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 1.1.0g and 1.1.1 [xx XXX xxxx]
12 *) QNX support removed (cannot find contributors to get their approval
13 for the license change).
16 *) TLSv1.3 replay protection for early data has been implemented. See the
17 SSL_read_early_data() man page for further details.
20 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
21 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
22 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
23 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
24 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
25 configuraton has been separated out. See the ciphers man page or the
26 SSL_CTX_set_ciphersuites() man page for more information.
29 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
30 in responder mode now supports the new "-multi" option, which
31 spawns the specified number of child processes to handle OCSP
32 requests. The "-timeout" option now also limits the OCSP
33 responder's patience to wait to receive the full client request
34 on a newly accepted connection. Child processes are respawned
35 as needed, and the CA index file is automatically reloaded
36 when changed. This makes it possible to run the "ocsp" responder
37 as a long-running service, making the OpenSSL CA somewhat more
38 feature-complete. In this mode, most diagnostic messages logged
39 after entering the event loop are logged via syslog(3) rather than
43 *) Added support for X448 and Ed448. Heavily based on original work by
47 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
48 objects loaded. This adds the functions OSSL_STORE_expect() and
49 OSSL_STORE_find() as well as needed tools to construct searches and
50 get the search data out of them.
53 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
54 version of OpenSSL should review their configuration settings to ensure
55 that they are still appropriate for TLSv1.3. For further information see:
56 https://www.openssl.org/blog/blog/2018/02/08/tlsv1.3/
58 NOTE: In this pre-release of OpenSSL a draft version of the
59 TLSv1.3 standard has been implemented. Implementations of different draft
60 versions of the standard do not inter-operate, and this version will not
61 inter-operate with an implementation of the final standard when it is
62 eventually published. Different pre-release versions may implement
63 different versions of the draft. The final version of OpenSSL 1.1.1 will
64 implement the final version of the standard.
65 TODO(TLS1.3): Remove the above note before final release
68 *) Grand redesign of the OpenSSL random generator
70 The default RAND method now utilizes an AES-CTR DRBG according to
71 NIST standard SP 800-90Ar1. The new random generator is essentially
72 a port of the default random generator from the OpenSSL FIPS 2.0
73 object module. It is a hybrid deterministic random bit generator
74 using an AES-CTR bit stream and which seeds and reseeds itself
75 automatically using trusted system entropy sources.
77 Some of its new features are:
78 o Support for multiple DRBG instances with seed chaining.
79 o Add a public DRBG instance for the default RAND method.
80 o Add a dedicated DRBG instance for generating long term private keys.
81 o Make the DRBG instances fork-safe.
82 o Keep all global DRBG instances on the secure heap if it is enabled.
83 o Add a DRBG instance to every SSL instance for lock free operation
84 and to increase unpredictability.
85 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
87 *) Changed Configure so it only says what it does and doesn't dump
88 so much data. Instead, ./configdata.pm should be used as a script
89 to display all sorts of configuration data.
92 *) Added processing of "make variables" to Configure.
95 *) Added SHA512/224 and SHA512/256 algorithm support.
98 *) The last traces of Netware support, first removed in 1.1.0, have
102 *) Get rid of Makefile.shared, and in the process, make the processing
103 of certain files (rc.obj, or the .def/.map/.opt files produced from
104 the ordinal files) more visible and hopefully easier to trace and
105 debug (or make silent).
108 *) Make it possible to have environment variable assignments as
109 arguments to config / Configure.
112 *) Add multi-prime RSA (RFC 8017) support.
115 *) Add SM3 implemented according to GB/T 32905-2016
116 [ Jack Lloyd <jack.lloyd@ribose.com>,
117 Ronald Tse <ronald.tse@ribose.com>,
118 Erick Borsboom <erick.borsboom@ribose.com> ]
120 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
121 as documented in RFC6066.
122 Based on a patch from Tomasz Moń
123 [Filipe Raimundo da Silva]
125 *) Add SM4 implemented according to GB/T 32907-2016.
126 [ Jack Lloyd <jack.lloyd@ribose.com>,
127 Ronald Tse <ronald.tse@ribose.com>,
128 Erick Borsboom <erick.borsboom@ribose.com> ]
130 *) Reimplement -newreq-nodes and ERR_error_string_n; the
131 original author does not agree with the license change.
134 *) Add ARIA AEAD TLS support.
137 *) Some macro definitions to support VS6 have been removed. Visual
138 Studio 6 has not worked since 1.1.0
141 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
142 without clearing the errors.
145 *) Add "atfork" functions. If building on a system that without
146 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
147 requirements. The RAND facility now uses/requires this.
153 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
154 not possible to disable entirely. However, it's still possible to
155 disable the console reading UI method, UI_OpenSSL() (use UI_null()
158 To disable, configure with 'no-ui-console'. 'no-ui' is still
159 possible to use as an alias. Check at compile time with the
160 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
161 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
164 *) Add a STORE module, which implements a uniform and URI based reader of
165 stores that can contain keys, certificates, CRLs and numerous other
166 objects. The main API is loosely based on a few stdio functions,
167 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
168 OSSL_STORE_error and OSSL_STORE_close.
169 The implementation uses backends called "loaders" to implement arbitrary
170 URI schemes. There is one built in "loader" for the 'file' scheme.
173 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
174 then adjusted to work on FreeBSD 8.4 as well.
175 Enable by configuring with 'enable-devcryptoeng'. This is done by default
176 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
179 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
180 util/mkerr.pl, which is adapted to allow those prefixes, leading to
181 error code calls like this:
183 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
185 With this change, we claim the namespaces OSSL and OPENSSL in a manner
186 that can be encoded in C. For the foreseeable future, this will only
188 [Richard Levitte and Tim Hudson]
190 *) Removed BSD cryptodev engine.
193 *) Add a build target 'build_all_generated', to build all generated files
194 and only that. This can be used to prepare everything that requires
195 things like perl for a system that lacks perl and then move everything
196 to that system and do the rest of the build there.
199 *) In the UI interface, make it possible to duplicate the user data. This
200 can be used by engines that need to retain the data for a longer time
201 than just the call where this user data is passed.
204 *) Ignore the '-named_curve auto' value for compatibility of applications
206 [Tomas Mraz <tmraz@fedoraproject.org>]
208 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
209 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
210 alerts across multiple records (some of which could be empty). In practice
211 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
212 prohibts this altogether and other libraries (BoringSSL, NSS) do not
213 support this at all. Supporting it adds significant complexity to the
214 record layer, and its removal is unlikely to cause inter-operability
218 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
219 with Z. These are meant to replace LONG and ZLONG and to be size safe.
220 The use of LONG and ZLONG is discouraged and scheduled for deprecation
224 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
225 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
226 [Richard Levitte, Andy Polyakov]
228 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
232 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
233 platform rather than 'mingw'.
236 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
237 success if they are asked to add an object which already exists
238 in the store. This change cascades to other functions which load
239 certificates and CRLs.
242 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
243 facilitate stack unwinding even from assembly subroutines.
246 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
247 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
250 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
251 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
252 which is the minimum version we support.
255 *) Certificate time validation (X509_cmp_time) enforces stricter
256 compliance with RFC 5280. Fractional seconds and timezone offsets
257 are no longer allowed.
260 *) Add support for ARIA
263 *) s_client will now send the Server Name Indication (SNI) extension by
264 default unless the new "-noservername" option is used. The server name is
265 based on the host provided to the "-connect" option unless overridden by
269 *) Add support for SipHash
272 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
273 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
274 prevent issues where no progress is being made and the peer continually
275 sends unrecognised record types, using up resources processing them.
278 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
279 using the algorithm defined in
280 https://www.akkadia.org/drepper/SHA-crypt.txt
283 *) Heartbeat support has been removed; the ABI is changed for now.
284 [Richard Levitte, Rich Salz]
286 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
289 *) The RSA "null" method, which was partially supported to avoid patent
290 issues, has been replaced to always returns NULL.
293 Changes between 1.1.0g and 1.1.0h [xx XXX xxxx]
295 *) Constructed ASN.1 types with a recursive definition could exceed the stack
297 Constructed ASN.1 types with a recursive definition (such as can be found
298 in PKCS7) could eventually exceed the stack given malicious input with
299 excessive recursion. This could result in a Denial Of Service attack. There
300 are no such structures used within SSL/TLS that come from untrusted sources
301 so this is considered safe.
303 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
308 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
310 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
311 effectively reduced to only comparing the least significant bit of each
312 byte. This allows an attacker to forge messages that would be considered as
313 authenticated in an amount of tries lower than that guaranteed by the
314 security claims of the scheme. The module can only be compiled by the
315 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
317 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
322 *) Add a build target 'build_all_generated', to build all generated files
323 and only that. This can be used to prepare everything that requires
324 things like perl for a system that lacks perl and then move everything
325 to that system and do the rest of the build there.
328 *) Backport SSL_OP_NO_RENGOTIATION
330 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
331 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
332 changes this is no longer possible in 1.1.0. Therefore the new
333 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
334 1.1.0 to provide equivalent functionality.
336 Note that if an application built against 1.1.0h headers (or above) is run
337 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
338 accepted but nothing will happen, i.e. renegotiation will not be prevented.
341 *) Removed the OS390-Unix config target. It relied on a script that doesn't
345 *) rsaz_1024_mul_avx2 overflow bug on x86_64
347 There is an overflow bug in the AVX2 Montgomery multiplication procedure
348 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
349 Analysis suggests that attacks against RSA and DSA as a result of this
350 defect would be very difficult to perform and are not believed likely.
351 Attacks against DH1024 are considered just feasible, because most of the
352 work necessary to deduce information about a private key may be performed
353 offline. The amount of resources required for such an attack would be
354 significant. However, for an attack on TLS to be meaningful, the server
355 would have to share the DH1024 private key among multiple clients, which is
356 no longer an option since CVE-2016-0701.
358 This only affects processors that support the AVX2 but not ADX extensions
359 like Intel Haswell (4th generation).
361 This issue was reported to OpenSSL by David Benjamin (Google). The issue
362 was originally found via the OSS-Fuzz project.
366 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
368 *) bn_sqrx8x_internal carry bug on x86_64
370 There is a carry propagating bug in the x86_64 Montgomery squaring
371 procedure. No EC algorithms are affected. Analysis suggests that attacks
372 against RSA and DSA as a result of this defect would be very difficult to
373 perform and are not believed likely. Attacks against DH are considered just
374 feasible (although very difficult) because most of the work necessary to
375 deduce information about a private key may be performed offline. The amount
376 of resources required for such an attack would be very significant and
377 likely only accessible to a limited number of attackers. An attacker would
378 additionally need online access to an unpatched system using the target
379 private key in a scenario with persistent DH parameters and a private
380 key that is shared between multiple clients.
382 This only affects processors that support the BMI1, BMI2 and ADX extensions
383 like Intel Broadwell (5th generation) and later or AMD Ryzen.
385 This issue was reported to OpenSSL by the OSS-Fuzz project.
389 *) Malformed X.509 IPAddressFamily could cause OOB read
391 If an X.509 certificate has a malformed IPAddressFamily extension,
392 OpenSSL could do a one-byte buffer overread. The most likely result
393 would be an erroneous display of the certificate in text format.
395 This issue was reported to OpenSSL by the OSS-Fuzz project.
399 Changes between 1.1.0e and 1.1.0f [25 May 2017]
401 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
402 platform rather than 'mingw'.
405 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
406 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
407 which is the minimum version we support.
410 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
412 *) Encrypt-Then-Mac renegotiation crash
414 During a renegotiation handshake if the Encrypt-Then-Mac extension is
415 negotiated where it was not in the original handshake (or vice-versa) then
416 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
417 and servers are affected.
419 This issue was reported to OpenSSL by Joe Orton (Red Hat).
423 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
425 *) Truncated packet could crash via OOB read
427 If one side of an SSL/TLS path is running on a 32-bit host and a specific
428 cipher is being used, then a truncated packet can cause that host to
429 perform an out-of-bounds read, usually resulting in a crash.
431 This issue was reported to OpenSSL by Robert Święcki of Google.
435 *) Bad (EC)DHE parameters cause a client crash
437 If a malicious server supplies bad parameters for a DHE or ECDHE key
438 exchange then this can result in the client attempting to dereference a
439 NULL pointer leading to a client crash. This could be exploited in a Denial
442 This issue was reported to OpenSSL by Guido Vranken.
446 *) BN_mod_exp may produce incorrect results on x86_64
448 There is a carry propagating bug in the x86_64 Montgomery squaring
449 procedure. No EC algorithms are affected. Analysis suggests that attacks
450 against RSA and DSA as a result of this defect would be very difficult to
451 perform and are not believed likely. Attacks against DH are considered just
452 feasible (although very difficult) because most of the work necessary to
453 deduce information about a private key may be performed offline. The amount
454 of resources required for such an attack would be very significant and
455 likely only accessible to a limited number of attackers. An attacker would
456 additionally need online access to an unpatched system using the target
457 private key in a scenario with persistent DH parameters and a private
458 key that is shared between multiple clients. For example this can occur by
459 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
460 similar to CVE-2015-3193 but must be treated as a separate problem.
462 This issue was reported to OpenSSL by the OSS-Fuzz project.
466 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
468 *) ChaCha20/Poly1305 heap-buffer-overflow
470 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
471 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
472 crash. This issue is not considered to be exploitable beyond a DoS.
474 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
478 *) CMS Null dereference
480 Applications parsing invalid CMS structures can crash with a NULL pointer
481 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
482 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
483 structure callback if an attempt is made to free certain invalid encodings.
484 Only CHOICE structures using a callback which do not handle NULL value are
487 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
491 *) Montgomery multiplication may produce incorrect results
493 There is a carry propagating bug in the Broadwell-specific Montgomery
494 multiplication procedure that handles input lengths divisible by, but
495 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
496 and DH private keys are impossible. This is because the subroutine in
497 question is not used in operations with the private key itself and an input
498 of the attacker's direct choice. Otherwise the bug can manifest itself as
499 transient authentication and key negotiation failures or reproducible
500 erroneous outcome of public-key operations with specially crafted input.
501 Among EC algorithms only Brainpool P-512 curves are affected and one
502 presumably can attack ECDH key negotiation. Impact was not analyzed in
503 detail, because pre-requisites for attack are considered unlikely. Namely
504 multiple clients have to choose the curve in question and the server has to
505 share the private key among them, neither of which is default behaviour.
506 Even then only clients that chose the curve will be affected.
508 This issue was publicly reported as transient failures and was not
509 initially recognized as a security issue. Thanks to Richard Morgan for
510 providing reproducible case.
514 *) Removed automatic addition of RPATH in shared libraries and executables,
515 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
518 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
520 *) Fix Use After Free for large message sizes
522 The patch applied to address CVE-2016-6307 resulted in an issue where if a
523 message larger than approx 16k is received then the underlying buffer to
524 store the incoming message is reallocated and moved. Unfortunately a
525 dangling pointer to the old location is left which results in an attempt to
526 write to the previously freed location. This is likely to result in a
527 crash, however it could potentially lead to execution of arbitrary code.
529 This issue only affects OpenSSL 1.1.0a.
531 This issue was reported to OpenSSL by Robert Święcki.
535 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
537 *) OCSP Status Request extension unbounded memory growth
539 A malicious client can send an excessively large OCSP Status Request
540 extension. If that client continually requests renegotiation, sending a
541 large OCSP Status Request extension each time, then there will be unbounded
542 memory growth on the server. This will eventually lead to a Denial Of
543 Service attack through memory exhaustion. Servers with a default
544 configuration are vulnerable even if they do not support OCSP. Builds using
545 the "no-ocsp" build time option are not affected.
547 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
551 *) SSL_peek() hang on empty record
553 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
554 sends an empty record. This could be exploited by a malicious peer in a
555 Denial Of Service attack.
557 This issue was reported to OpenSSL by Alex Gaynor.
561 *) Excessive allocation of memory in tls_get_message_header() and
562 dtls1_preprocess_fragment()
564 A (D)TLS message includes 3 bytes for its length in the header for the
565 message. This would allow for messages up to 16Mb in length. Messages of
566 this length are excessive and OpenSSL includes a check to ensure that a
567 peer is sending reasonably sized messages in order to avoid too much memory
568 being consumed to service a connection. A flaw in the logic of version
569 1.1.0 means that memory for the message is allocated too early, prior to
570 the excessive message length check. Due to way memory is allocated in
571 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
572 to service a connection. This could lead to a Denial of Service through
573 memory exhaustion. However, the excessive message length check still takes
574 place, and this would cause the connection to immediately fail. Assuming
575 that the application calls SSL_free() on the failed connection in a timely
576 manner then the 21Mb of allocated memory will then be immediately freed
577 again. Therefore the excessive memory allocation will be transitory in
578 nature. This then means that there is only a security impact if:
580 1) The application does not call SSL_free() in a timely manner in the event
581 that the connection fails
583 2) The application is working in a constrained environment where there is
584 very little free memory
586 3) The attacker initiates multiple connection attempts such that there are
587 multiple connections in a state where memory has been allocated for the
588 connection; SSL_free() has not yet been called; and there is insufficient
589 memory to service the multiple requests.
591 Except in the instance of (1) above any Denial Of Service is likely to be
592 transitory because as soon as the connection fails the memory is
593 subsequently freed again in the SSL_free() call. However there is an
594 increased risk during this period of application crashes due to the lack of
595 memory - which would then mean a more serious Denial of Service.
597 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
598 (CVE-2016-6307 and CVE-2016-6308)
601 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
602 had to be removed. Primary reason is that vendor assembler can't
603 assemble our modules with -KPIC flag. As result it, assembly
604 support, was not even available as option. But its lack means
605 lack of side-channel resistant code, which is incompatible with
606 security by todays standards. Fortunately gcc is readily available
607 prepackaged option, which we firmly point at...
610 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
612 *) Windows command-line tool supports UTF-8 opt-in option for arguments
613 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
614 (to any value) allows Windows user to access PKCS#12 file generated
615 with Windows CryptoAPI and protected with non-ASCII password, as well
616 as files generated under UTF-8 locale on Linux also protected with
620 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
621 have been disabled by default and removed from DEFAULT, just like RC4.
622 See the RC4 item below to re-enable both.
625 *) The method for finding the storage location for the Windows RAND seed file
626 has changed. First we check %RANDFILE%. If that is not set then we check
627 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
628 all else fails we fall back to C:\.
631 *) The EVP_EncryptUpdate() function has had its return type changed from void
632 to int. A return of 0 indicates and error while a return of 1 indicates
636 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
637 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
638 off the constant time implementation for RSA, DSA and DH have been made
639 no-ops and deprecated.
642 *) Windows RAND implementation was simplified to only get entropy by
643 calling CryptGenRandom(). Various other RAND-related tickets
645 [Joseph Wylie Yandle, Rich Salz]
647 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
648 and OPENSSL_LH_, respectively. The old names are available
649 with API compatibility. They new names are now completely documented.
652 *) Unify TYPE_up_ref(obj) methods signature.
653 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
654 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
655 int (instead of void) like all others TYPE_up_ref() methods.
656 So now these methods also check the return value of CRYPTO_atomic_add(),
657 and the validity of object reference counter.
658 [fdasilvayy@gmail.com]
660 *) With Windows Visual Studio builds, the .pdb files are installed
661 alongside the installed libraries and executables. For a static
662 library installation, ossl_static.pdb is the associate compiler
663 generated .pdb file to be used when linking programs.
666 *) Remove openssl.spec. Packaging files belong with the packagers.
669 *) Automatic Darwin/OSX configuration has had a refresh, it will now
670 recognise x86_64 architectures automatically. You can still decide
671 to build for a different bitness with the environment variable
672 KERNEL_BITS (can be 32 or 64), for example:
674 KERNEL_BITS=32 ./config
678 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
679 256 bit AES and HMAC with SHA256.
682 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
685 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
688 *) To enable users to have their own config files and build file templates,
689 Configure looks in the directory indicated by the environment variable
690 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
691 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
692 name and is used as is.
695 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
696 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
697 X509_CERT_FILE_CTX was removed.
700 *) "shared" builds are now the default. To create only static libraries use
701 the "no-shared" Configure option.
704 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
705 All of these option have not worked for some while and are fundamental
709 *) Make various cleanup routines no-ops and mark them as deprecated. Most
710 global cleanup functions are no longer required because they are handled
711 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
712 Explicitly de-initing can cause problems (e.g. where a library that uses
713 OpenSSL de-inits, but an application is still using it). The affected
714 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
715 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
716 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
720 *) --strict-warnings no longer enables runtime debugging options
721 such as REF_DEBUG. Instead, debug options are automatically
722 enabled with '--debug' builds.
723 [Andy Polyakov, Emilia Käsper]
725 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
726 have been moved out of the public header files. New functions for managing
727 these have been added.
730 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
731 objects have been moved out of the public header files. New
732 functions for managing these have been added.
735 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
736 have been moved out of the public header files. New functions for managing
737 these have been added.
740 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
741 moved out of the public header files. New functions for managing these
745 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
748 *) Removed the mk1mf build scripts.
751 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
752 it is always safe to #include a header now.
755 *) Removed the aged BC-32 config and all its supporting scripts
758 *) Removed support for Ultrix, Netware, and OS/2.
761 *) Add support for HKDF.
764 *) Add support for blake2b and blake2s
767 *) Added support for "pipelining". Ciphers that have the
768 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
769 encryptions/decryptions simultaneously. There are currently no built-in
770 ciphers with this property but the expectation is that engines will be able
771 to offer it to significantly improve throughput. Support has been extended
772 into libssl so that multiple records for a single connection can be
773 processed in one go (for >=TLS 1.1).
776 *) Added the AFALG engine. This is an async capable engine which is able to
777 offload work to the Linux kernel. In this initial version it only supports
778 AES128-CBC. The kernel must be version 4.1.0 or greater.
781 *) OpenSSL now uses a new threading API. It is no longer necessary to
782 set locking callbacks to use OpenSSL in a multi-threaded environment. There
783 are two supported threading models: pthreads and windows threads. It is
784 also possible to configure OpenSSL at compile time for "no-threads". The
785 old threading API should no longer be used. The functions have been
786 replaced with "no-op" compatibility macros.
787 [Alessandro Ghedini, Matt Caswell]
789 *) Modify behavior of ALPN to invoke callback after SNI/servername
790 callback, such that updates to the SSL_CTX affect ALPN.
793 *) Add SSL_CIPHER queries for authentication and key-exchange.
796 *) Changes to the DEFAULT cipherlist:
797 - Prefer (EC)DHE handshakes over plain RSA.
798 - Prefer AEAD ciphers over legacy ciphers.
799 - Prefer ECDSA over RSA when both certificates are available.
800 - Prefer TLSv1.2 ciphers/PRF.
801 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
805 *) Change the ECC default curve list to be this, in order: x25519,
806 secp256r1, secp521r1, secp384r1.
809 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
810 disabled by default. They can be re-enabled using the
811 enable-weak-ssl-ciphers option to Configure.
814 *) If the server has ALPN configured, but supports no protocols that the
815 client advertises, send a fatal "no_application_protocol" alert.
816 This behaviour is SHALL in RFC 7301, though it isn't universally
817 implemented by other servers.
820 *) Add X25519 support.
821 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
822 for public and private key encoding using the format documented in
823 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
824 key generation and key derivation.
826 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
830 *) Deprecate SRP_VBASE_get_by_user.
831 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
832 In order to fix an unavoidable memory leak (CVE-2016-0798),
833 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
834 seed, even if the seed is configured.
836 Users should use SRP_VBASE_get1_by_user instead. Note that in
837 SRP_VBASE_get1_by_user, caller must free the returned value. Note
838 also that even though configuring the SRP seed attempts to hide
839 invalid usernames by continuing the handshake with fake
840 credentials, this behaviour is not constant time and no strong
841 guarantees are made that the handshake is indistinguishable from
842 that of a valid user.
845 *) Configuration change; it's now possible to build dynamic engines
846 without having to build shared libraries and vice versa. This
847 only applies to the engines in engines/, those in crypto/engine/
848 will always be built into libcrypto (i.e. "static").
850 Building dynamic engines is enabled by default; to disable, use
851 the configuration option "disable-dynamic-engine".
853 The only requirements for building dynamic engines are the
854 presence of the DSO module and building with position independent
855 code, so they will also automatically be disabled if configuring
856 with "disable-dso" or "disable-pic".
858 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
859 are also taken away from openssl/opensslconf.h, as they are
863 *) Configuration change; if there is a known flag to compile
864 position independent code, it will always be applied on the
865 libcrypto and libssl object files, and never on the application
866 object files. This means other libraries that use routines from
867 libcrypto / libssl can be made into shared libraries regardless
868 of how OpenSSL was configured.
870 If this isn't desirable, the configuration options "disable-pic"
871 or "no-pic" can be used to disable the use of PIC. This will
872 also disable building shared libraries and dynamic engines.
875 *) Removed JPAKE code. It was experimental and has no wide use.
878 *) The INSTALL_PREFIX Makefile variable has been renamed to
879 DESTDIR. That makes for less confusion on what this variable
880 is for. Also, the configuration option --install_prefix is
884 *) Heartbeat for TLS has been removed and is disabled by default
885 for DTLS; configure with enable-heartbeats. Code that uses the
886 old #define's might need to be updated.
887 [Emilia Käsper, Rich Salz]
889 *) Rename REF_CHECK to REF_DEBUG.
892 *) New "unified" build system
894 The "unified" build system is aimed to be a common system for all
895 platforms we support. With it comes new support for VMS.
897 This system builds supports building in a different directory tree
898 than the source tree. It produces one Makefile (for unix family
899 or lookalikes), or one descrip.mms (for VMS).
901 The source of information to make the Makefile / descrip.mms is
902 small files called 'build.info', holding the necessary
903 information for each directory with source to compile, and a
904 template in Configurations, like unix-Makefile.tmpl or
907 With this change, the library names were also renamed on Windows
908 and on VMS. They now have names that are closer to the standard
909 on Unix, and include the major version number, and in certain
910 cases, the architecture they are built for. See "Notes on shared
911 libraries" in INSTALL.
913 We rely heavily on the perl module Text::Template.
916 *) Added support for auto-initialisation and de-initialisation of the library.
917 OpenSSL no longer requires explicit init or deinit routines to be called,
918 except in certain circumstances. See the OPENSSL_init_crypto() and
919 OPENSSL_init_ssl() man pages for further information.
922 *) The arguments to the DTLSv1_listen function have changed. Specifically the
923 "peer" argument is now expected to be a BIO_ADDR object.
925 *) Rewrite of BIO networking library. The BIO library lacked consistent
926 support of IPv6, and adding it required some more extensive
927 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
928 which hold all types of addresses and chains of address information.
929 It also introduces a new API, with functions like BIO_socket,
930 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
931 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
932 have been adapted accordingly.
935 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
939 *) CRIME protection: disable compression by default, even if OpenSSL is
940 compiled with zlib enabled. Applications can still enable compression
941 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
942 using the SSL_CONF library to configure compression.
945 *) The signature of the session callback configured with
946 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
947 was explicitly marked as 'const unsigned char*' instead of
951 *) Always DPURIFY. Remove the use of uninitialized memory in the
952 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
955 *) Removed many obsolete configuration items, including
956 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
957 MD2_CHAR, MD2_INT, MD2_LONG
959 IDEA_SHORT, IDEA_LONG
960 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
961 [Rich Salz, with advice from Andy Polyakov]
963 *) Many BN internals have been moved to an internal header file.
964 [Rich Salz with help from Andy Polyakov]
966 *) Configuration and writing out the results from it has changed.
967 Files such as Makefile include/openssl/opensslconf.h and are now
968 produced through general templates, such as Makefile.in and
969 crypto/opensslconf.h.in and some help from the perl module
972 Also, the center of configuration information is no longer
973 Makefile. Instead, Configure produces a perl module in
974 configdata.pm which holds most of the config data (in the hash
975 table %config), the target data that comes from the target
976 configuration in one of the Configurations/*.conf files (in
980 *) To clarify their intended purposes, the Configure options
981 --prefix and --openssldir change their semantics, and become more
982 straightforward and less interdependent.
984 --prefix shall be used exclusively to give the location INSTALLTOP
985 where programs, scripts, libraries, include files and manuals are
986 going to be installed. The default is now /usr/local.
988 --openssldir shall be used exclusively to give the default
989 location OPENSSLDIR where certificates, private keys, CRLs are
990 managed. This is also where the default openssl.cnf gets
992 If the directory given with this option is a relative path, the
993 values of both the --prefix value and the --openssldir value will
994 be combined to become OPENSSLDIR.
995 The default for --openssldir is INSTALLTOP/ssl.
997 Anyone who uses --openssldir to specify where OpenSSL is to be
998 installed MUST change to use --prefix instead.
1001 *) The GOST engine was out of date and therefore it has been removed. An up
1002 to date GOST engine is now being maintained in an external repository.
1003 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1004 support for GOST ciphersuites (these are only activated if a GOST engine
1008 *) EGD is no longer supported by default; use enable-egd when
1010 [Ben Kaduk and Rich Salz]
1012 *) The distribution now has Makefile.in files, which are used to
1013 create Makefile's when Configure is run. *Configure must be run
1014 before trying to build now.*
1017 *) The return value for SSL_CIPHER_description() for error conditions
1021 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1023 Obtaining and performing DNSSEC validation of TLSA records is
1024 the application's responsibility. The application provides
1025 the TLSA records of its choice to OpenSSL, and these are then
1026 used to authenticate the peer.
1028 The TLSA records need not even come from DNS. They can, for
1029 example, be used to implement local end-entity certificate or
1030 trust-anchor "pinning", where the "pin" data takes the form
1031 of TLSA records, which can augment or replace verification
1032 based on the usual WebPKI public certification authorities.
1035 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1036 continues to support deprecated interfaces in default builds.
1037 However, applications are strongly advised to compile their
1038 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1039 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1040 or the 1.1.0 releases.
1042 In environments in which all applications have been ported to
1043 not use any deprecated interfaces OpenSSL's Configure script
1044 should be used with the --api=1.1.0 option to entirely remove
1045 support for the deprecated features from the library and
1046 unconditionally disable them in the installed headers.
1047 Essentially the same effect can be achieved with the "no-deprecated"
1048 argument to Configure, except that this will always restrict
1049 the build to just the latest API, rather than a fixed API
1052 As applications are ported to future revisions of the API,
1053 they should update their compile-time OPENSSL_API_COMPAT define
1054 accordingly, but in most cases should be able to continue to
1055 compile with later releases.
1057 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1058 0x10000000L and 0x00908000L, respectively. However those
1059 versions did not support the OPENSSL_API_COMPAT feature, and
1060 so applications are not typically tested for explicit support
1061 of just the undeprecated features of either release.
1064 *) Add support for setting the minimum and maximum supported protocol.
1065 It can bet set via the SSL_set_min_proto_version() and
1066 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1067 MaxProtocol. It's recommended to use the new APIs to disable
1068 protocols instead of disabling individual protocols using
1069 SSL_set_options() or SSL_CONF's Protocol. This change also
1070 removes support for disabling TLS 1.2 in the OpenSSL TLS
1071 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1074 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1077 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1078 and integrates ECDSA and ECDH functionality into EC. Implementations can
1079 now redirect key generation and no longer need to convert to or from
1082 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1083 include the ec.h header file instead.
1086 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1087 ciphers who are no longer supported and drops support the ephemeral RSA key
1088 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1091 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1092 opaque. For HMAC_CTX, the following constructors and destructors
1095 HMAC_CTX *HMAC_CTX_new(void);
1096 void HMAC_CTX_free(HMAC_CTX *ctx);
1098 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1099 destroy such methods has been added. See EVP_MD_meth_new(3) and
1100 EVP_CIPHER_meth_new(3) for documentation.
1103 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1104 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1105 EVP_MD_CTX_reset() should be called instead to reinitialise
1106 an already created structure.
1107 2) For consistency with the majority of our object creators and
1108 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1109 EVP_MD_CTX_(new|free). The old names are retained as macros
1110 for deprecated builds.
1113 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1114 cryptographic operations to be performed asynchronously as long as an
1115 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1116 further details. Libssl has also had this capability integrated with the
1117 introduction of the new mode SSL_MODE_ASYNC and associated error
1118 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1119 pages. This work was developed in partnership with Intel Corp.
1122 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1123 always enabled now. If you want to disable the support you should
1124 exclude it using the list of supported ciphers. This also means that the
1125 "-no_ecdhe" option has been removed from s_server.
1128 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1129 SSL_{CTX_}set1_curves() which can set a list.
1132 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1133 curve you want to support using SSL_{CTX_}set1_curves().
1136 *) State machine rewrite. The state machine code has been significantly
1137 refactored in order to remove much duplication of code and solve issues
1138 with the old code (see ssl/statem/README for further details). This change
1139 does have some associated API changes. Notably the SSL_state() function
1140 has been removed and replaced by SSL_get_state which now returns an
1141 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1142 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1146 *) All instances of the string "ssleay" in the public API were replaced
1147 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1148 Some error codes related to internal RSA_eay API's were renamed.
1151 *) The demo files in crypto/threads were moved to demo/threads.
1154 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1156 [Matt Caswell, Rich Salz]
1158 *) New ASN.1 embed macro.
1160 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1161 structure is not allocated: it is part of the parent. That is instead of
1169 This reduces memory fragmentation and make it impossible to accidentally
1170 set a mandatory field to NULL.
1172 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1173 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1174 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1178 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1181 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1182 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1183 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1184 DES and RC4 ciphersuites.
1187 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1188 This changes the decoding behaviour for some invalid messages,
1189 though the change is mostly in the more lenient direction, and
1190 legacy behaviour is preserved as much as possible.
1193 *) Fix no-stdio build.
1194 [ David Woodhouse <David.Woodhouse@intel.com> and also
1195 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1197 *) New testing framework
1198 The testing framework has been largely rewritten and is now using
1199 perl and the perl modules Test::Harness and an extended variant of
1200 Test::More called OpenSSL::Test to do its work. All test scripts in
1201 test/ have been rewritten into test recipes, and all direct calls to
1202 executables in test/Makefile have become individual recipes using the
1203 simplified testing OpenSSL::Test::Simple.
1205 For documentation on our testing modules, do:
1207 perldoc test/testlib/OpenSSL/Test/Simple.pm
1208 perldoc test/testlib/OpenSSL/Test.pm
1212 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1213 are used; the latter aborts on memory leaks (usually checked on exit).
1214 Some undocumented "set malloc, etc., hooks" functions were removed
1215 and others were changed. All are now documented.
1218 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1220 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1222 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1223 from RFC4279, RFC4785, RFC5487, RFC5489.
1225 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1226 original RSA_PSK patch.
1229 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1230 era flag was never set throughout the codebase (only read). Also removed
1231 SSL3_FLAGS_POP_BUFFER which was only used if
1232 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1235 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1236 to be "oneline" instead of "compat".
1239 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1240 not aware of clients that still exhibit this bug, and the workaround
1241 hasn't been working properly for a while.
1244 *) The return type of BIO_number_read() and BIO_number_written() as well as
1245 the corresponding num_read and num_write members in the BIO structure has
1246 changed from unsigned long to uint64_t. On platforms where an unsigned
1247 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1251 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1252 OpenSSL without support for them. It also means that maintaining
1253 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1254 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1257 *) Removed support for the two export grade static DH ciphersuites
1258 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1259 were newly added (along with a number of other static DH ciphersuites) to
1260 1.0.2. However the two export ones have *never* worked since they were
1261 introduced. It seems strange in any case to be adding new export
1262 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1265 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1266 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1267 and turned into macros which simply call the new preferred function names
1268 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1269 should use the new names instead. Also as part of this change the ssl23.h
1270 header file has been removed.
1273 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1274 code and the associated standard is no longer considered fit-for-purpose.
1277 *) RT2547 was closed. When generating a private key, try to make the
1278 output file readable only by the owner. This behavior change might
1279 be noticeable when interacting with other software.
1281 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1285 *) Added HTTP GET support to the ocsp command.
1288 *) Changed default digest for the dgst and enc commands from MD5 to
1292 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1295 *) Added support for TLS extended master secret from
1296 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1297 initial patch which was a great help during development.
1300 *) All libssl internal structures have been removed from the public header
1301 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1302 now redundant). Users should not attempt to access internal structures
1303 directly. Instead they should use the provided API functions.
1306 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1307 Access to deprecated functions can be re-enabled by running config with
1308 "enable-deprecated". In addition applications wishing to use deprecated
1309 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1310 will, by default, disable some transitive includes that previously existed
1311 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1314 *) Added support for OCB mode. OpenSSL has been granted a patent license
1315 compatible with the OpenSSL license for use of OCB. Details are available
1316 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1317 for OCB can be removed by calling config with no-ocb.
1320 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1321 compatible client hello.
1324 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1325 done while fixing the error code for the key-too-small case.
1326 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1328 *) CA.sh has been removed; use CA.pl instead.
1331 *) Removed old DES API.
1334 *) Remove various unsupported platforms:
1340 Sinix/ReliantUNIX RM400
1345 16-bit platforms such as WIN16
1348 *) Clean up OPENSSL_NO_xxx #define's
1349 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1350 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1351 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1352 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1353 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1354 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1355 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1356 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1357 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1358 Remove MS_STATIC; it's a relic from platforms <32 bits.
1361 *) Cleaned up dead code
1362 Remove all but one '#ifdef undef' which is to be looked at.
1365 *) Clean up calling of xxx_free routines.
1366 Just like free(), fix most of the xxx_free routines to accept
1367 NULL. Remove the non-null checks from callers. Save much code.
1370 *) Add secure heap for storage of private keys (when possible).
1371 Add BIO_s_secmem(), CBIGNUM, etc.
1372 Contributed by Akamai Technologies under our Corporate CLA.
1375 *) Experimental support for a new, fast, unbiased prime candidate generator,
1376 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1377 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1379 *) New output format NSS in the sess_id command line tool. This allows
1380 exporting the session id and the master key in NSS keylog format.
1381 [Martin Kaiser <martin@kaiser.cx>]
1383 *) Harmonize version and its documentation. -f flag is used to display
1385 [mancha <mancha1@zoho.com>]
1387 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1388 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1389 [mancha <mancha1@zoho.com>]
1391 *) Fix some double frees. These are not thought to be exploitable.
1392 [mancha <mancha1@zoho.com>]
1394 *) A missing bounds check in the handling of the TLS heartbeat extension
1395 can be used to reveal up to 64k of memory to a connected client or
1398 Thanks for Neel Mehta of Google Security for discovering this bug and to
1399 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1400 preparing the fix (CVE-2014-0160)
1401 [Adam Langley, Bodo Moeller]
1403 *) Fix for the attack described in the paper "Recovering OpenSSL
1404 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1405 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1406 http://eprint.iacr.org/2014/140
1408 Thanks to Yuval Yarom and Naomi Benger for discovering this
1409 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1410 [Yuval Yarom and Naomi Benger]
1412 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1413 this fixes a limitation in previous versions of OpenSSL.
1416 *) Experimental encrypt-then-mac support.
1418 Experimental support for encrypt then mac from
1419 draft-gutmann-tls-encrypt-then-mac-02.txt
1421 To enable it set the appropriate extension number (0x42 for the test
1422 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1424 For non-compliant peers (i.e. just about everything) this should have no
1427 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1431 *) Add EVP support for key wrapping algorithms, to avoid problems with
1432 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1433 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1434 algorithms and include tests cases.
1437 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1441 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1442 MGF1 digest and OAEP label.
1445 *) Make openssl verify return errors.
1446 [Chris Palmer <palmer@google.com> and Ben Laurie]
1448 *) New function ASN1_TIME_diff to calculate the difference between two
1449 ASN1_TIME structures or one structure and the current time.
1452 *) Update fips_test_suite to support multiple command line options. New
1453 test to induce all self test errors in sequence and check expected
1457 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1458 sign or verify all in one operation.
1461 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1462 test programs and fips_test_suite. Includes functionality to parse
1463 the minimal script output of fipsalgest.pl directly.
1466 *) Add authorisation parameter to FIPS_module_mode_set().
1469 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1472 *) Use separate DRBG fields for internal and external flags. New function
1473 FIPS_drbg_health_check() to perform on demand health checking. Add
1474 generation tests to fips_test_suite with reduced health check interval to
1475 demonstrate periodic health checking. Add "nodh" option to
1476 fips_test_suite to skip very slow DH test.
1479 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1483 *) More extensive health check for DRBG checking many more failure modes.
1484 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1485 combination: call this in fips_test_suite.
1488 *) Add support for canonical generation of DSA parameter 'g'. See
1491 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1492 POST to handle HMAC cases.
1495 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1496 to return numerical and string versions of the FIPS module number.
1499 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1500 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1501 outside the validated module in the FIPS capable OpenSSL.
1504 *) Minor change to DRBG entropy callback semantics. In some cases
1505 there is no multiple of the block length between min_len and
1506 max_len. Allow the callback to return more than max_len bytes
1507 of entropy but discard any extra: it is the callback's responsibility
1508 to ensure that the extra data discarded does not impact the
1509 requested amount of entropy.
1512 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1513 information in FIPS186-3, SP800-57 and SP800-131A.
1516 *) CCM support via EVP. Interface is very similar to GCM case except we
1517 must supply all data in one chunk (i.e. no update, final) and the
1518 message length must be supplied if AAD is used. Add algorithm test
1522 *) Initial version of POST overhaul. Add POST callback to allow the status
1523 of POST to be monitored and/or failures induced. Modify fips_test_suite
1524 to use callback. Always run all selftests even if one fails.
1527 *) XTS support including algorithm test driver in the fips_gcmtest program.
1528 Note: this does increase the maximum key length from 32 to 64 bytes but
1529 there should be no binary compatibility issues as existing applications
1530 will never use XTS mode.
1533 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1534 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1535 performs algorithm blocking for unapproved PRNG types. Also do not
1536 set PRNG type in FIPS_mode_set(): leave this to the application.
1537 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1538 the standard OpenSSL PRNG: set additional data to a date time vector.
1541 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1542 This shouldn't present any incompatibility problems because applications
1543 shouldn't be using these directly and any that are will need to rethink
1544 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1547 *) Extensive self tests and health checking required by SP800-90 DRBG.
1548 Remove strength parameter from FIPS_drbg_instantiate and always
1549 instantiate at maximum supported strength.
1552 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1555 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1558 *) New function DH_compute_key_padded() to compute a DH key and pad with
1559 leading zeroes if needed: this complies with SP800-56A et al.
1562 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1563 anything, incomplete, subject to change and largely untested at present.
1566 *) Modify fipscanisteronly build option to only build the necessary object
1567 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1570 *) Add experimental option FIPSSYMS to give all symbols in
1571 fipscanister.o and FIPS or fips prefix. This will avoid
1572 conflicts with future versions of OpenSSL. Add perl script
1573 util/fipsas.pl to preprocess assembly language source files
1574 and rename any affected symbols.
1577 *) Add selftest checks and algorithm block of non-fips algorithms in
1578 FIPS mode. Remove DES2 from selftests.
1581 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1582 return internal method without any ENGINE dependencies. Add new
1583 tiny fips sign and verify functions.
1586 *) New build option no-ec2m to disable characteristic 2 code.
1589 *) New build option "fipscanisteronly". This only builds fipscanister.o
1590 and (currently) associated fips utilities. Uses the file Makefile.fips
1591 instead of Makefile.org as the prototype.
1594 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1595 Update fips_gcmtest to use IV generator.
1598 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1599 setting output buffer to NULL. The *Final function must be
1600 called although it will not retrieve any additional data. The tag
1601 can be set or retrieved with a ctrl. The IV length is by default 12
1602 bytes (96 bits) but can be set to an alternative value. If the IV
1603 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1607 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1608 underlying do_cipher function handles all cipher semantics itself
1609 including padding and finalisation. This is useful if (for example)
1610 an ENGINE cipher handles block padding itself. The behaviour of
1611 do_cipher is subtly changed if this flag is set: the return value
1612 is the number of characters written to the output buffer (zero is
1613 no longer an error code) or a negative error code. Also if the
1614 input buffer is NULL and length 0 finalisation should be performed.
1617 *) If a candidate issuer certificate is already part of the constructed
1618 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1621 *) Improve forward-security support: add functions
1623 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1624 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1626 for use by SSL/TLS servers; the callback function will be called whenever a
1627 new session is created, and gets to decide whether the session may be
1628 cached to make it resumable (return 0) or not (return 1). (As by the
1629 SSL/TLS protocol specifications, the session_id sent by the server will be
1630 empty to indicate that the session is not resumable; also, the server will
1631 not generate RFC 4507 (RFC 5077) session tickets.)
1633 A simple reasonable callback implementation is to return is_forward_secure.
1634 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1635 by the SSL/TLS server library, indicating whether it can provide forward
1637 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1639 *) New -verify_name option in command line utilities to set verification
1643 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1644 Add CMAC pkey methods.
1647 *) Experimental renegotiation in s_server -www mode. If the client
1648 browses /reneg connection is renegotiated. If /renegcert it is
1649 renegotiated requesting a certificate.
1652 *) Add an "external" session cache for debugging purposes to s_server. This
1653 should help trace issues which normally are only apparent in deployed
1654 multi-process servers.
1657 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1658 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1659 BIO_set_cipher() and some obscure PEM functions were changed so they
1660 can now return an error. The RAND changes required a change to the
1661 RAND_METHOD structure.
1664 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1665 a gcc attribute to warn if the result of a function is ignored. This
1666 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1667 whose return value is often ignored.
1670 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1671 These allow SCTs (signed certificate timestamps) to be requested and
1672 validated when establishing a connection.
1673 [Rob Percival <robpercival@google.com>]
1675 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1677 *) Prevent padding oracle in AES-NI CBC MAC check
1679 A MITM attacker can use a padding oracle attack to decrypt traffic
1680 when the connection uses an AES CBC cipher and the server support
1683 This issue was introduced as part of the fix for Lucky 13 padding
1684 attack (CVE-2013-0169). The padding check was rewritten to be in
1685 constant time by making sure that always the same bytes are read and
1686 compared against either the MAC or padding bytes. But it no longer
1687 checked that there was enough data to have both the MAC and padding
1690 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1694 *) Fix EVP_EncodeUpdate overflow
1696 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1697 Base64 encoding of binary data. If an attacker is able to supply very large
1698 amounts of input data then a length check can overflow resulting in a heap
1701 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1702 the PEM_write_bio* family of functions. These are mainly used within the
1703 OpenSSL command line applications, so any application which processes data
1704 from an untrusted source and outputs it as a PEM file should be considered
1705 vulnerable to this issue. User applications that call these APIs directly
1706 with large amounts of untrusted data may also be vulnerable.
1708 This issue was reported by Guido Vranken.
1712 *) Fix EVP_EncryptUpdate overflow
1714 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1715 is able to supply very large amounts of input data after a previous call to
1716 EVP_EncryptUpdate() with a partial block then a length check can overflow
1717 resulting in a heap corruption. Following an analysis of all OpenSSL
1718 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1719 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1720 the first called function after an EVP_EncryptInit(), and therefore that
1721 specific call must be safe. The second form is where the length passed to
1722 EVP_EncryptUpdate() can be seen from the code to be some small value and
1723 therefore there is no possibility of an overflow. Since all instances are
1724 one of these two forms, it is believed that there can be no overflows in
1725 internal code due to this problem. It should be noted that
1726 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1727 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1728 of these calls have also been analysed too and it is believed there are no
1729 instances in internal usage where an overflow could occur.
1731 This issue was reported by Guido Vranken.
1735 *) Prevent ASN.1 BIO excessive memory allocation
1737 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1738 a short invalid encoding can cause allocation of large amounts of memory
1739 potentially consuming excessive resources or exhausting memory.
1741 Any application parsing untrusted data through d2i BIO functions is
1742 affected. The memory based functions such as d2i_X509() are *not* affected.
1743 Since the memory based functions are used by the TLS library, TLS
1744 applications are not affected.
1746 This issue was reported by Brian Carpenter.
1752 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1753 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1754 in arbitrary stack data being returned in the buffer.
1756 This issue was reported by Guido Vranken.
1760 *) Modify behavior of ALPN to invoke callback after SNI/servername
1761 callback, such that updates to the SSL_CTX affect ALPN.
1764 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1768 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1769 methods are enabled and ssl2 is disabled the methods return NULL.
1772 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1774 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1775 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1776 provide any "EXPORT" or "LOW" strength ciphers.
1779 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1780 is by default disabled at build-time. Builds that are not configured with
1781 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1782 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1783 will need to explicitly call either of:
1785 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1787 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1789 as appropriate. Even if either of those is used, or the application
1790 explicitly uses the version-specific SSLv2_method() or its client and
1791 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1792 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1793 ciphers, and SSLv2 56-bit DES are no longer available.
1797 *) Fix a double-free in DSA code
1799 A double free bug was discovered when OpenSSL parses malformed DSA private
1800 keys and could lead to a DoS attack or memory corruption for applications
1801 that receive DSA private keys from untrusted sources. This scenario is
1804 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1809 *) Disable SRP fake user seed to address a server memory leak.
1811 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1813 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1814 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1815 was changed to ignore the "fake user" SRP seed, even if the seed
1818 Users should use SRP_VBASE_get1_by_user instead. Note that in
1819 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1820 also that even though configuring the SRP seed attempts to hide
1821 invalid usernames by continuing the handshake with fake
1822 credentials, this behaviour is not constant time and no strong
1823 guarantees are made that the handshake is indistinguishable from
1824 that of a valid user.
1828 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1830 In the BN_hex2bn function the number of hex digits is calculated using an
1831 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1832 large values of |i| this can result in |bn_expand| not allocating any
1833 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1834 field as NULL leading to a subsequent NULL ptr deref. For very large values
1835 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1836 In this case memory is allocated to the internal BIGNUM data field, but it
1837 is insufficiently sized leading to heap corruption. A similar issue exists
1838 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1839 is ever called by user applications with very large untrusted hex/dec data.
1840 This is anticipated to be a rare occurrence.
1842 All OpenSSL internal usage of these functions use data that is not expected
1843 to be untrusted, e.g. config file data or application command line
1844 arguments. If user developed applications generate config file data based
1845 on untrusted data then it is possible that this could also lead to security
1846 consequences. This is also anticipated to be rare.
1848 This issue was reported to OpenSSL by Guido Vranken.
1852 *) Fix memory issues in BIO_*printf functions
1854 The internal |fmtstr| function used in processing a "%s" format string in
1855 the BIO_*printf functions could overflow while calculating the length of a
1856 string and cause an OOB read when printing very long strings.
1858 Additionally the internal |doapr_outch| function can attempt to write to an
1859 OOB memory location (at an offset from the NULL pointer) in the event of a
1860 memory allocation failure. In 1.0.2 and below this could be caused where
1861 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1862 could be in processing a very long "%s" format string. Memory leaks can
1865 The first issue may mask the second issue dependent on compiler behaviour.
1866 These problems could enable attacks where large amounts of untrusted data
1867 is passed to the BIO_*printf functions. If applications use these functions
1868 in this way then they could be vulnerable. OpenSSL itself uses these
1869 functions when printing out human-readable dumps of ASN.1 data. Therefore
1870 applications that print this data could be vulnerable if the data is from
1871 untrusted sources. OpenSSL command line applications could also be
1872 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1873 as command line arguments.
1875 Libssl is not considered directly vulnerable. Additionally certificates etc
1876 received via remote connections via libssl are also unlikely to be able to
1877 trigger these issues because of message size limits enforced within libssl.
1879 This issue was reported to OpenSSL Guido Vranken.
1883 *) Side channel attack on modular exponentiation
1885 A side-channel attack was found which makes use of cache-bank conflicts on
1886 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1887 of RSA keys. The ability to exploit this issue is limited as it relies on
1888 an attacker who has control of code in a thread running on the same
1889 hyper-threaded core as the victim thread which is performing decryptions.
1891 This issue was reported to OpenSSL by Yuval Yarom, The University of
1892 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1893 Nadia Heninger, University of Pennsylvania with more information at
1894 http://cachebleed.info.
1898 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1899 if no keysize is specified with default_bits. This fixes an
1900 omission in an earlier change that changed all RSA/DSA key generation
1901 apps to use 2048 bits by default.
1904 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1905 *) DH small subgroups
1907 Historically OpenSSL only ever generated DH parameters based on "safe"
1908 primes. More recently (in version 1.0.2) support was provided for
1909 generating X9.42 style parameter files such as those required for RFC 5114
1910 support. The primes used in such files may not be "safe". Where an
1911 application is using DH configured with parameters based on primes that are
1912 not "safe" then an attacker could use this fact to find a peer's private
1913 DH exponent. This attack requires that the attacker complete multiple
1914 handshakes in which the peer uses the same private DH exponent. For example
1915 this could be used to discover a TLS server's private DH exponent if it's
1916 reusing the private DH exponent or it's using a static DH ciphersuite.
1918 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1919 TLS. It is not on by default. If the option is not set then the server
1920 reuses the same private DH exponent for the life of the server process and
1921 would be vulnerable to this attack. It is believed that many popular
1922 applications do set this option and would therefore not be at risk.
1924 The fix for this issue adds an additional check where a "q" parameter is
1925 available (as is the case in X9.42 based parameters). This detects the
1926 only known attack, and is the only possible defense for static DH
1927 ciphersuites. This could have some performance impact.
1929 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1930 default and cannot be disabled. This could have some performance impact.
1932 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1936 *) SSLv2 doesn't block disabled ciphers
1938 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1939 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1940 been disabled, provided that the SSLv2 protocol was not also disabled via
1943 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1944 and Sebastian Schinzel.
1948 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1950 *) BN_mod_exp may produce incorrect results on x86_64
1952 There is a carry propagating bug in the x86_64 Montgomery squaring
1953 procedure. No EC algorithms are affected. Analysis suggests that attacks
1954 against RSA and DSA as a result of this defect would be very difficult to
1955 perform and are not believed likely. Attacks against DH are considered just
1956 feasible (although very difficult) because most of the work necessary to
1957 deduce information about a private key may be performed offline. The amount
1958 of resources required for such an attack would be very significant and
1959 likely only accessible to a limited number of attackers. An attacker would
1960 additionally need online access to an unpatched system using the target
1961 private key in a scenario with persistent DH parameters and a private
1962 key that is shared between multiple clients. For example this can occur by
1963 default in OpenSSL DHE based SSL/TLS ciphersuites.
1965 This issue was reported to OpenSSL by Hanno Böck.
1969 *) Certificate verify crash with missing PSS parameter
1971 The signature verification routines will crash with a NULL pointer
1972 dereference if presented with an ASN.1 signature using the RSA PSS
1973 algorithm and absent mask generation function parameter. Since these
1974 routines are used to verify certificate signature algorithms this can be
1975 used to crash any certificate verification operation and exploited in a
1976 DoS attack. Any application which performs certificate verification is
1977 vulnerable including OpenSSL clients and servers which enable client
1980 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1984 *) X509_ATTRIBUTE memory leak
1986 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1987 memory. This structure is used by the PKCS#7 and CMS routines so any
1988 application which reads PKCS#7 or CMS data from untrusted sources is
1989 affected. SSL/TLS is not affected.
1991 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1996 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1997 This changes the decoding behaviour for some invalid messages,
1998 though the change is mostly in the more lenient direction, and
1999 legacy behaviour is preserved as much as possible.
2002 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2004 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2006 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2008 *) Alternate chains certificate forgery
2010 During certificate verification, OpenSSL will attempt to find an
2011 alternative certificate chain if the first attempt to build such a chain
2012 fails. An error in the implementation of this logic can mean that an
2013 attacker could cause certain checks on untrusted certificates to be
2014 bypassed, such as the CA flag, enabling them to use a valid leaf
2015 certificate to act as a CA and "issue" an invalid certificate.
2017 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2021 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2023 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2024 incompatibility in the handling of HMAC. The previous ABI has now been
2028 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2030 *) Malformed ECParameters causes infinite loop
2032 When processing an ECParameters structure OpenSSL enters an infinite loop
2033 if the curve specified is over a specially malformed binary polynomial
2036 This can be used to perform denial of service against any
2037 system which processes public keys, certificate requests or
2038 certificates. This includes TLS clients and TLS servers with
2039 client authentication enabled.
2041 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2045 *) Exploitable out-of-bounds read in X509_cmp_time
2047 X509_cmp_time does not properly check the length of the ASN1_TIME
2048 string and can read a few bytes out of bounds. In addition,
2049 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2052 An attacker can use this to craft malformed certificates and CRLs of
2053 various sizes and potentially cause a segmentation fault, resulting in
2054 a DoS on applications that verify certificates or CRLs. TLS clients
2055 that verify CRLs are affected. TLS clients and servers with client
2056 authentication enabled may be affected if they use custom verification
2059 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2060 independently by Hanno Böck.
2064 *) PKCS7 crash with missing EnvelopedContent
2066 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2067 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2068 with missing content and trigger a NULL pointer dereference on parsing.
2070 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2071 structures from untrusted sources are affected. OpenSSL clients and
2072 servers are not affected.
2074 This issue was reported to OpenSSL by Michal Zalewski (Google).
2078 *) CMS verify infinite loop with unknown hash function
2080 When verifying a signedData message the CMS code can enter an infinite loop
2081 if presented with an unknown hash function OID. This can be used to perform
2082 denial of service against any system which verifies signedData messages using
2084 This issue was reported to OpenSSL by Johannes Bauer.
2088 *) Race condition handling NewSessionTicket
2090 If a NewSessionTicket is received by a multi-threaded client when attempting to
2091 reuse a previous ticket then a race condition can occur potentially leading to
2092 a double free of the ticket data.
2096 *) Only support 256-bit or stronger elliptic curves with the
2097 'ecdh_auto' setting (server) or by default (client). Of supported
2098 curves, prefer P-256 (both).
2101 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2103 *) ClientHello sigalgs DoS fix
2105 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2106 invalid signature algorithms extension a NULL pointer dereference will
2107 occur. This can be exploited in a DoS attack against the server.
2109 This issue was was reported to OpenSSL by David Ramos of Stanford
2112 [Stephen Henson and Matt Caswell]
2114 *) Multiblock corrupted pointer fix
2116 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2117 feature only applies on 64 bit x86 architecture platforms that support AES
2118 NI instructions. A defect in the implementation of "multiblock" can cause
2119 OpenSSL's internal write buffer to become incorrectly set to NULL when
2120 using non-blocking IO. Typically, when the user application is using a
2121 socket BIO for writing, this will only result in a failed connection.
2122 However if some other BIO is used then it is likely that a segmentation
2123 fault will be triggered, thus enabling a potential DoS attack.
2125 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2129 *) Segmentation fault in DTLSv1_listen fix
2131 The DTLSv1_listen function is intended to be stateless and processes the
2132 initial ClientHello from many peers. It is common for user code to loop
2133 over the call to DTLSv1_listen until a valid ClientHello is received with
2134 an associated cookie. A defect in the implementation of DTLSv1_listen means
2135 that state is preserved in the SSL object from one invocation to the next
2136 that can lead to a segmentation fault. Errors processing the initial
2137 ClientHello can trigger this scenario. An example of such an error could be
2138 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2141 This issue was reported to OpenSSL by Per Allansson.
2145 *) Segmentation fault in ASN1_TYPE_cmp fix
2147 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2148 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2149 certificate signature algorithm consistency this can be used to crash any
2150 certificate verification operation and exploited in a DoS attack. Any
2151 application which performs certificate verification is vulnerable including
2152 OpenSSL clients and servers which enable client authentication.
2156 *) Segmentation fault for invalid PSS parameters fix
2158 The signature verification routines will crash with a NULL pointer
2159 dereference if presented with an ASN.1 signature using the RSA PSS
2160 algorithm and invalid parameters. Since these routines are used to verify
2161 certificate signature algorithms this can be used to crash any
2162 certificate verification operation and exploited in a DoS attack. Any
2163 application which performs certificate verification is vulnerable including
2164 OpenSSL clients and servers which enable client authentication.
2166 This issue was was reported to OpenSSL by Brian Carpenter.
2170 *) ASN.1 structure reuse memory corruption fix
2172 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2173 memory corruption via an invalid write. Such reuse is and has been
2174 strongly discouraged and is believed to be rare.
2176 Applications that parse structures containing CHOICE or ANY DEFINED BY
2177 components may be affected. Certificate parsing (d2i_X509 and related
2178 functions) are however not affected. OpenSSL clients and servers are
2183 *) PKCS7 NULL pointer dereferences fix
2185 The PKCS#7 parsing code does not handle missing outer ContentInfo
2186 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2187 missing content and trigger a NULL pointer dereference on parsing.
2189 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2190 otherwise parse PKCS#7 structures from untrusted sources are
2191 affected. OpenSSL clients and servers are not affected.
2193 This issue was reported to OpenSSL by Michal Zalewski (Google).
2197 *) DoS via reachable assert in SSLv2 servers fix
2199 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2200 servers that both support SSLv2 and enable export cipher suites by sending
2201 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2203 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2204 (OpenSSL development team).
2208 *) Empty CKE with client auth and DHE fix
2210 If client auth is used then a server can seg fault in the event of a DHE
2211 ciphersuite being selected and a zero length ClientKeyExchange message
2212 being sent by the client. This could be exploited in a DoS attack.
2216 *) Handshake with unseeded PRNG fix
2218 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2219 with an unseeded PRNG. The conditions are:
2220 - The client is on a platform where the PRNG has not been seeded
2221 automatically, and the user has not seeded manually
2222 - A protocol specific client method version has been used (i.e. not
2223 SSL_client_methodv23)
2224 - A ciphersuite is used that does not require additional random data from
2225 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2227 If the handshake succeeds then the client random that has been used will
2228 have been generated from a PRNG with insufficient entropy and therefore the
2229 output may be predictable.
2231 For example using the following command with an unseeded openssl will
2232 succeed on an unpatched platform:
2234 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2238 *) Use After Free following d2i_ECPrivatekey error fix
2240 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2241 could cause a use after free condition. This, in turn, could cause a double
2242 free in several private key parsing functions (such as d2i_PrivateKey
2243 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2244 for applications that receive EC private keys from untrusted
2245 sources. This scenario is considered rare.
2247 This issue was discovered by the BoringSSL project and fixed in their
2252 *) X509_to_X509_REQ NULL pointer deref fix
2254 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2255 the certificate key is invalid. This function is rarely used in practice.
2257 This issue was discovered by Brian Carpenter.
2261 *) Removed the export ciphers from the DEFAULT ciphers
2264 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2266 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2267 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2268 So far those who have to target multiple platforms would compromise
2269 and argue that binary targeting say ARMv5 would still execute on
2270 ARMv8. "Universal" build resolves this compromise by providing
2271 near-optimal performance even on newer platforms.
2274 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2275 (other platforms pending).
2276 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2278 *) Add support for the SignedCertificateTimestampList certificate and
2279 OCSP response extensions from RFC6962.
2282 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2283 for corner cases. (Certain input points at infinity could lead to
2284 bogus results, with non-infinity inputs mapped to infinity too.)
2287 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2288 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2289 common cases are optimized and there still is room for further
2290 improvements. Vector Permutation AES for Altivec is also added.
2293 *) Add support for little-endian ppc64 Linux target.
2294 [Marcelo Cerri (IBM)]
2296 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2297 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2298 are optimized and there still is room for further improvements.
2299 Both 32- and 64-bit modes are supported.
2300 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2302 *) Improved ARMv7 NEON support.
2305 *) Support for SPARC Architecture 2011 crypto extensions, first
2306 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2307 SHA256/512, MD5, GHASH and modular exponentiation.
2308 [Andy Polyakov, David Miller]
2310 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2312 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2314 *) Support for new and upcoming Intel processors, including AVX2,
2315 BMI and SHA ISA extensions. This includes additional "stitched"
2316 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2319 This work was sponsored by Intel Corp.
2322 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2323 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2324 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2327 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2328 this fixes a limitation in previous versions of OpenSSL.
2331 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2332 MGF1 digest and OAEP label.
2335 *) Add EVP support for key wrapping algorithms, to avoid problems with
2336 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2337 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2338 algorithms and include tests cases.
2341 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2343 [Douglas E. Engert, Steve Henson]
2345 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2346 difference in days and seconds between two tm or ASN1_TIME structures.
2349 *) Add -rev test option to s_server to just reverse order of characters
2350 received by client and send back to server. Also prints an abbreviated
2351 summary of the connection parameters.
2354 *) New option -brief for s_client and s_server to print out a brief summary
2355 of connection parameters.
2358 *) Add callbacks for arbitrary TLS extensions.
2359 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2361 *) New option -crl_download in several openssl utilities to download CRLs
2362 from CRLDP extension in certificates.
2365 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2368 *) New function X509_CRL_diff to generate a delta CRL from the difference
2369 of two full CRLs. Add support to "crl" utility.
2372 *) New functions to set lookup_crls function and to retrieve
2373 X509_STORE from X509_STORE_CTX.
2376 *) Print out deprecated issuer and subject unique ID fields in
2380 *) Extend OCSP I/O functions so they can be used for simple general purpose
2381 HTTP as well as OCSP. New wrapper function which can be used to download
2382 CRLs using the OCSP API.
2385 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2388 *) SSL_CONF* functions. These provide a common framework for application
2389 configuration using configuration files or command lines.
2392 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2393 message callback and prints the results. Needs compile time option
2394 "enable-ssl-trace". New options to s_client and s_server to enable
2398 *) New ctrl and macro to retrieve supported points extensions.
2399 Print out extension in s_server and s_client.
2402 *) New functions to retrieve certificate signature and signature
2406 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2410 *) New Suite B modes for TLS code. These use and enforce the requirements
2411 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2412 only use Suite B curves. The Suite B modes can be set by using the
2413 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2416 *) New chain verification flags for Suite B levels of security. Check
2417 algorithms are acceptable when flags are set in X509_verify_cert.
2420 *) Make tls1_check_chain return a set of flags indicating checks passed
2421 by a certificate chain. Add additional tests to handle client
2422 certificates: checks for matching certificate type and issuer name
2426 *) If an attempt is made to use a signature algorithm not in the peer
2427 preference list abort the handshake. If client has no suitable
2428 signature algorithms in response to a certificate request do not
2429 use the certificate.
2432 *) If server EC tmp key is not in client preference list abort handshake.
2435 *) Add support for certificate stores in CERT structure. This makes it
2436 possible to have different stores per SSL structure or one store in
2437 the parent SSL_CTX. Include distinct stores for certificate chain
2438 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2439 to build and store a certificate chain in CERT structure: returning
2440 an error if the chain cannot be built: this will allow applications
2441 to test if a chain is correctly configured.
2443 Note: if the CERT based stores are not set then the parent SSL_CTX
2444 store is used to retain compatibility with existing behaviour.
2448 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2449 mask based on the current session, check mask when sending client
2450 hello and checking the requested ciphersuite.
2453 *) New ctrls to retrieve and set certificate types in a certificate
2454 request message. Print out received values in s_client. If certificate
2455 types is not set with custom values set sensible values based on
2456 supported signature algorithms.
2459 *) Support for distinct client and server supported signature algorithms.
2462 *) Add certificate callback. If set this is called whenever a certificate
2463 is required by client or server. An application can decide which
2464 certificate chain to present based on arbitrary criteria: for example
2465 supported signature algorithms. Add very simple example to s_server.
2466 This fixes many of the problems and restrictions of the existing client
2467 certificate callback: for example you can now clear an existing
2468 certificate and specify the whole chain.
2471 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2472 the certificate can be used for (if anything). Set valid_flags field
2473 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2474 to have similar checks in it.
2476 Add new "cert_flags" field to CERT structure and include a "strict mode".
2477 This enforces some TLS certificate requirements (such as only permitting
2478 certificate signature algorithms contained in the supported algorithms
2479 extension) which some implementations ignore: this option should be used
2480 with caution as it could cause interoperability issues.
2483 *) Update and tidy signature algorithm extension processing. Work out
2484 shared signature algorithms based on preferences and peer algorithms
2485 and print them out in s_client and s_server. Abort handshake if no
2486 shared signature algorithms.
2489 *) Add new functions to allow customised supported signature algorithms
2490 for SSL and SSL_CTX structures. Add options to s_client and s_server
2494 *) New function SSL_certs_clear() to delete all references to certificates
2495 from an SSL structure. Before this once a certificate had been added
2496 it couldn't be removed.
2499 *) Integrate hostname, email address and IP address checking with certificate
2500 verification. New verify options supporting checking in openssl utility.
2503 *) Fixes and wildcard matching support to hostname and email checking
2504 functions. Add manual page.
2505 [Florian Weimer (Red Hat Product Security Team)]
2507 *) New functions to check a hostname email or IP address against a
2508 certificate. Add options x509 utility to print results of checks against
2512 *) Fix OCSP checking.
2513 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2515 *) Initial experimental support for explicitly trusted non-root CAs.
2516 OpenSSL still tries to build a complete chain to a root but if an
2517 intermediate CA has a trust setting included that is used. The first
2518 setting is used: whether to trust (e.g., -addtrust option to the x509
2522 *) Add -trusted_first option which attempts to find certificates in the
2523 trusted store even if an untrusted chain is also supplied.
2526 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2527 platform support for Linux and Android.
2530 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2533 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2534 When in FIPS mode the approved implementations are used as normal,
2535 when not in FIPS mode the internal unapproved versions are used instead.
2536 This means that the FIPS capable OpenSSL isn't forced to use the
2537 (often lower performance) FIPS implementations outside FIPS mode.
2540 *) Transparently support X9.42 DH parameters when calling
2541 PEM_read_bio_DHparameters. This means existing applications can handle
2542 the new parameter format automatically.
2545 *) Initial experimental support for X9.42 DH parameter format: mainly
2546 to support use of 'q' parameter for RFC5114 parameters.
2549 *) Add DH parameters from RFC5114 including test data to dhtest.
2552 *) Support for automatic EC temporary key parameter selection. If enabled
2553 the most preferred EC parameters are automatically used instead of
2554 hardcoded fixed parameters. Now a server just has to call:
2555 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2556 support ECDH and use the most appropriate parameters.
2559 *) Enhance and tidy EC curve and point format TLS extension code. Use
2560 static structures instead of allocation if default values are used.
2561 New ctrls to set curves we wish to support and to retrieve shared curves.
2562 Print out shared curves in s_server. New options to s_server and s_client
2563 to set list of supported curves.
2566 *) New ctrls to retrieve supported signature algorithms and
2567 supported curve values as an array of NIDs. Extend openssl utility
2568 to print out received values.
2571 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2572 between NIDs and the more common NIST names such as "P-256". Enhance
2573 ecparam utility and ECC method to recognise the NIST names for curves.
2576 *) Enhance SSL/TLS certificate chain handling to support different
2577 chains for each certificate instead of one chain in the parent SSL_CTX.
2580 *) Support for fixed DH ciphersuite client authentication: where both
2581 server and client use DH certificates with common parameters.
2584 *) Support for fixed DH ciphersuites: those requiring DH server
2588 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2590 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2591 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2592 X509_CINF_get_signature were reverted post internal team review.
2594 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2596 *) Build fixes for the Windows and OpenVMS platforms
2597 [Matt Caswell and Richard Levitte]
2599 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2601 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2602 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2603 dereference. This could lead to a Denial Of Service attack. Thanks to
2604 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2608 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2609 dtls1_buffer_record function under certain conditions. In particular this
2610 could occur if an attacker sent repeated DTLS records with the same
2611 sequence number but for the next epoch. The memory leak could be exploited
2612 by an attacker in a Denial of Service attack through memory exhaustion.
2613 Thanks to Chris Mueller for reporting this issue.
2617 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2618 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2619 method would be set to NULL which could later result in a NULL pointer
2620 dereference. Thanks to Frank Schmirler for reporting this issue.
2624 *) Abort handshake if server key exchange message is omitted for ephemeral
2627 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2628 reporting this issue.
2632 *) Remove non-export ephemeral RSA code on client and server. This code
2633 violated the TLS standard by allowing the use of temporary RSA keys in
2634 non-export ciphersuites and could be used by a server to effectively
2635 downgrade the RSA key length used to a value smaller than the server
2636 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2637 INRIA or reporting this issue.
2641 *) Fixed issue where DH client certificates are accepted without verification.
2642 An OpenSSL server will accept a DH certificate for client authentication
2643 without the certificate verify message. This effectively allows a client to
2644 authenticate without the use of a private key. This only affects servers
2645 which trust a client certificate authority which issues certificates
2646 containing DH keys: these are extremely rare and hardly ever encountered.
2647 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2652 *) Ensure that the session ID context of an SSL is updated when its
2653 SSL_CTX is updated via SSL_set_SSL_CTX.
2655 The session ID context is typically set from the parent SSL_CTX,
2656 and can vary with the CTX.
2659 *) Fix various certificate fingerprint issues.
2661 By using non-DER or invalid encodings outside the signed portion of a
2662 certificate the fingerprint can be changed without breaking the signature.
2663 Although no details of the signed portion of the certificate can be changed
2664 this can cause problems with some applications: e.g. those using the
2665 certificate fingerprint for blacklists.
2667 1. Reject signatures with non zero unused bits.
2669 If the BIT STRING containing the signature has non zero unused bits reject
2670 the signature. All current signature algorithms require zero unused bits.
2672 2. Check certificate algorithm consistency.
2674 Check the AlgorithmIdentifier inside TBS matches the one in the
2675 certificate signature. NB: this will result in signature failure
2676 errors for some broken certificates.
2678 Thanks to Konrad Kraszewski from Google for reporting this issue.
2680 3. Check DSA/ECDSA signatures use DER.
2682 Re-encode DSA/ECDSA signatures and compare with the original received
2683 signature. Return an error if there is a mismatch.
2685 This will reject various cases including garbage after signature
2686 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2687 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2688 (negative or with leading zeroes).
2690 Further analysis was conducted and fixes were developed by Stephen Henson
2691 of the OpenSSL core team.
2696 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2697 results on some platforms, including x86_64. This bug occurs at random
2698 with a very low probability, and is not known to be exploitable in any
2699 way, though its exact impact is difficult to determine. Thanks to Pieter
2700 Wuille (Blockstream) who reported this issue and also suggested an initial
2701 fix. Further analysis was conducted by the OpenSSL development team and
2702 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2703 the OpenSSL core team.
2707 *) Do not resume sessions on the server if the negotiated protocol
2708 version does not match the session's version. Resuming with a different
2709 version, while not strictly forbidden by the RFC, is of questionable
2710 sanity and breaks all known clients.
2711 [David Benjamin, Emilia Käsper]
2713 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2714 early CCS messages during renegotiation. (Note that because
2715 renegotiation is encrypted, this early CCS was not exploitable.)
2718 *) Tighten client-side session ticket handling during renegotiation:
2719 ensure that the client only accepts a session ticket if the server sends
2720 the extension anew in the ServerHello. Previously, a TLS client would
2721 reuse the old extension state and thus accept a session ticket if one was
2722 announced in the initial ServerHello.
2724 Similarly, ensure that the client requires a session ticket if one
2725 was advertised in the ServerHello. Previously, a TLS client would
2726 ignore a missing NewSessionTicket message.
2729 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2731 *) SRTP Memory Leak.
2733 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2734 sends a carefully crafted handshake message, to cause OpenSSL to fail
2735 to free up to 64k of memory causing a memory leak. This could be
2736 exploited in a Denial Of Service attack. This issue affects OpenSSL
2737 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2738 whether SRTP is used or configured. Implementations of OpenSSL that
2739 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2741 The fix was developed by the OpenSSL team.
2745 *) Session Ticket Memory Leak.
2747 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2748 integrity of that ticket is first verified. In the event of a session
2749 ticket integrity check failing, OpenSSL will fail to free memory
2750 causing a memory leak. By sending a large number of invalid session
2751 tickets an attacker could exploit this issue in a Denial Of Service
2756 *) Build option no-ssl3 is incomplete.
2758 When OpenSSL is configured with "no-ssl3" as a build option, servers
2759 could accept and complete a SSL 3.0 handshake, and clients could be
2760 configured to send them.
2762 [Akamai and the OpenSSL team]
2764 *) Add support for TLS_FALLBACK_SCSV.
2765 Client applications doing fallback retries should call
2766 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2768 [Adam Langley, Bodo Moeller]
2770 *) Add additional DigestInfo checks.
2772 Re-encode DigestInto in DER and check against the original when
2773 verifying RSA signature: this will reject any improperly encoded
2774 DigestInfo structures.
2776 Note: this is a precautionary measure and no attacks are currently known.
2780 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2782 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2783 SRP code can be overrun an internal buffer. Add sanity check that
2784 g, A, B < N to SRP code.
2786 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2787 Group for discovering this issue.
2791 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2792 TLS 1.0 instead of higher protocol versions when the ClientHello message
2793 is badly fragmented. This allows a man-in-the-middle attacker to force a
2794 downgrade to TLS 1.0 even if both the server and the client support a
2795 higher protocol version, by modifying the client's TLS records.
2797 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2798 researching this issue.
2802 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2803 to a denial of service attack. A malicious server can crash the client
2804 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2805 ciphersuite and sending carefully crafted handshake messages.
2807 Thanks to Felix Gröbert (Google) for discovering and researching this
2812 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2813 to leak memory. This can be exploited through a Denial of Service attack.
2814 Thanks to Adam Langley for discovering and researching this issue.
2818 *) An attacker can force openssl to consume large amounts of memory whilst
2819 processing DTLS handshake messages. This can be exploited through a
2820 Denial of Service attack.
2821 Thanks to Adam Langley for discovering and researching this issue.
2825 *) An attacker can force an error condition which causes openssl to crash
2826 whilst processing DTLS packets due to memory being freed twice. This
2827 can be exploited through a Denial of Service attack.
2828 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2833 *) If a multithreaded client connects to a malicious server using a resumed
2834 session and the server sends an ec point format extension it could write
2835 up to 255 bytes to freed memory.
2837 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2842 *) A malicious server can crash an OpenSSL client with a null pointer
2843 dereference (read) by specifying an SRP ciphersuite even though it was not
2844 properly negotiated with the client. This can be exploited through a
2845 Denial of Service attack.
2847 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2848 discovering and researching this issue.
2852 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2853 X509_name_oneline, X509_name_print_ex et al. to leak some information
2854 from the stack. Applications may be affected if they echo pretty printing
2855 output to the attacker.
2857 Thanks to Ivan Fratric (Google) for discovering this issue.
2859 [Emilia Käsper, and Steve Henson]
2861 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2862 for corner cases. (Certain input points at infinity could lead to
2863 bogus results, with non-infinity inputs mapped to infinity too.)
2866 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2868 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2869 handshake can force the use of weak keying material in OpenSSL
2870 SSL/TLS clients and servers.
2872 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2873 researching this issue. (CVE-2014-0224)
2874 [KIKUCHI Masashi, Steve Henson]
2876 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2877 OpenSSL DTLS client the code can be made to recurse eventually crashing
2880 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2882 [Imre Rad, Steve Henson]
2884 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2885 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2886 client or server. This is potentially exploitable to run arbitrary
2887 code on a vulnerable client or server.
2889 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2890 [Jüri Aedla, Steve Henson]
2892 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2893 are subject to a denial of service attack.
2895 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2896 this issue. (CVE-2014-3470)
2897 [Felix Gröbert, Ivan Fratric, Steve Henson]
2899 *) Harmonize version and its documentation. -f flag is used to display
2901 [mancha <mancha1@zoho.com>]
2903 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2904 in i2d_ECPrivateKey.
2905 [mancha <mancha1@zoho.com>]
2907 *) Fix some double frees. These are not thought to be exploitable.
2908 [mancha <mancha1@zoho.com>]
2910 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2912 *) A missing bounds check in the handling of the TLS heartbeat extension
2913 can be used to reveal up to 64k of memory to a connected client or
2916 Thanks for Neel Mehta of Google Security for discovering this bug and to
2917 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2918 preparing the fix (CVE-2014-0160)
2919 [Adam Langley, Bodo Moeller]
2921 *) Fix for the attack described in the paper "Recovering OpenSSL
2922 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2923 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2924 http://eprint.iacr.org/2014/140
2926 Thanks to Yuval Yarom and Naomi Benger for discovering this
2927 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2928 [Yuval Yarom and Naomi Benger]
2930 *) TLS pad extension: draft-agl-tls-padding-03
2932 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2933 TLS client Hello record length value would otherwise be > 255 and
2934 less that 512 pad with a dummy extension containing zeroes so it
2935 is at least 512 bytes long.
2937 [Adam Langley, Steve Henson]
2939 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2941 *) Fix for TLS record tampering bug. A carefully crafted invalid
2942 handshake could crash OpenSSL with a NULL pointer exception.
2943 Thanks to Anton Johansson for reporting this issues.
2946 *) Keep original DTLS digest and encryption contexts in retransmission
2947 structures so we can use the previous session parameters if they need
2948 to be resent. (CVE-2013-6450)
2951 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2952 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2953 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2954 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2955 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2956 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2957 [Rob Stradling, Adam Langley]
2959 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2961 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2962 supporting platforms or when small records were transferred.
2963 [Andy Polyakov, Steve Henson]
2965 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2967 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2969 This addresses the flaw in CBC record processing discovered by
2970 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2971 at: http://www.isg.rhul.ac.uk/tls/
2973 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2974 Security Group at Royal Holloway, University of London
2975 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2976 Emilia Käsper for the initial patch.
2978 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2980 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2981 ciphersuites which can be exploited in a denial of service attack.
2982 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2983 and detecting this bug and to Wolfgang Ettlinger
2984 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2988 *) Return an error when checking OCSP signatures when key is NULL.
2989 This fixes a DoS attack. (CVE-2013-0166)
2992 *) Make openssl verify return errors.
2993 [Chris Palmer <palmer@google.com> and Ben Laurie]
2995 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2996 the right response is stapled. Also change SSL_get_certificate()
2997 so it returns the certificate actually sent.
2998 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2999 [Rob Stradling <rob.stradling@comodo.com>]
3001 *) Fix possible deadlock when decoding public keys.
3004 *) Don't use TLS 1.0 record version number in initial client hello
3008 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3010 *) Sanity check record length before skipping explicit IV in TLS
3011 1.2, 1.1 and DTLS to fix DoS attack.
3013 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3014 fuzzing as a service testing platform.
3018 *) Initialise tkeylen properly when encrypting CMS messages.
3019 Thanks to Solar Designer of Openwall for reporting this issue.
3022 *) In FIPS mode don't try to use composite ciphers as they are not
3026 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3028 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3029 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3030 mean any application compiled against OpenSSL 1.0.0 headers setting
3031 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3032 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3033 0x10000000L Any application which was previously compiled against
3034 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3035 will need to be recompiled as a result. Letting be results in
3036 inability to disable specifically TLS 1.1 and in client context,
3037 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3040 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3041 disable just protocol X, but all protocols above X *if* there are
3042 protocols *below* X still enabled. In more practical terms it means
3043 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3044 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3045 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3049 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3051 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3052 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3053 in CRYPTO_realloc_clean.
3055 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3056 issue and to Adam Langley <agl@chromium.org> for fixing it.
3058 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3060 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3063 *) Workarounds for some broken servers that "hang" if a client hello
3064 record length exceeds 255 bytes.
3066 1. Do not use record version number > TLS 1.0 in initial client
3067 hello: some (but not all) hanging servers will now work.
3068 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3069 the number of ciphers sent in the client hello. This should be
3070 set to an even number, such as 50, for example by passing:
3071 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3072 Most broken servers should now work.
3073 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3074 TLS 1.2 client support entirely.
3077 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3080 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3082 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3083 STRING form instead of a DigestInfo.
3086 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3087 and the RSA_sign/RSA_verify functions. This was made more apparent when
3088 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3089 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3090 the correct format in RSA_verify so both forms transparently work.
3093 *) Some servers which support TLS 1.0 can choke if we initially indicate
3094 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3095 encrypted premaster secret. As a workaround use the maximum permitted
3096 client version in client hello, this should keep such servers happy
3097 and still work with previous versions of OpenSSL.
3100 *) Add support for TLS/DTLS heartbeats.
3101 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3103 *) Add support for SCTP.
3104 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3106 *) Improved PRNG seeding for VOS.
3107 [Paul Green <Paul.Green@stratus.com>]
3109 *) Extensive assembler packs updates, most notably:
3111 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3112 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3113 - x86_64: bit-sliced AES implementation;
3114 - ARM: NEON support, contemporary platforms optimizations;
3115 - s390x: z196 support;
3116 - *: GHASH and GF(2^m) multiplication implementations;
3120 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3121 (removal of unnecessary code)
3122 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3124 *) Add TLS key material exporter from RFC 5705.
3127 *) Add DTLS-SRTP negotiation from RFC 5764.
3130 *) Add Next Protocol Negotiation,
3131 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3132 disabled with a no-npn flag to config or Configure. Code donated
3134 [Adam Langley <agl@google.com> and Ben Laurie]
3136 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3137 NIST-P256, NIST-P521, with constant-time single point multiplication on
3138 typical inputs. Compiler support for the nonstandard type __uint128_t is
3139 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3140 Code made available under Apache License version 2.0.
3142 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3143 line to include this in your build of OpenSSL, and run "make depend" (or
3144 "make update"). This enables the following EC_METHODs:
3146 EC_GFp_nistp224_method()
3147 EC_GFp_nistp256_method()
3148 EC_GFp_nistp521_method()
3150 EC_GROUP_new_by_curve_name() will automatically use these (while
3151 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3153 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3155 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3156 all platforms. Move ssize_t definition from e_os.h to the public
3157 header file e_os2.h as it now appears in public header file cms.h
3160 *) New -sigopt option to the ca, req and x509 utilities. Additional
3161 signature parameters can be passed using this option and in
3165 *) Add RSA PSS signing function. This will generate and set the
3166 appropriate AlgorithmIdentifiers for PSS based on those in the
3167 corresponding EVP_MD_CTX structure. No application support yet.
3170 *) Support for companion algorithm specific ASN1 signing routines.
3171 New function ASN1_item_sign_ctx() signs a pre-initialised
3172 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3173 the appropriate parameters.
3176 *) Add new algorithm specific ASN1 verification initialisation function
3177 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3178 handling will be the same no matter what EVP_PKEY_METHOD is used.
3179 Add a PSS handler to support verification of PSS signatures: checked
3180 against a number of sample certificates.
3183 *) Add signature printing for PSS. Add PSS OIDs.
3184 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3186 *) Add algorithm specific signature printing. An individual ASN1 method
3187 can now print out signatures instead of the standard hex dump.
3189 More complex signatures (e.g. PSS) can print out more meaningful
3190 information. Include DSA version that prints out the signature
3194 *) Password based recipient info support for CMS library: implementing
3198 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3199 neatly separates the code into cipher and PBE sections and is required
3200 for some algorithms that split PBES2 into separate pieces (such as
3201 password based CMS).
3204 *) Session-handling fixes:
3205 - Fix handling of connections that are resuming with a session ID,
3206 but also support Session Tickets.
3207 - Fix a bug that suppressed issuing of a new ticket if the client
3208 presented a ticket with an expired session.
3209 - Try to set the ticket lifetime hint to something reasonable.
3210 - Make tickets shorter by excluding irrelevant information.
3211 - On the client side, don't ignore renewed tickets.
3212 [Adam Langley, Bodo Moeller (Google)]
3214 *) Fix PSK session representation.
3217 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3219 This work was sponsored by Intel.
3222 *) Add GCM support to TLS library. Some custom code is needed to split
3223 the IV between the fixed (from PRF) and explicit (from TLS record)
3224 portions. This adds all GCM ciphersuites supported by RFC5288 and
3225 RFC5289. Generalise some AES* cipherstrings to include GCM and
3226 add a special AESGCM string for GCM only.
3229 *) Expand range of ctrls for AES GCM. Permit setting invocation
3230 field on decrypt and retrieval of invocation field only on encrypt.
3233 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3234 As required by RFC5289 these ciphersuites cannot be used if for
3235 versions of TLS earlier than 1.2.
3238 *) For FIPS capable OpenSSL interpret a NULL default public key method
3239 as unset and return the appropriate default but do *not* set the default.
3240 This means we can return the appropriate method in applications that
3241 switch between FIPS and non-FIPS modes.
3244 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3245 ENGINE is used then we cannot handle that in the FIPS module so we
3246 keep original code iff non-FIPS operations are allowed.
3249 *) Add -attime option to openssl utilities.
3250 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3252 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3255 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3256 FIPS EC methods unconditionally for now.
3259 *) New build option no-ec2m to disable characteristic 2 code.
3262 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3263 all cases can be covered as some introduce binary incompatibilities.
3266 *) Redirect RSA operations to FIPS module including keygen,
3267 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3270 *) Add similar low level API blocking to ciphers.
3273 *) Low level digest APIs are not approved in FIPS mode: any attempt
3274 to use these will cause a fatal error. Applications that *really* want
3275 to use them can use the private_* version instead.
3278 *) Redirect cipher operations to FIPS module for FIPS builds.
3281 *) Redirect digest operations to FIPS module for FIPS builds.
3284 *) Update build system to add "fips" flag which will link in fipscanister.o
3285 for static and shared library builds embedding a signature if needed.
3288 *) Output TLS supported curves in preference order instead of numerical
3289 order. This is currently hardcoded for the highest order curves first.
3290 This should be configurable so applications can judge speed vs strength.
3293 *) Add TLS v1.2 server support for client authentication.
3296 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3300 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3301 FIPS modules versions.
3304 *) Add TLS v1.2 client side support for client authentication. Keep cache
3305 of handshake records longer as we don't know the hash algorithm to use
3306 until after the certificate request message is received.
3309 *) Initial TLS v1.2 client support. Add a default signature algorithms
3310 extension including all the algorithms we support. Parse new signature
3311 format in client key exchange. Relax some ECC signing restrictions for
3312 TLS v1.2 as indicated in RFC5246.
3315 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3316 to new signature format when needed using client digest preference.
3317 All server ciphersuites should now work correctly in TLS v1.2. No client
3318 support yet and no support for client certificates.
3321 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3322 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3323 ciphersuites. At present only RSA key exchange ciphersuites work with
3324 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3325 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3326 and version checking.
3329 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3330 with this defined it will not be affected by any changes to ssl internal
3331 structures. Add several utility functions to allow openssl application
3332 to work with OPENSSL_NO_SSL_INTERN defined.
3335 *) A long standing patch to add support for SRP from EdelWeb (Peter
3336 Sylvester and Christophe Renou) was integrated.
3337 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3338 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3341 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3344 *) Permit abbreviated handshakes when renegotiating using the function
3345 SSL_renegotiate_abbreviated().
3346 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3348 *) Add call to ENGINE_register_all_complete() to
3349 ENGINE_load_builtin_engines(), so some implementations get used
3350 automatically instead of needing explicit application support.
3353 *) Add support for TLS key exporter as described in RFC5705.
3354 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3356 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3357 a few changes are required:
3359 Add SSL_OP_NO_TLSv1_1 flag.
3360 Add TLSv1_1 methods.
3361 Update version checking logic to handle version 1.1.
3362 Add explicit IV handling (ported from DTLS code).
3363 Add command line options to s_client/s_server.
3366 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3368 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3369 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3370 content decryption and always return the same error. Note: this attack
3371 needs on average 2^20 messages so it only affects automated senders. The
3372 old behaviour can be re-enabled in the CMS code by setting the
3373 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3374 an MMA defence is not necessary.
3375 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3376 this issue. (CVE-2012-0884)
3379 *) Fix CVE-2011-4619: make sure we really are receiving a
3380 client hello before rejecting multiple SGC restarts. Thanks to
3381 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3384 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3386 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3387 Thanks to Antonio Martin, Enterprise Secure Access Research and
3388 Development, Cisco Systems, Inc. for discovering this bug and
3389 preparing a fix. (CVE-2012-0050)
3392 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3394 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3395 of the Vaudenay padding oracle attack on CBC mode encryption
3396 which enables an efficient plaintext recovery attack against
3397 the OpenSSL implementation of DTLS. Their attack exploits timing
3398 differences arising during decryption processing. A research
3399 paper describing this attack can be found at:
3400 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3401 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3402 Security Group at Royal Holloway, University of London
3403 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3404 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3405 for preparing the fix. (CVE-2011-4108)
3406 [Robin Seggelmann, Michael Tuexen]
3408 *) Clear bytes used for block padding of SSL 3.0 records.
3410 [Adam Langley (Google)]
3412 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3413 Kadianakis <desnacked@gmail.com> for discovering this issue and
3414 Adam Langley for preparing the fix. (CVE-2011-4619)
3415 [Adam Langley (Google)]
3417 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3418 [Andrey Kulikov <amdeich@gmail.com>]
3420 *) Prevent malformed RFC3779 data triggering an assertion failure.
3421 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3422 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3423 [Rob Austein <sra@hactrn.net>]
3425 *) Improved PRNG seeding for VOS.
3426 [Paul Green <Paul.Green@stratus.com>]
3428 *) Fix ssl_ciph.c set-up race.
3429 [Adam Langley (Google)]
3431 *) Fix spurious failures in ecdsatest.c.
3432 [Emilia Käsper (Google)]
3434 *) Fix the BIO_f_buffer() implementation (which was mixing different
3435 interpretations of the '..._len' fields).
3436 [Adam Langley (Google)]
3438 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3439 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3440 threads won't reuse the same blinding coefficients.
3442 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3443 lock to call BN_BLINDING_invert_ex, and avoids one use of
3444 BN_BLINDING_update for each BN_BLINDING structure (previously,
3445 the last update always remained unused).
3446 [Emilia Käsper (Google)]
3448 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3449 [Bob Buckholz (Google)]
3451 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3453 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3454 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3455 [Kaspar Brand <ossl@velox.ch>]
3457 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3458 for multi-threaded use of ECDH. (CVE-2011-3210)
3459 [Adam Langley (Google)]
3461 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3464 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3465 signature public key algorithm by using OID xref utilities instead.
3466 Before this you could only use some ECC ciphersuites with SHA1 only.
3469 *) Add protection against ECDSA timing attacks as mentioned in the paper
3470 by Billy Bob Brumley and Nicola Tuveri, see:
3472 http://eprint.iacr.org/2011/232.pdf
3474 [Billy Bob Brumley and Nicola Tuveri]
3476 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3478 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3479 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3481 *) Fix bug in string printing code: if *any* escaping is enabled we must
3482 escape the escape character (backslash) or the resulting string is
3486 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3488 *) Disable code workaround for ancient and obsolete Netscape browsers
3489 and servers: an attacker can use it in a ciphersuite downgrade attack.
3490 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3493 *) Fixed J-PAKE implementation error, originally discovered by
3494 Sebastien Martini, further info and confirmation from Stefan
3495 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3498 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3500 *) Fix extension code to avoid race conditions which can result in a buffer
3501 overrun vulnerability: resumed sessions must not be modified as they can
3502 be shared by multiple threads. CVE-2010-3864
3505 *) Fix WIN32 build system to correctly link an ENGINE directory into
3509 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3511 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3513 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3515 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3517 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3518 context. The operation can be customised via the ctrl mechanism in
3519 case ENGINEs want to include additional functionality.
3522 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3525 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3526 output hashes compatible with older versions of OpenSSL.
3527 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3529 *) Fix compression algorithm handling: if resuming a session use the
3530 compression algorithm of the resumed session instead of determining
3531 it from client hello again. Don't allow server to change algorithm.
3534 *) Add load_crls() function to apps tidying load_certs() too. Add option
3535 to verify utility to allow additional CRLs to be included.
3538 *) Update OCSP request code to permit adding custom headers to the request:
3539 some responders need this.
3542 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3544 [Julia Lawall <julia@diku.dk>]
3546 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3547 needlessly dereferenced structures, used obsolete functions and
3548 didn't handle all updated verify codes correctly.
3551 *) Disable MD2 in the default configuration.
3554 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3555 indicate the initial BIO being pushed or popped. This makes it possible
3556 to determine whether the BIO is the one explicitly called or as a result
3557 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3558 it handles reference counts correctly and doesn't zero out the I/O bio
3559 when it is not being explicitly popped. WARNING: applications which
3560 included workarounds for the old buggy behaviour will need to be modified
3561 or they could free up already freed BIOs.
3564 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3565 renaming to all platforms (within the 0.9.8 branch, this was
3566 done conditionally on Netware platforms to avoid a name clash).
3567 [Guenter <lists@gknw.net>]
3569 *) Add ECDHE and PSK support to DTLS.
3570 [Michael Tuexen <tuexen@fh-muenster.de>]
3572 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3576 *) Add "missing" function EVP_MD_flags() (without this the only way to
3577 retrieve a digest flags is by accessing the structure directly. Update
3578 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3579 or cipher is registered as in the "from" argument. Print out all
3580 registered digests in the dgst usage message instead of manually
3581 attempting to work them out.
3584 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3585 this allows the use of compression and extensions. Change default cipher
3586 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3587 by default unless an application cipher string requests it.
3590 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3591 key ids to find matching certificates and keys but some PKCS#12 files
3592 don't follow the (somewhat unwritten) rules and this strategy fails.
3593 Now just gather all certificates together and the first private key
3594 then look for the first certificate that matches the key.
3597 *) Support use of registered digest and cipher names for dgst and cipher
3598 commands instead of having to add each one as a special case. So now
3605 openssl dgst -sha256 foo
3607 and this works for ENGINE based algorithms too.
3611 *) Update Gost ENGINE to support parameter files.
3612 [Victor B. Wagner <vitus@cryptocom.ru>]
3614 *) Support GeneralizedTime in ca utility.
3615 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3617 *) Enhance the hash format used for certificate directory links. The new
3618 form uses the canonical encoding (meaning equivalent names will work
3619 even if they aren't identical) and uses SHA1 instead of MD5. This form
3620 is incompatible with the older format and as a result c_rehash should
3621 be used to rebuild symbolic links.
3624 *) Make PKCS#8 the default write format for private keys, replacing the
3625 traditional format. This form is standardised, more secure and doesn't
3626 include an implicit MD5 dependency.
3629 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3630 committed to OpenSSL should pass this lot as a minimum.
3633 *) Add session ticket override functionality for use by EAP-FAST.
3634 [Jouni Malinen <j@w1.fi>]
3636 *) Modify HMAC functions to return a value. Since these can be implemented
3637 in an ENGINE errors can occur.
3640 *) Type-checked OBJ_bsearch_ex.
3643 *) Type-checked OBJ_bsearch. Also some constification necessitated
3644 by type-checking. Still to come: TXT_DB, bsearch(?),
3645 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3649 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3650 seconds to a tm structure directly, instead of going through OS
3651 specific date routines. This avoids any issues with OS routines such
3652 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3653 and X509_time_adj_ex() to cover the extended range. The existing
3654 X509_time_adj() is still usable and will no longer have any date issues.
3657 *) Delta CRL support. New use deltas option which will attempt to locate
3658 and search any appropriate delta CRLs available.
3660 This work was sponsored by Google.
3663 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3664 code and add additional score elements. Validate alternate CRL paths
3665 as part of the CRL checking and indicate a new error "CRL path validation
3666 error" in this case. Applications wanting additional details can use
3667 the verify callback and check the new "parent" field. If this is not
3668 NULL CRL path validation is taking place. Existing applications won't
3669 see this because it requires extended CRL support which is off by
3672 This work was sponsored by Google.
3675 *) Support for freshest CRL extension.
3677 This work was sponsored by Google.
3680 *) Initial indirect CRL support. Currently only supported in the CRLs
3681 passed directly and not via lookup. Process certificate issuer
3682 CRL entry extension and lookup CRL entries by bother issuer name
3683 and serial number. Check and process CRL issuer entry in IDP extension.
3685 This work was sponsored by Google.
3688 *) Add support for distinct certificate and CRL paths. The CRL issuer
3689 certificate is validated separately in this case. Only enabled if
3690 an extended CRL support flag is set: this flag will enable additional
3691 CRL functionality in future.
3693 This work was sponsored by Google.
3696 *) Add support for policy mappings extension.
3698 This work was sponsored by Google.
3701 *) Fixes to pathlength constraint, self issued certificate handling,
3702 policy processing to align with RFC3280 and PKITS tests.
3704 This work was sponsored by Google.
3707 *) Support for name constraints certificate extension. DN, email, DNS
3708 and URI types are currently supported.
3710 This work was sponsored by Google.
3713 *) To cater for systems that provide a pointer-based thread ID rather
3714 than numeric, deprecate the current numeric thread ID mechanism and
3715 replace it with a structure and associated callback type. This
3716 mechanism allows a numeric "hash" to be extracted from a thread ID in
3717 either case, and on platforms where pointers are larger than 'long',
3718 mixing is done to help ensure the numeric 'hash' is usable even if it
3719 can't be guaranteed unique. The default mechanism is to use "&errno"
3720 as a pointer-based thread ID to distinguish between threads.
3722 Applications that want to provide their own thread IDs should now use
3723 CRYPTO_THREADID_set_callback() to register a callback that will call
3724 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3726 Note that ERR_remove_state() is now deprecated, because it is tied
3727 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3728 to free the current thread's error state should be replaced by
3729 ERR_remove_thread_state(NULL).
3731 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3732 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3733 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3734 application was previously providing a numeric thread callback that
3735 was inappropriate for distinguishing threads, then uniqueness might
3736 have been obtained with &errno that happened immediately in the
3737 intermediate development versions of OpenSSL; this is no longer the
3738 case, the numeric thread callback will now override the automatic use
3740 [Geoff Thorpe, with help from Bodo Moeller]
3742 *) Initial support for different CRL issuing certificates. This covers a
3743 simple case where the self issued certificates in the chain exist and
3744 the real CRL issuer is higher in the existing chain.
3746 This work was sponsored by Google.
3749 *) Removed effectively defunct crypto/store from the build.
3752 *) Revamp of STACK to provide stronger type-checking. Still to come:
3753 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3754 ASN1_STRING, CONF_VALUE.
3757 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3758 RAM on SSL connections. This option can save about 34k per idle SSL.
3761 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3762 STACK, TXT_DB, bsearch, qsort.
3765 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3766 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3767 support for data, signedData, compressedData, digestedData and
3768 encryptedData, envelopedData types included. Scripts to check against
3769 RFC4134 examples draft and interop and consistency checks of many
3770 content types and variants.
3773 *) Add options to enc utility to support use of zlib compression BIO.
3776 *) Extend mk1mf to support importing of options and assembly language
3777 files from Configure script, currently only included in VC-WIN32.
3778 The assembly language rules can now optionally generate the source
3779 files from the associated perl scripts.
3782 *) Implement remaining functionality needed to support GOST ciphersuites.
3783 Interop testing has been performed using CryptoPro implementations.
3784 [Victor B. Wagner <vitus@cryptocom.ru>]
3786 *) s390x assembler pack.
3789 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3793 *) Implement Opaque PRF Input TLS extension as specified in
3794 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3795 official specification yet and no extension type assignment by
3796 IANA exists, this extension (for now) will have to be explicitly
3797 enabled when building OpenSSL by providing the extension number
3798 to use. For example, specify an option
3800 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3802 to the "config" or "Configure" script to enable the extension,
3803 assuming extension number 0x9527 (which is a completely arbitrary
3804 and unofficial assignment based on the MD5 hash of the Internet
3805 Draft). Note that by doing so, you potentially lose
3806 interoperability with other TLS implementations since these might
3807 be using the same extension number for other purposes.
3809 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3810 opaque PRF input value to use in the handshake. This will create
3811 an internal copy of the length-'len' string at 'src', and will
3812 return non-zero for success.
3814 To get more control and flexibility, provide a callback function
3817 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3818 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3822 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3825 Callback function 'cb' will be called in handshakes, and is
3826 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3827 Argument 'arg' is for application purposes (the value as given to
3828 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3829 be provided to the callback function). The callback function
3830 has to return non-zero to report success: usually 1 to use opaque
3831 PRF input just if possible, or 2 to enforce use of the opaque PRF
3832 input. In the latter case, the library will abort the handshake
3833 if opaque PRF input is not successfully negotiated.
3835 Arguments 'peerinput' and 'len' given to the callback function
3836 will always be NULL and 0 in the case of a client. A server will
3837 see the client's opaque PRF input through these variables if
3838 available (NULL and 0 otherwise). Note that if the server
3839 provides an opaque PRF input, the length must be the same as the
3840 length of the client's opaque PRF input.
3842 Note that the callback function will only be called when creating
3843 a new session (session resumption can resume whatever was
3844 previously negotiated), and will not be called in SSL 2.0
3845 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3846 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3847 for applications that need to enforce opaque PRF input.
3851 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3854 [Victor B. Wagner <vitus@cryptocom.ru>]
3856 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3857 RFC4507bis. The encrypted ticket format is an encrypted encoded
3858 SSL_SESSION structure, that way new session features are automatically
3861 If a client application caches session in an SSL_SESSION structure
3862 support is transparent because tickets are now stored in the encoded
3865 The SSL_CTX structure automatically generates keys for ticket
3866 protection in servers so again support should be possible
3867 with no application modification.
3869 If a client or server wishes to disable RFC4507 support then the option
3870 SSL_OP_NO_TICKET can be set.
3872 Add a TLS extension debugging callback to allow the contents of any client
3873 or server extensions to be examined.
3875 This work was sponsored by Google.
3878 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3879 OpenSSL should now compile cleanly on gcc 4.2
3880 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3882 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3883 support including streaming MAC support: this is required for GOST
3884 ciphersuite support.
3885 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3887 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3888 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3889 to output in BER and PEM format.
3892 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3893 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3894 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3895 ENGINE support for HMAC keys which are unextractable. New -mac and
3896 -macopt options to dgst utility.
3899 *) New option -sigopt to dgst utility. Update dgst to use
3900 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3901 alternative signing parameters such as X9.31 or PSS in the dgst
3905 *) Change ssl_cipher_apply_rule(), the internal function that does
3906 the work each time a ciphersuite string requests enabling
3907 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3908 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3909 the order of disabled ciphersuites such that those ciphersuites
3910 that most recently went from enabled to disabled not only stay
3911 in order with respect to each other, but also have higher priority
3912 than other disabled ciphersuites the next time ciphersuites are
3915 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3916 the same ciphersuites as with "HIGH" alone, but in a specific
3917 order where the PSK ciphersuites come first (since they are the
3918 most recently disabled ciphersuites when "HIGH" is parsed).
3920 Also, change ssl_create_cipher_list() (using this new
3921 functionality) such that between otherwise identical
3922 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
3926 *) Change ssl_create_cipher_list() so that it automatically
3927 arranges the ciphersuites in reasonable order before starting
3928 to process the rule string. Thus, the definition for "DEFAULT"
3929 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3930 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3931 This makes it much easier to arrive at a reasonable default order
3932 in applications for which anonymous ciphers are OK (meaning
3933 that you can't actually use DEFAULT).
3934 [Bodo Moeller; suggested by Victor Duchovni]
3936 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3937 processing) into multiple integers instead of setting
3938 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3939 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3940 (These masks as well as the individual bit definitions are hidden
3941 away into the non-exported interface ssl/ssl_locl.h, so this
3942 change to the definition of the SSL_CIPHER structure shouldn't
3943 affect applications.) This give us more bits for each of these
3944 categories, so there is no longer a need to coagulate AES128 and
3945 AES256 into a single algorithm bit, and to coagulate Camellia128
3946 and Camellia256 into a single algorithm bit, which has led to all
3949 Thus, among other things, the kludge introduced in 0.9.7m and
3950 0.9.8e for masking out AES256 independently of AES128 or masking
3951 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3953 With the change, we also introduce new ciphersuite aliases that
3954 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3958 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3959 Use the leftmost N bytes of the signature input if the input is
3960 larger than the prime q (with N being the size in bytes of q).
3963 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3964 it yet and it is largely untested.
3967 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3970 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3971 some compilers (gcc 4.2 and later) reject their use. Safestack is
3972 reimplemented. Update ASN1 to avoid use of legacy functions.
3975 *) Win32/64 targets are linked with Winsock2.
3978 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3979 to external functions. This can be used to increase CRL handling
3980 efficiency especially when CRLs are very large by (for example) storing
3981 the CRL revoked certificates in a database.
3984 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3985 new CRLs added to a directory can be used. New command line option
3986 -verify_return_error to s_client and s_server. This causes real errors
3987 to be returned by the verify callback instead of carrying on no matter
3988 what. This reflects the way a "real world" verify callback would behave.
3991 *) GOST engine, supporting several GOST algorithms and public key formats.
3992 Kindly donated by Cryptocom.
3995 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3996 partitioned by DP are handled but no indirect CRL or reason partitioning
3997 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3998 selected via a scoring technique which handles IDP and AKID in CRLs.
4001 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4002 will ultimately be used for all verify operations: this will remove the
4003 X509_STORE dependency on certificate verification and allow alternative
4004 lookup methods. X509_STORE based implementations of these two callbacks.
4007 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4008 Modify get_crl() to find a valid (unexpired) CRL if possible.
4011 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4012 this would be called X509_CRL_cmp() but that name is already used by
4013 a function that just compares CRL issuer names. Cache several CRL
4014 extensions in X509_CRL structure and cache CRLDP in X509.
4017 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4018 this maps equivalent X509_NAME structures into a consistent structure.
4019 Name comparison can then be performed rapidly using memcmp().
4022 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4026 *) Allow digests to supply their own micalg string for S/MIME type using
4027 the ctrl EVP_MD_CTRL_MICALG.
4030 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4031 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4032 ctrl. It can then customise the structure before and/or after signing
4036 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4037 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4038 to free up any added signature OIDs.
4041 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4042 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4043 digest and cipher tables. New options added to openssl utility:
4044 list-message-digest-algorithms and list-cipher-algorithms.
4047 *) Change the array representation of binary polynomials: the list
4048 of degrees of non-zero coefficients is now terminated with -1.
4049 Previously it was terminated with 0, which was also part of the
4050 value; thus, the array representation was not applicable to
4051 polynomials where t^0 has coefficient zero. This change makes
4052 the array representation useful in a more general context.
4055 *) Various modifications and fixes to SSL/TLS cipher string
4056 handling. For ECC, the code now distinguishes between fixed ECDH
4057 with RSA certificates on the one hand and with ECDSA certificates
4058 on the other hand, since these are separate ciphersuites. The
4059 unused code for Fortezza ciphersuites has been removed.
4061 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4062 (not "ECDHE"). For consistency with the code for DH
4063 certificates, use of ECDH certificates is now considered ECDH
4064 authentication, not RSA or ECDSA authentication (the latter is
4065 merely the CA's signing algorithm and not actively used in the
4068 The temporary ciphersuite alias "ECCdraft" is no longer
4069 available, and ECC ciphersuites are no longer excluded from "ALL"
4070 and "DEFAULT". The following aliases now exist for RFC 4492
4071 ciphersuites, most of these by analogy with the DH case:
4073 kECDHr - ECDH cert, signed with RSA
4074 kECDHe - ECDH cert, signed with ECDSA
4075 kECDH - ECDH cert (signed with either RSA or ECDSA)
4076 kEECDH - ephemeral ECDH
4077 ECDH - ECDH cert or ephemeral ECDH
4083 AECDH - anonymous ECDH
4084 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4088 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4089 Use correct micalg parameters depending on digest(s) in signed message.
4092 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4093 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4096 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4097 an engine to register a method. Add ENGINE lookups for methods and
4098 functional reference processing.
4101 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4102 EVP_{Sign,Verify}* which allow an application to customise the signature
4106 *) New -resign option to smime utility. This adds one or more signers
4107 to an existing PKCS#7 signedData structure. Also -md option to use an
4108 alternative message digest algorithm for signing.
4111 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4112 create PKCS7 structures containing multiple signers. Update smime
4113 application to support multiple signers.
4116 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4120 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4121 Reorganize PBE internals to lookup from a static table using NIDs,
4122 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4123 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4124 PRF which will be automatically used with PBES2.
4127 *) Replace the algorithm specific calls to generate keys in "req" with the
4131 *) Update PKCS#7 enveloped data routines to use new API. This is now
4132 supported by any public key method supporting the encrypt operation. A
4133 ctrl is added to allow the public key algorithm to examine or modify
4134 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4138 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4139 a default digest type to use. In most cases this will be SHA1 but some
4140 algorithms (such as GOST) need to specify an alternative digest. The
4141 return value indicates how strong the preference is 1 means optional and
4142 2 is mandatory (that is it is the only supported type). Modify
4143 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4144 use the default md. Update openssl utilities to use the default digest
4145 type for signing if it is not explicitly indicated.
4148 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4149 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4150 signing method from the key type. This effectively removes the link
4151 between digests and public key types.
4154 *) Add an OID cross reference table and utility functions. Its purpose is to
4155 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4156 rsaEncryption. This will allow some of the algorithm specific hackery
4157 needed to use the correct OID to be removed.
4160 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4161 structures for PKCS7_sign(). They are now set up by the relevant public
4165 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4168 *) Add support for key derivation (agreement) in the API, DH method and
4172 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4173 public and private key formats. As a side effect these add additional
4174 command line functionality not previously available: DSA signatures can be
4175 generated and verified using pkeyutl and DH key support and generation in
4180 [Oliver Tappe <zooey@hirschkaefer.de>]
4182 *) New make target "install_html_docs" installs HTML renditions of the
4184 [Oliver Tappe <zooey@hirschkaefer.de>]
4186 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4187 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4188 support key and parameter generation and add initial key generation
4189 functionality for RSA.
4192 *) Add functions for main EVP_PKEY_method operations. The undocumented
4193 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4194 EVP_PKEY_{encrypt,decrypt}_old.
4197 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4198 key API, doesn't do much yet.
4201 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4202 public key algorithms. New option to openssl utility:
4203 "list-public-key-algorithms" to print out info.
4206 *) Implement the Supported Elliptic Curves Extension for
4207 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4210 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4211 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4214 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4215 utilities such as rsa, dsa, dsaparam etc except they process any key
4219 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4220 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4221 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4225 *) Initial support for pluggable public key ASN1.
4226 De-spaghettify the public key ASN1 handling. Move public and private
4227 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4228 algorithm specific handling to a single module within the relevant
4229 algorithm directory. Add functions to allow (near) opaque processing
4230 of public and private key structures.
4233 *) Implement the Supported Point Formats Extension for
4234 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4237 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4238 for the psk identity [hint] and the psk callback functions to the
4239 SSL_SESSION, SSL and SSL_CTX structure.
4242 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4246 SSL_CTX_use_psk_identity_hint
4247 SSL_get_psk_identity_hint
4248 SSL_get_psk_identity
4249 SSL_use_psk_identity_hint
4251 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4253 *) Add RFC 3161 compliant time stamp request creation, response generation
4254 and response verification functionality.
4255 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4257 *) Add initial support for TLS extensions, specifically for the server_name
4258 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4259 have new members for a host name. The SSL data structure has an
4260 additional member SSL_CTX *initial_ctx so that new sessions can be
4261 stored in that context to allow for session resumption, even after the
4262 SSL has been switched to a new SSL_CTX in reaction to a client's
4263 server_name extension.
4265 New functions (subject to change):
4267 SSL_get_servername()
4268 SSL_get_servername_type()
4271 New CTRL codes and macros (subject to change):
4273 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4274 - SSL_CTX_set_tlsext_servername_callback()
4275 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4276 - SSL_CTX_set_tlsext_servername_arg()
4277 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4279 openssl s_client has a new '-servername ...' option.
4281 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4282 '-key2 ...', '-servername_fatal' (subject to change). This allows
4283 testing the HostName extension for a specific single host name ('-cert'
4284 and '-key' remain fallbacks for handshakes without HostName
4285 negotiation). If the unrecognized_name alert has to be sent, this by
4286 default is a warning; it becomes fatal with the '-servername_fatal'
4289 [Peter Sylvester, Remy Allais, Christophe Renou]
4291 *) Whirlpool hash implementation is added.
4294 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4295 bn(64,32). Because of instruction set limitations it doesn't have
4296 any negative impact on performance. This was done mostly in order
4297 to make it possible to share assembler modules, such as bn_mul_mont
4298 implementations, between 32- and 64-bit builds without hassle.
4301 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4302 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4306 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4307 dedicated Montgomery multiplication procedure, is introduced.
4308 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4309 "64-bit" performance on certain 32-bit targets.
4312 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4313 in SSL structures. New SSL ctrl to set maximum send fragment size.
4314 Save memory by setting the I/O buffer sizes dynamically instead of
4315 using the maximum available value.
4318 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4319 in addition to the text details.
4322 *) Very, very preliminary EXPERIMENTAL support for printing of general
4323 ASN1 structures. This currently produces rather ugly output and doesn't
4324 handle several customised structures at all.
4327 *) Integrated support for PVK file format and some related formats such
4328 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4329 these in the 'rsa' and 'dsa' utilities.
4332 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4335 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4336 place for the (very old) "NETSCAPE" format certificates which are now
4337 handled using new ASN1 code equivalents.
4340 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4341 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4342 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4345 *) Modify CRL distribution points extension code to print out previously
4346 unsupported fields. Enhance extension setting code to allow setting of
4350 *) Add print and set support for Issuing Distribution Point CRL extension.
4353 *) Change 'Configure' script to enable Camellia by default.
4356 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4358 *) When rejecting SSL/TLS records due to an incorrect version number, never
4359 update s->server with a new major version number. As of
4360 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4361 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4362 the previous behavior could result in a read attempt at NULL when
4363 receiving specific incorrect SSL/TLS records once record payload
4364 protection is active. (CVE-2010-0740)
4365 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4367 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4368 could be crashed if the relevant tables were not present (e.g. chrooted).
4369 [Tomas Hoger <thoger@redhat.com>]
4371 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4373 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4374 [Martin Olsson, Neel Mehta]
4376 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4377 accommodate for stack sorting, always a write lock!).
4380 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4381 excessive delays in the RAND_poll(): over a minute. As a workaround
4382 include a time check in the inner Heap32Next loop too.
4385 *) The code that handled flushing of data in SSL/TLS originally used the
4386 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4387 the problem outlined in PR#1949. The fix suggested there however can
4388 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4389 of Apache). So instead simplify the code to flush unconditionally.
4390 This should be fine since flushing with no data to flush is a no op.
4393 *) Handle TLS versions 2.0 and later properly and correctly use the
4394 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4395 off ancient servers have a habit of sticking around for a while...
4398 *) Modify compression code so it frees up structures without using the
4399 ex_data callbacks. This works around a problem where some applications
4400 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4401 restarting) then use compression (e.g. SSL with compression) later.
4402 This results in significant per-connection memory leaks and
4403 has caused some security issues including CVE-2008-1678 and
4407 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4408 change when encrypting or decrypting.
4411 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4412 connect and renegotiate with servers which do not support RI.
4413 Until RI is more widely deployed this option is enabled by default.
4416 *) Add "missing" ssl ctrls to clear options and mode.
4419 *) If client attempts to renegotiate and doesn't support RI respond with
4420 a no_renegotiation alert as required by RFC5746. Some renegotiating
4421 TLS clients will continue a connection gracefully when they receive
4422 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4423 waiting for a server hello which it will never receive. Now we treat a
4424 received no_renegotiation alert as a fatal error. This is because
4425 applications requesting a renegotiation might well expect it to succeed
4426 and would have no code in place to handle the server denying it so the
4427 only safe thing to do is to terminate the connection.
4430 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4431 peer supports secure renegotiation and 0 otherwise. Print out peer
4432 renegotiation support in s_client/s_server.
4435 *) Replace the highly broken and deprecated SPKAC certification method with
4436 the updated NID creation version. This should correctly handle UTF8.
4439 *) Implement RFC5746. Re-enable renegotiation but require the extension
4440 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4441 turns out to be a bad idea. It has been replaced by
4442 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4443 SSL_CTX_set_options(). This is really not recommended unless you
4444 know what you are doing.
4445 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4447 *) Fixes to stateless session resumption handling. Use initial_ctx when
4448 issuing and attempting to decrypt tickets in case it has changed during
4449 servername handling. Use a non-zero length session ID when attempting
4450 stateless session resumption: this makes it possible to determine if
4451 a resumption has occurred immediately after receiving server hello
4452 (several places in OpenSSL subtly assume this) instead of later in
4456 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4457 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4458 fixes for a few places where the return code is not checked
4460 [Julia Lawall <julia@diku.dk>]
4462 *) Add --strict-warnings option to Configure script to include devteam
4463 warnings in other configurations.
4466 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4467 makes it possible to install openssl libraries in locations which
4468 have names other than "lib", for example "/usr/lib64" which some
4470 [Steve Henson, based on patch from Jeremy Utley]
4472 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4473 X690 8.9.12 and can produce some misleading textual output of OIDs.
4474 [Steve Henson, reported by Dan Kaminsky]
4476 *) Delete MD2 from algorithm tables. This follows the recommendation in
4477 several standards that it is not used in new applications due to
4478 several cryptographic weaknesses. For binary compatibility reasons
4479 the MD2 API is still compiled in by default.
4482 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4486 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4487 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4489 [Guenter <lists@gknw.net>]
4491 *) Fix the server certificate chain building code to use X509_verify_cert(),
4492 it used to have an ad-hoc builder which was unable to cope with anything
4493 other than a simple chain.
4494 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4496 *) Don't check self signed certificate signatures in X509_verify_cert()
4497 by default (a flag can override this): it just wastes time without
4498 adding any security. As a useful side effect self signed root CAs
4499 with non-FIPS digests are now usable in FIPS mode.
4502 *) In dtls1_process_out_of_seq_message() the check if the current message
4503 is already buffered was missing. For every new message was memory
4504 allocated, allowing an attacker to perform an denial of service attack
4505 with sending out of seq handshake messages until there is no memory
4506 left. Additionally every future message was buffered, even if the
4507 sequence number made no sense and would be part of another handshake.
4508 So only messages with sequence numbers less than 10 in advance will be
4509 buffered. (CVE-2009-1378)
4510 [Robin Seggelmann, discovered by Daniel Mentz]
4512 *) Records are buffered if they arrive with a future epoch to be
4513 processed after finishing the corresponding handshake. There is
4514 currently no limitation to this buffer allowing an attacker to perform
4515 a DOS attack with sending records with future epochs until there is no
4516 memory left. This patch adds the pqueue_size() function to determine
4517 the size of a buffer and limits the record buffer to 100 entries.
4519 [Robin Seggelmann, discovered by Daniel Mentz]
4521 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4522 parent structure is freed. (CVE-2009-1379)
4525 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4526 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4529 [Ilya O. <vrghost@gmail.com>]
4531 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4533 *) Disable renegotiation completely - this fixes a severe security
4534 problem (CVE-2009-3555) at the cost of breaking all
4535 renegotiation. Renegotiation can be re-enabled by setting
4536 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4537 run-time. This is really not recommended unless you know what
4541 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4543 *) Don't set val to NULL when freeing up structures, it is freed up by
4544 underlying code. If sizeof(void *) > sizeof(long) this can result in
4545 zeroing past the valid field. (CVE-2009-0789)
4546 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4548 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4549 checked correctly. This would allow some invalid signed attributes to
4550 appear to verify correctly. (CVE-2009-0591)
4551 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4553 *) Reject UniversalString and BMPString types with invalid lengths. This
4554 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4555 a legal length. (CVE-2009-0590)
4558 *) Set S/MIME signing as the default purpose rather than setting it
4559 unconditionally. This allows applications to override it at the store
4563 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4564 to handle some structures.
4567 *) Improve efficiency of mem_gets: don't search whole buffer each time
4569 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4571 *) New -hex option for openssl rand.
4574 *) Print out UTF8String and NumericString when parsing ASN1.
4577 *) Support NumericString type for name components.
4580 *) Allow CC in the environment to override the automatically chosen
4581 compiler. Note that nothing is done to ensure flags work with the
4585 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4587 *) Properly check EVP_VerifyFinal() and similar return values
4589 [Ben Laurie, Bodo Moeller, Google Security Team]
4591 *) Enable TLS extensions by default.
4594 *) Allow the CHIL engine to be loaded, whether the application is
4595 multithreaded or not. (This does not release the developer from the
4596 obligation to set up the dynamic locking callbacks.)
4597 [Sander Temme <sander@temme.net>]
4599 *) Use correct exit code if there is an error in dgst command.
4600 [Steve Henson; problem pointed out by Roland Dirlewanger]
4602 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4603 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4606 *) Add experimental JPAKE support, including demo authentication in
4607 s_client and s_server.
4610 *) Set the comparison function in v3_addr_canonize().
4611 [Rob Austein <sra@hactrn.net>]
4613 *) Add support for XMPP STARTTLS in s_client.
4614 [Philip Paeps <philip@freebsd.org>]
4616 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4617 to ensure that even with this option, only ciphersuites in the
4618 server's preference list will be accepted. (Note that the option
4619 applies only when resuming a session, so the earlier behavior was
4620 just about the algorithm choice for symmetric cryptography.)
4623 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4625 *) Fix NULL pointer dereference if a DTLS server received
4626 ChangeCipherSpec as first record (CVE-2009-1386).
4629 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4630 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4633 *) The fix in 0.9.8c that supposedly got rid of unsafe
4634 double-checked locking was incomplete for RSA blinding,
4635 addressing just one layer of what turns out to have been
4636 doubly unsafe triple-checked locking.
4638 So now fix this for real by retiring the MONT_HELPER macro
4639 in crypto/rsa/rsa_eay.c.
4641 [Bodo Moeller; problem pointed out by Marius Schilder]
4643 *) Various precautionary measures:
4645 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4647 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4648 (NB: This would require knowledge of the secret session ticket key
4649 to exploit, in which case you'd be SOL either way.)
4651 - Change bn_nist.c so that it will properly handle input BIGNUMs
4652 outside the expected range.
4654 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4657 [Neel Mehta, Bodo Moeller]
4659 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4660 the load fails. Useful for distros.
4661 [Ben Laurie and the FreeBSD team]
4663 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4666 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4669 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4671 This work was sponsored by Logica.
4674 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4675 keystores. Support for SSL/TLS client authentication too.
4676 Not compiled unless enable-capieng specified to Configure.
4678 This work was sponsored by Logica.
4681 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4682 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4683 attribute creation routines such as certificate requests and PKCS#12
4687 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4689 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4690 handshake which could lead to a client crash as found using the
4691 Codenomicon TLS test suite (CVE-2008-1672)
4692 [Steve Henson, Mark Cox]
4694 *) Fix double free in TLS server name extensions which could lead to
4695 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4698 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4700 Clear the error queue to ensure that error entries left from
4701 older function calls do not interfere with the correct operation.
4702 [Lutz Jaenicke, Erik de Castro Lopo]
4704 *) Remove root CA certificates of commercial CAs:
4706 The OpenSSL project does not recommend any specific CA and does not
4707 have any policy with respect to including or excluding any CA.
4708 Therefore it does not make any sense to ship an arbitrary selection
4709 of root CA certificates with the OpenSSL software.
4712 *) RSA OAEP patches to fix two separate invalid memory reads.
4713 The first one involves inputs when 'lzero' is greater than
4714 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4715 before the beginning of from). The second one involves inputs where
4716 the 'db' section contains nothing but zeroes (there is a one-byte
4717 invalid read after the end of 'db').
4718 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4720 *) Partial backport from 0.9.9-dev:
4722 Introduce bn_mul_mont (dedicated Montgomery multiplication
4723 procedure) as a candidate for BIGNUM assembler implementation.
4724 While 0.9.9-dev uses assembler for various architectures, only
4725 x86_64 is available by default here in the 0.9.8 branch, and
4726 32-bit x86 is available through a compile-time setting.
4728 To try the 32-bit x86 assembler implementation, use Configure
4729 option "enable-montasm" (which exists only for this backport).
4731 As "enable-montasm" for 32-bit x86 disclaims code stability
4732 anyway, in this constellation we activate additional code
4733 backported from 0.9.9-dev for further performance improvements,
4734 namely BN_from_montgomery_word. (To enable this otherwise,
4735 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4737 [Andy Polyakov (backport partially by Bodo Moeller)]
4739 *) Add TLS session ticket callback. This allows an application to set
4740 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4741 values. This is useful for key rollover for example where several key
4742 sets may exist with different names.
4745 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4746 This was broken until now in 0.9.8 releases, such that the only way
4747 a registered ENGINE could be used (assuming it initialises
4748 successfully on the host) was to explicitly set it as the default
4749 for the relevant algorithms. This is in contradiction with 0.9.7
4750 behaviour and the documentation. With this fix, when an ENGINE is
4751 registered into a given algorithm's table of implementations, the
4752 'uptodate' flag is reset so that auto-discovery will be used next
4753 time a new context for that algorithm attempts to select an
4755 [Ian Lister (tweaked by Geoff Thorpe)]
4757 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4758 implementation in the following ways:
4760 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4763 Lack of BER streaming support means one pass streaming processing is
4764 only supported if data is detached: setting the streaming flag is
4765 ignored for embedded content.
4767 CMS support is disabled by default and must be explicitly enabled
4768 with the enable-cms configuration option.
4771 *) Update the GMP engine glue to do direct copies between BIGNUM and
4772 mpz_t when openssl and GMP use the same limb size. Otherwise the
4773 existing "conversion via a text string export" trick is still used.
4774 [Paul Sheer <paulsheer@gmail.com>]
4776 *) Zlib compression BIO. This is a filter BIO which compressed and
4777 uncompresses any data passed through it.
4780 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4781 RFC3394 compatible AES key wrapping.
4784 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4785 sets string data without copying. X509_ALGOR_set0() and
4786 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4787 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4788 from an X509_ATTRIBUTE structure optionally checking it occurs only
4789 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4793 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4794 to get the expected BN_FLG_CONSTTIME behavior.
4795 [Bodo Moeller (Google)]
4799 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4800 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4801 - added some more tests to do_tests.pl
4802 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4803 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4804 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4805 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4806 - various changes to netware.pl to enable gcc-cross builds on Win32
4808 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4809 - various changes to fix missing prototype warnings
4810 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4811 - added AES, WHIRLPOOL and CPUID assembler code to build files
4812 - added missing AES assembler make rules to mk1mf.pl
4813 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4814 [Guenter Knauf <eflash@gmx.net>]
4816 *) Implement certificate status request TLS extension defined in RFC3546.
4817 A client can set the appropriate parameters and receive the encoded
4818 OCSP response via a callback. A server can query the supplied parameters
4819 and set the encoded OCSP response in the callback. Add simplified examples
4820 to s_client and s_server.
4823 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4825 *) Fix various bugs:
4826 + Binary incompatibility of ssl_ctx_st structure
4827 + DTLS interoperation with non-compliant servers
4828 + Don't call get_session_cb() without proposed session
4829 + Fix ia64 assembler code
4830 [Andy Polyakov, Steve Henson]
4832 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4834 *) DTLS Handshake overhaul. There were longstanding issues with
4835 OpenSSL DTLS implementation, which were making it impossible for
4836 RFC 4347 compliant client to communicate with OpenSSL server.
4837 Unfortunately just fixing these incompatibilities would "cut off"
4838 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4839 server keeps tolerating non RFC compliant syntax. The opposite is
4840 not true, 0.9.8f client can not communicate with earlier server.
4841 This update even addresses CVE-2007-4995.
4844 *) Changes to avoid need for function casts in OpenSSL: some compilers
4845 (gcc 4.2 and later) reject their use.
4846 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4849 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4850 RFC4507bis. The encrypted ticket format is an encrypted encoded
4851 SSL_SESSION structure, that way new session features are automatically
4854 If a client application caches session in an SSL_SESSION structure
4855 support is transparent because tickets are now stored in the encoded
4858 The SSL_CTX structure automatically generates keys for ticket
4859 protection in servers so again support should be possible
4860 with no application modification.
4862 If a client or server wishes to disable RFC4507 support then the option
4863 SSL_OP_NO_TICKET can be set.
4865 Add a TLS extension debugging callback to allow the contents of any client
4866 or server extensions to be examined.
4868 This work was sponsored by Google.
4871 *) Add initial support for TLS extensions, specifically for the server_name
4872 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4873 have new members for a host name. The SSL data structure has an
4874 additional member SSL_CTX *initial_ctx so that new sessions can be
4875 stored in that context to allow for session resumption, even after the
4876 SSL has been switched to a new SSL_CTX in reaction to a client's
4877 server_name extension.
4879 New functions (subject to change):
4881 SSL_get_servername()
4882 SSL_get_servername_type()
4885 New CTRL codes and macros (subject to change):
4887 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4888 - SSL_CTX_set_tlsext_servername_callback()
4889 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4890 - SSL_CTX_set_tlsext_servername_arg()
4891 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4893 openssl s_client has a new '-servername ...' option.
4895 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4896 '-key2 ...', '-servername_fatal' (subject to change). This allows
4897 testing the HostName extension for a specific single host name ('-cert'
4898 and '-key' remain fallbacks for handshakes without HostName
4899 negotiation). If the unrecognized_name alert has to be sent, this by
4900 default is a warning; it becomes fatal with the '-servername_fatal'
4903 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4905 *) Add AES and SSE2 assembly language support to VC++ build.
4908 *) Mitigate attack on final subtraction in Montgomery reduction.
4911 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4912 (which previously caused an internal error).
4915 *) Squeeze another 10% out of IGE mode when in != out.
4918 *) AES IGE mode speedup.
4919 [Dean Gaudet (Google)]
4921 *) Add the Korean symmetric 128-bit cipher SEED (see
4922 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4923 add SEED ciphersuites from RFC 4162:
4925 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4926 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4927 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4928 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4930 To minimize changes between patchlevels in the OpenSSL 0.9.8
4931 series, SEED remains excluded from compilation unless OpenSSL
4932 is configured with 'enable-seed'.
4933 [KISA, Bodo Moeller]
4935 *) Mitigate branch prediction attacks, which can be practical if a
4936 single processor is shared, allowing a spy process to extract
4937 information. For detailed background information, see
4938 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4939 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4940 and Necessary Software Countermeasures"). The core of the change
4941 are new versions BN_div_no_branch() and
4942 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4943 respectively, which are slower, but avoid the security-relevant
4944 conditional branches. These are automatically called by BN_div()
4945 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4946 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4947 remove a conditional branch.
4949 BN_FLG_CONSTTIME is the new name for the previous
4950 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4951 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4952 in the exponent causes BN_mod_exp_mont() to use the alternative
4953 implementation in BN_mod_exp_mont_consttime().) The old name
4954 remains as a deprecated alias.
4956 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4957 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4958 constant-time implementations for more than just exponentiation.
4959 Here too the old name is kept as a deprecated alias.
4961 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4962 the BN_BLINDING structure gets an independent copy of the
4963 modulus. This means that the previous "BIGNUM *m" argument to
4964 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4965 essentially becomes "const BIGNUM *m", although we can't actually
4966 change this in the header file before 0.9.9. It allows
4967 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4968 enable BN_FLG_CONSTTIME.
4970 [Matthew D Wood (Intel Corp)]
4972 *) In the SSL/TLS server implementation, be strict about session ID
4973 context matching (which matters if an application uses a single
4974 external cache for different purposes). Previously,
4975 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4976 set. This did ensure strict client verification, but meant that,
4977 with applications using a single external cache for quite
4978 different requirements, clients could circumvent ciphersuite
4979 restrictions for a given session ID context by starting a session
4980 in a different context.
4983 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4984 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4985 authentication-only ciphersuites.
4988 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4989 not complete and could lead to a possible single byte overflow
4990 (CVE-2007-5135) [Ben Laurie]
4992 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4994 *) Since AES128 and AES256 (and similarly Camellia128 and
4995 Camellia256) share a single mask bit in the logic of
4996 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4997 kludge to work properly if AES128 is available and AES256 isn't
4998 (or if Camellia128 is available and Camellia256 isn't).
5001 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5002 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5003 When a point or a seed is encoded in a BIT STRING, we need to
5004 prevent the removal of trailing zero bits to get the proper DER
5005 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5006 of a NamedBitList, for which trailing 0 bits need to be removed.)
5009 *) Have SSL/TLS server implementation tolerate "mismatched" record
5010 protocol version while receiving ClientHello even if the
5011 ClientHello is fragmented. (The server can't insist on the
5012 particular protocol version it has chosen before the ServerHello
5013 message has informed the client about his choice.)
5016 *) Add RFC 3779 support.
5017 [Rob Austein for ARIN, Ben Laurie]
5019 *) Load error codes if they are not already present instead of using a
5020 static variable. This allows them to be cleanly unloaded and reloaded.
5021 Improve header file function name parsing.
5024 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5025 or CAPABILITY handshake as required by RFCs.
5028 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5030 *) Introduce limits to prevent malicious keys being able to
5031 cause a denial of service. (CVE-2006-2940)
5032 [Steve Henson, Bodo Moeller]
5034 *) Fix ASN.1 parsing of certain invalid structures that can result
5035 in a denial of service. (CVE-2006-2937) [Steve Henson]
5037 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5038 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5040 *) Fix SSL client code which could crash if connecting to a
5041 malicious SSLv2 server. (CVE-2006-4343)
5042 [Tavis Ormandy and Will Drewry, Google Security Team]
5044 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5045 match only those. Before that, "AES256-SHA" would be interpreted
5046 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5047 the same strength classification in 0.9.7h) as we currently only
5048 have a single AES bit in the ciphersuite description bitmap.
5049 That change, however, also applied to ciphersuite strings such as
5050 "RC4-MD5" that intentionally matched multiple ciphersuites --
5051 namely, SSL 2.0 ciphersuites in addition to the more common ones
5052 from SSL 3.0/TLS 1.0.
5054 So we change the selection algorithm again: Naming an explicit
5055 ciphersuite selects this one ciphersuite, and any other similar
5056 ciphersuite (same bitmap) from *other* protocol versions.
5057 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5058 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5060 Since SSL 2.0 does not have any ciphersuites for which the
5061 128/256 bit distinction would be relevant, this works for now.
5062 The proper fix will be to use different bits for AES128 and
5063 AES256, which would have avoided the problems from the beginning;
5064 however, bits are scarce, so we can only do this in a new release
5065 (not just a patchlevel) when we can change the SSL_CIPHER
5066 definition to split the single 'unsigned long mask' bitmap into
5067 multiple values to extend the available space.
5071 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5073 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5074 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5076 *) Add AES IGE and biIGE modes.
5079 *) Change the Unix randomness entropy gathering to use poll() when
5080 possible instead of select(), since the latter has some
5081 undesirable limitations.
5082 [Darryl Miles via Richard Levitte and Bodo Moeller]
5084 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5085 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5086 cannot be implicitly activated as part of, e.g., the "AES" alias.
5087 However, please upgrade to OpenSSL 0.9.9[-dev] for
5088 non-experimental use of the ECC ciphersuites to get TLS extension
5089 support, which is required for curve and point format negotiation
5090 to avoid potential handshake problems.
5093 *) Disable rogue ciphersuites:
5095 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5096 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5097 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5099 The latter two were purportedly from
5100 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5103 Also deactivate the remaining ciphersuites from
5104 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5105 unofficial, and the ID has long expired.
5108 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5109 dual-core machines) and other potential thread-safety issues.
5112 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5113 versions), which is now available for royalty-free use
5114 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5115 Also, add Camellia TLS ciphersuites from RFC 4132.
5117 To minimize changes between patchlevels in the OpenSSL 0.9.8
5118 series, Camellia remains excluded from compilation unless OpenSSL
5119 is configured with 'enable-camellia'.
5122 *) Disable the padding bug check when compression is in use. The padding
5123 bug check assumes the first packet is of even length, this is not
5124 necessarily true if compression is enabled and can result in false
5125 positives causing handshake failure. The actual bug test is ancient
5126 code so it is hoped that implementations will either have fixed it by
5127 now or any which still have the bug do not support compression.
5130 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5132 *) When applying a cipher rule check to see if string match is an explicit
5133 cipher suite and only match that one cipher suite if it is.
5136 *) Link in manifests for VC++ if needed.
5137 [Austin Ziegler <halostatue@gmail.com>]
5139 *) Update support for ECC-based TLS ciphersuites according to
5140 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5141 TLS extensions, which are supported starting with the 0.9.9
5142 branch, not in the OpenSSL 0.9.8 branch).
5145 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5146 opaque EVP_CIPHER_CTX handling.
5149 *) Fixes and enhancements to zlib compression code. We now only use
5150 "zlib1.dll" and use the default __cdecl calling convention on Win32
5151 to conform with the standards mentioned here:
5152 http://www.zlib.net/DLL_FAQ.txt
5153 Static zlib linking now works on Windows and the new --with-zlib-include
5154 --with-zlib-lib options to Configure can be used to supply the location
5155 of the headers and library. Gracefully handle case where zlib library
5159 *) Several fixes and enhancements to the OID generation code. The old code
5160 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5161 handle numbers larger than ULONG_MAX, truncated printing and had a
5162 non standard OBJ_obj2txt() behaviour.
5165 *) Add support for building of engines under engine/ as shared libraries
5166 under VC++ build system.
5169 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5170 Hopefully, we will not see any false combination of paths any more.
5173 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5175 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5176 (part of SSL_OP_ALL). This option used to disable the
5177 countermeasure against man-in-the-middle protocol-version
5178 rollback in the SSL 2.0 server implementation, which is a bad
5179 idea. (CVE-2005-2969)
5181 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5182 for Information Security, National Institute of Advanced Industrial
5183 Science and Technology [AIST], Japan)]
5185 *) Add two function to clear and return the verify parameter flags.
5188 *) Keep cipherlists sorted in the source instead of sorting them at
5189 runtime, thus removing the need for a lock.
5192 *) Avoid some small subgroup attacks in Diffie-Hellman.
5193 [Nick Mathewson and Ben Laurie]
5195 *) Add functions for well-known primes.
5198 *) Extended Windows CE support.
5199 [Satoshi Nakamura and Andy Polyakov]
5201 *) Initialize SSL_METHOD structures at compile time instead of during
5202 runtime, thus removing the need for a lock.
5205 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5206 attempting to decrypt each encrypted key in turn. Add support to
5210 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5212 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5215 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5218 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5219 key into the same file any more.
5222 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5225 *) Add -utf8 command line and config file option to 'ca'.
5226 [Stefan <stf@udoma.org]
5228 *) Removed the macro des_crypt(), as it seems to conflict with some
5229 libraries. Use DES_crypt().
5232 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5233 involves renaming the source and generated shared-libs for
5234 both. The engines will accept the corrected or legacy ids
5235 ('ncipher' and '4758_cca' respectively) when binding. NB,
5236 this only applies when building 'shared'.
5237 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5239 *) Add attribute functions to EVP_PKEY structure. Modify
5240 PKCS12_create() to recognize a CSP name attribute and
5241 use it. Make -CSP option work again in pkcs12 utility.
5244 *) Add new functionality to the bn blinding code:
5245 - automatic re-creation of the BN_BLINDING parameters after
5246 a fixed number of uses (currently 32)
5247 - add new function for parameter creation
5248 - introduce flags to control the update behaviour of the
5249 BN_BLINDING parameters
5250 - hide BN_BLINDING structure
5251 Add a second BN_BLINDING slot to the RSA structure to improve
5252 performance when a single RSA object is shared among several
5256 *) Add support for DTLS.
5257 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5259 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5260 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5263 *) Remove buggy and incomplete DH cert support from
5264 ssl/ssl_rsa.c and ssl/s3_both.c
5267 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5268 the apps/openssl applications.
5271 *) Compile clean with "-Wall -Wmissing-prototypes
5272 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5273 DEBUG_SAFESTACK must also be set.
5276 *) Change ./Configure so that certain algorithms can be disabled by default.
5277 The new counterpiece to "no-xxx" is "enable-xxx".
5279 The patented RC5 and MDC2 algorithms will now be disabled unless
5280 "enable-rc5" and "enable-mdc2", respectively, are specified.
5282 (IDEA remains enabled despite being patented. This is because IDEA
5283 is frequently required for interoperability, and there is no license
5284 fee for non-commercial use. As before, "no-idea" can be used to
5285 avoid this algorithm.)
5289 *) Add processing of proxy certificates (see RFC 3820). This work was
5290 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5291 EGEE (Enabling Grids for E-science in Europe).
5294 *) RC4 performance overhaul on modern architectures/implementations, such
5295 as Intel P4, IA-64 and AMD64.
5298 *) New utility extract-section.pl. This can be used specify an alternative
5299 section number in a pod file instead of having to treat each file as
5300 a separate case in Makefile. This can be done by adding two lines to the
5303 =for comment openssl_section:XXX
5305 The blank line is mandatory.
5309 *) New arguments -certform, -keyform and -pass for s_client and s_server
5310 to allow alternative format key and certificate files and passphrase
5314 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5315 update associated structures and add various utility functions.
5317 Add new policy related verify parameters, include policy checking in
5318 standard verify code. Enhance 'smime' application with extra parameters
5319 to support policy checking and print out.
5322 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5323 Nehemiah processors. These extensions support AES encryption in hardware
5324 as well as RNG (though RNG support is currently disabled).
5325 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5327 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5330 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5331 [Andy Polyakov and a number of other people]
5333 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5334 implementation contributed by IBM.
5335 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5337 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5338 exponent rather than 'unsigned long'. There is a corresponding change to
5339 the new 'rsa_keygen' element of the RSA_METHOD structure.
5340 [Jelte Jansen, Geoff Thorpe]
5342 *) Functionality for creating the initial serial number file is now
5343 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5345 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5346 number file to 1, which is bound to cause problems. To avoid
5347 the problems while respecting compatibility between different 0.9.7
5348 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5349 CA.pl for serial number initialization. With the new release 0.9.8,
5350 we can fix the problem directly in the 'ca' utility.)
5353 *) Reduced header interdependencies by declaring more opaque objects in
5354 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5355 give fewer recursive includes, which could break lazy source code - so
5356 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5357 developers should define this symbol when building and using openssl to
5358 ensure they track the recommended behaviour, interfaces, [etc], but
5359 backwards-compatible behaviour prevails when this isn't defined.
5362 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5365 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5366 This will generate a random key of the appropriate length based on the
5367 cipher context. The EVP_CIPHER can provide its own random key generation
5368 routine to support keys of a specific form. This is used in the des and
5369 3des routines to generate a key of the correct parity. Update S/MIME
5370 code to use new functions and hence generate correct parity DES keys.
5371 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5372 valid (weak or incorrect parity).
5375 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5376 as looking them up. This is useful when the verified structure may contain
5377 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5378 present unless the new PKCS7_NO_CRL flag is asserted.
5381 *) Extend ASN1 oid configuration module. It now additionally accepts the
5384 shortName = some long name, 1.2.3.4
5387 *) Reimplemented the BN_CTX implementation. There is now no more static
5388 limitation on the number of variables it can handle nor the depth of the
5389 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5390 information can now expand as required, and rather than having a single
5391 static array of bignums, BN_CTX now uses a linked-list of such arrays
5392 allowing it to expand on demand whilst maintaining the usefulness of
5393 BN_CTX's "bundling".
5396 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5397 to allow all RSA operations to function using a single BN_CTX.
5400 *) Preliminary support for certificate policy evaluation and checking. This
5401 is initially intended to pass the tests outlined in "Conformance Testing
5402 of Relying Party Client Certificate Path Processing Logic" v1.07.
5405 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5406 remained unused and not that useful. A variety of other little bignum
5407 tweaks and fixes have also been made continuing on from the audit (see
5411 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5412 associated ASN1, EVP and SSL functions and old ASN1 macros.
5415 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5416 and this should never fail. So the return value from the use of
5417 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5418 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5421 *) BN_CTX_get() should return zero-valued bignums, providing the same
5422 initialised value as BN_new().
5423 [Geoff Thorpe, suggested by Ulf Möller]
5425 *) Support for inhibitAnyPolicy certificate extension.
5428 *) An audit of the BIGNUM code is underway, for which debugging code is
5429 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5430 is considered valid when processing BIGNUMs, and causes execution to
5431 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5432 further steps are taken to deliberately pollute unused data in BIGNUM
5433 structures to try and expose faulty code further on. For now, openssl will
5434 (in its default mode of operation) continue to tolerate the inconsistent
5435 forms that it has tolerated in the past, but authors and packagers should
5436 consider trying openssl and their own applications when compiled with
5437 these debugging symbols defined. It will help highlight potential bugs in
5438 their own code, and will improve the test coverage for OpenSSL itself. At
5439 some point, these tighter rules will become openssl's default to improve
5440 maintainability, though the assert()s and other overheads will remain only
5441 in debugging configurations. See bn.h for more details.
5442 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5444 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5445 that can only be obtained through BN_CTX_new() (which implicitly
5446 initialises it). The presence of this function only made it possible
5447 to overwrite an existing structure (and cause memory leaks).
5450 *) Because of the callback-based approach for implementing LHASH as a
5451 template type, lh_insert() adds opaque objects to hash-tables and
5452 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5453 to clean up those corresponding objects before destroying the hash table
5454 (and losing the object pointers). So some over-zealous constifications in
5455 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5456 objects as "const" and the lh_doall[_arg] callback wrappers are not
5457 prototyped to have "const" restrictions on the object pointers they are
5458 given (and so aren't required to cast them away any more).
5461 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5462 (speed) prefers to use its own implementation. The two implementations
5463 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5464 its object type properly exposed (MS_TM) instead of casting to/from "char
5465 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5466 aren't necessarily the greatest nomenclatures - but this is what was used
5467 internally to the implementation so I've used that for now.
5470 *) Ensure that deprecated functions do not get compiled when
5471 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5472 the self-tests were still using deprecated key-generation functions so
5473 these have been updated also.
5476 *) Reorganise PKCS#7 code to separate the digest location functionality
5477 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5478 New function PKCS7_set_digest() to set the digest type for PKCS#7
5479 digestedData type. Add additional code to correctly generate the
5480 digestedData type and add support for this type in PKCS7 initialization
5484 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5485 structure of type "other".
5488 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5489 sure the loop does correctly stop and breaking ("division by zero")
5490 modulus operations are not performed. The (pre-generated) prime
5491 table crypto/bn/bn_prime.h was already correct, but it could not be
5492 re-generated on some platforms because of the "division by zero"
5493 situation in the script.
5494 [Ralf S. Engelschall]
5496 *) Update support for ECC-based TLS ciphersuites according to
5497 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5498 SHA-1 now is only used for "small" curves (where the
5499 representation of a field element takes up to 24 bytes); for
5500 larger curves, the field element resulting from ECDH is directly
5501 used as premaster secret.
5502 [Douglas Stebila (Sun Microsystems Laboratories)]
5504 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5505 curve secp160r1 to the tests.
5506 [Douglas Stebila (Sun Microsystems Laboratories)]
5508 *) Add the possibility to load symbols globally with DSO.
5509 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5511 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5512 control of the error stack.
5515 *) Add support for STORE in ENGINE.
5518 *) Add the STORE type. The intention is to provide a common interface
5519 to certificate and key stores, be they simple file-based stores, or
5520 HSM-type store, or LDAP stores, or...
5521 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5524 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5525 pass a list of arguments to any function as well as provide a way
5526 for a function to pass data back to the caller.
5529 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5530 works like BUF_strdup() but can be used to duplicate a portion of
5531 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5535 *) Add the function sk_find_ex() which works like sk_find(), but will
5536 return an index to an element even if an exact match couldn't be
5537 found. The index is guaranteed to point at the element where the
5538 searched-for key would be inserted to preserve sorting order.
5541 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5542 takes an extra flags argument for optional functionality. Currently,
5543 the following flags are defined:
5545 OBJ_BSEARCH_VALUE_ON_NOMATCH
5546 This one gets OBJ_bsearch_ex() to return a pointer to the first
5547 element where the comparing function returns a negative or zero
5550 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5551 This one gets OBJ_bsearch_ex() to return a pointer to the first
5552 element where the comparing function returns zero. This is useful
5553 if there are more than one element where the comparing function
5557 *) Make it possible to create self-signed certificates with 'openssl ca'
5558 in such a way that the self-signed certificate becomes part of the
5559 CA database and uses the same mechanisms for serial number generation
5560 as all other certificate signing. The new flag '-selfsign' enables
5561 this functionality. Adapt CA.sh and CA.pl.in.
5564 *) Add functionality to check the public key of a certificate request
5565 against a given private. This is useful to check that a certificate
5566 request can be signed by that key (self-signing).
5569 *) Make it possible to have multiple active certificates with the same
5570 subject in the CA index file. This is done only if the keyword
5571 'unique_subject' is set to 'no' in the main CA section (default
5572 if 'CA_default') of the configuration file. The value is saved
5573 with the database itself in a separate index attribute file,
5574 named like the index file with '.attr' appended to the name.
5577 *) Generate multi-valued AVAs using '+' notation in config files for
5581 *) Support for nameConstraints certificate extension.
5584 *) Support for policyConstraints certificate extension.
5587 *) Support for policyMappings certificate extension.
5590 *) Make sure the default DSA_METHOD implementation only uses its
5591 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5592 and change its own handlers to be NULL so as to remove unnecessary
5593 indirection. This lets alternative implementations fallback to the
5594 default implementation more easily.
5597 *) Support for directoryName in GeneralName related extensions
5601 *) Make it possible to link applications using Makefile.shared.
5602 Make that possible even when linking against static libraries!
5605 *) Support for single pass processing for S/MIME signing. This now
5606 means that S/MIME signing can be done from a pipe, in addition
5607 cleartext signing (multipart/signed type) is effectively streaming
5608 and the signed data does not need to be all held in memory.
5610 This is done with a new flag PKCS7_STREAM. When this flag is set
5611 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5612 is done after the data is output (and digests calculated) in
5613 SMIME_write_PKCS7().
5616 *) Add full support for -rpath/-R, both in shared libraries and
5617 applications, at least on the platforms where it's known how
5621 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5622 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5623 will now compute a table of multiples of the generator that
5624 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5625 faster (notably in the case of a single point multiplication,
5626 scalar * generator).
5627 [Nils Larsch, Bodo Moeller]
5629 *) IPv6 support for certificate extensions. The various extensions
5630 which use the IP:a.b.c.d can now take IPv6 addresses using the
5631 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5635 *) Added an ENGINE that implements RSA by performing private key
5636 exponentiations with the GMP library. The conversions to and from
5637 GMP's mpz_t format aren't optimised nor are any montgomery forms
5638 cached, and on x86 it appears OpenSSL's own performance has caught up.
5639 However there are likely to be other architectures where GMP could
5640 provide a boost. This ENGINE is not built in by default, but it can be
5641 specified at Configure time and should be accompanied by the necessary
5642 linker additions, eg;
5643 ./config -DOPENSSL_USE_GMP -lgmp
5646 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5647 testing availability of engines with "-t" - the old behaviour is
5648 produced by increasing the feature's verbosity with "-tt".
5651 *) ECDSA routines: under certain error conditions uninitialized BN objects
5652 could be freed. Solution: make sure initialization is performed early
5653 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5657 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5658 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5659 software implementations. For DSA and DH, parameter generation can
5660 also be overridden by providing the appropriate method callbacks.
5663 *) Change the "progress" mechanism used in key-generation and
5664 primality testing to functions that take a new BN_GENCB pointer in
5665 place of callback/argument pairs. The new API functions have "_ex"
5666 postfixes and the older functions are reimplemented as wrappers for
5667 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5668 declarations of the old functions to help (graceful) attempts to
5669 migrate to the new functions. Also, the new key-generation API
5670 functions operate on a caller-supplied key-structure and return
5671 success/failure rather than returning a key or NULL - this is to
5672 help make "keygen" another member function of RSA_METHOD etc.
5674 Example for using the new callback interface:
5676 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5680 BN_GENCB_set(&my_cb, my_callback, my_arg);
5682 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5683 /* For the meaning of a, b in calls to my_callback(), see the
5684 * documentation of the function that calls the callback.
5685 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5686 * my_callback should return 1 if it wants BN_is_prime_ex()
5687 * to continue, or 0 to stop.
5692 *) Change the ZLIB compression method to be stateful, and make it
5693 available to TLS with the number defined in
5694 draft-ietf-tls-compression-04.txt.
5697 *) Add the ASN.1 structures and functions for CertificatePair, which
5698 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5700 CertificatePair ::= SEQUENCE {
5701 forward [0] Certificate OPTIONAL,
5702 reverse [1] Certificate OPTIONAL,
5703 -- at least one of the pair shall be present -- }
5705 Also implement the PEM functions to read and write certificate
5706 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5708 This needed to be defined, mostly for the sake of the LDAP
5709 attribute crossCertificatePair, but may prove useful elsewhere as
5713 *) Make it possible to inhibit symlinking of shared libraries in
5714 Makefile.shared, for Cygwin's sake.
5717 *) Extend the BIGNUM API by creating a function
5718 void BN_set_negative(BIGNUM *a, int neg);
5719 and a macro that behave like
5720 int BN_is_negative(const BIGNUM *a);
5722 to avoid the need to access 'a->neg' directly in applications.
5725 *) Implement fast modular reduction for pseudo-Mersenne primes
5726 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5727 EC_GROUP_new_curve_GFp() will now automatically use this
5729 [Nils Larsch <nla@trustcenter.de>]
5731 *) Add new lock type (CRYPTO_LOCK_BN).
5734 *) Change the ENGINE framework to automatically load engines
5735 dynamically from specific directories unless they could be
5736 found to already be built in or loaded. Move all the
5737 current engines except for the cryptodev one to a new
5739 The engines in engines/ are built as shared libraries if
5740 the "shared" options was given to ./Configure or ./config.
5741 Otherwise, they are inserted in libcrypto.a.
5742 /usr/local/ssl/engines is the default directory for dynamic
5743 engines, but that can be overridden at configure time through
5744 the usual use of --prefix and/or --openssldir, and at run
5745 time with the environment variable OPENSSL_ENGINES.
5746 [Geoff Thorpe and Richard Levitte]
5748 *) Add Makefile.shared, a helper makefile to build shared
5749 libraries. Adapt Makefile.org.
5752 *) Add version info to Win32 DLLs.
5753 [Peter 'Luna' Runestig" <peter@runestig.com>]
5755 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5756 can be added using this API to created arbitrary PKCS#12
5757 files while avoiding the low level API.
5759 New options to PKCS12_create(), key or cert can be NULL and
5760 will then be omitted from the output file. The encryption
5761 algorithm NIDs can be set to -1 for no encryption, the mac
5762 iteration count can be set to 0 to omit the mac.
5764 Enhance pkcs12 utility by making the -nokeys and -nocerts
5765 options work when creating a PKCS#12 file. New option -nomac
5766 to omit the mac, NONE can be set for an encryption algorithm.
5767 New code is modified to use the enhanced PKCS12_create()
5768 instead of the low level API.
5771 *) Extend ASN1 encoder to support indefinite length constructed
5772 encoding. This can output sequences tags and octet strings in
5773 this form. Modify pk7_asn1.c to support indefinite length
5774 encoding. This is experimental and needs additional code to
5775 be useful, such as an ASN1 bio and some enhanced streaming
5778 Extend template encode functionality so that tagging is passed
5779 down to the template encoder.
5782 *) Let 'openssl req' fail if an argument to '-newkey' is not
5783 recognized instead of using RSA as a default.
5786 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5787 As these are not official, they are not included in "ALL";
5788 the "ECCdraft" ciphersuite group alias can be used to select them.
5789 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5791 *) Add ECDH engine support.
5792 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5794 *) Add ECDH in new directory crypto/ecdh/.
5795 [Douglas Stebila (Sun Microsystems Laboratories)]
5797 *) Let BN_rand_range() abort with an error after 100 iterations
5798 without success (which indicates a broken PRNG).
5801 *) Change BN_mod_sqrt() so that it verifies that the input value
5802 is really the square of the return value. (Previously,
5803 BN_mod_sqrt would show GIGO behaviour.)
5806 *) Add named elliptic curves over binary fields from X9.62, SECG,
5807 and WAP/WTLS; add OIDs that were still missing.
5809 [Sheueling Chang Shantz and Douglas Stebila
5810 (Sun Microsystems Laboratories)]
5812 *) Extend the EC library for elliptic curves over binary fields
5813 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5816 EC_GF2m_simple_method
5820 EC_GROUP_new_curve_GF2m
5821 EC_GROUP_set_curve_GF2m
5822 EC_GROUP_get_curve_GF2m
5823 EC_POINT_set_affine_coordinates_GF2m
5824 EC_POINT_get_affine_coordinates_GF2m
5825 EC_POINT_set_compressed_coordinates_GF2m
5827 Point compression for binary fields is disabled by default for
5828 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5831 As binary polynomials are represented as BIGNUMs, various members
5832 of the EC_GROUP and EC_POINT data structures can be shared
5833 between the implementations for prime fields and binary fields;
5834 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5835 are essentially identical to their ..._GFp counterparts.
5836 (For simplicity, the '..._GFp' prefix has been dropped from
5837 various internal method names.)
5839 An internal 'field_div' method (similar to 'field_mul' and
5840 'field_sqr') has been added; this is used only for binary fields.
5842 [Sheueling Chang Shantz and Douglas Stebila
5843 (Sun Microsystems Laboratories)]
5845 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5846 through methods ('mul', 'precompute_mult').
5848 The generic implementations (now internally called 'ec_wNAF_mul'
5849 and 'ec_wNAF_precomputed_mult') remain the default if these
5850 methods are undefined.
5852 [Sheueling Chang Shantz and Douglas Stebila
5853 (Sun Microsystems Laboratories)]
5855 *) New function EC_GROUP_get_degree, which is defined through
5856 EC_METHOD. For curves over prime fields, this returns the bit
5857 length of the modulus.
5859 [Sheueling Chang Shantz and Douglas Stebila
5860 (Sun Microsystems Laboratories)]
5862 *) New functions EC_GROUP_dup, EC_POINT_dup.
5863 (These simply call ..._new and ..._copy).
5865 [Sheueling Chang Shantz and Douglas Stebila
5866 (Sun Microsystems Laboratories)]
5868 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5869 Polynomials are represented as BIGNUMs (where the sign bit is not
5870 used) in the following functions [macros]:
5873 BN_GF2m_sub [= BN_GF2m_add]
5874 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5875 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5876 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5878 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5879 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5880 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5881 BN_GF2m_cmp [= BN_ucmp]
5883 (Note that only the 'mod' functions are actually for fields GF(2^m).
5884 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5886 For some functions, an the irreducible polynomial defining a
5887 field can be given as an 'unsigned int[]' with strictly
5888 decreasing elements giving the indices of those bits that are set;
5889 i.e., p[] represents the polynomial
5890 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5892 p[0] > p[1] > ... > p[k] = 0.
5893 This applies to the following functions:
5898 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5899 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5901 BN_GF2m_mod_sqrt_arr
5902 BN_GF2m_mod_solve_quad_arr
5906 Conversion can be performed by the following functions:
5911 bntest.c has additional tests for binary polynomial arithmetic.
5913 Two implementations for BN_GF2m_mod_div() are available.
5914 The default algorithm simply uses BN_GF2m_mod_inv() and
5915 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5916 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5917 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5919 [Sheueling Chang Shantz and Douglas Stebila
5920 (Sun Microsystems Laboratories)]
5922 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5923 functionality is disabled at compile-time.
5924 [Douglas Stebila <douglas.stebila@sun.com>]
5926 *) Change default behaviour of 'openssl asn1parse' so that more
5927 information is visible when viewing, e.g., a certificate:
5929 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5930 mode the content of non-printable OCTET STRINGs is output in a
5931 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5932 avoid the appearance of a printable string.
5933 [Nils Larsch <nla@trustcenter.de>]
5935 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5937 EC_GROUP_set_asn1_flag()
5938 EC_GROUP_get_asn1_flag()
5939 EC_GROUP_set_point_conversion_form()
5940 EC_GROUP_get_point_conversion_form()
5941 These control ASN1 encoding details:
5942 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5943 has been set to OPENSSL_EC_NAMED_CURVE.
5944 - Points are encoded in uncompressed form by default; options for
5945 asn1_for are as for point2oct, namely
5946 POINT_CONVERSION_COMPRESSED
5947 POINT_CONVERSION_UNCOMPRESSED
5948 POINT_CONVERSION_HYBRID
5950 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5953 EC_GROUP_get0_seed()
5954 EC_GROUP_get_seed_len()
5955 This is used only for ASN1 purposes (so far).
5956 [Nils Larsch <nla@trustcenter.de>]
5958 *) Add 'field_type' member to EC_METHOD, which holds the NID
5959 of the appropriate field type OID. The new function
5960 EC_METHOD_get_field_type() returns this value.
5961 [Nils Larsch <nla@trustcenter.de>]
5966 EC_POINT_point2hex()
5967 EC_POINT_hex2point()
5968 providing useful interfaces to EC_POINT_point2oct() and
5969 EC_POINT_oct2point().
5970 [Nils Larsch <nla@trustcenter.de>]
5972 *) Change internals of the EC library so that the functions
5973 EC_GROUP_set_generator()
5974 EC_GROUP_get_generator()
5975 EC_GROUP_get_order()
5976 EC_GROUP_get_cofactor()
5977 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5978 to methods, which would lead to unnecessary code duplication when
5979 adding different types of curves.
5980 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5982 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5983 arithmetic, and such that modified wNAFs are generated
5984 (which avoid length expansion in many cases).
5987 *) Add a function EC_GROUP_check_discriminant() (defined via
5988 EC_METHOD) that verifies that the curve discriminant is non-zero.
5990 Add a function EC_GROUP_check() that makes some sanity tests
5991 on a EC_GROUP, its generator and order. This includes
5992 EC_GROUP_check_discriminant().
5993 [Nils Larsch <nla@trustcenter.de>]
5995 *) Add ECDSA in new directory crypto/ecdsa/.
5997 Add applications 'openssl ecparam' and 'openssl ecdsa'
5998 (these are based on 'openssl dsaparam' and 'openssl dsa').
6000 ECDSA support is also included in various other files across the
6001 library. Most notably,
6002 - 'openssl req' now has a '-newkey ecdsa:file' option;
6003 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6004 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6005 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6006 them suitable for ECDSA where domain parameters must be
6007 extracted before the specific public key;
6008 - ECDSA engine support has been added.
6009 [Nils Larsch <nla@trustcenter.de>]
6011 *) Include some named elliptic curves, and add OIDs from X9.62,
6012 SECG, and WAP/WTLS. Each curve can be obtained from the new
6014 EC_GROUP_new_by_curve_name(),
6015 and the list of available named curves can be obtained with
6016 EC_get_builtin_curves().
6017 Also add a 'curve_name' member to EC_GROUP objects, which can be
6019 EC_GROUP_set_curve_name()
6020 EC_GROUP_get_curve_name()
6021 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6023 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6024 was actually never needed) and in BN_mul(). The removal in BN_mul()
6025 required a small change in bn_mul_part_recursive() and the addition
6026 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6027 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6028 bn_sub_words() and bn_add_words() except they take arrays with
6032 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6034 *) Cleanse PEM buffers before freeing them since they may contain
6036 [Benjamin Bennett <ben@psc.edu>]
6038 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6039 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6040 authentication-only ciphersuites.
6043 *) Since AES128 and AES256 share a single mask bit in the logic of
6044 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6045 kludge to work properly if AES128 is available and AES256 isn't.
6048 *) Expand security boundary to match 1.1.1 module.
6051 *) Remove redundant features: hash file source, editing of test vectors
6052 modify fipsld to use external fips_premain.c signature.
6055 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6056 run algorithm test programs.
6059 *) Make algorithm test programs more tolerant of whitespace.
6062 *) Have SSL/TLS server implementation tolerate "mismatched" record
6063 protocol version while receiving ClientHello even if the
6064 ClientHello is fragmented. (The server can't insist on the
6065 particular protocol version it has chosen before the ServerHello
6066 message has informed the client about his choice.)
6069 *) Load error codes if they are not already present instead of using a
6070 static variable. This allows them to be cleanly unloaded and reloaded.
6073 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6075 *) Introduce limits to prevent malicious keys being able to
6076 cause a denial of service. (CVE-2006-2940)
6077 [Steve Henson, Bodo Moeller]
6079 *) Fix ASN.1 parsing of certain invalid structures that can result
6080 in a denial of service. (CVE-2006-2937) [Steve Henson]
6082 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6083 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6085 *) Fix SSL client code which could crash if connecting to a
6086 malicious SSLv2 server. (CVE-2006-4343)
6087 [Tavis Ormandy and Will Drewry, Google Security Team]
6089 *) Change ciphersuite string processing so that an explicit
6090 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6091 will no longer include "AES128-SHA"), and any other similar
6092 ciphersuite (same bitmap) from *other* protocol versions (so that
6093 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6094 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6095 changes from 0.9.8b and 0.9.8d.
6098 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6100 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6101 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6103 *) Change the Unix randomness entropy gathering to use poll() when
6104 possible instead of select(), since the latter has some
6105 undesirable limitations.
6106 [Darryl Miles via Richard Levitte and Bodo Moeller]
6108 *) Disable rogue ciphersuites:
6110 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6111 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6112 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6114 The latter two were purportedly from
6115 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6118 Also deactivate the remaining ciphersuites from
6119 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6120 unofficial, and the ID has long expired.
6123 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6124 dual-core machines) and other potential thread-safety issues.
6127 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6129 *) Adapt fipsld and the build system to link against the validated FIPS
6130 module in FIPS mode.
6133 *) Fixes for VC++ 2005 build under Windows.
6136 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6137 from a Windows bash shell such as MSYS. It is autodetected from the
6138 "config" script when run from a VC++ environment. Modify standard VC++
6139 build to use fipscanister.o from the GNU make build.
6142 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6144 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6145 The value now differs depending on if you build for FIPS or not.
6146 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6147 safely run with a non-FIPSed libcrypto, as it may crash because of
6148 the difference induced by this change.
6151 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6153 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6154 (part of SSL_OP_ALL). This option used to disable the
6155 countermeasure against man-in-the-middle protocol-version
6156 rollback in the SSL 2.0 server implementation, which is a bad
6157 idea. (CVE-2005-2969)
6159 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6160 for Information Security, National Institute of Advanced Industrial
6161 Science and Technology [AIST], Japan)]
6163 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6164 mainly for FIPS compliance and not fully integrated at this stage.
6167 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6168 the exponentiation using a fixed-length exponent. (Otherwise,
6169 the information leaked through timing could expose the secret key
6170 after many signatures; cf. Bleichenbacher's attack on DSA with
6174 *) Make a new fixed-window mod_exp implementation the default for
6175 RSA, DSA, and DH private-key operations so that the sequence of
6176 squares and multiplies and the memory access pattern are
6177 independent of the particular secret key. This will mitigate
6178 cache-timing and potential related attacks.
6180 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6181 and this is automatically used by BN_mod_exp_mont() if the new flag
6182 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6183 will use this BN flag for private exponents unless the flag
6184 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6185 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6187 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6189 *) Change the client implementation for SSLv23_method() and
6190 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6191 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6192 (Previously, the SSL 2.0 backwards compatible Client Hello
6193 message format would be used even with SSL_OP_NO_SSLv2.)
6196 *) Add support for smime-type MIME parameter in S/MIME messages which some
6200 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6201 a threadsafe manner. Modify rsa code to use new function and add calls
6202 to dsa and dh code (which had race conditions before).
6205 *) Include the fixed error library code in the C error file definitions
6206 instead of fixing them up at runtime. This keeps the error code
6207 structures constant.
6210 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6212 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6215 *) Fixes for newer kerberos headers. NB: the casts are needed because
6216 the 'length' field is signed on one version and unsigned on another
6217 with no (?) obvious way to tell the difference, without these VC++
6218 complains. Also the "definition" of FAR (blank) is no longer included
6219 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6220 some needed definitions.
6223 *) Undo Cygwin change.
6226 *) Added support for proxy certificates according to RFC 3820.
6227 Because they may be a security thread to unaware applications,
6228 they must be explicitly allowed in run-time. See
6229 docs/HOWTO/proxy_certificates.txt for further information.
6232 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6234 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6235 server and client random values. Previously
6236 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6237 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6239 This change has negligible security impact because:
6241 1. Server and client random values still have 24 bytes of pseudo random
6244 2. Server and client random values are sent in the clear in the initial
6247 3. The master secret is derived using the premaster secret (48 bytes in
6248 size for static RSA ciphersuites) as well as client server and random
6251 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6254 [Stephen Henson, reported by UK NISCC]
6256 *) Use Windows randomness collection on Cygwin.
6259 *) Fix hang in EGD/PRNGD query when communication socket is closed
6260 prematurely by EGD/PRNGD.
6261 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6263 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6266 *) Back-port of selected performance improvements from development
6267 branch, as well as improved support for PowerPC platforms.
6270 *) Add lots of checks for memory allocation failure, error codes to indicate
6271 failure and freeing up memory if a failure occurs.
6272 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6274 *) Add new -passin argument to dgst.
6277 *) Perform some character comparisons of different types in X509_NAME_cmp:
6278 this is needed for some certificates that re-encode DNs into UTF8Strings
6279 (in violation of RFC3280) and can't or won't issue name rollover
6283 *) Make an explicit check during certificate validation to see that
6284 the CA setting in each certificate on the chain is correct. As a
6285 side effect always do the following basic checks on extensions,
6286 not just when there's an associated purpose to the check:
6288 - if there is an unhandled critical extension (unless the user
6289 has chosen to ignore this fault)
6290 - if the path length has been exceeded (if one is set at all)
6291 - that certain extensions fit the associated purpose (if one has
6295 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6297 *) Avoid a race condition when CRLs are checked in a multi threaded
6298 environment. This would happen due to the reordering of the revoked
6299 entries during signature checking and serial number lookup. Now the
6300 encoding is cached and the serial number sort performed under a lock.
6301 Add new STACK function sk_is_sorted().
6304 *) Add Delta CRL to the extension code.
6307 *) Various fixes to s3_pkt.c so alerts are sent properly.
6308 [David Holmes <d.holmes@f5.com>]
6310 *) Reduce the chances of duplicate issuer name and serial numbers (in
6311 violation of RFC3280) using the OpenSSL certificate creation utilities.
6312 This is done by creating a random 64 bit value for the initial serial
6313 number when a serial number file is created or when a self signed
6314 certificate is created using 'openssl req -x509'. The initial serial
6315 number file is created using 'openssl x509 -next_serial' in CA.pl
6316 rather than being initialized to 1.
6319 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6321 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6322 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6323 [Joe Orton, Steve Henson]
6325 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6327 [Joe Orton, Steve Henson]
6329 *) Make it possible to have multiple active certificates with the same
6330 subject in the CA index file. This is done only if the keyword
6331 'unique_subject' is set to 'no' in the main CA section (default
6332 if 'CA_default') of the configuration file. The value is saved
6333 with the database itself in a separate index attribute file,
6334 named like the index file with '.attr' appended to the name.
6337 *) X509 verify fixes. Disable broken certificate workarounds when
6338 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6339 keyUsage extension present. Don't accept CRLs with unhandled critical
6340 extensions: since verify currently doesn't process CRL extensions this
6341 rejects a CRL with *any* critical extensions. Add new verify error codes
6345 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6346 A clarification of RFC2560 will require the use of OCTET STRINGs and
6347 some implementations cannot handle the current raw format. Since OpenSSL
6348 copies and compares OCSP nonces as opaque blobs without any attempt at
6349 parsing them this should not create any compatibility issues.
6352 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6353 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6354 this HMAC (and other) operations are several times slower than OpenSSL
6358 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6359 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6361 *) Use the correct content when signing type "other".
6364 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6366 *) Fix various bugs revealed by running the NISCC test suite:
6368 Stop out of bounds reads in the ASN1 code when presented with
6369 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6371 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6373 If verify callback ignores invalid public key errors don't try to check
6374 certificate signature with the NULL public key.
6378 *) New -ignore_err option in ocsp application to stop the server
6379 exiting on the first error in a request.
6382 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6383 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6387 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6388 extra data after the compression methods not only for TLS 1.0
6389 but also for SSL 3.0 (as required by the specification).
6390 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6392 *) Change X509_certificate_type() to mark the key as exported/exportable
6393 when it's 512 *bits* long, not 512 bytes.
6396 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6397 blocks during encryption.
6400 *) Various fixes to base64 BIO and non blocking I/O. On write
6401 flushes were not handled properly if the BIO retried. On read
6402 data was not being buffered properly and had various logic bugs.
6403 This also affects blocking I/O when the data being decoded is a
6407 *) Various S/MIME bugfixes and compatibility changes:
6408 output correct application/pkcs7 MIME type if
6409 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6410 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6411 of files as .eml work). Correctly handle very long lines in MIME
6415 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6417 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6418 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6419 a protocol version number mismatch like a decryption error
6420 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6423 *) Turn on RSA blinding by default in the default implementation
6424 to avoid a timing attack. Applications that don't want it can call
6425 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6426 They would be ill-advised to do so in most cases.
6427 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6429 *) Change RSA blinding code so that it works when the PRNG is not
6430 seeded (in this case, the secret RSA exponent is abused as
6431 an unpredictable seed -- if it is not unpredictable, there
6432 is no point in blinding anyway). Make RSA blinding thread-safe
6433 by remembering the creator's thread ID in rsa->blinding and
6434 having all other threads use local one-time blinding factors
6435 (this requires more computation than sharing rsa->blinding, but
6436 avoids excessive locking; and if an RSA object is not shared
6437 between threads, blinding will still be very fast).
6440 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6441 ENGINE as defaults for all supported algorithms irrespective of
6442 the 'flags' parameter. 'flags' is now honoured, so applications
6443 should make sure they are passing it correctly.
6446 *) Target "mingw" now allows native Windows code to be generated in
6447 the Cygwin environment as well as with the MinGW compiler.
6450 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6452 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6453 via timing by performing a MAC computation even if incorrect
6454 block cipher padding has been found. This is a countermeasure
6455 against active attacks where the attacker has to distinguish
6456 between bad padding and a MAC verification error. (CVE-2003-0078)
6458 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6459 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6460 Martin Vuagnoux (EPFL, Ilion)]
6462 *) Make the no-err option work as intended. The intention with no-err
6463 is not to have the whole error stack handling routines removed from
6464 libcrypto, it's only intended to remove all the function name and
6465 reason texts, thereby removing some of the footprint that may not
6466 be interesting if those errors aren't displayed anyway.
6468 NOTE: it's still possible for any application or module to have it's
6469 own set of error texts inserted. The routines are there, just not
6470 used by default when no-err is given.
6473 *) Add support for FreeBSD on IA64.
6474 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6476 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6477 Kerberos function mit_des_cbc_cksum(). Before this change,
6478 the value returned by DES_cbc_cksum() was like the one from
6479 mit_des_cbc_cksum(), except the bytes were swapped.
6480 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6482 *) Allow an application to disable the automatic SSL chain building.
6483 Before this a rather primitive chain build was always performed in
6484 ssl3_output_cert_chain(): an application had no way to send the
6485 correct chain if the automatic operation produced an incorrect result.
6487 Now the chain builder is disabled if either:
6489 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6491 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6493 The reasoning behind this is that an application would not want the
6494 auto chain building to take place if extra chain certificates are
6495 present and it might also want a means of sending no additional
6496 certificates (for example the chain has two certificates and the
6500 *) Add the possibility to build without the ENGINE framework.
6501 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6503 *) Under Win32 gmtime() can return NULL: check return value in
6504 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6507 *) DSA routines: under certain error conditions uninitialized BN objects
6508 could be freed. Solution: make sure initialization is performed early
6509 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6510 Nils Larsch <nla@trustcenter.de> via PR#459)
6513 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6514 checked on reconnect on the client side, therefore session resumption
6515 could still fail with a "ssl session id is different" error. This
6516 behaviour is masked when SSL_OP_ALL is used due to
6517 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6518 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6519 followup to PR #377.
6522 *) IA-32 assembler support enhancements: unified ELF targets, support
6523 for SCO/Caldera platforms, fix for Cygwin shared build.
6526 *) Add support for FreeBSD on sparc64. As a consequence, support for
6527 FreeBSD on non-x86 processors is separate from x86 processors on
6528 the config script, much like the NetBSD support.
6529 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6531 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6533 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6536 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6537 code (06) was taken as the first octet of the session ID and the last
6538 octet was ignored consequently. As a result SSLv2 client side session
6539 caching could not have worked due to the session ID mismatch between
6541 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6545 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6546 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6550 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6551 seems that in spite of existing for more than a year, many application
6552 author have done nothing to provide the necessary callbacks, which
6553 means that this particular engine will not work properly anywhere.
6554 This is a very unfortunate situation which forces us, in the name
6555 of usability, to give the hw_ncipher.c a static lock, which is part
6557 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6558 appear in 0.9.8 or later. We EXPECT application authors to have
6559 dealt properly with this when 0.9.8 is released (unless we actually
6560 make such changes in the libcrypto locking code that changes will
6561 have to be made anyway).
6564 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6565 octets have been read, EOF or an error occurs. Without this change
6566 some truncated ASN1 structures will not produce an error.
6569 *) Disable Heimdal support, since it hasn't been fully implemented.
6570 Still give the possibility to force the use of Heimdal, but with
6571 warnings and a request that patches get sent to openssl-dev.
6574 *) Add the VC-CE target, introduce the WINCE sysname, and add
6575 INSTALL.WCE and appropriate conditionals to make it build.
6576 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6578 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6579 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6580 edit numbers of the version.
6581 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6583 *) Introduce safe string copy and catenation functions
6584 (BUF_strlcpy() and BUF_strlcat()).
6585 [Ben Laurie (CHATS) and Richard Levitte]
6587 *) Avoid using fixed-size buffers for one-line DNs.
6588 [Ben Laurie (CHATS)]
6590 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6591 resizing buffers containing secrets, and use where appropriate.
6592 [Ben Laurie (CHATS)]
6594 *) Avoid using fixed size buffers for configuration file location.
6595 [Ben Laurie (CHATS)]
6597 *) Avoid filename truncation for various CA files.
6598 [Ben Laurie (CHATS)]
6600 *) Use sizeof in preference to magic numbers.
6601 [Ben Laurie (CHATS)]
6603 *) Avoid filename truncation in cert requests.
6604 [Ben Laurie (CHATS)]
6606 *) Add assertions to check for (supposedly impossible) buffer
6608 [Ben Laurie (CHATS)]
6610 *) Don't cache truncated DNS entries in the local cache (this could
6611 potentially lead to a spoofing attack).
6612 [Ben Laurie (CHATS)]
6614 *) Fix various buffers to be large enough for hex/decimal
6615 representations in a platform independent manner.
6616 [Ben Laurie (CHATS)]
6618 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6619 resizing buffers containing secrets, and use where appropriate.
6620 [Ben Laurie (CHATS)]
6622 *) Add BIO_indent() to avoid much slightly worrying code to do
6624 [Ben Laurie (CHATS)]
6626 *) Convert sprintf()/BIO_puts() to BIO_printf().
6627 [Ben Laurie (CHATS)]
6629 *) buffer_gets() could terminate with the buffer only half
6631 [Ben Laurie (CHATS)]
6633 *) Add assertions to prevent user-supplied crypto functions from
6634 overflowing internal buffers by having large block sizes, etc.
6635 [Ben Laurie (CHATS)]
6637 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6639 [Ben Laurie (CHATS)]
6641 *) Eliminate unused copy of key in RC4.
6642 [Ben Laurie (CHATS)]
6644 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6645 [Ben Laurie (CHATS)]
6647 *) Fix off-by-one error in EGD path.
6648 [Ben Laurie (CHATS)]
6650 *) If RANDFILE path is too long, ignore instead of truncating.
6651 [Ben Laurie (CHATS)]
6653 *) Eliminate unused and incorrectly sized X.509 structure
6655 [Ben Laurie (CHATS)]
6657 *) Eliminate unused and dangerous function knumber().
6658 [Ben Laurie (CHATS)]
6660 *) Eliminate unused and dangerous structure, KSSL_ERR.
6661 [Ben Laurie (CHATS)]
6663 *) Protect against overlong session ID context length in an encoded
6664 session object. Since these are local, this does not appear to be
6666 [Ben Laurie (CHATS)]
6668 *) Change from security patch (see 0.9.6e below) that did not affect
6669 the 0.9.6 release series:
6671 Remote buffer overflow in SSL3 protocol - an attacker could
6672 supply an oversized master key in Kerberos-enabled versions.
6674 [Ben Laurie (CHATS)]
6676 *) Change the SSL kerb5 codes to match RFC 2712.
6679 *) Make -nameopt work fully for req and add -reqopt switch.
6680 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6682 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6683 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6685 *) Make sure tests can be performed even if the corresponding algorithms
6686 have been removed entirely. This was also the last step to make
6687 OpenSSL compilable with DJGPP under all reasonable conditions.
6688 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6690 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6691 to allow version independent disabling of normally unselected ciphers,
6692 which may be activated as a side-effect of selecting a single cipher.
6694 (E.g., cipher list string "RSA" enables ciphersuites that are left
6695 out of "ALL" because they do not provide symmetric encryption.
6696 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6697 [Lutz Jaenicke, Bodo Moeller]
6699 *) Add appropriate support for separate platform-dependent build
6700 directories. The recommended way to make a platform-dependent
6701 build directory is the following (tested on Linux), maybe with
6704 # Place yourself outside of the OpenSSL source tree. In
6705 # this example, the environment variable OPENSSL_SOURCE
6706 # is assumed to contain the absolute OpenSSL source directory.
6707 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6708 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6709 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6710 mkdir -p `dirname $F`
6711 ln -s $OPENSSL_SOURCE/$F $F
6714 To be absolutely sure not to disturb the source tree, a "make clean"
6715 is a good thing. If it isn't successful, don't worry about it,
6716 it probably means the source directory is very clean.
6719 *) Make sure any ENGINE control commands make local copies of string
6720 pointers passed to them whenever necessary. Otherwise it is possible
6721 the caller may have overwritten (or deallocated) the original string
6722 data when a later ENGINE operation tries to use the stored values.
6723 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6725 *) Improve diagnostics in file reading and command-line digests.
6726 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6728 *) Add AES modes CFB and OFB to the object database. Correct an
6729 error in AES-CFB decryption.
6732 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6733 allows existing EVP_CIPHER_CTX structures to be reused after
6734 calling EVP_*Final(). This behaviour is used by encryption
6735 BIOs and some applications. This has the side effect that
6736 applications must explicitly clean up cipher contexts with
6737 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6740 *) Check the values of dna and dnb in bn_mul_recursive before calling
6741 bn_mul_comba (a non zero value means the a or b arrays do not contain
6742 n2 elements) and fallback to bn_mul_normal if either is not zero.
6745 *) Fix escaping of non-ASCII characters when using the -subj option
6746 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6749 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6750 form for "surname", serialNumber has no short form.
6751 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6752 therefore remove "mail" short name for "internet 7".
6753 The OID for unique identifiers in X509 certificates is
6754 x500UniqueIdentifier, not uniqueIdentifier.
6755 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6758 *) Add an "init" command to the ENGINE config module and auto initialize
6759 ENGINEs. Without any "init" command the ENGINE will be initialized
6760 after all ctrl commands have been executed on it. If init=1 the
6761 ENGINE is initialized at that point (ctrls before that point are run
6762 on the uninitialized ENGINE and after on the initialized one). If
6763 init=0 then the ENGINE will not be initialized at all.
6766 *) Fix the 'app_verify_callback' interface so that the user-defined
6767 argument is actually passed to the callback: In the
6768 SSL_CTX_set_cert_verify_callback() prototype, the callback
6769 declaration has been changed from
6772 int (*cb)(X509_STORE_CTX *,void *);
6773 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6774 i=s->ctx->app_verify_callback(&ctx)
6775 has been changed into
6776 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6778 To update applications using SSL_CTX_set_cert_verify_callback(),
6779 a dummy argument can be added to their callback functions.
6780 [D. K. Smetters <smetters@parc.xerox.com>]
6782 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6783 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6785 *) Add and OPENSSL_LOAD_CONF define which will cause
6786 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6787 This allows older applications to transparently support certain
6788 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6789 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6790 load the config file and OPENSSL_add_all_algorithms_conf() which will
6791 always load it have also been added.
6794 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6795 Adjust NIDs and EVP layer.
6796 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6798 *) Config modules support in openssl utility.
6800 Most commands now load modules from the config file,
6801 though in a few (such as version) this isn't done
6802 because it couldn't be used for anything.
6804 In the case of ca and req the config file used is
6805 the same as the utility itself: that is the -config
6806 command line option can be used to specify an
6810 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6811 use "openssl_conf" if filename is NULL use default openssl config file.
6814 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6815 config section name. Add a new flag to tolerate a missing config file
6816 and move code to CONF_modules_load_file().
6819 *) Support for crypto accelerator cards from Accelerated Encryption
6820 Processing, www.aep.ie. (Use engine 'aep')
6821 The support was copied from 0.9.6c [engine] and adapted/corrected
6822 to work with the new engine framework.
6823 [AEP Inc. and Richard Levitte]
6825 *) Support for SureWare crypto accelerator cards from Baltimore
6826 Technologies. (Use engine 'sureware')
6827 The support was copied from 0.9.6c [engine] and adapted
6828 to work with the new engine framework.
6831 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6832 make the newer ENGINE framework commands for the CHIL engine work.
6833 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6835 *) Make it possible to produce shared libraries on ReliantUNIX.
6836 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6838 *) Add the configuration target debug-linux-ppro.
6839 Make 'openssl rsa' use the general key loading routines
6840 implemented in apps.c, and make those routines able to
6841 handle the key format FORMAT_NETSCAPE and the variant
6843 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6845 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6846 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6848 *) Add -keyform to rsautl, and document -engine.
6849 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6851 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6852 BIO_R_NO_SUCH_FILE error code rather than the generic
6853 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6856 *) Add new functions
6858 ERR_peek_last_error_line
6859 ERR_peek_last_error_line_data.
6860 These are similar to
6863 ERR_peek_error_line_data,
6864 but report on the latest error recorded rather than the first one
6865 still in the error queue.
6866 [Ben Laurie, Bodo Moeller]
6868 *) default_algorithms option in ENGINE config module. This allows things
6870 default_algorithms = ALL
6871 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6874 *) Preliminary ENGINE config module.
6877 *) New experimental application configuration code.
6880 *) Change the AES code to follow the same name structure as all other
6881 symmetric ciphers, and behave the same way. Move everything to
6882 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6883 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6885 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6886 [Ben Laurie and Theo de Raadt]
6888 *) Add option to output public keys in req command.
6889 [Massimiliano Pala madwolf@openca.org]
6891 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6892 (up to about 10% better than before for P-192 and P-224).
6895 *) New functions/macros
6897 SSL_CTX_set_msg_callback(ctx, cb)
6898 SSL_CTX_set_msg_callback_arg(ctx, arg)
6899 SSL_set_msg_callback(ssl, cb)
6900 SSL_set_msg_callback_arg(ssl, arg)
6902 to request calling a callback function
6904 void cb(int write_p, int version, int content_type,
6905 const void *buf, size_t len, SSL *ssl, void *arg)
6907 whenever a protocol message has been completely received
6908 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6909 protocol version according to which the SSL library interprets
6910 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6911 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6912 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6913 specification (change_cipher_spec(20), alert(21), handshake(22)).
6914 'buf' and 'len' point to the actual message, 'ssl' to the
6915 SSL object, and 'arg' is the application-defined value set by
6916 SSL[_CTX]_set_msg_callback_arg().
6918 'openssl s_client' and 'openssl s_server' have new '-msg' options
6919 to enable a callback that displays all protocol messages.
6922 *) Change the shared library support so shared libraries are built as
6923 soon as the corresponding static library is finished, and thereby get
6924 openssl and the test programs linked against the shared library.
6925 This still only happens when the keyword "shard" has been given to
6926 the configuration scripts.
6928 NOTE: shared library support is still an experimental thing, and
6929 backward binary compatibility is still not guaranteed.
6930 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6932 *) Add support for Subject Information Access extension.
6933 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6935 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6936 additional bytes when new memory had to be allocated, not just
6937 when reusing an existing buffer.
6940 *) New command line and configuration option 'utf8' for the req command.
6941 This allows field values to be specified as UTF8 strings.
6944 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6945 runs for the former and machine-readable output for the latter.
6948 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6949 of the e-mail address in the DN (i.e., it will go into a certificate
6950 extension only). The new configuration file option 'email_in_dn = no'
6951 has the same effect.
6952 [Massimiliano Pala madwolf@openca.org]
6954 *) Change all functions with names starting with des_ to be starting
6955 with DES_ instead. Add wrappers that are compatible with libdes,
6956 but are named _ossl_old_des_*. Finally, add macros that map the
6957 des_* symbols to the corresponding _ossl_old_des_* if libdes
6958 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6959 desired, the des_* symbols will be mapped to DES_*, with one
6962 Since we provide two compatibility mappings, the user needs to
6963 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6964 compatibility is desired. The default (i.e., when that macro
6965 isn't defined) is OpenSSL 0.9.6c compatibility.
6967 There are also macros that enable and disable the support of old
6968 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6969 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6970 are defined, the default will apply: to support the old des routines.
6972 In either case, one must include openssl/des.h to get the correct
6973 definitions. Do not try to just include openssl/des_old.h, that
6976 NOTE: This is a major break of an old API into a new one. Software
6977 authors are encouraged to switch to the DES_ style functions. Some
6978 time in the future, des_old.h and the libdes compatibility functions
6979 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6980 default), and then completely removed.
6983 *) Test for certificates which contain unsupported critical extensions.
6984 If such a certificate is found during a verify operation it is
6985 rejected by default: this behaviour can be overridden by either
6986 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6987 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6988 X509_supported_extension() has also been added which returns 1 if a
6989 particular extension is supported.
6992 *) Modify the behaviour of EVP cipher functions in similar way to digests
6993 to retain compatibility with existing code.
6996 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6997 compatibility with existing code. In particular the 'ctx' parameter does
6998 not have to be to be initialized before the call to EVP_DigestInit() and
6999 it is tidied up after a call to EVP_DigestFinal(). New function
7000 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7001 EVP_MD_CTX_copy() changed to not require the destination to be
7002 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7003 requires the destination to be valid.
7005 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7006 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7009 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7010 so that complete 'Handshake' protocol structures are kept in memory
7011 instead of overwriting 'msg_type' and 'length' with 'body' data.
7014 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7015 [Massimo Santin via Richard Levitte]
7017 *) Major restructuring to the underlying ENGINE code. This includes
7018 reduction of linker bloat, separation of pure "ENGINE" manipulation
7019 (initialisation, etc) from functionality dealing with implementations
7020 of specific crypto interfaces. This change also introduces integrated
7021 support for symmetric ciphers and digest implementations - so ENGINEs
7022 can now accelerate these by providing EVP_CIPHER and EVP_MD
7023 implementations of their own. This is detailed in crypto/engine/README
7024 as it couldn't be adequately described here. However, there are a few
7025 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7026 were changed in the original introduction of ENGINE code have now
7027 reverted back - the hooking from this code to ENGINE is now a good
7028 deal more passive and at run-time, operations deal directly with
7029 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7030 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7031 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7032 they were not being used by the framework as there is no concept of a
7033 BIGNUM_METHOD and they could not be generalised to the new
7034 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7035 ENGINE_cpy() has been removed as it cannot be consistently defined in
7039 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7042 *) Change mkdef.pl to sort symbols that get the same entry number,
7043 and make sure the automatically generated functions ERR_load_*
7044 become part of libeay.num as well.
7047 *) New function SSL_renegotiate_pending(). This returns true once
7048 renegotiation has been requested (either SSL_renegotiate() call
7049 or HelloRequest/ClientHello received from the peer) and becomes
7050 false once a handshake has been completed.
7051 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7052 sends a HelloRequest, but does not ensure that a handshake takes
7053 place. SSL_renegotiate_pending() is useful for checking if the
7054 client has followed the request.)
7057 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7058 By default, clients may request session resumption even during
7059 renegotiation (if session ID contexts permit); with this option,
7060 session resumption is possible only in the first handshake.
7062 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7063 more bits available for options that should not be part of
7064 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7067 *) Add some demos for certificate and certificate request creation.
7070 *) Make maximum certificate chain size accepted from the peer application
7071 settable (SSL*_get/set_max_cert_list()), as proposed by
7072 "Douglas E. Engert" <deengert@anl.gov>.
7075 *) Add support for shared libraries for Unixware-7
7076 (Boyd Lynn Gerber <gerberb@zenez.com>).
7079 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7080 be done prior to destruction. Use this to unload error strings from
7081 ENGINEs that load their own error strings. NB: This adds two new API
7082 functions to "get" and "set" this destroy handler in an ENGINE.
7085 *) Alter all existing ENGINE implementations (except "openssl" and
7086 "openbsd") to dynamically instantiate their own error strings. This
7087 makes them more flexible to be built both as statically-linked ENGINEs
7088 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7089 Also, add stub code to each that makes building them as self-contained
7090 shared-libraries easier (see README.ENGINE).
7093 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7094 implementations into applications that are completely implemented in
7095 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7096 commands that can be used to configure what shared-library to load and
7097 to control aspects of the way it is handled. Also, made an update to
7098 the README.ENGINE file that brings its information up-to-date and
7099 provides some information and instructions on the "dynamic" ENGINE
7100 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7103 *) Make it possible to unload ranges of ERR strings with a new
7104 "ERR_unload_strings" function.
7107 *) Add a copy() function to EVP_MD.
7110 *) Make EVP_MD routines take a context pointer instead of just the
7111 md_data void pointer.
7114 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7115 that the digest can only process a single chunk of data
7116 (typically because it is provided by a piece of
7117 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7118 is only going to provide a single chunk of data, and hence the
7119 framework needn't accumulate the data for oneshot drivers.
7122 *) As with "ERR", make it possible to replace the underlying "ex_data"
7123 functions. This change also alters the storage and management of global
7124 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7125 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7126 index counters. The API functions that use this state have been changed
7127 to take a "class_index" rather than pointers to the class's local STACK
7128 and counter, and there is now an API function to dynamically create new
7129 classes. This centralisation allows us to (a) plug a lot of the
7130 thread-safety problems that existed, and (b) makes it possible to clean
7131 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7132 such data would previously have always leaked in application code and
7133 workarounds were in place to make the memory debugging turn a blind eye
7134 to it. Application code that doesn't use this new function will still
7135 leak as before, but their memory debugging output will announce it now
7136 rather than letting it slide.
7138 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7139 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7140 has a return value to indicate success or failure.
7143 *) Make it possible to replace the underlying "ERR" functions such that the
7144 global state (2 LHASH tables and 2 locks) is only used by the "default"
7145 implementation. This change also adds two functions to "get" and "set"
7146 the implementation prior to it being automatically set the first time
7147 any other ERR function takes place. Ie. an application can call "get",
7148 pass the return value to a module it has just loaded, and that module
7149 can call its own "set" function using that value. This means the
7150 module's "ERR" operations will use (and modify) the error state in the
7151 application and not in its own statically linked copy of OpenSSL code.
7154 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7155 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7156 the operation, and provides a more encapsulated way for external code
7157 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7158 to use these functions rather than manually incrementing the counts.
7160 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7163 *) Add EVP test program.
7166 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7169 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7170 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7171 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7172 These allow a CRL to be built without having to access X509_CRL fields
7173 directly. Modify 'ca' application to use new functions.
7176 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7177 bug workarounds. Rollback attack detection is a security feature.
7178 The problem will only arise on OpenSSL servers when TLSv1 is not
7179 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7180 Software authors not wanting to support TLSv1 will have special reasons
7181 for their choice and can explicitly enable this option.
7182 [Bodo Moeller, Lutz Jaenicke]
7184 *) Rationalise EVP so it can be extended: don't include a union of
7185 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7186 (similar to those existing for EVP_CIPHER_CTX).
7191 EVP_MD_CTX_init(&md); /* new function call */
7192 EVP_DigestInit(&md, EVP_sha1());
7193 EVP_DigestUpdate(&md, in, len);
7194 EVP_DigestFinal(&md, out, NULL);
7195 EVP_MD_CTX_cleanup(&md); /* new function call */
7199 *) Make DES key schedule conform to the usual scheme, as well as
7200 correcting its structure. This means that calls to DES functions
7201 now have to pass a pointer to a des_key_schedule instead of a
7202 plain des_key_schedule (which was actually always a pointer
7205 des_key_schedule ks;
7207 des_set_key_checked(..., &ks);
7208 des_ncbc_encrypt(..., &ks, ...);
7210 (Note that a later change renames 'des_...' into 'DES_...'.)
7213 *) Initial reduction of linker bloat: the use of some functions, such as
7214 PEM causes large amounts of unused functions to be linked in due to
7215 poor organisation. For example pem_all.c contains every PEM function
7216 which has a knock on effect of linking in large amounts of (unused)
7217 ASN1 code. Grouping together similar functions and splitting unrelated
7218 functions prevents this.
7221 *) Cleanup of EVP macros.
7224 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7225 correct _ecb suffix.
7228 *) Add initial OCSP responder support to ocsp application. The
7229 revocation information is handled using the text based index
7230 use by the ca application. The responder can either handle
7231 requests generated internally, supplied in files (for example
7232 via a CGI script) or using an internal minimal server.
7235 *) Add configuration choices to get zlib compression for TLS.
7238 *) Changes to Kerberos SSL for RFC 2712 compliance:
7239 1. Implemented real KerberosWrapper, instead of just using
7240 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7241 2. Implemented optional authenticator field of KerberosWrapper.
7243 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7244 and authenticator structs; see crypto/krb5/.
7246 Generalized Kerberos calls to support multiple Kerberos libraries.
7247 [Vern Staats <staatsvr@asc.hpc.mil>,
7248 Jeffrey Altman <jaltman@columbia.edu>
7249 via Richard Levitte]
7251 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7252 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7253 values for each of the key sizes rather than having just
7254 parameters (and 'speed' generating keys each time).
7257 *) Speed up EVP routines.
7260 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7261 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7262 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7263 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7265 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7266 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7267 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7270 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7272 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7275 *) Added the OS2-EMX target.
7276 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7278 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7279 to support NCONF routines in extension code. New function CONF_set_nconf()
7280 to allow functions which take an NCONF to also handle the old LHASH
7281 structure: this means that the old CONF compatible routines can be
7282 retained (in particular wrt extensions) without having to duplicate the
7283 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7286 *) Enhance the general user interface with mechanisms for inner control
7287 and with possibilities to have yes/no kind of prompts.
7290 *) Change all calls to low level digest routines in the library and
7291 applications to use EVP. Add missing calls to HMAC_cleanup() and
7292 don't assume HMAC_CTX can be copied using memcpy().
7293 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7295 *) Add the possibility to control engines through control names but with
7296 arbitrary arguments instead of just a string.
7297 Change the key loaders to take a UI_METHOD instead of a callback
7298 function pointer. NOTE: this breaks binary compatibility with earlier
7299 versions of OpenSSL [engine].
7300 Adapt the nCipher code for these new conditions and add a card insertion
7304 *) Enhance the general user interface with mechanisms to better support
7305 dialog box interfaces, application-defined prompts, the possibility
7306 to use defaults (for example default passwords from somewhere else)
7307 and interrupts/cancellations.
7310 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7311 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7314 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7315 tidy up some unnecessarily weird code in 'sk_new()').
7316 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7318 *) Change the key loading routines for ENGINEs to use the same kind
7319 callback (pem_password_cb) as all other routines that need this
7323 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7324 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7325 than this minimum value is recommended.
7328 *) New random seeder for OpenVMS, using the system process statistics
7329 that are easily reachable.
7332 *) Windows apparently can't transparently handle global
7333 variables defined in DLLs. Initialisations such as:
7335 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7337 won't compile. This is used by the any applications that need to
7338 declare their own ASN1 modules. This was fixed by adding the option
7339 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7340 needed for static libraries under Win32.
7343 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7344 setting of purpose and trust fields. New X509_STORE trust and
7345 purpose functions and tidy up setting in other SSL functions.
7348 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7349 structure. These are inherited by X509_STORE_CTX when it is
7350 initialised. This allows various defaults to be set in the
7351 X509_STORE structure (such as flags for CRL checking and custom
7352 purpose or trust settings) for functions which only use X509_STORE_CTX
7353 internally such as S/MIME.
7355 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7356 trust settings if they are not set in X509_STORE. This allows X509_STORE
7357 purposes and trust (in S/MIME for example) to override any set by default.
7359 Add command line options for CRL checking to smime, s_client and s_server
7363 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7364 are set then the CRL is looked up in the X509_STORE structure and
7365 its validity and signature checked, then if the certificate is found
7366 in the CRL the verify fails with a revoked error.
7368 Various new CRL related callbacks added to X509_STORE_CTX structure.
7370 Command line options added to 'verify' application to support this.
7372 This needs some additional work, such as being able to handle multiple
7373 CRLs with different times, extension based lookup (rather than just
7374 by subject name) and ultimately more complete V2 CRL extension
7378 *) Add a general user interface API (crypto/ui/). This is designed
7379 to replace things like des_read_password and friends (backward
7380 compatibility functions using this new API are provided).
7381 The purpose is to remove prompting functions from the DES code
7382 section as well as provide for prompting through dialog boxes in
7383 a window system and the like.
7386 *) Add "ex_data" support to ENGINE so implementations can add state at a
7387 per-structure level rather than having to store it globally.
7390 *) Make it possible for ENGINE structures to be copied when retrieved by
7391 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7392 This causes the "original" ENGINE structure to act like a template,
7393 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7394 operational state can be localised to each ENGINE structure, despite the
7395 fact they all share the same "methods". New ENGINE structures returned in
7396 this case have no functional references and the return value is the single
7397 structural reference. This matches the single structural reference returned
7398 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7402 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7403 needs to match any other type at all we need to manually clear the
7407 *) Changes to the "openssl engine" utility to include;
7408 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7409 about an ENGINE's available control commands.
7410 - executing control commands from command line arguments using the
7411 '-pre' and '-post' switches. '-post' is only used if '-t' is
7412 specified and the ENGINE is successfully initialised. The syntax for
7413 the individual commands are colon-separated, for example;
7414 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7417 *) New dynamic control command support for ENGINEs. ENGINEs can now
7418 declare their own commands (numbers), names (strings), descriptions,
7419 and input types for run-time discovery by calling applications. A
7420 subset of these commands are implicitly classed as "executable"
7421 depending on their input type, and only these can be invoked through
7422 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7423 can be based on user input, config files, etc). The distinction is
7424 that "executable" commands cannot return anything other than a boolean
7425 result and can only support numeric or string input, whereas some
7426 discoverable commands may only be for direct use through
7427 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7428 pointers, or other custom uses. The "executable" commands are to
7429 support parameterisations of ENGINE behaviour that can be
7430 unambiguously defined by ENGINEs and used consistently across any
7431 OpenSSL-based application. Commands have been added to all the
7432 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7433 control over shared-library paths without source code alterations.
7436 *) Changed all ENGINE implementations to dynamically allocate their
7437 ENGINEs rather than declaring them statically. Apart from this being
7438 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7439 this also allows the implementations to compile without using the
7440 internal engine_int.h header.
7443 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7444 'const' value. Any code that should be able to modify a RAND_METHOD
7445 should already have non-const pointers to it (ie. they should only
7446 modify their own ones).
7449 *) Made a variety of little tweaks to the ENGINE code.
7450 - "atalla" and "ubsec" string definitions were moved from header files
7451 to C code. "nuron" string definitions were placed in variables
7452 rather than hard-coded - allowing parameterisation of these values
7453 later on via ctrl() commands.
7454 - Removed unused "#if 0"'d code.
7455 - Fixed engine list iteration code so it uses ENGINE_free() to release
7456 structural references.
7457 - Constified the RAND_METHOD element of ENGINE structures.
7458 - Constified various get/set functions as appropriate and added
7459 missing functions (including a catch-all ENGINE_cpy that duplicates
7460 all ENGINE values onto a new ENGINE except reference counts/state).
7461 - Removed NULL parameter checks in get/set functions. Setting a method
7462 or function to NULL is a way of cancelling out a previously set
7463 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7464 and doesn't justify the extra error symbols and code.
7465 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7466 flags from engine_int.h to engine.h.
7467 - Changed prototypes for ENGINE handler functions (init(), finish(),
7468 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7471 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7472 to the algorithm using long division. The binary algorithm can be
7473 used only if the modulus is odd. On 32-bit systems, it is faster
7474 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7475 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7476 up to 450 bits. In 64-bit environments, the binary algorithm
7477 appears to be advantageous for much longer moduli; here we use it
7478 for moduli up to 2048 bits.
7481 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7482 could not support the combine flag in choice fields.
7485 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7486 extensions from a certificate request to the certificate.
7489 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7490 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7491 file: this allows the display of the certificate about to be
7492 signed to be customised, to allow certain fields to be included
7493 or excluded and extension details. The old system didn't display
7494 multicharacter strings properly, omitted fields not in the policy
7495 and couldn't display additional details such as extensions.
7498 *) Function EC_POINTs_mul for multiple scalar multiplication
7499 of an arbitrary number of elliptic curve points
7500 \sum scalars[i]*points[i],
7501 optionally including the generator defined for the EC_GROUP:
7502 scalar*generator + \sum scalars[i]*points[i].
7504 EC_POINT_mul is a simple wrapper function for the typical case
7505 that the point list has just one item (besides the optional
7509 *) First EC_METHODs for curves over GF(p):
7511 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7512 operations and provides various method functions that can also
7513 operate with faster implementations of modular arithmetic.
7515 EC_GFp_mont_method() reuses most functions that are part of
7516 EC_GFp_simple_method, but uses Montgomery arithmetic.
7518 [Bodo Moeller; point addition and point doubling
7519 implementation directly derived from source code provided by
7520 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7522 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7523 crypto/ec/ec_lib.c):
7525 Curves are EC_GROUP objects (with an optional group generator)
7526 based on EC_METHODs that are built into the library.
7528 Points are EC_POINT objects based on EC_GROUP objects.
7530 Most of the framework would be able to handle curves over arbitrary
7531 finite fields, but as there are no obvious types for fields other
7532 than GF(p), some functions are limited to that for now.
7535 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7536 that the file contains a complete HTTP response.
7539 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7540 change the def and num file printf format specifier from "%-40sXXX"
7541 to "%-39s XXX". The latter will always guarantee a space after the
7542 field while the former will cause them to run together if the field
7543 is 40 of more characters long.
7546 *) Constify the cipher and digest 'method' functions and structures
7547 and modify related functions to take constant EVP_MD and EVP_CIPHER
7551 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7552 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7555 *) Modify EVP_Digest*() routines so they now return values. Although the
7556 internal software routines can never fail additional hardware versions
7560 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7562 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7563 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7566 ERR_R_NESTED_ASN1_ERROR
7568 ERR_R_MISSING_ASN1_EOS
7569 were 4 .. 9, conflicting with
7570 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7572 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7573 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7575 Add new error code 'ERR_R_INTERNAL_ERROR'.
7578 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7582 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7583 sets the subject name for a new request or supersedes the
7584 subject name in a given request. Formats that can be parsed are
7585 'CN=Some Name, OU=myOU, C=IT'
7587 'CN=Some Name/OU=myOU/C=IT'.
7589 Add options '-batch' and '-verbose' to 'openssl req'.
7590 [Massimiliano Pala <madwolf@hackmasters.net>]
7592 *) Introduce the possibility to access global variables through
7593 functions on platform were that's the best way to handle exporting
7594 global variables in shared libraries. To enable this functionality,
7595 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7596 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7597 is normally done by Configure or something similar).
7599 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7600 in the source file (foo.c) like this:
7602 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7603 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7605 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7606 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7608 OPENSSL_DECLARE_GLOBAL(int,foo);
7609 #define foo OPENSSL_GLOBAL_REF(foo)
7610 OPENSSL_DECLARE_GLOBAL(double,bar);
7611 #define bar OPENSSL_GLOBAL_REF(bar)
7613 The #defines are very important, and therefore so is including the
7614 header file everywhere where the defined globals are used.
7616 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7617 of ASN.1 items, but that structure is a bit different.
7619 The largest change is in util/mkdef.pl which has been enhanced with
7620 better and easier to understand logic to choose which symbols should
7621 go into the Windows .def files as well as a number of fixes and code
7622 cleanup (among others, algorithm keywords are now sorted
7623 lexicographically to avoid constant rewrites).
7626 *) In BN_div() keep a copy of the sign of 'num' before writing the
7627 result to 'rm' because if rm==num the value will be overwritten
7628 and produce the wrong result if 'num' is negative: this caused
7629 problems with BN_mod() and BN_nnmod().
7632 *) Function OCSP_request_verify(). This checks the signature on an
7633 OCSP request and verifies the signer certificate. The signer
7634 certificate is just checked for a generic purpose and OCSP request
7638 *) Add OCSP_check_validity() function to check the validity of OCSP
7639 responses. OCSP responses are prepared in real time and may only
7640 be a few seconds old. Simply checking that the current time lies
7641 between thisUpdate and nextUpdate max reject otherwise valid responses
7642 caused by either OCSP responder or client clock inaccuracy. Instead
7643 we allow thisUpdate and nextUpdate to fall within a certain period of
7644 the current time. The age of the response can also optionally be
7645 checked. Two new options -validity_period and -status_age added to
7649 *) If signature or public key algorithm is unrecognized print out its
7650 OID rather that just UNKNOWN.
7653 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7654 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7655 ID to be generated from the issuer certificate alone which can then be
7656 passed to OCSP_id_issuer_cmp().
7659 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7660 ASN1 modules to export functions returning ASN1_ITEM pointers
7661 instead of the ASN1_ITEM structures themselves. This adds several
7662 new macros which allow the underlying ASN1 function/structure to
7663 be accessed transparently. As a result code should not use ASN1_ITEM
7664 references directly (such as &X509_it) but instead use the relevant
7665 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7666 use of the new ASN1 code on platforms where exporting structures
7667 is problematical (for example in shared libraries) but exporting
7668 functions returning pointers to structures is not.
7671 *) Add support for overriding the generation of SSL/TLS session IDs.
7672 These callbacks can be registered either in an SSL_CTX or per SSL.
7673 The purpose of this is to allow applications to control, if they wish,
7674 the arbitrary values chosen for use as session IDs, particularly as it
7675 can be useful for session caching in multiple-server environments. A
7676 command-line switch for testing this (and any client code that wishes
7677 to use such a feature) has been added to "s_server".
7678 [Geoff Thorpe, Lutz Jaenicke]
7680 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7681 of the form '#if defined(...) || defined(...) || ...' and
7682 '#if !defined(...) && !defined(...) && ...'. This also avoids
7683 the growing number of special cases it was previously handling.
7686 *) Make all configuration macros available for application by making
7687 sure they are available in opensslconf.h, by giving them names starting
7688 with "OPENSSL_" to avoid conflicts with other packages and by making
7689 sure e_os2.h will cover all platform-specific cases together with
7691 Additionally, it is now possible to define configuration/platform-
7692 specific names (called "system identities"). In the C code, these
7693 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7694 macro with the name beginning with "OPENSSL_SYS_", which is determined
7695 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7699 *) New option -set_serial to 'req' and 'x509' this allows the serial
7700 number to use to be specified on the command line. Previously self
7701 signed certificates were hard coded with serial number 0 and the
7702 CA options of 'x509' had to use a serial number in a file which was
7706 *) New options to 'ca' utility to support V2 CRL entry extensions.
7707 Currently CRL reason, invalidity date and hold instruction are
7708 supported. Add new CRL extensions to V3 code and some new objects.
7711 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7712 disable standard block padding (aka PKCS#5 padding) in the EVP
7713 API, which was previously mandatory. This means that the data is
7714 not padded in any way and so the total length much be a multiple
7715 of the block size, otherwise an error occurs.
7718 *) Initial (incomplete) OCSP SSL support.
7721 *) New function OCSP_parse_url(). This splits up a URL into its host,
7722 port and path components: primarily to parse OCSP URLs. New -url
7723 option to ocsp utility.
7726 *) New nonce behavior. The return value of OCSP_check_nonce() now
7727 reflects the various checks performed. Applications can decide
7728 whether to tolerate certain situations such as an absent nonce
7729 in a response when one was present in a request: the ocsp application
7730 just prints out a warning. New function OCSP_add1_basic_nonce()
7731 this is to allow responders to include a nonce in a response even if
7732 the request is nonce-less.
7735 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7736 skipped when using openssl x509 multiple times on a single input file,
7737 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7740 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7741 set string type: to handle setting ASN1_TIME structures. Fix ca
7742 utility to correctly initialize revocation date of CRLs.
7745 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7746 the clients preferred ciphersuites and rather use its own preferences.
7747 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7748 Internet Explorer by ensuring unchanged hash method during stepup.
7749 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7752 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7753 to aes and add a new 'exist' option to print out symbols that don't
7757 *) Additional options to ocsp utility to allow flags to be set and
7758 additional certificates supplied.
7761 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7762 OCSP client a number of certificate to only verify the response
7766 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7767 handle the new API. Currently only ECB, CBC modes supported. Add new
7770 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7771 Encryption Standard (AES) Ciphersuites for Transport Layer
7772 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7773 not enabled by default and were not part of the "ALL" ciphersuite
7774 alias because they were not yet official; they could be
7775 explicitly requested by specifying the "AESdraft" ciphersuite
7776 group alias. In the final release of OpenSSL 0.9.7, the group
7777 alias is called "AES" and is part of "ALL".)
7778 [Ben Laurie, Steve Henson, Bodo Moeller]
7780 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7781 request to response.
7784 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7785 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7786 extract information from a certificate request. OCSP_response_create()
7787 creates a response and optionally adds a basic response structure.
7788 OCSP_basic_add1_status() adds a complete single response to a basic
7789 response and returns the OCSP_SINGLERESP structure just added (to allow
7790 extensions to be included for example). OCSP_basic_add1_cert() adds a
7791 certificate to a basic response and OCSP_basic_sign() signs a basic
7792 response with various flags. New helper functions ASN1_TIME_check()
7793 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7794 (converts ASN1_TIME to GeneralizedTime).
7797 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7798 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7799 structure from a certificate. X509_pubkey_digest() digests the public_key
7800 contents: this is used in various key identifiers.
7803 *) Make sk_sort() tolerate a NULL argument.
7804 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7806 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7807 passed by the function are trusted implicitly. If any of them signed the
7808 response then it is assumed to be valid and is not verified.
7811 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7812 to data. This was previously part of the PKCS7 ASN1 code. This
7813 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7814 [Steve Henson, reported by Kenneth R. Robinette
7815 <support@securenetterm.com>]
7817 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7818 routines: without these tracing memory leaks is very painful.
7819 Fix leaks in PKCS12 and PKCS7 routines.
7822 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7823 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7824 effectively meant GeneralizedTime would never be used. Now it
7825 is initialised to -1 but X509_time_adj() now has to check the value
7826 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7827 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7828 [Steve Henson, reported by Kenneth R. Robinette
7829 <support@securenetterm.com>]
7831 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7832 result in a zero length in the ASN1_INTEGER structure which was
7833 not consistent with the structure when d2i_ASN1_INTEGER() was used
7834 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7835 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7836 where it did not print out a minus for negative ASN1_INTEGER.
7839 *) Add summary printout to ocsp utility. The various functions which
7840 convert status values to strings have been renamed to:
7841 OCSP_response_status_str(), OCSP_cert_status_str() and
7842 OCSP_crl_reason_str() and are no longer static. New options
7843 to verify nonce values and to disable verification. OCSP response
7844 printout format cleaned up.
7847 *) Add additional OCSP certificate checks. These are those specified
7848 in RFC2560. This consists of two separate checks: the CA of the
7849 certificate being checked must either be the OCSP signer certificate
7850 or the issuer of the OCSP signer certificate. In the latter case the
7851 OCSP signer certificate must contain the OCSP signing extended key
7852 usage. This check is performed by attempting to match the OCSP
7853 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7854 in the OCSP_CERTID structures of the response.
7857 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7858 and related routines. This uses the standard OpenSSL certificate
7859 verify routines to perform initial checks (just CA validity) and
7860 to obtain the certificate chain. Then additional checks will be
7861 performed on the chain. Currently the root CA is checked to see
7862 if it is explicitly trusted for OCSP signing. This is used to set
7863 a root CA as a global signing root: that is any certificate that
7864 chains to that CA is an acceptable OCSP signing certificate.
7867 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7868 extensions from a separate configuration file.
7869 As when reading extensions from the main configuration file,
7870 the '-extensions ...' option may be used for specifying the
7872 [Massimiliano Pala <madwolf@comune.modena.it>]
7874 *) New OCSP utility. Allows OCSP requests to be generated or
7875 read. The request can be sent to a responder and the output
7876 parsed, outputed or printed in text form. Not complete yet:
7877 still needs to check the OCSP response validity.
7880 *) New subcommands for 'openssl ca':
7881 'openssl ca -status <serial>' prints the status of the cert with
7882 the given serial number (according to the index file).
7883 'openssl ca -updatedb' updates the expiry status of certificates
7885 [Massimiliano Pala <madwolf@comune.modena.it>]
7887 *) New '-newreq-nodes' command option to CA.pl. This is like
7888 '-newreq', but calls 'openssl req' with the '-nodes' option
7889 so that the resulting key is not encrypted.
7890 [Damien Miller <djm@mindrot.org>]
7892 *) New configuration for the GNU Hurd.
7893 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7895 *) Initial code to implement OCSP basic response verify. This
7896 is currently incomplete. Currently just finds the signer's
7897 certificate and verifies the signature on the response.
7900 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7901 value of OPENSSLDIR. This is available via the new '-d' option
7902 to 'openssl version', and is also included in 'openssl version -a'.
7905 *) Allowing defining memory allocation callbacks that will be given
7906 file name and line number information in additional arguments
7907 (a const char* and an int). The basic functionality remains, as
7908 well as the original possibility to just replace malloc(),
7909 realloc() and free() by functions that do not know about these
7910 additional arguments. To register and find out the current
7911 settings for extended allocation functions, the following
7912 functions are provided:
7914 CRYPTO_set_mem_ex_functions
7915 CRYPTO_set_locked_mem_ex_functions
7916 CRYPTO_get_mem_ex_functions
7917 CRYPTO_get_locked_mem_ex_functions
7919 These work the same way as CRYPTO_set_mem_functions and friends.
7920 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7921 extended allocation function is enabled.
7922 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7923 a conventional allocation function is enabled.
7924 [Richard Levitte, Bodo Moeller]
7926 *) Finish off removing the remaining LHASH function pointer casts.
7927 There should no longer be any prototype-casting required when using
7928 the LHASH abstraction, and any casts that remain are "bugs". See
7929 the callback types and macros at the head of lhash.h for details
7930 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7933 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7934 If /dev/[u]random devices are not available or do not return enough
7935 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7937 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7938 /etc/entropy will be queried once each in this sequence, querying stops
7939 when enough entropy was collected without querying more sockets.
7942 *) Change the Unix RAND_poll() variant to be able to poll several
7943 random devices, as specified by DEVRANDOM, until a sufficient amount
7944 of data has been collected. We spend at most 10 ms on each file
7945 (select timeout) and read in non-blocking mode. DEVRANDOM now
7946 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7947 (previously it was just the string "/dev/urandom"), so on typical
7948 platforms the 10 ms delay will never occur.
7949 Also separate out the Unix variant to its own file, rand_unix.c.
7950 For VMS, there's a currently-empty rand_vms.c.
7953 *) Move OCSP client related routines to ocsp_cl.c. These
7954 provide utility functions which an application needing
7955 to issue a request to an OCSP responder and analyse the
7956 response will typically need: as opposed to those which an
7957 OCSP responder itself would need which will be added later.
7959 OCSP_request_sign() signs an OCSP request with an API similar
7960 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7961 response. OCSP_response_get1_basic() extracts basic response
7962 from response. OCSP_resp_find_status(): finds and extracts status
7963 information from an OCSP_CERTID structure (which will be created
7964 when the request structure is built). These are built from lower
7965 level functions which work on OCSP_SINGLERESP structures but
7966 won't normally be used unless the application wishes to examine
7967 extensions in the OCSP response for example.
7969 Replace nonce routines with a pair of functions.
7970 OCSP_request_add1_nonce() adds a nonce value and optionally
7971 generates a random value. OCSP_check_nonce() checks the
7972 validity of the nonce in an OCSP response.
7975 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7976 This doesn't copy the supplied OCSP_CERTID and avoids the
7977 need to free up the newly created id. Change return type
7978 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7979 This can then be used to add extensions to the request.
7980 Deleted OCSP_request_new(), since most of its functionality
7981 is now in OCSP_REQUEST_new() (and the case insensitive name
7982 clash) apart from the ability to set the request name which
7983 will be added elsewhere.
7986 *) Update OCSP API. Remove obsolete extensions argument from
7987 various functions. Extensions are now handled using the new
7988 OCSP extension code. New simple OCSP HTTP function which
7989 can be used to send requests and parse the response.
7992 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7993 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7994 uses the special reorder version of SET OF to sort the attributes
7995 and reorder them to match the encoded order. This resolves a long
7996 standing problem: a verify on a PKCS7 structure just after signing
7997 it used to fail because the attribute order did not match the
7998 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7999 it uses the received order. This is necessary to tolerate some broken
8000 software that does not order SET OF. This is handled by encoding
8001 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8002 to produce the required SET OF.
8005 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8006 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8007 files to get correct declarations of the ASN.1 item variables.
8010 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8011 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8012 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8013 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8014 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8015 ASN1_ITEM and no wrapper functions.
8018 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8019 replace the old function pointer based I/O routines. Change most of
8020 the *_d2i_bio() and *_d2i_fp() functions to use these.
8023 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8024 lines, recognize more "algorithms" that can be deselected, and make
8025 it complain about algorithm deselection that isn't recognised.
8028 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8029 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8030 to use new functions. Add NO_ASN1_OLD which can be set to remove
8031 some old style ASN1 functions: this can be used to determine if old
8032 code will still work when these eventually go away.
8035 *) New extension functions for OCSP structures, these follow the
8036 same conventions as certificates and CRLs.
8039 *) New function X509V3_add1_i2d(). This automatically encodes and
8040 adds an extension. Its behaviour can be customised with various
8041 flags to append, replace or delete. Various wrappers added for
8042 certificates and CRLs.
8045 *) Fix to avoid calling the underlying ASN1 print routine when
8046 an extension cannot be parsed. Correct a typo in the
8047 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8050 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8051 entries for variables.
8054 *) Add functionality to apps/openssl.c for detecting locking
8055 problems: As the program is single-threaded, all we have
8056 to do is register a locking callback using an array for
8057 storing which locks are currently held by the program.
8060 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8061 SSL_get_ex_data_X509_STORE_idx(), which is used in
8062 ssl_verify_cert_chain() and thus can be called at any time
8063 during TLS/SSL handshakes so that thread-safety is essential.
8064 Unfortunately, the ex_data design is not at all suited
8065 for multi-threaded use, so it probably should be abolished.
8068 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8069 [Broadcom, tweaked and integrated by Geoff Thorpe]
8071 *) Move common extension printing code to new function
8072 X509V3_print_extensions(). Reorganise OCSP print routines and
8073 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8076 *) New function X509_signature_print() to remove duplication in some
8080 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8081 set (this was treated exactly the same as SET OF previously). This
8082 is used to reorder the STACK representing the structure to match the
8083 encoding. This will be used to get round a problem where a PKCS7
8084 structure which was signed could not be verified because the STACK
8085 order did not reflect the encoded order.
8088 *) Reimplement the OCSP ASN1 module using the new code.
8091 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8092 for its ASN1 operations. The old style function pointers still exist
8093 for now but they will eventually go away.
8096 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8097 completely replaces the old ASN1 functionality with a table driven
8098 encoder and decoder which interprets an ASN1_ITEM structure describing
8099 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8100 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8101 has also been converted to the new form.
8104 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8105 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8106 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8107 for negative moduli.
8110 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8111 of not touching the result's sign bit.
8114 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8118 *) Changed the LHASH code to use prototypes for callbacks, and created
8119 macros to declare and implement thin (optionally static) functions
8120 that provide type-safety and avoid function pointer casting for the
8121 type-specific callbacks.
8124 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8126 [Veers Staats <staatsvr@asc.hpc.mil>,
8127 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8129 *) Reformat the FAQ so the different questions and answers can be divided
8130 in sections depending on the subject.
8133 *) Have the zlib compression code load ZLIB.DLL dynamically under
8137 *) New function BN_mod_sqrt for computing square roots modulo a prime
8138 (using the probabilistic Tonelli-Shanks algorithm unless
8139 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8140 be handled deterministically).
8141 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8143 *) Make BN_mod_inverse faster by explicitly handling small quotients
8144 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8145 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8148 *) New function BN_kronecker.
8151 *) Fix BN_gcd so that it works on negative inputs; the result is
8152 positive unless both parameters are zero.
8153 Previously something reasonably close to an infinite loop was
8154 possible because numbers could be growing instead of shrinking
8155 in the implementation of Euclid's algorithm.
8158 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8159 sign of the number in question.
8161 Fix BN_is_word(a,w) to work correctly for w == 0.
8163 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8164 because its test if the absolute value of 'a' equals 'w'.
8165 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8166 it exists mostly for use in the implementations of BN_is_zero(),
8167 BN_is_one(), and BN_is_word().
8170 *) New function BN_swap.
8173 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8174 the exponentiation functions are more likely to produce reasonable
8175 results on negative inputs.
8178 *) Change BN_mod_mul so that the result is always non-negative.
8179 Previously, it could be negative if one of the factors was negative;
8180 I don't think anyone really wanted that behaviour.
8183 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8184 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8185 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8186 and add new functions:
8195 BN_mod_lshift1_quick
8199 These functions always generate non-negative results.
8201 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8202 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8204 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8205 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8206 be reduced modulo m.
8207 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8210 The following entry accidentally appeared in the CHANGES file
8211 distributed with OpenSSL 0.9.7. The modifications described in
8212 it do *not* apply to OpenSSL 0.9.7.
8214 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8215 was actually never needed) and in BN_mul(). The removal in BN_mul()
8216 required a small change in bn_mul_part_recursive() and the addition
8217 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8218 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8219 bn_sub_words() and bn_add_words() except they take arrays with
8224 *) In 'openssl passwd', verify passwords read from the terminal
8225 unless the '-salt' option is used (which usually means that
8226 verification would just waste user's time since the resulting
8227 hash is going to be compared with some given password hash)
8228 or the new '-noverify' option is used.
8230 This is an incompatible change, but it does not affect
8231 non-interactive use of 'openssl passwd' (passwords on the command
8232 line, '-stdin' option, '-in ...' option) and thus should not
8236 *) Remove all references to RSAref, since there's no more need for it.
8239 *) Make DSO load along a path given through an environment variable
8240 (SHLIB_PATH) with shl_load().
8243 *) Constify the ENGINE code as a result of BIGNUM constification.
8244 Also constify the RSA code and most things related to it. In a
8245 few places, most notable in the depth of the ASN.1 code, ugly
8246 casts back to non-const were required (to be solved at a later
8250 *) Make it so the openssl application has all engines loaded by default.
8253 *) Constify the BIGNUM routines a little more.
8256 *) Add the following functions:
8258 ENGINE_load_cswift()
8260 ENGINE_load_atalla()
8262 ENGINE_load_builtin_engines()
8264 That way, an application can itself choose if external engines that
8265 are built-in in OpenSSL shall ever be used or not. The benefit is
8266 that applications won't have to be linked with libdl or other dso
8267 libraries unless it's really needed.
8269 Changed 'openssl engine' to load all engines on demand.
8270 Changed the engine header files to avoid the duplication of some
8271 declarations (they differed!).
8274 *) 'openssl engine' can now list capabilities.
8277 *) Better error reporting in 'openssl engine'.
8280 *) Never call load_dh_param(NULL) in s_server.
8283 *) Add engine application. It can currently list engines by name and
8284 identity, and test if they are actually available.
8287 *) Improve RPM specification file by forcing symbolic linking and making
8288 sure the installed documentation is also owned by root.root.
8289 [Damien Miller <djm@mindrot.org>]
8291 *) Give the OpenSSL applications more possibilities to make use of
8292 keys (public as well as private) handled by engines.
8295 *) Add OCSP code that comes from CertCo.
8298 *) Add VMS support for the Rijndael code.
8301 *) Added untested support for Nuron crypto accelerator.
8304 *) Add support for external cryptographic devices. This code was
8305 previously distributed separately as the "engine" branch.
8306 [Geoff Thorpe, Richard Levitte]
8308 *) Rework the filename-translation in the DSO code. It is now possible to
8309 have far greater control over how a "name" is turned into a filename
8310 depending on the operating environment and any oddities about the
8311 different shared library filenames on each system.
8314 *) Support threads on FreeBSD-elf in Configure.
8317 *) Fix for SHA1 assembly problem with MASM: it produces
8318 warnings about corrupt line number information when assembling
8319 with debugging information. This is caused by the overlapping
8321 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8324 NCONF_get_number() has no error checking at all. As a replacement,
8325 NCONF_get_number_e() is defined (_e for "error checking") and is
8326 promoted strongly. The old NCONF_get_number is kept around for
8327 binary backward compatibility.
8328 Make it possible for methods to load from something other than a BIO,
8329 by providing a function pointer that is given a name instead of a BIO.
8330 For example, this could be used to load configuration data from an
8334 *) Fix for non blocking accept BIOs. Added new I/O special reason
8335 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8336 with non blocking I/O was not possible because no retry code was
8337 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8341 *) Added the beginnings of Rijndael support.
8344 *) Fix for bug in DirectoryString mask setting. Add support for
8345 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8346 to allow certificate printing to more controllable, additional
8347 'certopt' option to 'x509' to allow new printing options to be
8351 *) Clean old EAY MD5 hack from e_os.h.
8354 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8356 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8357 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8358 [Joe Orton, Steve Henson]
8360 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8362 *) Fix additional bug revealed by the NISCC test suite:
8364 Stop bug triggering large recursion when presented with
8365 certain ASN.1 tags (CVE-2003-0851)
8368 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8370 *) Fix various bugs revealed by running the NISCC test suite:
8372 Stop out of bounds reads in the ASN1 code when presented with
8373 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8375 If verify callback ignores invalid public key errors don't try to check
8376 certificate signature with the NULL public key.
8380 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8381 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8385 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8386 extra data after the compression methods not only for TLS 1.0
8387 but also for SSL 3.0 (as required by the specification).
8388 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8390 *) Change X509_certificate_type() to mark the key as exported/exportable
8391 when it's 512 *bits* long, not 512 bytes.
8394 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8396 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8397 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8398 a protocol version number mismatch like a decryption error
8399 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8402 *) Turn on RSA blinding by default in the default implementation
8403 to avoid a timing attack. Applications that don't want it can call
8404 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8405 They would be ill-advised to do so in most cases.
8406 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8408 *) Change RSA blinding code so that it works when the PRNG is not
8409 seeded (in this case, the secret RSA exponent is abused as
8410 an unpredictable seed -- if it is not unpredictable, there
8411 is no point in blinding anyway). Make RSA blinding thread-safe
8412 by remembering the creator's thread ID in rsa->blinding and
8413 having all other threads use local one-time blinding factors
8414 (this requires more computation than sharing rsa->blinding, but
8415 avoids excessive locking; and if an RSA object is not shared
8416 between threads, blinding will still be very fast).
8419 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8421 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8422 via timing by performing a MAC computation even if incorrect
8423 block cipher padding has been found. This is a countermeasure
8424 against active attacks where the attacker has to distinguish
8425 between bad padding and a MAC verification error. (CVE-2003-0078)
8427 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8428 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8429 Martin Vuagnoux (EPFL, Ilion)]
8431 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8433 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8434 memory from it's contents. This is done with a counter that will
8435 place alternating values in each byte. This can be used to solve
8436 two issues: 1) the removal of calls to memset() by highly optimizing
8437 compilers, and 2) cleansing with other values than 0, since those can
8438 be read through on certain media, for example a swap space on disk.
8441 *) Bugfix: client side session caching did not work with external caching,
8442 because the session->cipher setting was not restored when reloading
8443 from the external cache. This problem was masked, when
8444 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8445 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8448 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8449 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8450 [Zeev Lieber <zeev-l@yahoo.com>]
8452 *) Undo an undocumented change introduced in 0.9.6e which caused
8453 repeated calls to OpenSSL_add_all_ciphers() and
8454 OpenSSL_add_all_digests() to be ignored, even after calling
8458 *) Change the default configuration reader to deal with last line not
8459 being properly terminated.
8462 *) Change X509_NAME_cmp() so it applies the special rules on handling
8463 DN values that are of type PrintableString, as well as RDNs of type
8464 emailAddress where the value has the type ia5String.
8465 [stefank@valicert.com via Richard Levitte]
8467 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8468 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8469 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8470 the bitwise-OR of the two for use by the majority of applications
8471 wanting this behaviour, and update the docs. The documented
8472 behaviour and actual behaviour were inconsistent and had been
8473 changing anyway, so this is more a bug-fix than a behavioural
8475 [Geoff Thorpe, diagnosed by Nadav Har'El]
8477 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8478 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8481 *) Fix initialization code race conditions in
8482 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8483 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8484 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8485 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8486 ssl2_get_cipher_by_char(),
8487 ssl3_get_cipher_by_char().
8488 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8490 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8491 the cached sessions are flushed, as the remove_cb() might use ex_data
8492 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8493 (see [openssl.org #212]).
8494 [Geoff Thorpe, Lutz Jaenicke]
8496 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8497 length, instead of the encoding length to d2i_ASN1_OBJECT.
8500 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8502 *) [In 0.9.6g-engine release:]
8503 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8504 [Lynn Gazis <lgazis@rainbow.com>]
8506 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8508 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8509 and get fix the header length calculation.
8510 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8511 Alon Kantor <alonk@checkpoint.com> (and others),
8514 *) Use proper error handling instead of 'assertions' in buffer
8515 overflow checks added in 0.9.6e. This prevents DoS (the
8516 assertions could call abort()).
8517 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8519 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8521 *) Add various sanity checks to asn1_get_length() to reject
8522 the ASN1 length bytes if they exceed sizeof(long), will appear
8523 negative or the content length exceeds the length of the
8525 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8527 *) Fix cipher selection routines: ciphers without encryption had no flags
8528 for the cipher strength set and where therefore not handled correctly
8529 by the selection routines (PR #130).
8532 *) Fix EVP_dsa_sha macro.
8536 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8537 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8538 that was added in OpenSSL 0.9.6d.
8540 As the countermeasure turned out to be incompatible with some
8541 broken SSL implementations, the new option is part of SSL_OP_ALL.
8542 SSL_OP_ALL is usually employed when compatibility with weird SSL
8543 implementations is desired (e.g. '-bugs' option to 's_client' and
8544 's_server'), so the new option is automatically set in many
8548 *) Changes in security patch:
8550 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8551 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8552 Air Force Materiel Command, USAF, under agreement number
8555 *) Add various sanity checks to asn1_get_length() to reject
8556 the ASN1 length bytes if they exceed sizeof(long), will appear
8557 negative or the content length exceeds the length of the
8558 supplied buffer. (CVE-2002-0659)
8559 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8561 *) Assertions for various potential buffer overflows, not known to
8563 [Ben Laurie (CHATS)]
8565 *) Various temporary buffers to hold ASCII versions of integers were
8566 too small for 64 bit platforms. (CVE-2002-0655)
8567 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8569 *) Remote buffer overflow in SSL3 protocol - an attacker could
8570 supply an oversized session ID to a client. (CVE-2002-0656)
8571 [Ben Laurie (CHATS)]
8573 *) Remote buffer overflow in SSL2 protocol - an attacker could
8574 supply an oversized client master key. (CVE-2002-0656)
8575 [Ben Laurie (CHATS)]
8577 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8579 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8580 encoded as NULL) with id-dsa-with-sha1.
8581 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8583 *) Check various X509_...() return values in apps/req.c.
8584 [Nils Larsch <nla@trustcenter.de>]
8586 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8587 an end-of-file condition would erroneously be flagged, when the CRLF
8588 was just at the end of a processed block. The bug was discovered when
8589 processing data through a buffering memory BIO handing the data to a
8590 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8591 <ptsekov@syntrex.com> and Nedelcho Stanev.
8594 *) Implement a countermeasure against a vulnerability recently found
8595 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8596 before application data chunks to avoid the use of known IVs
8597 with data potentially chosen by the attacker.
8600 *) Fix length checks in ssl3_get_client_hello().
8603 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8604 to prevent ssl3_read_internal() from incorrectly assuming that
8605 ssl3_read_bytes() found application data while handshake
8606 processing was enabled when in fact s->s3->in_read_app_data was
8607 merely automatically cleared during the initial handshake.
8608 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8610 *) Fix object definitions for Private and Enterprise: they were not
8611 recognized in their shortname (=lowercase) representation. Extend
8612 obj_dat.pl to issue an error when using undefined keywords instead
8613 of silently ignoring the problem (Svenning Sorensen
8614 <sss@sss.dnsalias.net>).
8617 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8618 generators, i.e. generators other than 2 and 5. (Previously, the
8619 code did not properly initialise the 'add' and 'rem' values to
8620 BN_generate_prime().)
8622 In the new general case, we do not insist that 'generator' is
8623 actually a primitive root: This requirement is rather pointless;
8624 a generator of the order-q subgroup is just as good, if not
8628 *) Map new X509 verification errors to alerts. Discovered and submitted by
8629 Tom Wu <tom@arcot.com>.
8632 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8633 returning non-zero before the data has been completely received
8634 when using non-blocking I/O.
8635 [Bodo Moeller; problem pointed out by John Hughes]
8637 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8638 [Ben Laurie, Lutz Jaenicke]
8640 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8641 Yoram Zahavi <YoramZ@gilian.com>).
8644 *) Add information about CygWin 1.3 and on, and preserve proper
8645 configuration for the versions before that.
8646 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8648 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8649 check whether we deal with a copy of a session and do not delete from
8650 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8651 <izhar@checkpoint.com>.
8654 *) Do not store session data into the internal session cache, if it
8655 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8656 flag is set). Proposed by Aslam <aslam@funk.com>.
8659 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8663 *) [In 0.9.6d-engine release:]
8664 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8665 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8667 *) Add the configuration target linux-s390x.
8668 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8670 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8671 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8672 variable as an indication that a ClientHello message has been
8673 received. As the flag value will be lost between multiple
8674 invocations of ssl3_accept when using non-blocking I/O, the
8675 function may not be aware that a handshake has actually taken
8676 place, thus preventing a new session from being added to the
8679 To avoid this problem, we now set s->new_session to 2 instead of
8680 using a local variable.
8681 [Lutz Jaenicke, Bodo Moeller]
8683 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8684 if the SSL_R_LENGTH_MISMATCH error is detected.
8685 [Geoff Thorpe, Bodo Moeller]
8687 *) New 'shared_ldflag' column in Configure platform table.
8690 *) Fix EVP_CIPHER_mode macro.
8691 ["Dan S. Camper" <dan@bti.net>]
8693 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8694 type, we must throw them away by setting rr->length to 0.
8695 [D P Chang <dpc@qualys.com>]
8697 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8699 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8700 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8701 worked incorrectly for those cases where range = 10..._2 and
8702 3*range is two bits longer than range.)
8705 *) Only add signing time to PKCS7 structures if it is not already
8709 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8710 OBJ_ld_ce should be OBJ_id_ce.
8711 Also some ip-pda OIDs in crypto/objects/objects.txt were
8712 incorrect (cf. RFC 3039).
8713 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8715 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8716 returns early because it has nothing to do.
8717 [Andy Schneider <andy.schneider@bjss.co.uk>]
8719 *) [In 0.9.6c-engine release:]
8720 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8721 [Andy Schneider <andy.schneider@bjss.co.uk>]
8723 *) [In 0.9.6c-engine release:]
8724 Add support for Cryptographic Appliance's keyserver technology.
8725 (Use engine 'keyclient')
8726 [Cryptographic Appliances and Geoff Thorpe]
8728 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8729 is called via tools/c89.sh because arguments have to be
8730 rearranged (all '-L' options must appear before the first object
8732 [Richard Shapiro <rshapiro@abinitio.com>]
8734 *) [In 0.9.6c-engine release:]
8735 Add support for Broadcom crypto accelerator cards, backported
8737 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8739 *) [In 0.9.6c-engine release:]
8740 Add support for SureWare crypto accelerator cards from
8741 Baltimore Technologies. (Use engine 'sureware')
8742 [Baltimore Technologies and Mark Cox]
8744 *) [In 0.9.6c-engine release:]
8745 Add support for crypto accelerator cards from Accelerated
8746 Encryption Processing, www.aep.ie. (Use engine 'aep')
8747 [AEP Inc. and Mark Cox]
8749 *) Add a configuration entry for gcc on UnixWare.
8750 [Gary Benson <gbenson@redhat.com>]
8752 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8753 messages are stored in a single piece (fixed-length part and
8754 variable-length part combined) and fix various bugs found on the way.
8757 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8758 instead. BIO_gethostbyname() does not know what timeouts are
8759 appropriate, so entries would stay in cache even when they have
8761 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8763 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8764 faced with a pathologically small ClientHello fragment that does
8765 not contain client_version: Instead of aborting with an error,
8766 simply choose the highest available protocol version (i.e.,
8767 TLS 1.0 unless it is disabled). In practice, ClientHello
8768 messages are never sent like this, but this change gives us
8769 strictly correct behaviour at least for TLS.
8772 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8773 never resets s->method to s->ctx->method when called from within
8774 one of the SSL handshake functions.
8775 [Bodo Moeller; problem pointed out by Niko Baric]
8777 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8778 (sent using the client's version number) if client_version is
8779 smaller than the protocol version in use. Also change
8780 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8781 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8782 the client will at least see that alert.
8785 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8789 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8790 client receives HelloRequest while in a handshake.
8791 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8793 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8794 should end in 'break', not 'goto end' which circumvents various
8795 cleanups done in state SSL_ST_OK. But session related stuff
8796 must be disabled for SSL_ST_OK in the case that we just sent a
8799 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8800 before just sending a HelloRequest.
8801 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8803 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8804 reveal whether illegal block cipher padding was found or a MAC
8805 verification error occurred. (Neither SSLerr() codes nor alerts
8806 are directly visible to potential attackers, but the information
8807 may leak via logfiles.)
8809 Similar changes are not required for the SSL 2.0 implementation
8810 because the number of padding bytes is sent in clear for SSL 2.0,
8811 and the extra bytes are just ignored. However ssl/s2_pkt.c
8812 failed to verify that the purported number of padding bytes is in
8816 *) Add OpenUNIX-8 support including shared libraries
8817 (Boyd Lynn Gerber <gerberb@zenez.com>).
8820 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8821 'wristwatch attack' using huge encoding parameters (cf.
8822 James H. Manger's CRYPTO 2001 paper). Note that the
8823 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8824 encoding parameters and hence was not vulnerable.
8827 *) BN_sqr() bug fix.
8828 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8830 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8831 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8832 followed by modular reduction.
8833 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8835 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8836 equivalent based on BN_pseudo_rand() instead of BN_rand().
8839 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8840 This function was broken, as the check for a new client hello message
8841 to handle SGC did not allow these large messages.
8842 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8845 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8848 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8849 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8852 *) Rework the configuration and shared library support for Tru64 Unix.
8853 The configuration part makes use of modern compiler features and
8854 still retains old compiler behavior for those that run older versions
8855 of the OS. The shared library support part includes a variant that
8856 uses the RPATH feature, and is available through the special
8857 configuration target "alpha-cc-rpath", which will never be selected
8859 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8861 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8862 with the same message size as in ssl3_get_certificate_request().
8863 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8864 messages might inadvertently be reject as too long.
8865 [Petr Lampa <lampa@fee.vutbr.cz>]
8867 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8870 *) Modified SSL library such that the verify_callback that has been set
8871 specificly for an SSL object with SSL_set_verify() is actually being
8872 used. Before the change, a verify_callback set with this function was
8873 ignored and the verify_callback() set in the SSL_CTX at the time of
8874 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8875 to allow the necessary settings.
8878 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8879 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8880 done automatically (in contradiction to the requirements of the C
8881 standard). This made problems when used from OpenSSH.
8884 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8885 dh->length and always used
8887 BN_rand_range(priv_key, dh->p).
8889 BN_rand_range() is not necessary for Diffie-Hellman, and this
8890 specific range makes Diffie-Hellman unnecessarily inefficient if
8891 dh->length (recommended exponent length) is much smaller than the
8892 length of dh->p. We could use BN_rand_range() if the order of
8893 the subgroup was stored in the DH structure, but we only have
8898 BN_rand(priv_key, l, ...)
8900 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8906 RSA_eay_public_encrypt
8907 RSA_eay_private_decrypt
8908 RSA_eay_private_encrypt (signing)
8909 RSA_eay_public_decrypt (signature verification)
8911 (default implementations for RSA_public_encrypt,
8912 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8913 always reject numbers >= n.
8916 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8917 to synchronize access to 'locking_thread'. This is necessary on
8918 systems where access to 'locking_thread' (an 'unsigned long'
8919 variable) is not atomic.
8922 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8923 *before* setting the 'crypto_lock_rand' flag. The previous code had
8924 a race condition if 0 is a valid thread ID.
8925 [Travis Vitek <vitek@roguewave.com>]
8927 *) Add support for shared libraries under Irix.
8928 [Albert Chin-A-Young <china@thewrittenword.com>]
8930 *) Add configuration option to build on Linux on both big-endian and
8932 [Ralf Baechle <ralf@uni-koblenz.de>]
8934 *) Add the possibility to create shared libraries on HP-UX.
8937 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8939 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8940 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8941 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8942 PRNG state recovery was possible based on the output of
8943 one PRNG request appropriately sized to gain knowledge on
8944 'md' followed by enough consecutive 1-byte PRNG requests
8945 to traverse all of 'state'.
8947 1. When updating 'md_local' (the current thread's copy of 'md')
8948 during PRNG output generation, hash all of the previous
8949 'md_local' value, not just the half used for PRNG output.
8951 2. Make the number of bytes from 'state' included into the hash
8952 independent from the number of PRNG bytes requested.
8954 The first measure alone would be sufficient to avoid
8955 Markku-Juhani's attack. (Actually it had never occurred
8956 to me that the half of 'md_local' used for chaining was the
8957 half from which PRNG output bytes were taken -- I had always
8958 assumed that the secret half would be used.) The second
8959 measure makes sure that additional data from 'state' is never
8960 mixed into 'md_local' in small portions; this heuristically
8961 further strengthens the PRNG.
8964 *) Fix crypto/bn/asm/mips3.s.
8967 *) When only the key is given to "enc", the IV is undefined. Print out
8968 an error message in this case.
8971 *) Handle special case when X509_NAME is empty in X509 printing routines.
8974 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8975 positive and less than q.
8978 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8979 used: it isn't thread safe and the add_lock_callback should handle
8981 [Paul Rose <Paul.Rose@bridge.com>]
8983 *) Verify that incoming data obeys the block size in
8984 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8988 [Ulf Möller, Bodo Möller]
8990 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8991 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8992 when fixing the server behaviour for backwards-compatible 'client
8993 hello' messages. (Note that the attack is impractical against
8994 SSL 3.0 and TLS 1.0 anyway because length and version checking
8995 means that the probability of guessing a valid ciphertext is
8996 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8999 Before 0.9.5, the countermeasure (hide the error by generating a
9000 random 'decryption result') did not work properly because
9001 ERR_clear_error() was missing, meaning that SSL_get_error() would
9002 detect the supposedly ignored error.
9004 Both problems are now fixed.
9007 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9008 (previously it was 1024).
9011 *) Fix for compatibility mode trust settings: ignore trust settings
9012 unless some valid trust or reject settings are present.
9015 *) Fix for blowfish EVP: its a variable length cipher.
9018 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9019 parameters in DSA public key structures and return an error in the
9020 DSA routines if parameters are absent.
9023 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9024 in the current directory if neither $RANDFILE nor $HOME was set.
9025 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9026 caused some confusion to Windows users who haven't defined $HOME.
9027 Thus RAND_file_name() is changed again: e_os.h can define a
9028 DEFAULT_HOME, which will be used if $HOME is not set.
9029 For Windows, we use "C:"; on other platforms, we still require
9030 environment variables.
9032 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9033 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9034 having multiple threads call RAND_poll() concurrently.
9037 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9038 combination of a flag and a thread ID variable.
9039 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9040 flag), *other* threads can enter ssleay_add_bytes without obeying
9041 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9042 that they do not hold after the first thread unsets add_do_not_lock).
9045 *) Change bctest again: '-x' expressions are not available in all
9049 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9051 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9052 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9054 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9055 the default extension for executables, if any. Also, make the perl
9056 scripts that use symlink() to test if it really exists and use "cp"
9057 if it doesn't. All this made OpenSSL compilable and installable in
9061 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9062 If SEQUENCE is length is indefinite just set c->slen to the total
9063 amount of data available.
9064 [Steve Henson, reported by shige@FreeBSD.org]
9065 [This change does not apply to 0.9.7.]
9067 *) Change bctest to avoid here-documents inside command substitution
9068 (workaround for FreeBSD /bin/sh bug).
9069 For compatibility with Ultrix, avoid shell functions (introduced
9070 in the bctest version that searches along $PATH).
9073 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9074 with des_encrypt() defined on some operating systems, like Solaris
9078 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9079 On the Importance of Eliminating Errors in Cryptographic
9080 Computations, J. Cryptology 14 (2001) 2, 101-119,
9081 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9084 *) MIPS assembler BIGNUM division bug fix.
9087 *) Disabled incorrect Alpha assembler code.
9090 *) Fix PKCS#7 decode routines so they correctly update the length
9091 after reading an EOC for the EXPLICIT tag.
9093 [This change does not apply to 0.9.7.]
9095 *) Fix bug in PKCS#12 key generation routines. This was triggered
9096 if a 3DES key was generated with a 0 initial byte. Include
9097 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9098 (but broken) behaviour.
9101 *) Enhance bctest to search for a working bc along $PATH and print
9103 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9105 *) Fix memory leaks in err.c: free err_data string if necessary;
9106 don't write to the wrong index in ERR_set_error_data.
9109 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9113 *) Replace rdtsc with _emit statements for VC++ version 5.
9114 [Jeremy Cooper <jeremy@baymoo.org>]
9116 *) Make it possible to reuse SSLv2 sessions.
9119 *) In copy_email() check for >= 0 as a return value for
9120 X509_NAME_get_index_by_NID() since 0 is a valid index.
9121 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9123 *) Avoid coredump with unsupported or invalid public keys by checking if
9124 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9125 PKCS7_verify() fails with non detached data.
9128 *) Don't use getenv in library functions when run as setuid/setgid.
9129 New function OPENSSL_issetugid().
9132 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9133 due to incorrect handling of multi-threading:
9135 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9137 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9139 3. Count how many times MemCheck_off() has been called so that
9140 nested use can be treated correctly. This also avoids
9141 inband-signalling in the previous code (which relied on the
9142 assumption that thread ID 0 is impossible).
9145 *) Add "-rand" option also to s_client and s_server.
9148 *) Fix CPU detection on Irix 6.x.
9149 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9150 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9152 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9155 [This change does not apply to 0.9.7.]
9157 *) Use the cached encoding of an X509_NAME structure rather than
9158 copying it. This is apparently the reason for the libsafe "errors"
9159 but the code is actually correct.
9162 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9163 Bleichenbacher's DSA attack.
9164 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9165 to be set and top=0 forces the highest bit to be set; top=-1 is new
9166 and leaves the highest bit random.
9167 [Ulf Moeller, Bodo Moeller]
9169 *) In the NCONF_...-based implementations for CONF_... queries
9170 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9171 a temporary CONF structure with the data component set to NULL
9172 (which gives segmentation faults in lh_retrieve).
9173 Instead, use NULL for the CONF pointer in CONF_get_string and
9174 CONF_get_number (which may use environment variables) and directly
9175 return NULL from CONF_get_section.
9178 *) Fix potential buffer overrun for EBCDIC.
9181 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9182 keyUsage if basicConstraints absent for a CA.
9185 *) Make SMIME_write_PKCS7() write mail header values with a format that
9186 is more generally accepted (no spaces before the semicolon), since
9187 some programs can't parse those values properly otherwise. Also make
9188 sure BIO's that break lines after each write do not create invalid
9192 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9193 macros previously used would not encode an empty SEQUENCE OF
9194 and break the signature.
9196 [This change does not apply to 0.9.7.]
9198 *) Zero the premaster secret after deriving the master secret in
9202 *) Add some EVP_add_digest_alias registrations (as found in
9203 OpenSSL_add_all_digests()) to SSL_library_init()
9204 aka OpenSSL_add_ssl_algorithms(). This provides improved
9205 compatibility with peers using X.509 certificates
9206 with unconventional AlgorithmIdentifier OIDs.
9209 *) Fix for Irix with NO_ASM.
9210 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9212 *) ./config script fixes.
9213 [Ulf Moeller, Richard Levitte]
9215 *) Fix 'openssl passwd -1'.
9218 *) Change PKCS12_key_gen_asc() so it can cope with non null
9219 terminated strings whose length is passed in the passlen
9220 parameter, for example from PEM callbacks. This was done
9221 by adding an extra length parameter to asc2uni().
9222 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9224 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9225 call failed, free the DSA structure.
9228 *) Fix to uni2asc() to cope with zero length Unicode strings.
9229 These are present in some PKCS#12 files.
9232 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9233 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9234 when writing a 32767 byte record.
9235 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9237 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9238 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9240 (RSA objects have a reference count access to which is protected
9241 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9242 so they are meant to be shared between threads.)
9243 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9244 "Reddie, Steven" <Steven.Reddie@ca.com>]
9246 *) Fix a deadlock in CRYPTO_mem_leaks().
9249 *) Use better test patterns in bntest.
9252 *) rand_win.c fix for Borland C.
9255 *) BN_rshift bugfix for n == 0.
9258 *) Add a 'bctest' script that checks for some known 'bc' bugs
9259 so that 'make test' does not abort just because 'bc' is broken.
9262 *) Store verify_result within SSL_SESSION also for client side to
9263 avoid potential security hole. (Re-used sessions on the client side
9264 always resulted in verify_result==X509_V_OK, not using the original
9265 result of the server certificate verification.)
9268 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9269 SSL3_RT_APPLICATION_DATA, return 0.
9270 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9274 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9275 releases, have been re-implemented by renaming the previous
9276 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9277 and ssl3_read_internal, respectively, and adding 'peek' parameters
9278 to them. The new ssl[23]_{read,peek} functions are calls to
9279 ssl[23]_read_internal with the 'peek' flag set appropriately.
9280 A 'peek' parameter has also been added to ssl3_read_bytes, which
9281 does the actual work for ssl3_read_internal.
9284 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9285 the method-specific "init()" handler. Also clean up ex_data after
9286 calling the method-specific "finish()" handler. Previously, this was
9287 happening the other way round.
9290 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9291 The previous value, 12, was not always sufficient for BN_mod_exp().
9294 *) Make sure that shared libraries get the internal name engine with
9295 the full version number and not just 0. This should mark the
9296 shared libraries as not backward compatible. Of course, this should
9297 be changed again when we can guarantee backward binary compatibility.
9300 *) Fix typo in get_cert_by_subject() in by_dir.c
9301 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9303 *) Rework the system to generate shared libraries:
9305 - Make note of the expected extension for the shared libraries and
9306 if there is a need for symbolic links from for example libcrypto.so.0
9307 to libcrypto.so.0.9.7. There is extended info in Configure for
9310 - Make as few rebuilds of the shared libraries as possible.
9312 - Still avoid linking the OpenSSL programs with the shared libraries.
9314 - When installing, install the shared libraries separately from the
9318 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9320 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9321 and not in SSL_clear because the latter is also used by the
9322 accept/connect functions; previously, the settings made by
9323 SSL_set_read_ahead would be lost during the handshake.
9324 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9326 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9327 Previously, it would create entries for disabled algorithms no
9331 *) Added several new manual pages for SSL_* function.
9334 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9336 *) In ssl23_get_client_hello, generate an error message when faced
9337 with an initial SSL 3.0/TLS record that is too small to contain the
9338 first two bytes of the ClientHello message, i.e. client_version.
9339 (Note that this is a pathologic case that probably has never happened
9340 in real life.) The previous approach was to use the version number
9341 from the record header as a substitute; but our protocol choice
9342 should not depend on that one because it is not authenticated
9343 by the Finished messages.
9346 *) More robust randomness gathering functions for Windows.
9347 [Jeffrey Altman <jaltman@columbia.edu>]
9349 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9350 not set then we don't setup the error code for issuer check errors
9351 to avoid possibly overwriting other errors which the callback does
9352 handle. If an application does set the flag then we assume it knows
9353 what it is doing and can handle the new informational codes
9357 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9358 a general "ANY" type, as such it should be able to decode anything
9359 including tagged types. However it didn't check the class so it would
9360 wrongly interpret tagged types in the same way as their universal
9361 counterpart and unknown types were just rejected. Changed so that the
9362 tagged and unknown types are handled in the same way as a SEQUENCE:
9363 that is the encoding is stored intact. There is also a new type
9364 "V_ASN1_OTHER" which is used when the class is not universal, in this
9365 case we have no idea what the actual type is so we just lump them all
9369 *) On VMS, stdout may very well lead to a file that is written to
9370 in a record-oriented fashion. That means that every write() will
9371 write a separate record, which will be read separately by the
9372 programs trying to read from it. This can be very confusing.
9374 The solution is to put a BIO filter in the way that will buffer
9375 text until a linefeed is reached, and then write everything a
9376 line at a time, so every record written will be an actual line,
9377 not chunks of lines and not (usually doesn't happen, but I've
9378 seen it once) several lines in one record. BIO_f_linebuffer() is
9381 Currently, it's a VMS-only method, because that's where it has
9382 been tested well enough.
9385 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9386 it can return incorrect results.
9387 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9388 but it was in 0.9.6-beta[12].)
9391 *) Disable the check for content being present when verifying detached
9392 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9393 include zero length content when signing messages.
9396 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9397 BIO_ctrl (for BIO pairs).
9400 *) Add DSO method for VMS.
9403 *) Bug fix: Montgomery multiplication could produce results with the
9407 *) Add RPM specification openssl.spec and modify it to build three
9408 packages. The default package contains applications, application
9409 documentation and run-time libraries. The devel package contains
9410 include files, static libraries and function documentation. The
9411 doc package contains the contents of the doc directory. The original
9412 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9415 *) Add a large number of documentation files for many SSL routines.
9416 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9418 *) Add a configuration entry for Sony News 4.
9419 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9421 *) Don't set the two most significant bits to one when generating a
9422 random number < q in the DSA library.
9425 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9426 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9427 the underlying transport is blocking) if a handshake took place.
9428 (The default behaviour is needed by applications such as s_client
9429 and s_server that use select() to determine when to use SSL_read;
9430 but for applications that know in advance when to expect data, it
9431 just makes things more complicated.)
9434 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9438 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9439 work better on such systems.
9440 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9442 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9443 Update PKCS12_parse() so it copies the friendlyName and the
9444 keyid to the certificates aux info.
9447 *) Fix bug in PKCS7_verify() which caused an infinite loop
9448 if there was more than one signature.
9449 [Sven Uszpelkat <su@celocom.de>]
9451 *) Major change in util/mkdef.pl to include extra information
9452 about each symbol, as well as presenting variables as well
9453 as functions. This change means that there's n more need
9454 to rebuild the .num files when some algorithms are excluded.
9457 *) Allow the verify time to be set by an application,
9458 rather than always using the current time.
9461 *) Phase 2 verify code reorganisation. The certificate
9462 verify code now looks up an issuer certificate by a
9463 number of criteria: subject name, authority key id
9464 and key usage. It also verifies self signed certificates
9465 by the same criteria. The main comparison function is
9466 X509_check_issued() which performs these checks.
9468 Lot of changes were necessary in order to support this
9469 without completely rewriting the lookup code.
9471 Authority and subject key identifier are now cached.
9473 The LHASH 'certs' is X509_STORE has now been replaced
9474 by a STACK_OF(X509_OBJECT). This is mainly because an
9475 LHASH can't store or retrieve multiple objects with
9476 the same hash value.
9478 As a result various functions (which were all internal
9479 use only) have changed to handle the new X509_STORE
9480 structure. This will break anything that messed round
9481 with X509_STORE internally.
9483 The functions X509_STORE_add_cert() now checks for an
9484 exact match, rather than just subject name.
9486 The X509_STORE API doesn't directly support the retrieval
9487 of multiple certificates matching a given criteria, however
9488 this can be worked round by performing a lookup first
9489 (which will fill the cache with candidate certificates)
9490 and then examining the cache for matches. This is probably
9491 the best we can do without throwing out X509_LOOKUP
9492 entirely (maybe later...).
9494 The X509_VERIFY_CTX structure has been enhanced considerably.
9496 All certificate lookup operations now go via a get_issuer()
9497 callback. Although this currently uses an X509_STORE it
9498 can be replaced by custom lookups. This is a simple way
9499 to bypass the X509_STORE hackery necessary to make this
9500 work and makes it possible to use more efficient techniques
9501 in future. A very simple version which uses a simple
9502 STACK for its trusted certificate store is also provided
9503 using X509_STORE_CTX_trusted_stack().
9505 The verify_cb() and verify() callbacks now have equivalents
9506 in the X509_STORE_CTX structure.
9508 X509_STORE_CTX also has a 'flags' field which can be used
9509 to customise the verify behaviour.
9512 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9513 excludes S/MIME capabilities.
9516 *) When a certificate request is read in keep a copy of the
9517 original encoding of the signed data and use it when outputting
9518 again. Signatures then use the original encoding rather than
9519 a decoded, encoded version which may cause problems if the
9520 request is improperly encoded.
9523 *) For consistency with other BIO_puts implementations, call
9524 buffer_write(b, ...) directly in buffer_puts instead of calling
9527 In BIO_puts, increment b->num_write as in BIO_write.
9528 [Peter.Sylvester@EdelWeb.fr]
9530 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9531 BN_zero, we may not return a BIGNUM with an array consisting of
9535 *) Avoid calling abort() from within the library when problems are
9536 detected, except if preprocessor symbols have been defined
9537 (such as REF_CHECK, BN_DEBUG etc.).
9540 *) New openssl application 'rsautl'. This utility can be
9541 used for low level RSA operations. DER public key
9542 BIO/fp routines also added.
9545 *) New Configure entry and patches for compiling on QNX 4.
9546 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9548 *) A demo state-machine implementation was sponsored by
9549 Nuron (http://www.nuron.com/) and is now available in
9550 demos/state_machine.
9553 *) New options added to the 'dgst' utility for signature
9554 generation and verification.
9557 *) Unrecognized PKCS#7 content types are now handled via a
9558 catch all ASN1_TYPE structure. This allows unsupported
9559 types to be stored as a "blob" and an application can
9560 encode and decode it manually.
9563 *) Fix various signed/unsigned issues to make a_strex.c
9565 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9567 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9568 length if passed a buffer. ASN1_INTEGER_to_BN failed
9569 if passed a NULL BN and its argument was negative.
9570 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9572 *) Modification to PKCS#7 encoding routines to output definite
9573 length encoding. Since currently the whole structures are in
9574 memory there's not real point in using indefinite length
9575 constructed encoding. However if OpenSSL is compiled with
9576 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9579 *) Added BIO_vprintf() and BIO_vsnprintf().
9582 *) Added more prefixes to parse for in the strings written
9583 through a logging bio, to cover all the levels that are available
9584 through syslog. The prefixes are now:
9586 PANIC, EMERG, EMR => LOG_EMERG
9587 ALERT, ALR => LOG_ALERT
9588 CRIT, CRI => LOG_CRIT
9589 ERROR, ERR => LOG_ERR
9590 WARNING, WARN, WAR => LOG_WARNING
9591 NOTICE, NOTE, NOT => LOG_NOTICE
9592 INFO, INF => LOG_INFO
9593 DEBUG, DBG => LOG_DEBUG
9595 and as before, if none of those prefixes are present at the
9596 beginning of the string, LOG_ERR is chosen.
9598 On Win32, the LOG_* levels are mapped according to this:
9600 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9601 LOG_WARNING => EVENTLOG_WARNING_TYPE
9602 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9606 *) Made it possible to reconfigure with just the configuration
9607 argument "reconf" or "reconfigure". The command line arguments
9608 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9609 and are retrieved from there when reconfiguring.
9613 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9615 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9618 *) The obj_dat.pl script was messing up the sorting of object
9619 names. The reason was that it compared the quoted version
9620 of strings as a result "OCSP" > "OCSP Signing" because
9621 " > SPACE. Changed script to store unquoted versions of
9622 names and add quotes on output. It was also omitting some
9623 names from the lookup table if they were given a default
9624 value (that is if SN is missing it is given the same
9625 value as LN and vice versa), these are now added on the
9626 grounds that if an object has a name we should be able to
9627 look it up. Finally added warning output when duplicate
9628 short or long names are found.
9631 *) Changes needed for Tandem NSK.
9632 [Scott Uroff <scott@xypro.com>]
9634 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9635 RSA_padding_check_SSLv23(), special padding was never detected
9636 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9637 version rollback attacks was not effective.
9639 In s23_clnt.c, don't use special rollback-attack detection padding
9640 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9641 client; similarly, in s23_srvr.c, don't do the rollback check if
9642 SSL 2.0 is the only protocol enabled in the server.
9645 *) Make it possible to get hexdumps of unprintable data with 'openssl
9646 asn1parse'. By implication, the functions ASN1_parse_dump() and
9647 BIO_dump_indent() are added.
9650 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9651 these print out strings and name structures based on various
9652 flags including RFC2253 support and proper handling of
9653 multibyte characters. Added options to the 'x509' utility
9654 to allow the various flags to be set.
9657 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9658 Also change the functions X509_cmp_current_time() and
9659 X509_gmtime_adj() work with an ASN1_TIME structure,
9660 this will enable certificates using GeneralizedTime in validity
9661 dates to be checked.
9664 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9665 negative public key encodings) on by default,
9666 NO_NEG_PUBKEY_BUG can be set to disable it.
9669 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9670 content octets. An i2c_ASN1_OBJECT is unnecessary because
9671 the encoding can be trivially obtained from the structure.
9674 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9675 not read locks (CRYPTO_r_[un]lock).
9678 *) A first attempt at creating official support for shared
9679 libraries through configuration. I've kept it so the
9680 default is static libraries only, and the OpenSSL programs
9681 are always statically linked for now, but there are
9682 preparations for dynamic linking in place.
9683 This has been tested on Linux and Tru64.
9686 *) Randomness polling function for Win9x, as described in:
9687 Peter Gutmann, Software Generation of Practically Strong
9691 *) Fix so PRNG is seeded in req if using an already existing
9695 *) New options to smime application. -inform and -outform
9696 allow alternative formats for the S/MIME message including
9697 PEM and DER. The -content option allows the content to be
9698 specified separately. This should allow things like Netscape
9699 form signing output easier to verify.
9702 *) Fix the ASN1 encoding of tags using the 'long form'.
9705 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9706 STRING types. These convert content octets to and from the
9707 underlying type. The actual tag and length octets are
9708 already assumed to have been read in and checked. These
9709 are needed because all other string types have virtually
9710 identical handling apart from the tag. By having versions
9711 of the ASN1 functions that just operate on content octets
9712 IMPLICIT tagging can be handled properly. It also allows
9713 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9714 and ASN1_INTEGER are identical apart from the tag.
9717 *) Change the handling of OID objects as follows:
9719 - New object identifiers are inserted in objects.txt, following
9720 the syntax given in objects.README.
9721 - objects.pl is used to process obj_mac.num and create a new
9723 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9726 This is currently kind of a hack, and the perl code in objects.pl
9727 isn't very elegant, but it works as I intended. The simplest way
9728 to check that it worked correctly is to look in obj_dat.h and
9729 check the array nid_objs and make sure the objects haven't moved
9730 around (this is important!). Additions are OK, as well as
9731 consistent name changes.
9734 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9737 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9738 The given file adds to whatever has already been seeded into the
9739 random pool through the RANDFILE configuration file option or
9740 environment variable, or the default random state file.
9743 *) mkstack.pl now sorts each macro group into lexical order.
9744 Previously the output order depended on the order the files
9745 appeared in the directory, resulting in needless rewriting
9749 *) Patches to make OpenSSL compile under Win32 again. Mostly
9750 work arounds for the VC++ problem that it treats func() as
9751 func(void). Also stripped out the parts of mkdef.pl that
9752 added extra typesafe functions: these no longer exist.
9755 *) Reorganisation of the stack code. The macros are now all
9756 collected in safestack.h . Each macro is defined in terms of
9757 a "stack macro" of the form SKM_<name>(type, a, b). The
9758 DEBUG_SAFESTACK is now handled in terms of function casts,
9759 this has the advantage of retaining type safety without the
9760 use of additional functions. If DEBUG_SAFESTACK is not defined
9761 then the non typesafe macros are used instead. Also modified the
9762 mkstack.pl script to handle the new form. Needs testing to see
9763 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9764 the default if no major problems. Similar behaviour for ASN1_SET_OF
9765 and PKCS12_STACK_OF.
9768 *) When some versions of IIS use the 'NET' form of private key the
9769 key derivation algorithm is different. Normally MD5(password) is
9770 used as a 128 bit RC4 key. In the modified case
9771 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9772 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9773 as the old Netscape_RSA functions except they have an additional
9774 'sgckey' parameter which uses the modified algorithm. Also added
9775 an -sgckey command line option to the rsa utility. Thanks to
9776 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9777 algorithm to openssl-dev.
9780 *) The evp_local.h macros were using 'c.##kname' which resulted in
9781 invalid expansion on some systems (SCO 5.0.5 for example).
9782 Corrected to 'c.kname'.
9783 [Phillip Porch <root@theporch.com>]
9785 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9786 a STACK of email addresses from a certificate or request, these look
9787 in the subject name and the subject alternative name extensions and
9788 omit any duplicate addresses.
9791 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9792 This makes DSA verification about 2 % faster.
9795 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9796 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9797 plus overhead for 1024 bit moduli).
9798 This makes exponentiations about 0.5 % faster for 1024 bit
9799 exponents (as measured by "openssl speed rsa2048").
9802 *) Rename memory handling macros to avoid conflicts with other
9804 Malloc => OPENSSL_malloc
9805 Malloc_locked => OPENSSL_malloc_locked
9806 Realloc => OPENSSL_realloc
9807 Free => OPENSSL_free
9810 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9811 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9814 *) CygWin32 support.
9815 [John Jarvie <jjarvie@newsguy.com>]
9817 *) The type-safe stack code has been rejigged. It is now only compiled
9818 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9819 by default all type-specific stack functions are "#define"d back to
9820 standard stack functions. This results in more streamlined output
9821 but retains the type-safety checking possibilities of the original
9825 *) The STACK code has been cleaned up, and certain type declarations
9826 that didn't make a lot of sense have been brought in line. This has
9827 also involved a cleanup of sorts in safestack.h to more correctly
9828 map type-safe stack functions onto their plain stack counterparts.
9829 This work has also resulted in a variety of "const"ifications of
9830 lots of the code, especially "_cmp" operations which should normally
9831 be prototyped with "const" parameters anyway.
9834 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9835 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9836 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9837 where all of 'md' is used each time the PRNG is used, but 'state'
9838 is used only indexed by a cyclic counter. As entropy may not be
9839 well distributed from the beginning, 'md' is important as a
9840 chaining variable. However, the output function chains only half
9841 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9842 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9843 in all of 'state' being rewritten, with the new values depending
9844 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9847 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9848 the handshake is continued after ssl_verify_cert_chain();
9849 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9850 can lead to 'unexplainable' connection aborts later.
9851 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9853 *) Major EVP API cipher revision.
9854 Add hooks for extra EVP features. This allows various cipher
9855 parameters to be set in the EVP interface. Support added for variable
9856 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9857 setting of RC2 and RC5 parameters.
9859 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9862 Remove lots of duplicated code from the EVP library. For example *every*
9863 cipher init() function handles the 'iv' in the same way according to the
9864 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9865 for CFB and OFB modes they zero ctx->num.
9867 New functionality allows removal of S/MIME code RC2 hack.
9869 Most of the routines have the same form and so can be declared in terms
9872 By shifting this to the top level EVP_CipherInit() it can be removed from
9873 all individual ciphers. If the cipher wants to handle IVs or keys
9874 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9877 Change lots of functions like EVP_EncryptUpdate() to now return a
9878 value: although software versions of the algorithms cannot fail
9879 any installed hardware versions can.
9882 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9883 this option is set, tolerate broken clients that send the negotiated
9884 protocol version number instead of the requested protocol version
9888 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9889 i.e. non-zero for export ciphersuites, zero otherwise.
9890 Previous versions had this flag inverted, inconsistent with
9891 rsa_tmp_cb (..._TMP_RSA_CB).
9892 [Bodo Moeller; problem reported by Amit Chopra]
9894 *) Add missing DSA library text string. Work around for some IIS
9895 key files with invalid SEQUENCE encoding.
9898 *) Add a document (doc/standards.txt) that list all kinds of standards
9899 and so on that are implemented in OpenSSL.
9902 *) Enhance c_rehash script. Old version would mishandle certificates
9903 with the same subject name hash and wouldn't handle CRLs at all.
9904 Added -fingerprint option to crl utility, to support new c_rehash
9908 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9911 *) Fix for SSL server purpose checking. Server checking was
9912 rejecting certificates which had extended key usage present
9913 but no ssl client purpose.
9914 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9916 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9917 is a little unclear about how a blank password is handled.
9918 Since the password in encoded as a BMPString with terminating
9919 double NULL a zero length password would end up as just the
9920 double NULL. However no password at all is different and is
9921 handled differently in the PKCS#12 key generation code. NS
9922 treats a blank password as zero length. MSIE treats it as no
9923 password on export: but it will try both on import. We now do
9924 the same: PKCS12_parse() tries zero length and no password if
9925 the password is set to "" or NULL (NULL is now a valid password:
9926 it wasn't before) as does the pkcs12 application.
9929 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9930 perror when PEM_read_bio_X509_REQ fails, the error message must
9931 be obtained from the error queue.
9934 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9935 it in ERR_remove_state if appropriate, and change ERR_get_state
9936 accordingly to avoid race conditions (this is necessary because
9937 thread_hash is no longer constant once set).
9940 *) Bugfix for linux-elf makefile.one.
9943 *) RSA_get_default_method() will now cause a default
9944 RSA_METHOD to be chosen if one doesn't exist already.
9945 Previously this was only set during a call to RSA_new()
9946 or RSA_new_method(NULL) meaning it was possible for
9947 RSA_get_default_method() to return NULL.
9950 *) Added native name translation to the existing DSO code
9951 that will convert (if the flag to do so is set) filenames
9952 that are sufficiently small and have no path information
9953 into a canonical native form. Eg. "blah" converted to
9954 "libblah.so" or "blah.dll" etc.
9957 *) New function ERR_error_string_n(e, buf, len) which is like
9958 ERR_error_string(e, buf), but writes at most 'len' bytes
9959 including the 0 terminator. For ERR_error_string_n, 'buf'
9961 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9963 *) CONF library reworked to become more general. A new CONF
9964 configuration file reader "class" is implemented as well as a
9965 new functions (NCONF_*, for "New CONF") to handle it. The now
9966 old CONF_* functions are still there, but are reimplemented to
9967 work in terms of the new functions. Also, a set of functions
9968 to handle the internal storage of the configuration data is
9969 provided to make it easier to write new configuration file
9970 reader "classes" (I can definitely see something reading a
9971 configuration file in XML format, for example), called _CONF_*,
9972 or "the configuration storage API"...
9974 The new configuration file reading functions are:
9976 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9977 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9979 NCONF_default, NCONF_WIN32
9981 NCONF_dump_fp, NCONF_dump_bio
9983 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9984 NCONF_new creates a new CONF object. This works in the same way
9985 as other interfaces in OpenSSL, like the BIO interface.
9986 NCONF_dump_* dump the internal storage of the configuration file,
9987 which is useful for debugging. All other functions take the same
9988 arguments as the old CONF_* functions wth the exception of the
9989 first that must be a `CONF *' instead of a `LHASH *'.
9991 To make it easer to use the new classes with the old CONF_* functions,
9992 the function CONF_set_default_method is provided.
9995 *) Add '-tls1' option to 'openssl ciphers', which was already
9996 mentioned in the documentation but had not been implemented.
9997 (This option is not yet really useful because even the additional
9998 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10001 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10002 OpenSSL-based applications) load shared libraries and bind to
10003 them in a portable way.
10004 [Geoff Thorpe, with contributions from Richard Levitte]
10006 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10008 *) Make sure _lrotl and _lrotr are only used with MSVC.
10010 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10011 (the default implementation of RAND_status).
10013 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10014 to '-clrext' (= clear extensions), as intended and documented.
10015 [Bodo Moeller; inconsistency pointed out by Michael Attili
10016 <attili@amaxo.com>]
10018 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10019 was larger than the MD block size.
10020 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10022 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10023 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10024 using the passed key: if the passed key was a private key the result
10025 of X509_print(), for example, would be to print out all the private key
10029 *) des_quad_cksum() byte order bug fix.
10030 [Ulf Möller, using the problem description in krb4-0.9.7, where
10031 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10033 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10035 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10037 *) For easily testing in shell scripts whether some command
10038 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10039 returns with exit code 0 iff no command of the given name is available.
10040 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10041 the output goes to stdout and nothing is printed to stderr.
10042 Additional arguments are always ignored.
10044 Since for each cipher there is a command of the same name,
10045 the 'no-cipher' compilation switches can be tested this way.
10047 ('openssl no-XXX' is not able to detect pseudo-commands such
10048 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10051 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10054 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10055 is set; it will be thrown away anyway because each handshake creates
10057 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10058 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10059 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10060 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10063 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10064 'Q' and 'R' lose their special meanings (quit/renegotiate).
10065 This is part of what -quiet does; unlike -quiet, -ign_eof
10066 does not suppress any output.
10069 *) Add compatibility options to the purpose and trust code. The
10070 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10071 accepts a certificate or CA, this was the previous behaviour,
10072 with all the associated security issues.
10074 X509_TRUST_COMPAT is the old trust behaviour: only and
10075 automatically trust self signed roots in certificate store. A
10076 new trust setting X509_TRUST_DEFAULT is used to specify that
10077 a purpose has no associated trust setting and it should instead
10078 use the value in the default purpose.
10081 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10082 and fix a memory leak.
10085 *) In util/mkerr.pl (which implements 'make errors'), preserve
10086 reason strings from the previous version of the .c file, as
10087 the default to have only downcase letters (and digits) in
10088 automatically generated reasons codes is not always appropriate.
10091 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10092 using strerror. Previously, ERR_reason_error_string() returned
10093 library names as reason strings for SYSerr; but SYSerr is a special
10094 case where small numbers are errno values, not library numbers.
10097 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10098 converts DSA parameters into DH parameters. (When creating parameters,
10099 DSA_generate_parameters is used.)
10102 *) Include 'length' (recommended exponent length) in C code generated
10103 by 'openssl dhparam -C'.
10106 *) The second argument to set_label in perlasm was already being used
10107 so couldn't be used as a "file scope" flag. Moved to third argument
10111 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10112 instead of RAND_bytes for encryption IVs and salts.
10115 *) Include RAND_status() into RAND_METHOD instead of implementing
10116 it only for md_rand.c Otherwise replacing the PRNG by calling
10117 RAND_set_rand_method would be impossible.
10120 *) Don't let DSA_generate_key() enter an infinite loop if the random
10121 number generation fails.
10124 *) New 'rand' application for creating pseudo-random output.
10127 *) Added configuration support for Linux/IA64
10128 [Rolf Haberrecker <rolf@suse.de>]
10130 *) Assembler module support for Mingw32.
10133 *) Shared library support for HPUX (in shlib/).
10134 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10136 *) Shared library support for Solaris gcc.
10137 [Lutz Behnke <behnke@trustcenter.de>]
10139 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10141 *) PKCS7_encrypt() was adding text MIME headers twice because they
10142 were added manually and by SMIME_crlf_copy().
10145 *) In bntest.c don't call BN_rand with zero bits argument.
10146 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10148 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10149 case was implemented. This caused BN_div_recp() to fail occasionally.
10152 *) Add an optional second argument to the set_label() in the perl
10153 assembly language builder. If this argument exists and is set
10154 to 1 it signals that the assembler should use a symbol whose
10155 scope is the entire file, not just the current function. This
10156 is needed with MASM which uses the format label:: for this scope.
10157 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10159 *) Change the ASN1 types so they are typedefs by default. Before
10160 almost all types were #define'd to ASN1_STRING which was causing
10161 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10165 *) Change names of new functions to the new get1/get0 naming
10166 convention: After 'get1', the caller owns a reference count
10167 and has to call ..._free; 'get0' returns a pointer to some
10168 data structure without incrementing reference counters.
10169 (Some of the existing 'get' functions increment a reference
10170 counter, some don't.)
10171 Similarly, 'set1' and 'add1' functions increase reference
10172 counters or duplicate objects.
10175 *) Allow for the possibility of temp RSA key generation failure:
10176 the code used to assume it always worked and crashed on failure.
10179 *) Fix potential buffer overrun problem in BIO_printf().
10180 [Ulf Möller, using public domain code by Patrick Powell; problem
10181 pointed out by David Sacerdote <das33@cornell.edu>]
10183 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10184 RAND_egd() and RAND_status(). In the command line application,
10185 the EGD socket can be specified like a seed file using RANDFILE
10189 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10190 Some CAs (e.g. Verisign) distribute certificates in this form.
10193 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10194 list to exclude them. This means that no special compilation option
10195 is needed to use anonymous DH: it just needs to be included in the
10199 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10200 EVP_MD_type. The old functionality is available in a new macro called
10201 EVP_MD_md(). Change code that uses it and update docs.
10204 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10205 where the 'void *' argument is replaced by a function pointer argument.
10206 Previously 'void *' was abused to point to functions, which works on
10207 many platforms, but is not correct. As these functions are usually
10208 called by macros defined in OpenSSL header files, most source code
10209 should work without changes.
10212 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10213 sections with information on -D... compiler switches used for
10214 compiling the library so that applications can see them. To enable
10215 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10216 must be defined. E.g.,
10217 #define OPENSSL_ALGORITHM_DEFINES
10218 #include <openssl/opensslconf.h>
10219 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10220 [Richard Levitte, Ulf and Bodo Möller]
10222 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10226 *) Change the 'other' type in certificate aux info to a STACK_OF
10227 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10228 the required ASN1 format: arbitrary types determined by an OID.
10231 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10232 argument to 'req'. This is not because the function is newer or
10233 better than others it just uses the work 'NEW' in the certificate
10234 request header lines. Some software needs this.
10237 *) Reorganise password command line arguments: now passwords can be
10238 obtained from various sources. Delete the PEM_cb function and make
10239 it the default behaviour: i.e. if the callback is NULL and the
10240 usrdata argument is not NULL interpret it as a null terminated pass
10241 phrase. If usrdata and the callback are NULL then the pass phrase
10242 is prompted for as usual.
10245 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10246 the support is automatically enabled. The resulting binaries will
10247 autodetect the card and use it if present.
10248 [Ben Laurie and Compaq Inc.]
10250 *) Work around for Netscape hang bug. This sends certificate request
10251 and server done in one record. Since this is perfectly legal in the
10252 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10253 the bugs/SSLv3 entry for more info.
10256 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10259 *) Add -rand argument to smime and pkcs12 applications and read/write
10263 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10266 *) Add command line password options to the remaining applications.
10269 *) Bug fix for BN_div_recp() for numerators with an even number of
10273 *) More tests in bntest.c, and changed test_bn output.
10276 *) ./config recognizes MacOS X now.
10279 *) Bug fix for BN_div() when the first words of num and divisor are
10280 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10283 *) Add support for various broken PKCS#8 formats, and command line
10284 options to produce them.
10287 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10288 get temporary BIGNUMs from a BN_CTX.
10291 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10295 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10296 include a #define from the old name to the new. The original intent
10297 was that statically linked binaries could for example just call
10298 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10299 link with digests. This never worked because SSLeay_add_all_digests()
10300 and SSLeay_add_all_ciphers() were in the same source file so calling
10301 one would link with the other. They are now in separate source files.
10304 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10307 *) Use a less unusual form of the Miller-Rabin primality test (it used
10308 a binary algorithm for exponentiation integrated into the Miller-Rabin
10309 loop, our standard modexp algorithms are faster).
10312 *) Support for the EBCDIC character set completed.
10313 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10315 *) Source code cleanups: use const where appropriate, eliminate casts,
10316 use void * instead of char * in lhash.
10319 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10320 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10321 this the server could overwrite ephemeral keys that the client
10325 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10326 using 50 iterations of the Rabin-Miller test.
10328 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10329 iterations of the Rabin-Miller test as required by the appendix
10330 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10331 As BN_is_prime_fasttest includes trial division, DSA parameter
10332 generation becomes much faster.
10334 This implies a change for the callback functions in DSA_is_prime
10335 and DSA_generate_parameters: The callback function is called once
10336 for each positive witness in the Rabin-Miller test, not just
10337 occasionally in the inner loop; and the parameters to the
10338 callback function now provide an iteration count for the outer
10339 loop rather than for the current invocation of the inner loop.
10340 DSA_generate_parameters additionally can call the callback
10341 function with an 'iteration count' of -1, meaning that a
10342 candidate has passed the trial division test (when q is generated
10343 from an application-provided seed, trial division is skipped).
10346 *) New function BN_is_prime_fasttest that optionally does trial
10347 division before starting the Rabin-Miller test and has
10348 an additional BN_CTX * argument (whereas BN_is_prime always
10349 has to allocate at least one BN_CTX).
10350 'callback(1, -1, cb_arg)' is called when a number has passed the
10351 trial division stage.
10354 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10358 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10361 *) New function BN_pseudo_rand().
10364 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10365 bignum version of BN_from_montgomery() with the working code from
10366 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10370 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10371 made it impossible to use the same SSL_SESSION data structure in
10372 SSL2 clients in multiple threads.
10375 *) The return value of RAND_load_file() no longer counts bytes obtained
10376 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10377 to seed the PRNG (previously an explicit byte count was required).
10378 [Ulf Möller, Bodo Möller]
10380 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10381 used (char *) instead of (void *) and had casts all over the place.
10384 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10387 *) Retain source code compatibility for BN_prime_checks macro:
10388 BN_is_prime(..., BN_prime_checks, ...) now uses
10389 BN_prime_checks_for_size to determine the appropriate number of
10390 Rabin-Miller iterations.
10393 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10394 DH_CHECK_P_NOT_SAFE_PRIME.
10395 (Check if this is true? OpenPGP calls them "strong".)
10398 *) Merge the functionality of "dh" and "gendh" programs into a new program
10399 "dhparam". The old programs are retained for now but will handle DH keys
10400 (instead of parameters) in future.
10403 *) Make the ciphers, s_server and s_client programs check the return values
10404 when a new cipher list is set.
10407 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10408 ciphers. Before when the 56bit ciphers were enabled the sorting was
10411 The syntax for the cipher sorting has been extended to support sorting by
10412 cipher-strength (using the strength_bits hard coded in the tables).
10413 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10415 Fix a bug in the cipher-command parser: when supplying a cipher command
10416 string with an "undefined" symbol (neither command nor alphanumeric
10417 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10418 an error is flagged.
10420 Due to the strength-sorting extension, the code of the
10421 ssl_create_cipher_list() function was completely rearranged. I hope that
10422 the readability was also increased :-)
10423 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10425 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10426 for the first serial number and places 2 in the serial number file. This
10427 avoids problems when the root CA is created with serial number zero and
10428 the first user certificate has the same issuer name and serial number
10432 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10433 the new code. Add documentation for this stuff.
10436 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10437 X509_*() to X509at_*() on the grounds that they don't handle X509
10438 structures and behave in an analogous way to the X509v3 functions:
10439 they shouldn't be called directly but wrapper functions should be used
10442 So we also now have some wrapper functions that call the X509at functions
10443 when passed certificate requests. (TO DO: similar things can be done with
10444 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10445 things. Some of these need some d2i or i2d and print functionality
10446 because they handle more complex structures.)
10449 *) Add missing #ifndefs that caused missing symbols when building libssl
10450 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10451 NO_RSA in ssl/s2*.c.
10452 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10454 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10455 has a return value which indicates the quality of the random data
10456 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10457 error queue. New function RAND_pseudo_bytes() generates output that is
10458 guaranteed to be unique but not unpredictable. RAND_add is like
10459 RAND_seed, but takes an extra argument for an entropy estimate
10460 (RAND_seed always assumes full entropy).
10463 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10464 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10465 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10466 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10467 false-positive rate of at most 2^-80 for random input.
10470 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10473 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10474 in the 0.9.5 release), this returns the chain
10475 from an X509_CTX structure with a dup of the stack and all
10476 the X509 reference counts upped: so the stack will exist
10477 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10480 Also make SSL_SESSION_print() print out the verify return
10484 *) Add manpage for the pkcs12 command. Also change the default
10485 behaviour so MAC iteration counts are used unless the new
10486 -nomaciter option is used. This improves file security and
10487 only older versions of MSIE (4.0 for example) need it.
10490 *) Honor the no-xxx Configure options when creating .DEF files.
10493 *) Add PKCS#10 attributes to field table: challengePassword,
10494 unstructuredName and unstructuredAddress. These are taken from
10495 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10496 international characters are used.
10498 More changes to X509_ATTRIBUTE code: allow the setting of types
10499 based on strings. Remove the 'loc' parameter when adding
10500 attributes because these will be a SET OF encoding which is sorted
10504 *) Initial changes to the 'req' utility to allow request generation
10505 automation. This will allow an application to just generate a template
10506 file containing all the field values and have req construct the
10509 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10510 used all over the place including certificate requests and PKCS#7
10511 structures. They are currently handled manually where necessary with
10512 some primitive wrappers for PKCS#7. The new functions behave in a
10513 manner analogous to the X509 extension functions: they allow
10514 attributes to be looked up by NID and added.
10516 Later something similar to the X509V3 code would be desirable to
10517 automatically handle the encoding, decoding and printing of the
10518 more complex types. The string types like challengePassword can
10519 be handled by the string table functions.
10521 Also modified the multi byte string table handling. Now there is
10522 a 'global mask' which masks out certain types. The table itself
10523 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10524 is useful when for example there is only one permissible type
10525 (as in countryName) and using the mask might result in no valid
10529 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10530 SSL_get_peer_finished to allow applications to obtain the latest
10531 Finished messages sent to the peer or expected from the peer,
10532 respectively. (SSL_get_peer_finished is usually the Finished message
10533 actually received from the peer, otherwise the protocol will be aborted.)
10535 As the Finished message are message digests of the complete handshake
10536 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10537 be used for external authentication procedures when the authentication
10538 provided by SSL/TLS is not desired or is not enough.
10541 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10542 the host supports BWX extension and if Compaq C is present on the
10543 $PATH. Just exploiting of the BWX extension results in 20-30%
10544 performance kick for some algorithms, e.g. DES and RC4 to mention
10545 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10549 *) Add support for MS "fast SGC". This is arguably a violation of the
10550 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10551 weak crypto and after checking the certificate is SGC a second one
10552 with strong crypto. MS SGC stops the first handshake after receiving
10553 the server certificate message and sends a second client hello. Since
10554 a server will typically do all the time consuming operations before
10555 expecting any further messages from the client (server key exchange
10556 is the most expensive) there is little difference between the two.
10558 To get OpenSSL to support MS SGC we have to permit a second client
10559 hello message after we have sent server done. In addition we have to
10560 reset the MAC if we do get this second client hello.
10563 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10564 if a DER encoded private key is RSA or DSA traditional format. Changed
10565 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10566 format DER encoded private key. Newer code should use PKCS#8 format which
10567 has the key type encoded in the ASN1 structure. Added DER private key
10568 support to pkcs8 application.
10571 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10572 ciphersuites has been selected (as required by the SSL 3/TLS 1
10573 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10574 is set, we interpret this as a request to violate the specification
10575 (the worst that can happen is a handshake failure, and 'correct'
10576 behaviour would result in a handshake failure anyway).
10579 *) In SSL_CTX_add_session, take into account that there might be multiple
10580 SSL_SESSION structures with the same session ID (e.g. when two threads
10581 concurrently obtain them from an external cache).
10582 The internal cache can handle only one SSL_SESSION with a given ID,
10583 so if there's a conflict, we now throw out the old one to achieve
10587 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10588 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10589 some routines that use cipher OIDs: some ciphers do not have OIDs
10590 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10594 *) Simplify the trust setting structure and code. Now we just have
10595 two sequences of OIDs for trusted and rejected settings. These will
10596 typically have values the same as the extended key usage extension
10597 and any application specific purposes.
10599 The trust checking code now has a default behaviour: it will just
10600 check for an object with the same NID as the passed id. Functions can
10601 be provided to override either the default behaviour or the behaviour
10602 for a given id. SSL client, server and email already have functions
10603 in place for compatibility: they check the NID and also return "trusted"
10604 if the certificate is self signed.
10607 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10608 traditional format into an EVP_PKEY structure.
10611 *) Add a password callback function PEM_cb() which either prompts for
10612 a password if usr_data is NULL or otherwise assumes it is a null
10613 terminated password. Allow passwords to be passed on command line
10614 environment or config files in a few more utilities.
10617 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10618 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10619 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10620 Update documentation.
10623 *) Support for ASN1 "NULL" type. This could be handled before by using
10624 ASN1_TYPE but there wasn't any function that would try to read a NULL
10625 and produce an error if it couldn't. For compatibility we also have
10626 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10627 don't allocate anything because they don't need to.
10630 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10632 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10634 *) Rebuild of the memory allocation routines used by OpenSSL code and
10635 possibly others as well. The purpose is to make an interface that
10636 provide hooks so anyone can build a separate set of allocation and
10637 deallocation routines to be used by OpenSSL, for example memory
10638 pool implementations, or something else, which was previously hard
10639 since Malloc(), Realloc() and Free() were defined as macros having
10640 the values malloc, realloc and free, respectively (except for Win32
10641 compilations). The same is provided for memory debugging code.
10642 OpenSSL already comes with functionality to find memory leaks, but
10643 this gives people a chance to debug other memory problems.
10645 With these changes, a new set of functions and macros have appeared:
10647 CRYPTO_set_mem_debug_functions() [F]
10648 CRYPTO_get_mem_debug_functions() [F]
10649 CRYPTO_dbg_set_options() [F]
10650 CRYPTO_dbg_get_options() [F]
10651 CRYPTO_malloc_debug_init() [M]
10653 The memory debug functions are NULL by default, unless the library
10654 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10655 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10656 gives the standard debugging functions that come with OpenSSL) or
10657 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10658 provided by the library user) must be used. When the standard
10659 debugging functions are used, CRYPTO_dbg_set_options can be used to
10660 request additional information:
10661 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10662 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10664 Also, things like CRYPTO_set_mem_functions will always give the
10665 expected result (the new set of functions is used for allocation
10666 and deallocation) at all times, regardless of platform and compiler
10669 To finish it up, some functions that were never use in any other
10670 way than through macros have a new API and new semantic:
10672 CRYPTO_dbg_malloc()
10673 CRYPTO_dbg_realloc()
10676 All macros of value have retained their old syntax.
10677 [Richard Levitte and Bodo Moeller]
10679 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10680 ordering of SMIMECapabilities wasn't in "strength order" and there
10681 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10685 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10686 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10687 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10689 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10690 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10691 functionality to handle multipart/signed properly) and a utility
10692 called 'smime' to call all this stuff. This is based on code I
10693 originally wrote for Celo who have kindly allowed it to be
10694 included in OpenSSL.
10697 *) Add variants des_set_key_checked and des_set_key_unchecked of
10698 des_set_key (aka des_key_sched). Global variable des_check_key
10699 decides which of these is called by des_set_key; this way
10700 des_check_key behaves as it always did, but applications and
10701 the library itself, which was buggy for des_check_key == 1,
10702 have a cleaner way to pick the version they need.
10705 *) New function PKCS12_newpass() which changes the password of a
10709 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10710 dynamic mix. In both cases the ids can be used as an index into the
10711 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10712 functions so they accept a list of the field values and the
10713 application doesn't need to directly manipulate the X509_TRUST
10717 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10721 *) Modify the way the V3 extension code looks up extensions. This now
10722 works in a similar way to the object code: we have some "standard"
10723 extensions in a static table which is searched with OBJ_bsearch()
10724 and the application can add dynamic ones if needed. The file
10725 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10726 updated whenever a new extension is added to the core code and kept
10727 in ext_nid order. There is a simple program 'tabtest.c' which checks
10728 this. New extensions are not added too often so this file can readily
10729 be maintained manually.
10731 There are two big advantages in doing things this way. The extensions
10732 can be looked up immediately and no longer need to be "added" using
10733 X509V3_add_standard_extensions(): this function now does nothing.
10734 [Side note: I get *lots* of email saying the extension code doesn't
10735 work because people forget to call this function]
10736 Also no dynamic allocation is done unless new extensions are added:
10737 so if we don't add custom extensions there is no need to call
10738 X509V3_EXT_cleanup().
10741 *) Modify enc utility's salting as follows: make salting the default. Add a
10742 magic header, so unsalted files fail gracefully instead of just decrypting
10743 to garbage. This is because not salting is a big security hole, so people
10744 should be discouraged from doing it.
10747 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10748 digest to be passed on the command line but it only used this
10749 parameter when signing a certificate. Modified so all relevant
10750 operations are affected by the digest parameter including the
10751 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10752 DSA key was used because it didn't fix the digest.
10755 *) Initial certificate chain verify code. Currently tests the untrusted
10756 certificates for consistency with the verify purpose (which is set
10757 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10759 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10760 this is because it will reject chains with invalid extensions whereas
10761 every previous version of OpenSSL and SSLeay made no checks at all.
10763 Trust code: checks the root CA for the relevant trust settings. Trust
10764 settings have an initial value consistent with the verify purpose: e.g.
10765 if the verify purpose is for SSL client use it expects the CA to be
10766 trusted for SSL client use. However the default value can be changed to
10767 permit custom trust settings: one example of this would be to only trust
10768 certificates from a specific "secure" set of CAs.
10770 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10771 which should be used for version portability: especially since the
10772 verify structure is likely to change more often now.
10774 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10775 to set them. If not set then assume SSL clients will verify SSL servers
10778 Two new options to the verify program: -untrusted allows a set of
10779 untrusted certificates to be passed in and -purpose which sets the
10780 intended purpose of the certificate. If a purpose is set then the
10781 new chain verify code is used to check extension consistency.
10784 *) Support for the authority information access extension.
10787 *) Modify RSA and DSA PEM read routines to transparently handle
10788 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10789 public keys in a format compatible with certificate
10790 SubjectPublicKeyInfo structures. Unfortunately there were already
10791 functions called *_PublicKey_* which used various odd formats so
10792 these are retained for compatibility: however the DSA variants were
10793 never in a public release so they have been deleted. Changed dsa/rsa
10794 utilities to handle the new format: note no releases ever handled public
10795 keys so we should be OK.
10797 The primary motivation for this change is to avoid the same fiasco
10798 that dogs private keys: there are several incompatible private key
10799 formats some of which are standard and some OpenSSL specific and
10800 require various evil hacks to allow partial transparent handling and
10801 even then it doesn't work with DER formats. Given the option anything
10802 other than PKCS#8 should be dumped: but the other formats have to
10803 stay in the name of compatibility.
10805 With public keys and the benefit of hindsight one standard format
10806 is used which works with EVP_PKEY, RSA or DSA structures: though
10807 it clearly returns an error if you try to read the wrong kind of key.
10809 Added a -pubkey option to the 'x509' utility to output the public key.
10810 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10811 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10812 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10813 that do the same as the EVP_PKEY_assign_*() except they up the
10814 reference count of the added key (they don't "swallow" the
10818 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10819 CRLs would fail if the file contained no certificates or no CRLs:
10820 added a new function to read in both types and return the number
10821 read: this means that if none are read it will be an error. The
10822 DER versions of the certificate and CRL reader would always fail
10823 because it isn't possible to mix certificates and CRLs in DER format
10824 without choking one or the other routine. Changed this to just read
10825 a certificate: this is the best we can do. Also modified the code
10826 in apps/verify.c to take notice of return codes: it was previously
10827 attempting to read in certificates from NULL pointers and ignoring
10828 any errors: this is one reason why the cert and CRL reader seemed
10829 to work. It doesn't check return codes from the default certificate
10830 routines: these may well fail if the certificates aren't installed.
10833 *) Code to support otherName option in GeneralName.
10836 *) First update to verify code. Change the verify utility
10837 so it warns if it is passed a self signed certificate:
10838 for consistency with the normal behaviour. X509_verify
10839 has been modified to it will now verify a self signed
10840 certificate if *exactly* the same certificate appears
10841 in the store: it was previously impossible to trust a
10842 single self signed certificate. This means that:
10843 openssl verify ss.pem
10844 now gives a warning about a self signed certificate but
10845 openssl verify -CAfile ss.pem ss.pem
10849 *) For servers, store verify_result in SSL_SESSION data structure
10850 (and add it to external session representation).
10851 This is needed when client certificate verifications fails,
10852 but an application-provided verification callback (set by
10853 SSL_CTX_set_cert_verify_callback) allows accepting the session
10854 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10855 but returns 1): When the session is reused, we have to set
10856 ssl->verify_result to the appropriate error code to avoid
10858 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10860 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10861 case in PKCS7_dataInit() where the signed PKCS7 structure
10862 didn't contain any existing data because it was being created.
10863 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10865 *) Add a salt to the key derivation routines in enc.c. This
10866 forms the first 8 bytes of the encrypted file. Also add a
10867 -S option to allow a salt to be input on the command line.
10870 *) New function X509_cmp(). Oddly enough there wasn't a function
10871 to compare two certificates. We do this by working out the SHA1
10872 hash and comparing that. X509_cmp() will be needed by the trust
10876 *) SSL_get1_session() is like SSL_get_session(), but increments
10877 the reference count in the SSL_SESSION returned.
10878 [Geoff Thorpe <geoff@eu.c2.net>]
10880 *) Fix for 'req': it was adding a null to request attributes.
10881 Also change the X509_LOOKUP and X509_INFO code to handle
10882 certificate auxiliary information.
10885 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10889 *) Add the possibility to add extra information to the memory leak
10890 detecting output, to form tracebacks, showing from where each
10891 allocation was originated: CRYPTO_push_info("constant string") adds
10892 the string plus current file name and line number to a per-thread
10893 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10894 is like calling CYRPTO_pop_info() until the stack is empty.
10895 Also updated memory leak detection code to be multi-thread-safe.
10898 *) Add options -text and -noout to pkcs7 utility and delete the
10899 encryption options which never did anything. Update docs.
10902 *) Add options to some of the utilities to allow the pass phrase
10903 to be included on either the command line (not recommended on
10904 OSes like Unix) or read from the environment. Update the
10905 manpages and fix a few bugs.
10908 *) Add a few manpages for some of the openssl commands.
10911 *) Fix the -revoke option in ca. It was freeing up memory twice,
10912 leaking and not finding already revoked certificates.
10915 *) Extensive changes to support certificate auxiliary information.
10916 This involves the use of X509_CERT_AUX structure and X509_AUX
10917 functions. An X509_AUX function such as PEM_read_X509_AUX()
10918 can still read in a certificate file in the usual way but it
10919 will also read in any additional "auxiliary information". By
10920 doing things this way a fair degree of compatibility can be
10921 retained: existing certificates can have this information added
10922 using the new 'x509' options.
10924 Current auxiliary information includes an "alias" and some trust
10925 settings. The trust settings will ultimately be used in enhanced
10926 certificate chain verification routines: currently a certificate
10927 can only be trusted if it is self signed and then it is trusted
10931 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10932 The problem was that one of the replacement routines had not been working
10933 since SSLeay releases. For now the offending routine has been replaced
10934 with non-optimised assembler. Even so, this now gives around 95%
10935 performance improvement for 1024 bit RSA signs.
10938 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10939 handling. Most clients have the effective key size in bits equal to
10940 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10941 A few however don't do this and instead use the size of the decrypted key
10942 to determine the RC2 key length and the AlgorithmIdentifier to determine
10943 the effective key length. In this case the effective key length can still
10944 be 40 bits but the key length can be 168 bits for example. This is fixed
10945 by manually forcing an RC2 key into the EVP_PKEY structure because the
10946 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10947 the key length and effective key length are equal.
10950 *) Add a bunch of functions that should simplify the creation of
10951 X509_NAME structures. Now you should be able to do:
10952 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10953 and have it automatically work out the correct field type and fill in
10954 the structures. The more adventurous can try:
10955 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10956 and it will (hopefully) work out the correct multibyte encoding.
10959 *) Change the 'req' utility to use the new field handling and multibyte
10960 copy routines. Before the DN field creation was handled in an ad hoc
10961 way in req, ca, and x509 which was rather broken and didn't support
10962 BMPStrings or UTF8Strings. Since some software doesn't implement
10963 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10964 using the dirstring_type option. See the new comment in the default
10965 openssl.cnf for more info.
10968 *) Make crypto/rand/md_rand.c more robust:
10969 - Assure unique random numbers after fork().
10970 - Make sure that concurrent threads access the global counter and
10971 md serializably so that we never lose entropy in them
10972 or use exactly the same state in multiple threads.
10973 Access to the large state is not always serializable because
10974 the additional locking could be a performance killer, and
10975 md should be large enough anyway.
10978 *) New file apps/app_rand.c with commonly needed functionality
10979 for handling the random seed file.
10981 Use the random seed file in some applications that previously did not:
10983 dsaparam -genkey (which also ignored its '-rand' option),
10986 x509 (when signing).
10987 Except on systems with /dev/urandom, it is crucial to have a random
10988 seed file at least for key creation, DSA signing, and for DH exchanges;
10989 for RSA signatures we could do without one.
10991 gendh and gendsa (unlike genrsa) used to read only the first byte
10992 of each file listed in the '-rand' option. The function as previously
10993 found in genrsa is now in app_rand.c and is used by all programs
10994 that support '-rand'.
10997 *) In RAND_write_file, use mode 0600 for creating files;
10998 don't just chmod when it may be too late.
11001 *) Report an error from X509_STORE_load_locations
11002 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11005 *) New function ASN1_mbstring_copy() this copies a string in either
11006 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11007 into an ASN1_STRING type. A mask of permissible types is passed
11008 and it chooses the "minimal" type to use or an error if not type
11012 *) Add function equivalents to the various macros in asn1.h. The old
11013 macros are retained with an M_ prefix. Code inside the library can
11014 use the M_ macros. External code (including the openssl utility)
11015 should *NOT* in order to be "shared library friendly".
11018 *) Add various functions that can check a certificate's extensions
11019 to see if it usable for various purposes such as SSL client,
11020 server or S/MIME and CAs of these types. This is currently
11021 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11022 verification. Also added a -purpose flag to x509 utility to
11023 print out all the purposes.
11026 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11030 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11031 for, obtain and decode and extension and obtain its critical flag.
11032 This allows all the necessary extension code to be handled in a
11033 single function call.
11036 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11037 platforms. See crypto/rc4/rc4_enc.c for further details.
11040 *) New -noout option to asn1parse. This causes no output to be produced
11041 its main use is when combined with -strparse and -out to extract data
11042 from a file (which may not be in ASN.1 format).
11045 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11046 when producing the local key id.
11047 [Richard Levitte <levitte@stacken.kth.se>]
11049 *) New option -dhparam in s_server. This allows a DH parameter file to be
11050 stated explicitly. If it is not stated then it tries the first server
11051 certificate file. The previous behaviour hard coded the filename
11055 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11056 a public key to be input or output. For example:
11057 openssl rsa -in key.pem -pubout -out pubkey.pem
11058 Also added necessary DSA public key functions to handle this.
11061 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11062 in the message. This was handled by allowing
11063 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11064 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11066 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11067 to the end of the strings whereas this didn't. This would cause problems
11068 if strings read with d2i_ASN1_bytes() were later modified.
11069 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11071 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11072 data and it contains EOF it will end up returning an error. This is
11073 caused by input 46 bytes long. The cause is due to the way base64
11074 BIOs find the start of base64 encoded data. They do this by trying a
11075 trial decode on each line until they find one that works. When they
11076 do a flag is set and it starts again knowing it can pass all the
11077 data directly through the decoder. Unfortunately it doesn't reset
11078 the context it uses. This means that if EOF is reached an attempt
11079 is made to pass two EOFs through the context and this causes the
11080 resulting error. This can also cause other problems as well. As is
11081 usual with these problems it takes *ages* to find and the fix is
11082 trivial: move one line.
11083 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11085 *) Ugly workaround to get s_client and s_server working under Windows. The
11086 old code wouldn't work because it needed to select() on sockets and the
11087 tty (for keypresses and to see if data could be written). Win32 only
11088 supports select() on sockets so we select() with a 1s timeout on the
11089 sockets and then see if any characters are waiting to be read, if none
11090 are present then we retry, we also assume we can always write data to
11091 the tty. This isn't nice because the code then blocks until we've
11092 received a complete line of data and it is effectively polling the
11093 keyboard at 1s intervals: however it's quite a bit better than not
11094 working at all :-) A dedicated Windows application might handle this
11095 with an event loop for example.
11098 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11099 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11100 will be called when RSA_sign() and RSA_verify() are used. This is useful
11101 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11102 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11103 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11104 This necessitated the support of an extra signature type NID_md5_sha1
11105 for SSL signatures and modifications to the SSL library to use it instead
11106 of calling RSA_public_decrypt() and RSA_private_encrypt().
11109 *) Add new -verify -CAfile and -CApath options to the crl program, these
11110 will lookup a CRL issuers certificate and verify the signature in a
11111 similar way to the verify program. Tidy up the crl program so it
11112 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11113 less strict. It will now permit CRL extensions even if it is not
11114 a V2 CRL: this will allow it to tolerate some broken CRLs.
11117 *) Initialize all non-automatic variables each time one of the openssl
11118 sub-programs is started (this is necessary as they may be started
11119 multiple times from the "OpenSSL>" prompt).
11120 [Lennart Bang, Bodo Moeller]
11122 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11123 removing all other RSA functionality (this is what NO_RSA does). This
11124 is so (for example) those in the US can disable those operations covered
11125 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11129 *) Non-copying interface to BIO pairs.
11130 (still largely untested)
11133 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11134 ASCII string. This was handled independently in various places before.
11137 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11138 UTF8 strings a character at a time.
11141 *) Use client_version from client hello to select the protocol
11142 (s23_srvr.c) and for RSA client key exchange verification
11143 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11146 *) Add various utility functions to handle SPKACs, these were previously
11147 handled by poking round in the structure internals. Added new function
11148 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11149 print, verify and generate SPKACs. Based on an original idea from
11150 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11153 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11156 *) Allow the config file extension section to be overwritten on the
11157 command line. Based on an original idea from Massimiliano Pala
11158 <madwolf@comune.modena.it>. The new option is called -extensions
11159 and can be applied to ca, req and x509. Also -reqexts to override
11160 the request extensions in req and -crlexts to override the crl extensions
11164 *) Add new feature to the SPKAC handling in ca. Now you can include
11165 the same field multiple times by preceding it by "XXXX." for example:
11168 this is the same syntax as used in the req config file.
11171 *) Allow certificate extensions to be added to certificate requests. These
11172 are specified in a 'req_extensions' option of the req section of the
11173 config file. They can be printed out with the -text option to req but
11174 are otherwise ignored at present.
11177 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11178 data read consists of only the final block it would not decrypted because
11179 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11180 A misplaced 'break' also meant the decrypted final block might not be
11181 copied until the next read.
11184 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11185 a few extra parameters to the DH structure: these will be useful if
11186 for example we want the value of 'q' or implement X9.42 DH.
11189 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11190 provides hooks that allow the default DSA functions or functions on a
11191 "per key" basis to be replaced. This allows hardware acceleration and
11192 hardware key storage to be handled without major modification to the
11193 library. Also added low level modexp hooks and CRYPTO_EX structure and
11194 associated functions.
11197 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11198 as "read only": it can't be written to and the buffer it points to will
11199 not be freed. Reading from a read only BIO is much more efficient than
11200 a normal memory BIO. This was added because there are several times when
11201 an area of memory needs to be read from a BIO. The previous method was
11202 to create a memory BIO and write the data to it, this results in two
11203 copies of the data and an O(n^2) reading algorithm. There is a new
11204 function BIO_new_mem_buf() which creates a read only memory BIO from
11205 an area of memory. Also modified the PKCS#7 routines to use read only
11209 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11210 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11211 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11212 but a retry condition occurred while trying to read the rest.
11215 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11216 NID_pkcs7_encrypted by default: this was wrong since this should almost
11217 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11218 the encrypted data type: this is a more sensible place to put it and it
11219 allows the PKCS#12 code to be tidied up that duplicated this
11223 *) Changed obj_dat.pl script so it takes its input and output files on
11224 the command line. This should avoid shell escape redirection problems
11228 *) Initial support for certificate extension requests, these are included
11229 in things like Xenroll certificate requests. Included functions to allow
11230 extensions to be obtained and added.
11233 *) -crlf option to s_client and s_server for sending newlines as
11234 CRLF (as required by many protocols).
11237 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11239 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11240 [Ralf S. Engelschall]
11242 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11243 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11245 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11249 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11250 DH parameters/keys (q is lost during that conversion, but the resulting
11251 DH parameters contain its length).
11253 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11254 much faster than DH_generate_parameters (which creates parameters
11255 where p = 2*q + 1), and also the smaller q makes DH computations
11256 much more efficient (160-bit exponentiation instead of 1024-bit
11257 exponentiation); so this provides a convenient way to support DHE
11258 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11259 utter importance to use
11260 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11262 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11263 when such DH parameters are used, because otherwise small subgroup
11264 attacks may become possible!
11267 *) Avoid memory leak in i2d_DHparams.
11270 *) Allow the -k option to be used more than once in the enc program:
11271 this allows the same encrypted message to be read by multiple recipients.
11274 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11275 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11276 it will always use the numerical form of the OID, even if it has a short
11280 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11281 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11282 otherwise bn_mod_exp was called. In the case of hardware keys for example
11283 no private key components need be present and it might store extra data
11284 in the RSA structure, which cannot be accessed from bn_mod_exp.
11285 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11286 private key operations.
11289 *) Added support for SPARC Linux.
11292 *) pem_password_cb function type incompatibly changed from
11293 typedef int pem_password_cb(char *buf, int size, int rwflag);
11295 ....(char *buf, int size, int rwflag, void *userdata);
11296 so that applications can pass data to their callbacks:
11297 The PEM[_ASN1]_{read,write}... functions and macros now take an
11298 additional void * argument, which is just handed through whenever
11299 the password callback is called.
11300 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11302 New function SSL_CTX_set_default_passwd_cb_userdata.
11304 Compatibility note: As many C implementations push function arguments
11305 onto the stack in reverse order, the new library version is likely to
11306 interoperate with programs that have been compiled with the old
11307 pem_password_cb definition (PEM_whatever takes some data that
11308 happens to be on the stack as its last argument, and the callback
11309 just ignores this garbage); but there is no guarantee whatsoever that
11312 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11313 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11314 problems not only on Windows, but also on some Unix platforms.
11315 To avoid problematic command lines, these definitions are now in an
11316 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11317 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11320 *) MIPS III/IV assembler module is reimplemented.
11323 *) More DES library cleanups: remove references to srand/rand and
11324 delete an unused file.
11327 *) Add support for the free Netwide assembler (NASM) under Win32,
11328 since not many people have MASM (ml) and it can be hard to obtain.
11329 This is currently experimental but it seems to work OK and pass all
11330 the tests. Check out INSTALL.W32 for info.
11333 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11334 without temporary keys kept an extra copy of the server key,
11335 and connections with temporary keys did not free everything in case
11339 *) New function RSA_check_key and new openssl rsa option -check
11340 for verifying the consistency of RSA keys.
11341 [Ulf Moeller, Bodo Moeller]
11343 *) Various changes to make Win32 compile work:
11344 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11345 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11346 comparison" warnings.
11347 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11350 *) Add a debugging option to PKCS#5 v2 key generation function: when
11351 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11352 derived keys are printed to stderr.
11355 *) Copy the flags in ASN1_STRING_dup().
11356 [Roman E. Pavlov <pre@mo.msk.ru>]
11358 *) The x509 application mishandled signing requests containing DSA
11359 keys when the signing key was also DSA and the parameters didn't match.
11361 It was supposed to omit the parameters when they matched the signing key:
11362 the verifying software was then supposed to automatically use the CA's
11363 parameters if they were absent from the end user certificate.
11365 Omitting parameters is no longer recommended. The test was also
11366 the wrong way round! This was probably due to unusual behaviour in
11367 EVP_cmp_parameters() which returns 1 if the parameters match.
11368 This meant that parameters were omitted when they *didn't* match and
11369 the certificate was useless. Certificates signed with 'ca' didn't have
11371 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11373 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11374 The interface is as follows:
11375 Applications can use
11376 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11377 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11378 "off" is now the default.
11379 The library internally uses
11380 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11381 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11382 to disable memory-checking temporarily.
11384 Some inconsistent states that previously were possible (and were
11385 even the default) are now avoided.
11387 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11388 with each memory chunk allocated; this is occasionally more helpful
11389 than just having a counter.
11391 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11393 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11397 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11398 which largely parallels "options", but is for changing API behaviour,
11399 whereas "options" are about protocol behaviour.
11400 Initial "mode" flags are:
11402 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11403 a single record has been written.
11404 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11405 retries use the same buffer location.
11406 (But all of the contents must be
11410 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11413 *) Fix problems with no-hmac etc.
11414 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11416 *) New functions RSA_get_default_method(), RSA_set_method() and
11417 RSA_get_method(). These allows replacement of RSA_METHODs without having
11418 to mess around with the internals of an RSA structure.
11421 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11422 Also really enable memory leak checks in openssl.c and in some
11424 [Chad C. Mulligan, Bodo Moeller]
11426 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11427 up the length of negative integers. This has now been simplified to just
11428 store the length when it is first determined and use it later, rather
11429 than trying to keep track of where data is copied and updating it to
11431 [Steve Henson, reported by Brien Wheeler
11432 <bwheeler@authentica-security.com>]
11434 *) Add a new function PKCS7_signatureVerify. This allows the verification
11435 of a PKCS#7 signature but with the signing certificate passed to the
11436 function itself. This contrasts with PKCS7_dataVerify which assumes the
11437 certificate is present in the PKCS#7 structure. This isn't always the
11438 case: certificates can be omitted from a PKCS#7 structure and be
11439 distributed by "out of band" means (such as a certificate database).
11442 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11443 function prototypes in pem.h, also change util/mkdef.pl to add the
11444 necessary function names.
11447 *) mk1mf.pl (used by Windows builds) did not properly read the
11448 options set by Configure in the top level Makefile, and Configure
11449 was not even able to write more than one option correctly.
11450 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11453 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11454 file to be loaded from a BIO or FILE pointer. The BIO version will
11455 for example allow memory BIOs to contain config info.
11458 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11459 Whoever hopes to achieve shared-library compatibility across versions
11460 must use this, not the compile-time macro.
11461 (Exercise 0.9.4: Which is the minimum library version required by
11463 Note: All this applies only to multi-threaded programs, others don't
11467 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11468 through a BIO pair triggered the default case, i.e.
11469 SSLerr(...,SSL_R_UNKNOWN_STATE).
11472 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11473 can use the SSL library even if none of the specific BIOs is
11477 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11478 for the encoded length.
11479 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11481 *) Add initial documentation of the X509V3 functions.
11484 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11485 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11486 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11487 secure PKCS#8 private key format with a high iteration count.
11490 *) Fix determination of Perl interpreter: A perl or perl5
11491 _directory_ in $PATH was also accepted as the interpreter.
11492 [Ralf S. Engelschall]
11494 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11495 wrong with it but it was very old and did things like calling
11496 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11497 unusual formatting.
11500 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11501 to use the new extension code.
11504 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11505 with macros. This should make it easier to change their form, add extra
11506 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11510 *) Add to configuration table a new entry that can specify an alternative
11511 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11512 according to Mark Crispin <MRC@Panda.COM>.
11516 *) DES CBC did not update the IV. Weird.
11519 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11520 Changing the behaviour of the former might break existing programs --
11521 where IV updating is needed, des_ncbc_encrypt can be used.
11524 *) When bntest is run from "make test" it drives bc to check its
11525 calculations, as well as internally checking them. If an internal check
11526 fails, it needs to cause bc to give a non-zero result or make test carries
11527 on without noticing the failure. Fixed.
11530 *) DES library cleanups.
11533 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11534 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11535 ciphers. NOTE: although the key derivation function has been verified
11536 against some published test vectors it has not been extensively tested
11537 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11541 *) Instead of "mkdir -p", which is not fully portable, use new
11542 Perl script "util/mkdir-p.pl".
11545 *) Rewrite the way password based encryption (PBE) is handled. It used to
11546 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11547 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11548 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11549 the 'parameter' field of the AlgorithmIdentifier is passed to the
11550 underlying key generation function so it must do its own ASN1 parsing.
11551 This has also changed the EVP_PBE_CipherInit() function which now has a
11552 'parameter' argument instead of literal salt and iteration count values
11553 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11556 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11557 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11558 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11559 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11560 value was just used as a "magic string" and not used directly its
11561 value doesn't matter.
11564 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11568 *) "linux-sparc64" configuration (ultrapenguin).
11569 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11570 "linux-sparc" configuration.
11571 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11573 *) config now generates no-xxx options for missing ciphers.
11576 *) Support the EBCDIC character set (work in progress).
11577 File ebcdic.c not yet included because it has a different license.
11578 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11580 *) Support BS2000/OSD-POSIX.
11581 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11583 *) Make callbacks for key generation use void * instead of char *.
11586 *) Make S/MIME samples compile (not yet tested).
11589 *) Additional typesafe stacks.
11592 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11596 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11598 *) New configuration variant "sco5-gcc".
11600 *) Updated some demos.
11601 [Sean O Riordain, Wade Scholine]
11603 *) Add missing BIO_free at exit of pkcs12 application.
11606 *) Fix memory leak in conf.c.
11609 *) Updates for Win32 to assembler version of MD5.
11612 *) Set #! path to perl in apps/der_chop to where we found it
11613 instead of using a fixed path.
11616 *) SHA library changes for irix64-mips4-cc.
11619 *) Improvements for VMS support.
11623 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11625 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11626 This also avoids the problems with SC4.2 and unpatched SC5.
11627 [Andy Polyakov <appro@fy.chalmers.se>]
11629 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11630 These are required because of the typesafe stack would otherwise break
11631 existing code. If old code used a structure member which used to be STACK
11632 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11633 sk_num or sk_value it would produce an error because the num, data members
11634 are not present in STACK_OF. Now it just produces a warning. sk_set
11635 replaces the old method of assigning a value to sk_value
11636 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11637 that does this will no longer work (and should use sk_set instead) but
11638 this could be regarded as a "questionable" behaviour anyway.
11641 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11642 correctly handle encrypted S/MIME data.
11645 *) Change type of various DES function arguments from des_cblock
11646 (which means, in function argument declarations, pointer to char)
11647 to des_cblock * (meaning pointer to array with 8 char elements),
11648 which allows the compiler to do more typechecking; it was like
11649 that back in SSLeay, but with lots of ugly casts.
11651 Introduce new type const_des_cblock.
11654 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11655 problems: find RecipientInfo structure that matches recipient certificate
11656 and initialise the ASN1 structures properly based on passed cipher.
11659 *) Belatedly make the BN tests actually check the results.
11662 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11663 to and from BNs: it was completely broken. New compilation option
11664 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11665 key elements as negative integers.
11668 *) Reorganize and speed up MD5.
11669 [Andy Polyakov <appro@fy.chalmers.se>]
11672 [Richard Levitte <richard@levitte.org>]
11674 *) New option -out to asn1parse to allow the parsed structure to be
11675 output to a file. This is most useful when combined with the -strparse
11676 option to examine the output of things like OCTET STRINGS.
11679 *) Make SSL library a little more fool-proof by not requiring any longer
11680 that SSL_set_{accept,connect}_state be called before
11681 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11682 in many applications because usually everything *appeared* to work as
11683 intended anyway -- now it really works as intended).
11686 *) Move openssl.cnf out of lib/.
11689 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11690 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11691 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11692 [Ralf S. Engelschall]
11694 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11695 handle PKCS#7 enveloped data properly.
11696 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11698 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11699 copying pointers. The cert_st handling is changed by this in
11700 various ways (and thus what used to be known as ctx->default_cert
11701 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11702 any longer when s->cert does not give us what we need).
11703 ssl_cert_instantiate becomes obsolete by this change.
11704 As soon as we've got the new code right (possibly it already is?),
11705 we have solved a couple of bugs of the earlier code where s->cert
11706 was used as if it could not have been shared with other SSL structures.
11708 Note that using the SSL API in certain dirty ways now will result
11709 in different behaviour than observed with earlier library versions:
11710 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11711 does not influence s as it used to.
11713 In order to clean up things more thoroughly, inside SSL_SESSION
11714 we don't use CERT any longer, but a new structure SESS_CERT
11715 that holds per-session data (if available); currently, this is
11716 the peer's certificate chain and, for clients, the server's certificate
11717 and temporary key. CERT holds only those values that can have
11718 meaningful defaults in an SSL_CTX.
11721 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11722 from the internal representation. Various PKCS#7 fixes: remove some
11723 evil casts and set the enc_dig_alg field properly based on the signing
11727 *) Allow PKCS#12 password to be set from the command line or the
11728 environment. Let 'ca' get its config file name from the environment
11729 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11733 *) Allow certificate policies extension to use an IA5STRING for the
11734 organization field. This is contrary to the PKIX definition but
11735 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11739 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11740 without disallowing inline assembler and the like for non-pedantic builds.
11743 *) Support Borland C++ builder.
11744 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11746 *) Support Mingw32.
11749 *) SHA-1 cleanups and performance enhancements.
11750 [Andy Polyakov <appro@fy.chalmers.se>]
11752 *) Sparc v8plus assembler for the bignum library.
11753 [Andy Polyakov <appro@fy.chalmers.se>]
11755 *) Accept any -xxx and +xxx compiler options in Configure.
11758 *) Update HPUX configuration.
11761 *) Add missing sk_<type>_unshift() function to safestack.h
11762 [Ralf S. Engelschall]
11764 *) New function SSL_CTX_use_certificate_chain_file that sets the
11765 "extra_cert"s in addition to the certificate. (This makes sense
11766 only for "PEM" format files, as chains as a whole are not
11770 *) Support verify_depth from the SSL API.
11771 x509_vfy.c had what can be considered an off-by-one-error:
11772 Its depth (which was not part of the external interface)
11773 was actually counting the number of certificates in a chain;
11774 now it really counts the depth.
11777 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11778 instead of X509err, which often resulted in confusing error
11779 messages since the error codes are not globally unique
11780 (e.g. an alleged error in ssl3_accept when a certificate
11781 didn't match the private key).
11783 *) New function SSL_CTX_set_session_id_context that allows to set a default
11784 value (so that you don't need SSL_set_session_id_context for each
11785 connection using the SSL_CTX).
11788 *) OAEP decoding bug fix.
11791 *) Support INSTALL_PREFIX for package builders, as proposed by
11795 *) New Configure options "threads" and "no-threads". For systems
11796 where the proper compiler options are known (currently Solaris
11797 and Linux), "threads" is the default.
11800 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11803 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11804 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11805 such as /usr/local/bin.
11808 *) "make linux-shared" to build shared libraries.
11809 [Niels Poppe <niels@netbox.org>]
11811 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11814 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11815 extension adding in x509 utility.
11818 *) Remove NOPROTO sections and error code comments.
11821 *) Partial rewrite of the DEF file generator to now parse the ANSI
11825 *) New Configure options --prefix=DIR and --openssldir=DIR.
11828 *) Complete rewrite of the error code script(s). It is all now handled
11829 by one script at the top level which handles error code gathering,
11830 header rewriting and C source file generation. It should be much better
11831 than the old method: it now uses a modified version of Ulf's parser to
11832 read the ANSI prototypes in all header files (thus the old K&R definitions
11833 aren't needed for error creation any more) and do a better job of
11834 translating function codes into names. The old 'ASN1 error code imbedded
11835 in a comment' is no longer necessary and it doesn't use .err files which
11836 have now been deleted. Also the error code call doesn't have to appear all
11837 on one line (which resulted in some large lines...).
11840 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11843 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11844 0 (which usually indicates a closed connection), but continue reading.
11847 *) Fix some race conditions.
11850 *) Add support for CRL distribution points extension. Add Certificate
11851 Policies and CRL distribution points documentation.
11854 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11857 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11858 8 of keying material. Merlin has also confirmed interop with this fix
11859 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11860 [Merlin Hughes <merlin@baltimore.ie>]
11862 *) Fix lots of warnings.
11863 [Richard Levitte <levitte@stacken.kth.se>]
11865 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11866 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11867 [Richard Levitte <levitte@stacken.kth.se>]
11869 *) Fix problems with sizeof(long) == 8.
11870 [Andy Polyakov <appro@fy.chalmers.se>]
11872 *) Change functions to ANSI C.
11875 *) Fix typos in error codes.
11876 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11878 *) Remove defunct assembler files from Configure.
11881 *) SPARC v8 assembler BIGNUM implementation.
11882 [Andy Polyakov <appro@fy.chalmers.se>]
11884 *) Support for Certificate Policies extension: both print and set.
11885 Various additions to support the r2i method this uses.
11888 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11889 return a const string when you are expecting an allocated buffer.
11892 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11893 types DirectoryString and DisplayText.
11896 *) Add code to allow r2i extensions to access the configuration database,
11897 add an LHASH database driver and add several ctx helper functions.
11900 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11901 fail when they extended the size of a BIGNUM.
11904 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11905 support typesafe stack.
11908 *) Fix typo in SSL_[gs]et_options().
11909 [Nils Frostberg <nils@medcom.se>]
11911 *) Delete various functions and files that belonged to the (now obsolete)
11912 old X509V3 handling code.
11915 *) New Configure option "rsaref".
11918 *) Don't auto-generate pem.h.
11921 *) Introduce type-safe ASN.1 SETs.
11924 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11925 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11927 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11928 that links with OpenSSL (well at least cause lots of warnings), but fear
11929 not: the conversion is trivial, and it eliminates loads of evil casts. A
11930 few STACKed things have been converted already. Feel free to convert more.
11931 In the fullness of time, I'll do away with the STACK type altogether.
11934 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11935 specified in <certfile> by updating the entry in the index.txt file.
11936 This way one no longer has to edit the index.txt file manually for
11937 revoking a certificate. The -revoke option does the gory details now.
11938 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11940 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11941 `-text' option at all and this way the `-noout -text' combination was
11942 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11943 [Ralf S. Engelschall]
11945 *) Make sure a corresponding plain text error message exists for the
11946 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11947 verify callback function determined that a certificate was revoked.
11948 [Ralf S. Engelschall]
11950 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11951 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11952 all available ciphers including rc5, which was forgotten until now.
11953 In order to let the testing shell script know which algorithms
11954 are available, a new (up to now undocumented) command
11955 "openssl list-cipher-commands" is used.
11958 *) Bugfix: s_client occasionally would sleep in select() when
11959 it should have checked SSL_pending() first.
11962 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11963 the raw DSA values prior to ASN.1 encoding.
11966 *) Tweaks to Configure
11967 [Niels Poppe <niels@netbox.org>]
11969 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11973 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11976 *) New config option to avoid instructions that are illegal on the 80386.
11977 The default code is faster, but requires at least a 486.
11980 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11981 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11982 same as SSL2_VERSION anyway.
11985 *) New "-showcerts" option for s_client.
11988 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11989 application. Various cleanups and fixes.
11992 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11993 modify error routines to work internally. Add error codes and PBE init
11994 to library startup routines.
11997 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11998 packing functions to asn1 and evp. Changed function names and error
11999 codes along the way.
12002 *) PKCS12 integration: and so it begins... First of several patches to
12003 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12004 objects to objects.h
12007 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12008 and display support for Thawte strong extranet extension.
12011 *) Add LinuxPPC support.
12012 [Jeff Dubrule <igor@pobox.org>]
12014 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12015 bn_div_words in alpha.s.
12016 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12018 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12019 OAEP isn't supported when OpenSSL is built with RSAref.
12020 [Ulf Moeller <ulf@fitug.de>]
12022 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12023 so they no longer are missing under -DNOPROTO.
12024 [Soren S. Jorvang <soren@t.dk>]
12027 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12029 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12030 doesn't work when the session is reused. Coming soon!
12033 *) Fix a security hole, that allows sessions to be reused in the wrong
12034 context thus bypassing client cert protection! All software that uses
12035 client certs and session caches in multiple contexts NEEDS PATCHING to
12036 allow session reuse! A fuller solution is in the works.
12037 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12039 *) Some more source tree cleanups (removed obsolete files
12040 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12041 permission on "config" script to be executable) and a fix for the INSTALL
12043 [Ulf Moeller <ulf@fitug.de>]
12045 *) Remove some legacy and erroneous uses of malloc, free instead of
12047 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12049 *) Make rsa_oaep_test return non-zero on error.
12050 [Ulf Moeller <ulf@fitug.de>]
12052 *) Add support for native Solaris shared libraries. Configure
12053 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12054 if someone would make that last step automatic.
12055 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12057 *) ctx_size was not built with the right compiler during "make links". Fixed.
12060 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12061 except NULL ciphers". This means the default cipher list will no longer
12062 enable NULL ciphers. They need to be specifically enabled e.g. with
12063 the string "DEFAULT:eNULL".
12066 *) Fix to RSA private encryption routines: if p < q then it would
12067 occasionally produce an invalid result. This will only happen with
12068 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12071 *) Be less restrictive and allow also `perl util/perlpath.pl
12072 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12073 because this way one can also use an interpreter named `perl5' (which is
12074 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12075 installed as `perl').
12076 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12078 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12079 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12081 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12082 advapi32.lib to Win32 build and change the pem test comparison
12083 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12084 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12085 and crypto/des/ede_cbcm_enc.c.
12088 *) DES quad checksum was broken on big-endian architectures. Fixed.
12091 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12092 Win32 test batch file so it (might) work again. The Win32 test batch file
12093 is horrible: I feel ill....
12096 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12097 in e_os.h. Audit of header files to check ANSI and non ANSI
12098 sections: 10 functions were absent from non ANSI section and not exported
12099 from Windows DLLs. Fixed up libeay.num for new functions.
12102 *) Make `openssl version' output lines consistent.
12103 [Ralf S. Engelschall]
12105 *) Fix Win32 symbol export lists for BIO functions: Added
12106 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12107 to ms/libeay{16,32}.def.
12108 [Ralf S. Engelschall]
12110 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12111 fine under Unix and passes some trivial tests I've now added. But the
12112 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12113 added to make sure no one expects that this stuff really works in the
12114 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12115 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12117 [Ralf S. Engelschall]
12119 *) Fix the generation of two part addresses in perl.
12120 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12122 *) Add config entry for Linux on MIPS.
12123 [John Tobey <jtobey@channel1.com>]
12125 *) Make links whenever Configure is run, unless we are on Windoze.
12128 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12129 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12133 *) Add a useful kludge to allow package maintainers to specify compiler and
12134 other platforms details on the command line without having to patch the
12135 Configure script everytime: One now can use ``perl Configure
12136 <id>:<details>'', i.e. platform ids are allowed to have details appended
12137 to them (separated by colons). This is treated as there would be a static
12138 pre-configured entry in Configure's %table under key <id> with value
12139 <details> and ``perl Configure <id>'' is called. So, when you want to
12140 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12141 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12142 now, which overrides the FreeBSD-elf entry on-the-fly.
12143 [Ralf S. Engelschall]
12145 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12148 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12149 on the `perl Configure ...' command line. This way one can compile
12150 OpenSSL libraries with Position Independent Code (PIC) which is needed
12151 for linking it into DSOs.
12152 [Ralf S. Engelschall]
12154 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12158 *) Cleaned up the LICENSE document: The official contact for any license
12159 questions now is the OpenSSL core team under openssl-core@openssl.org.
12160 And add a paragraph about the dual-license situation to make sure people
12161 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12162 to the OpenSSL toolkit.
12163 [Ralf S. Engelschall]
12165 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12166 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12167 Additionally cleaned up the `make links' target: Remove unnecessary
12168 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12169 to speed processing and no longer clutter the display with confusing
12170 stuff. Instead only the actually done links are displayed.
12171 [Ralf S. Engelschall]
12173 *) Permit null encryption ciphersuites, used for authentication only. It used
12174 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12175 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12179 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12180 signed attributes when verifying signatures (this would break them),
12181 the detached data encoding was wrong and public keys obtained using
12182 X509_get_pubkey() weren't freed.
12185 *) Add text documentation for the BUFFER functions. Also added a work around
12186 to a Win95 console bug. This was triggered by the password read stuff: the
12187 last character typed gets carried over to the next fread(). If you were
12188 generating a new cert request using 'req' for example then the last
12189 character of the passphrase would be CR which would then enter the first
12193 *) Added the new `Includes OpenSSL Cryptography Software' button as
12194 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12195 button and can be used by applications based on OpenSSL to show the
12196 relationship to the OpenSSL project.
12197 [Ralf S. Engelschall]
12199 *) Remove confusing variables in function signatures in files
12200 ssl/ssl_lib.c and ssl/ssl.h.
12201 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12203 *) Don't install bss_file.c under PREFIX/include/
12204 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12206 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12207 functions that return function pointers and has support for NT specific
12208 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12209 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12210 unsigned to signed types: this was killing the Win32 compile.
12213 *) Add new certificate file to stack functions,
12214 SSL_add_dir_cert_subjects_to_stack() and
12215 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12216 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12217 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12218 This means that Apache-SSL and similar packages don't have to mess around
12219 to add as many CAs as they want to the preferred list.
12222 *) Experiment with doxygen documentation. Currently only partially applied to
12224 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12225 openssl.doxy as the configuration file.
12228 *) Get rid of remaining C++-style comments which strict C compilers hate.
12229 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12231 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12232 compiled in by default: it has problems with large keys.
12235 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12236 DH private keys and/or callback functions which directly correspond to
12237 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12238 is needed for applications which have to configure certificates on a
12239 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12241 For the RSA certificate situation is makes no difference, but
12242 for the DSA certificate situation this fixes the "no shared cipher"
12243 problem where the OpenSSL cipher selection procedure failed because the
12244 temporary keys were not overtaken from the context and the API provided
12245 no way to reconfigure them.
12246 The new functions now let applications reconfigure the stuff and they
12247 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12248 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12249 non-public-API function ssl_cert_instantiate() is used as a helper
12250 function and also to reduce code redundancy inside ssl_rsa.c.
12251 [Ralf S. Engelschall]
12253 *) Move s_server -dcert and -dkey options out of the undocumented feature
12254 area because they are useful for the DSA situation and should be
12255 recognized by the users.
12256 [Ralf S. Engelschall]
12258 *) Fix the cipher decision scheme for export ciphers: the export bits are
12259 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12260 SSL_EXP_MASK. So, the original variable has to be used instead of the
12261 already masked variable.
12262 [Richard Levitte <levitte@stacken.kth.se>]
12264 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12265 [Richard Levitte <levitte@stacken.kth.se>]
12267 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12268 from `int' to `unsigned int' because it's a length and initialized by
12269 EVP_DigestFinal() which expects an `unsigned int *'.
12270 [Richard Levitte <levitte@stacken.kth.se>]
12272 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12273 script. Instead use the usual Shell->Perl transition trick.
12274 [Ralf S. Engelschall]
12276 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12277 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12278 -noout -modulus' as it's already the case for `openssl rsa -noout
12279 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12280 currently the public key is printed (a decision which was already done by
12281 `openssl dsa -modulus' in the past) which serves a similar purpose.
12282 Additionally the NO_RSA no longer completely removes the whole -modulus
12283 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12285 [Ralf S. Engelschall]
12287 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12288 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12289 [Arne Ansper <arne@ats.cyber.ee>]
12291 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12292 to be added. Now both 'req' and 'ca' can use new objects defined in the
12296 *) Add cool BIO that does syslog (or event log on NT).
12297 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12299 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12300 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12301 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12302 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12305 *) Add preliminary config info for new extension code.
12308 *) Make RSA_NO_PADDING really use no padding.
12309 [Ulf Moeller <ulf@fitug.de>]
12311 *) Generate errors when private/public key check is done.
12314 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12315 for some CRL extensions and new objects added.
12318 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12319 key usage extension and fuller support for authority key id.
12322 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12323 padding method for RSA, which is recommended for new applications in PKCS
12324 #1 v2.0 (RFC 2437, October 1998).
12325 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12326 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12327 against Bleichbacher's attack on RSA.
12328 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12331 *) Updates to the new SSL compression code
12332 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12334 *) Fix so that the version number in the master secret, when passed
12335 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12336 (because the server will not accept higher), that the version number
12337 is 0x03,0x01, not 0x03,0x00
12338 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12340 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12341 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12342 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12345 *) Support for RAW extensions where an arbitrary extension can be
12346 created by including its DER encoding. See apps/openssl.cnf for
12350 *) Make sure latest Perl versions don't interpret some generated C array
12351 code as Perl array code in the crypto/err/err_genc.pl script.
12352 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12354 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12355 not many people have the assembler. Various Win32 compilation fixes and
12356 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12357 build instructions.
12360 *) Modify configure script 'Configure' to automatically create crypto/date.h
12361 file under Win32 and also build pem.h from pem.org. New script
12362 util/mkfiles.pl to create the MINFO file on environments that can't do a
12363 'make files': perl util/mkfiles.pl >MINFO should work.
12366 *) Major rework of DES function declarations, in the pursuit of correctness
12367 and purity. As a result, many evil casts evaporated, and some weirdness,
12368 too. You may find this causes warnings in your code. Zapping your evil
12369 casts will probably fix them. Mostly.
12372 *) Fix for a typo in asn1.h. Bug fix to object creation script
12373 obj_dat.pl. It considered a zero in an object definition to mean
12374 "end of object": none of the objects in objects.h have any zeros
12375 so it wasn't spotted.
12376 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12378 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12379 Masking (CBCM). In the absence of test vectors, the best I have been able
12380 to do is check that the decrypt undoes the encrypt, so far. Send me test
12381 vectors if you have them.
12384 *) Correct calculation of key length for export ciphers (too much space was
12385 allocated for null ciphers). This has not been tested!
12388 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12389 message is now correct (it understands "crypto" and "ssl" on its
12390 command line). There is also now an "update" option. This will update
12391 the util/ssleay.num and util/libeay.num files with any new functions.
12393 perl util/mkdef.pl crypto ssl update
12394 it will update them.
12397 *) Overhauled the Perl interface (perl/*):
12398 - ported BN stuff to OpenSSL's different BN library
12399 - made the perl/ source tree CVS-aware
12400 - renamed the package from SSLeay to OpenSSL (the files still contain
12401 their history because I've copied them in the repository)
12402 - removed obsolete files (the test scripts will be replaced
12403 by better Test::Harness variants in the future)
12404 [Ralf S. Engelschall]
12406 *) First cut for a very conservative source tree cleanup:
12407 1. merge various obsolete readme texts into doc/ssleay.txt
12408 where we collect the old documents and readme texts.
12409 2. remove the first part of files where I'm already sure that we no
12410 longer need them because of three reasons: either they are just temporary
12411 files which were left by Eric or they are preserved original files where
12412 I've verified that the diff is also available in the CVS via "cvs diff
12413 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12414 the crypto/md/ stuff).
12415 [Ralf S. Engelschall]
12417 *) More extension code. Incomplete support for subject and issuer alt
12418 name, issuer and authority key id. Change the i2v function parameters
12419 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12420 what that's for :-) Fix to ASN1 macro which messed up
12421 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12424 *) Preliminary support for ENUMERATED type. This is largely copied from the
12428 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12429 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12431 *) Make sure `make rehash' target really finds the `openssl' program.
12432 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12434 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12435 like to hear about it if this slows down other processors.
12438 *) Add CygWin32 platform information to Configure script.
12439 [Alan Batie <batie@aahz.jf.intel.com>]
12441 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12442 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12444 *) New program nseq to manipulate netscape certificate sequences
12447 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12451 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12452 but the BN code had some problems that would cause failures when
12453 doing certificate verification and some other functions.
12454 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12456 *) Add ASN1 and PEM code to support netscape certificate sequences.
12459 *) Add ASN1 and PEM code to support netscape certificate sequences.
12462 *) Add several PKIX and private extended key usage OIDs.
12465 *) Modify the 'ca' program to handle the new extension code. Modify
12466 openssl.cnf for new extension format, add comments.
12469 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12470 and add a sample to openssl.cnf so req -x509 now adds appropriate
12474 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12475 error code, add initial support to X509_print() and x509 application.
12478 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12479 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12480 stuff is currently isolated and isn't even compiled yet.
12483 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12484 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12485 Removed the versions check from X509 routines when loading extensions:
12486 this allows certain broken certificates that don't set the version
12487 properly to be processed.
12490 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12491 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12492 can still be regenerated with "make depend".
12495 *) Spelling mistake in C version of CAST-128.
12496 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12498 *) Changes to the error generation code. The perl script err-code.pl
12499 now reads in the old error codes and retains the old numbers, only
12500 adding new ones if necessary. It also only changes the .err files if new
12501 codes are added. The makefiles have been modified to only insert errors
12502 when needed (to avoid needlessly modifying header files). This is done
12503 by only inserting errors if the .err file is newer than the auto generated
12504 C file. To rebuild all the error codes from scratch (the old behaviour)
12505 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12506 or delete all the .err files.
12509 *) CAST-128 was incorrectly implemented for short keys. The C version has
12510 been fixed, but is untested. The assembler versions are also fixed, but
12511 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12512 to regenerate it if needed.
12513 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12514 Hagino <itojun@kame.net>]
12516 *) File was opened incorrectly in randfile.c.
12517 [Ulf Möller <ulf@fitug.de>]
12519 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12520 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12521 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12522 al: it's just almost always a UTCTime. Note this patch adds new error
12523 codes so do a "make errors" if there are problems.
12526 *) Correct Linux 1 recognition in config.
12527 [Ulf Möller <ulf@fitug.de>]
12529 *) Remove pointless MD5 hash when using DSA keys in ca.
12530 [Anonymous <nobody@replay.com>]
12532 *) Generate an error if given an empty string as a cert directory. Also
12533 generate an error if handed NULL (previously returned 0 to indicate an
12534 error, but didn't set one).
12535 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12537 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12540 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12541 parameters. This was causing a warning which killed off the Win32 compile.
12544 *) Remove C++ style comments from crypto/bn/bn_local.h.
12545 [Neil Costigan <neil.costigan@celocom.com>]
12547 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12548 based on a text string, looking up short and long names and finally
12549 "dot" format. The "dot" format stuff didn't work. Added new function
12550 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12551 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12552 OID is not part of the table.
12555 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12556 X509_LOOKUP_by_alias().
12559 *) Sort openssl functions by name.
12562 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12563 encryption from sample DSA keys (in case anyone is interested the password
12567 *) Make _all_ *_free functions accept a NULL pointer.
12568 [Frans Heymans <fheymans@isaserver.be>]
12570 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12572 [Anonymous <nobody@replay.com>]
12574 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12575 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12577 *) Don't blow it for numeric -newkey arguments to apps/req.
12578 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12580 *) Temp key "for export" tests were wrong in s3_srvr.c.
12581 [Anonymous <nobody@replay.com>]
12583 *) Add prototype for temp key callback functions
12584 SSL_CTX_set_tmp_{rsa,dh}_callback().
12587 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12588 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12591 *) X509_name_add_entry() freed the wrong thing after an error.
12592 [Arne Ansper <arne@ats.cyber.ee>]
12594 *) rsa_eay.c would attempt to free a NULL context.
12595 [Arne Ansper <arne@ats.cyber.ee>]
12597 *) BIO_s_socket() had a broken should_retry() on Windoze.
12598 [Arne Ansper <arne@ats.cyber.ee>]
12600 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12601 [Arne Ansper <arne@ats.cyber.ee>]
12603 *) Make sure the already existing X509_STORE->depth variable is initialized
12604 in X509_STORE_new(), but document the fact that this variable is still
12605 unused in the certificate verification process.
12606 [Ralf S. Engelschall]
12608 *) Fix the various library and apps files to free up pkeys obtained from
12609 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12612 *) Fix reference counting in X509_PUBKEY_get(). This makes
12613 demos/maurice/example2.c work, amongst others, probably.
12614 [Steve Henson and Ben Laurie]
12616 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12617 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12618 are no longer created. This way we have a single and consistent command
12619 line interface `openssl <command>', similar to `cvs <command>'.
12620 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12622 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12623 BIT STRING wrapper always have zero unused bits.
12626 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12629 *) Make the top-level INSTALL documentation easier to understand.
12632 *) Makefiles updated to exit if an error occurs in a sub-directory
12633 make (including if user presses ^C) [Paul Sutton]
12635 *) Make Montgomery context stuff explicit in RSA data structure.
12638 *) Fix build order of pem and err to allow for generated pem.h.
12641 *) Fix renumbering bug in X509_NAME_delete_entry().
12644 *) Enhanced the err-ins.pl script so it makes the error library number
12645 global and can add a library name. This is needed for external ASN1 and
12646 other error libraries.
12649 *) Fixed sk_insert which never worked properly.
12652 *) Fix ASN1 macros so they can handle indefinite length constructed
12653 EXPLICIT tags. Some non standard certificates use these: they can now
12657 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12658 into a single doc/ssleay.txt bundle. This way the information is still
12659 preserved but no longer messes up this directory. Now it's new room for
12660 the new set of documentation files.
12661 [Ralf S. Engelschall]
12663 *) SETs were incorrectly DER encoded. This was a major pain, because they
12664 shared code with SEQUENCEs, which aren't coded the same. This means that
12665 almost everything to do with SETs or SEQUENCEs has either changed name or
12666 number of arguments.
12667 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12669 *) Fix test data to work with the above.
12672 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12673 was already fixed by Eric for 0.9.1 it seems.
12674 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12676 *) Autodetect FreeBSD3.
12679 *) Fix various bugs in Configure. This affects the following platforms:
12683 unixware-2.0-pentium
12687 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12688 before they are needed.
12691 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12695 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12697 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12698 changed SSLeay to OpenSSL in version strings.
12699 [Ralf S. Engelschall]
12701 *) Some fixups to the top-level documents.
12704 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12705 because the symlink to include/ was missing.
12706 [Ralf S. Engelschall]
12708 *) Incorporated the popular no-RSA/DSA-only patches
12709 which allow to compile a RSA-free SSLeay.
12710 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12712 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12713 when "ssleay" is still not found.
12714 [Ralf S. Engelschall]
12716 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12717 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12719 *) Updated the README file.
12720 [Ralf S. Engelschall]
12722 *) Added various .cvsignore files in the CVS repository subdirs
12723 to make a "cvs update" really silent.
12724 [Ralf S. Engelschall]
12726 *) Recompiled the error-definition header files and added
12727 missing symbols to the Win32 linker tables.
12728 [Ralf S. Engelschall]
12730 *) Cleaned up the top-level documents;
12731 o new files: CHANGES and LICENSE
12732 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12733 o merged COPYRIGHT into LICENSE
12734 o removed obsolete TODO file
12735 o renamed MICROSOFT to INSTALL.W32
12736 [Ralf S. Engelschall]
12738 *) Removed dummy files from the 0.9.1b source tree:
12739 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12740 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12741 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12742 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12743 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12744 [Ralf S. Engelschall]
12746 *) Added various platform portability fixes.
12749 *) The Genesis of the OpenSSL rpject:
12750 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12751 Young and Tim J. Hudson created while they were working for C2Net until
12753 [The OpenSSL Project]
12756 Changes between 0.9.0b and 0.9.1b [not released]
12758 *) Updated a few CA certificates under certs/
12761 *) Changed some BIGNUM api stuff.
12764 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12765 DGUX x86, Linux Alpha, etc.
12768 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12769 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12773 *) Add -strparse option to asn1pars program which parses nested
12775 [Dr Stephen Henson <shenson@bigfoot.com>]
12777 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12780 *) DSA fix for "ca" program.
12783 *) Added "-genkey" option to "dsaparam" program.
12786 *) Added RIPE MD160 (rmd160) message digest.
12789 *) Added -a (all) option to "ssleay version" command.
12792 *) Added PLATFORM define which is the id given to Configure.
12795 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12798 *) Extended the ASN.1 parser routines.
12801 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12804 *) Added a BN_CTX to the BN library.
12807 *) Fixed the weak key values in DES library
12810 *) Changed API in EVP library for cipher aliases.
12813 *) Added support for RC2/64bit cipher.
12816 *) Converted the lhash library to the crypto/mem.c functions.
12819 *) Added more recognized ASN.1 object ids.
12822 *) Added more RSA padding checks for SSL/TLS.
12825 *) Added BIO proxy/filter functionality.
12828 *) Added extra_certs to SSL_CTX which can be used
12829 send extra CA certificates to the client in the CA cert chain sending
12830 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12833 *) Now Fortezza is denied in the authentication phase because
12834 this is key exchange mechanism is not supported by SSLeay at all.
12837 *) Additional PKCS1 checks.
12840 *) Support the string "TLSv1" for all TLS v1 ciphers.
12843 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12844 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12847 *) Fixed a few memory leaks.
12850 *) Fixed various code and comment typos.
12853 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12854 bytes sent in the client random.
12855 [Edward Bishop <ebishop@spyglass.com>]