One now can use ``perl Configure
12401 <id>:<details>'', i.e. platform ids are allowed to have details appended
12402 to them (separated by colons). This is treated as there would be a static
12403 pre-configured entry in Configure's %table under key <id> with value
12404 <details> and ``perl Configure <id>'' is called. So, when you want to
12405 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12406 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12407 now, which overrides the FreeBSD-elf entry on-the-fly.
12408 [Ralf S. Engelschall]
12410 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12413 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12414 on the `perl Configure ...' command line. This way one can compile
12415 OpenSSL libraries with Position Independent Code (PIC) which is needed
12416 for linking it into DSOs.
12417 [Ralf S. Engelschall]
12419 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12423 *) Cleaned up the LICENSE document: The official contact for any license
12424 questions now is the OpenSSL core team under openssl-core@openssl.org.
12425 And add a paragraph about the dual-license situation to make sure people
12426 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12427 to the OpenSSL toolkit.
12428 [Ralf S. Engelschall]
12430 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12431 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12432 Additionally cleaned up the `make links' target: Remove unnecessary
12433 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12434 to speed processing and no longer clutter the display with confusing
12435 stuff. Instead only the actually done links are displayed.
12436 [Ralf S. Engelschall]
12438 *) Permit null encryption ciphersuites, used for authentication only. It used
12439 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12440 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12444 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12445 signed attributes when verifying signatures (this would break them),
12446 the detached data encoding was wrong and public keys obtained using
12447 X509_get_pubkey() weren't freed.
12450 *) Add text documentation for the BUFFER functions. Also added a work around
12451 to a Win95 console bug. This was triggered by the password read stuff: the
12452 last character typed gets carried over to the next fread(). If you were
12453 generating a new cert request using 'req' for example then the last
12454 character of the passphrase would be CR which would then enter the first
12458 *) Added the new `Includes OpenSSL Cryptography Software' button as
12459 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12460 button and can be used by applications based on OpenSSL to show the
12461 relationship to the OpenSSL project.
12462 [Ralf S. Engelschall]
12464 *) Remove confusing variables in function signatures in files
12465 ssl/ssl_lib.c and ssl/ssl.h.
12466 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12468 *) Don't install bss_file.c under PREFIX/include/
12469 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12471 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12472 functions that return function pointers and has support for NT specific
12473 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12474 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12475 unsigned to signed types: this was killing the Win32 compile.
12478 *) Add new certificate file to stack functions,
12479 SSL_add_dir_cert_subjects_to_stack() and
12480 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12481 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12482 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12483 This means that Apache-SSL and similar packages don't have to mess around
12484 to add as many CAs as they want to the preferred list.
12487 *) Experiment with doxygen documentation. Currently only partially applied to
12489 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12490 openssl.doxy as the configuration file.
12493 *) Get rid of remaining C++-style comments which strict C compilers hate.
12494 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12496 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12497 compiled in by default: it has problems with large keys.
12500 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12501 DH private keys and/or callback functions which directly correspond to
12502 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12503 is needed for applications which have to configure certificates on a
12504 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12506 For the RSA certificate situation is makes no difference, but
12507 for the DSA certificate situation this fixes the "no shared cipher"
12508 problem where the OpenSSL cipher selection procedure failed because the
12509 temporary keys were not overtaken from the context and the API provided
12510 no way to reconfigure them.
12511 The new functions now let applications reconfigure the stuff and they
12512 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12513 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12514 non-public-API function ssl_cert_instantiate() is used as a helper
12515 function and also to reduce code redundancy inside ssl_rsa.c.
12516 [Ralf S. Engelschall]
12518 *) Move s_server -dcert and -dkey options out of the undocumented feature
12519 area because they are useful for the DSA situation and should be
12520 recognized by the users.
12521 [Ralf S. Engelschall]
12523 *) Fix the cipher decision scheme for export ciphers: the export bits are
12524 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12525 SSL_EXP_MASK. So, the original variable has to be used instead of the
12526 already masked variable.
12527 [Richard Levitte <levitte@stacken.kth.se>]
12529 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12530 [Richard Levitte <levitte@stacken.kth.se>]
12532 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12533 from `int' to `unsigned int' because it's a length and initialized by
12534 EVP_DigestFinal() which expects an `unsigned int *'.
12535 [Richard Levitte <levitte@stacken.kth.se>]
12537 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12538 script. Instead use the usual Shell->Perl transition trick.
12539 [Ralf S. Engelschall]
12541 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12542 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12543 -noout -modulus' as it's already the case for `openssl rsa -noout
12544 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12545 currently the public key is printed (a decision which was already done by
12546 `openssl dsa -modulus' in the past) which serves a similar purpose.
12547 Additionally the NO_RSA no longer completely removes the whole -modulus
12548 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12550 [Ralf S. Engelschall]
12552 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12553 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12554 [Arne Ansper <arne@ats.cyber.ee>]
12556 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12557 to be added. Now both 'req' and 'ca' can use new objects defined in the
12561 *) Add cool BIO that does syslog (or event log on NT).
12562 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12564 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12565 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12566 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12567 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12570 *) Add preliminary config info for new extension code.
12573 *) Make RSA_NO_PADDING really use no padding.
12574 [Ulf Moeller <ulf@fitug.de>]
12576 *) Generate errors when private/public key check is done.
12579 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12580 for some CRL extensions and new objects added.
12583 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12584 key usage extension and fuller support for authority key id.
12587 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12588 padding method for RSA, which is recommended for new applications in PKCS
12589 #1 v2.0 (RFC 2437, October 1998).
12590 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12591 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12592 against Bleichbacher's attack on RSA.
12593 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12596 *) Updates to the new SSL compression code
12597 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12599 *) Fix so that the version number in the master secret, when passed
12600 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12601 (because the server will not accept higher), that the version number
12602 is 0x03,0x01, not 0x03,0x00
12603 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12605 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12606 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12607 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12610 *) Support for RAW extensions where an arbitrary extension can be
12611 created by including its DER encoding. See apps/openssl.cnf for
12615 *) Make sure latest Perl versions don't interpret some generated C array
12616 code as Perl array code in the crypto/err/err_genc.pl script.
12617 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12619 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12620 not many people have the assembler. Various Win32 compilation fixes and
12621 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12622 build instructions.
12625 *) Modify configure script 'Configure' to automatically create crypto/date.h
12626 file under Win32 and also build pem.h from pem.org. New script
12627 util/mkfiles.pl to create the MINFO file on environments that can't do a
12628 'make files': perl util/mkfiles.pl >MINFO should work.
12631 *) Major rework of DES function declarations, in the pursuit of correctness
12632 and purity. As a result, many evil casts evaporated, and some weirdness,
12633 too. You may find this causes warnings in your code. Zapping your evil
12634 casts will probably fix them. Mostly.
12637 *) Fix for a typo in asn1.h. Bug fix to object creation script
12638 obj_dat.pl. It considered a zero in an object definition to mean
12639 "end of object": none of the objects in objects.h have any zeros
12640 so it wasn't spotted.
12641 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12643 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12644 Masking (CBCM). In the absence of test vectors, the best I have been able
12645 to do is check that the decrypt undoes the encrypt, so far. Send me test
12646 vectors if you have them.
12649 *) Correct calculation of key length for export ciphers (too much space was
12650 allocated for null ciphers). This has not been tested!
12653 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12654 message is now correct (it understands "crypto" and "ssl" on its
12655 command line). There is also now an "update" option. This will update
12656 the util/ssleay.num and util/libeay.num files with any new functions.
12658 perl util/mkdef.pl crypto ssl update
12659 it will update them.
12662 *) Overhauled the Perl interface (perl/*):
12663 - ported BN stuff to OpenSSL's different BN library
12664 - made the perl/ source tree CVS-aware
12665 - renamed the package from SSLeay to OpenSSL (the files still contain
12666 their history because I've copied them in the repository)
12667 - removed obsolete files (the test scripts will be replaced
12668 by better Test::Harness variants in the future)
12669 [Ralf S. Engelschall]
12671 *) First cut for a very conservative source tree cleanup:
12672 1. merge various obsolete readme texts into doc/ssleay.txt
12673 where we collect the old documents and readme texts.
12674 2. remove the first part of files where I'm already sure that we no
12675 longer need them because of three reasons: either they are just temporary
12676 files which were left by Eric or they are preserved original files where
12677 I've verified that the diff is also available in the CVS via "cvs diff
12678 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12679 the crypto/md/ stuff).
12680 [Ralf S. Engelschall]
12682 *) More extension code. Incomplete support for subject and issuer alt
12683 name, issuer and authority key id. Change the i2v function parameters
12684 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12685 what that's for :-) Fix to ASN1 macro which messed up
12686 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12689 *) Preliminary support for ENUMERATED type. This is largely copied from the
12693 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12694 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12696 *) Make sure `make rehash' target really finds the `openssl' program.
12697 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12699 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12700 like to hear about it if this slows down other processors.
12703 *) Add CygWin32 platform information to Configure script.
12704 [Alan Batie <batie@aahz.jf.intel.com>]
12706 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12707 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12709 *) New program nseq to manipulate netscape certificate sequences
12712 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12716 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12717 but the BN code had some problems that would cause failures when
12718 doing certificate verification and some other functions.
12719 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12721 *) Add ASN1 and PEM code to support netscape certificate sequences.
12724 *) Add ASN1 and PEM code to support netscape certificate sequences.
12727 *) Add several PKIX and private extended key usage OIDs.
12730 *) Modify the 'ca' program to handle the new extension code. Modify
12731 openssl.cnf for new extension format, add comments.
12734 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12735 and add a sample to openssl.cnf so req -x509 now adds appropriate
12739 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12740 error code, add initial support to X509_print() and x509 application.
12743 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12744 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12745 stuff is currently isolated and isn't even compiled yet.
12748 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12749 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12750 Removed the versions check from X509 routines when loading extensions:
12751 this allows certain broken certificates that don't set the version
12752 properly to be processed.
12755 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12756 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12757 can still be regenerated with "make depend".
12760 *) Spelling mistake in C version of CAST-128.
12761 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12763 *) Changes to the error generation code. The perl script err-code.pl
12764 now reads in the old error codes and retains the old numbers, only
12765 adding new ones if necessary. It also only changes the .err files if new
12766 codes are added. The makefiles have been modified to only insert errors
12767 when needed (to avoid needlessly modifying header files). This is done
12768 by only inserting errors if the .err file is newer than the auto generated
12769 C file. To rebuild all the error codes from scratch (the old behaviour)
12770 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12771 or delete all the .err files.
12774 *) CAST-128 was incorrectly implemented for short keys. The C version has
12775 been fixed, but is untested. The assembler versions are also fixed, but
12776 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12777 to regenerate it if needed.
12778 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12779 Hagino <itojun@kame.net>]
12781 *) File was opened incorrectly in randfile.c.
12782 [Ulf Möller <ulf@fitug.de>]
12784 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12785 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12786 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12787 al: it's just almost always a UTCTime. Note this patch adds new error
12788 codes so do a "make errors" if there are problems.
12791 *) Correct Linux 1 recognition in config.
12792 [Ulf Möller <ulf@fitug.de>]
12794 *) Remove pointless MD5 hash when using DSA keys in ca.
12795 [Anonymous <nobody@replay.com>]
12797 *) Generate an error if given an empty string as a cert directory. Also
12798 generate an error if handed NULL (previously returned 0 to indicate an
12799 error, but didn't set one).
12800 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12802 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12805 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12806 parameters. This was causing a warning which killed off the Win32 compile.
12809 *) Remove C++ style comments from crypto/bn/bn_local.h.
12810 [Neil Costigan <neil.costigan@celocom.com>]
12812 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12813 based on a text string, looking up short and long names and finally
12814 "dot" format. The "dot" format stuff didn't work. Added new function
12815 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12816 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12817 OID is not part of the table.
12820 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12821 X509_LOOKUP_by_alias().
12824 *) Sort openssl functions by name.
12827 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12828 encryption from sample DSA keys (in case anyone is interested the password
12832 *) Make _all_ *_free functions accept a NULL pointer.
12833 [Frans Heymans <fheymans@isaserver.be>]
12835 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12837 [Anonymous <nobody@replay.com>]
12839 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12840 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12842 *) Don't blow it for numeric -newkey arguments to apps/req.
12843 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12845 *) Temp key "for export" tests were wrong in s3_srvr.c.
12846 [Anonymous <nobody@replay.com>]
12848 *) Add prototype for temp key callback functions
12849 SSL_CTX_set_tmp_{rsa,dh}_callback().
12852 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12853 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12856 *) X509_name_add_entry() freed the wrong thing after an error.
12857 [Arne Ansper <arne@ats.cyber.ee>]
12859 *) rsa_eay.c would attempt to free a NULL context.
12860 [Arne Ansper <arne@ats.cyber.ee>]
12862 *) BIO_s_socket() had a broken should_retry() on Windoze.
12863 [Arne Ansper <arne@ats.cyber.ee>]
12865 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12866 [Arne Ansper <arne@ats.cyber.ee>]
12868 *) Make sure the already existing X509_STORE->depth variable is initialized
12869 in X509_STORE_new(), but document the fact that this variable is still
12870 unused in the certificate verification process.
12871 [Ralf S. Engelschall]
12873 *) Fix the various library and apps files to free up pkeys obtained from
12874 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12877 *) Fix reference counting in X509_PUBKEY_get(). This makes
12878 demos/maurice/example2.c work, amongst others, probably.
12879 [Steve Henson and Ben Laurie]
12881 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12882 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12883 are no longer created. This way we have a single and consistent command
12884 line interface `openssl <command>', similar to `cvs <command>'.
12885 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12887 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12888 BIT STRING wrapper always have zero unused bits.
12891 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12894 *) Make the top-level INSTALL documentation easier to understand.
12897 *) Makefiles updated to exit if an error occurs in a sub-directory
12898 make (including if user presses ^C) [Paul Sutton]
12900 *) Make Montgomery context stuff explicit in RSA data structure.
12903 *) Fix build order of pem and err to allow for generated pem.h.
12906 *) Fix renumbering bug in X509_NAME_delete_entry().
12909 *) Enhanced the err-ins.pl script so it makes the error library number
12910 global and can add a library name. This is needed for external ASN1 and
12911 other error libraries.
12914 *) Fixed sk_insert which never worked properly.
12917 *) Fix ASN1 macros so they can handle indefinite length constructed
12918 EXPLICIT tags. Some non standard certificates use these: they can now
12922 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12923 into a single doc/ssleay.txt bundle. This way the information is still
12924 preserved but no longer messes up this directory. Now it's new room for
12925 the new set of documentation files.
12926 [Ralf S. Engelschall]
12928 *) SETs were incorrectly DER encoded. This was a major pain, because they
12929 shared code with SEQUENCEs, which aren't coded the same. This means that
12930 almost everything to do with SETs or SEQUENCEs has either changed name or
12931 number of arguments.
12932 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12934 *) Fix test data to work with the above.
12937 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12938 was already fixed by Eric for 0.9.1 it seems.
12939 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12941 *) Autodetect FreeBSD3.
12944 *) Fix various bugs in Configure. This affects the following platforms:
12948 unixware-2.0-pentium
12952 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12953 before they are needed.
12956 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12960 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12962 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12963 changed SSLeay to OpenSSL in version strings.
12964 [Ralf S. Engelschall]
12966 *) Some fixups to the top-level documents.
12969 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12970 because the symlink to include/ was missing.
12971 [Ralf S. Engelschall]
12973 *) Incorporated the popular no-RSA/DSA-only patches
12974 which allow to compile a RSA-free SSLeay.
12975 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12977 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12978 when "ssleay" is still not found.
12979 [Ralf S. Engelschall]
12981 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12982 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12984 *) Updated the README file.
12985 [Ralf S. Engelschall]
12987 *) Added various .cvsignore files in the CVS repository subdirs
12988 to make a "cvs update" really silent.
12989 [Ralf S. Engelschall]
12991 *) Recompiled the error-definition header files and added
12992 missing symbols to the Win32 linker tables.
12993 [Ralf S. Engelschall]
12995 *) Cleaned up the top-level documents;
12996 o new files: CHANGES and LICENSE
12997 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12998 o merged COPYRIGHT into LICENSE
12999 o removed obsolete TODO file
13000 o renamed MICROSOFT to INSTALL.W32
13001 [Ralf S. Engelschall]
13003 *) Removed dummy files from the 0.9.1b source tree:
13004 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13005 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13006 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13007 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13008 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13009 [Ralf S. Engelschall]
13011 *) Added various platform portability fixes.
13014 *) The Genesis of the OpenSSL rpject:
13015 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13016 Young and Tim J. Hudson created while they were working for C2Net until
13018 [The OpenSSL Project]
13021 Changes between 0.9.0b and 0.9.1b [not released]
13023 *) Updated a few CA certificates under certs/
13026 *) Changed some BIGNUM api stuff.
13029 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13030 DGUX x86, Linux Alpha, etc.
13033 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13034 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13038 *) Add -strparse option to asn1pars program which parses nested
13040 [Dr Stephen Henson <shenson@bigfoot.com>]
13042 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13045 *) DSA fix for "ca" program.
13048 *) Added "-genkey" option to "dsaparam" program.
13051 *) Added RIPE MD160 (rmd160) message digest.
13054 *) Added -a (all) option to "ssleay version" command.
13057 *) Added PLATFORM define which is the id given to Configure.
13060 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13063 *) Extended the ASN.1 parser routines.
13066 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13069 *) Added a BN_CTX to the BN library.
13072 *) Fixed the weak key values in DES library
13075 *) Changed API in EVP library for cipher aliases.
13078 *) Added support for RC2/64bit cipher.
13081 *) Converted the lhash library to the crypto/mem.c functions.
13084 *) Added more recognized ASN.1 object ids.
13087 *) Added more RSA padding checks for SSL/TLS.
13090 *) Added BIO proxy/filter functionality.
13093 *) Added extra_certs to SSL_CTX which can be used
13094 send extra CA certificates to the client in the CA cert chain sending
13095 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13098 *) Now Fortezza is denied in the authentication phase because
13099 this is key exchange mechanism is not supported by SSLeay at all.
13102 *) Additional PKCS1 checks.
13105 *) Support the string "TLSv1" for all TLS v1 ciphers.
13108 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13109 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13112 *) Fixed a few memory leaks.
13115 *) Fixed various code and comment typos.
13118 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13119 bytes sent in the client random.
13120 [Edward Bishop <ebishop@spyglass.com>]