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 *) Added new public header file <openssl/rand_drbg.h> and documentation
13 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
16 *) QNX support removed (cannot find contributors to get their approval
17 for the license change).
20 *) TLSv1.3 replay protection for early data has been implemented. See the
21 SSL_read_early_data() man page for further details.
24 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
25 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
26 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
27 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
28 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
29 configuraton has been separated out. See the ciphers man page or the
30 SSL_CTX_set_ciphersuites() man page for more information.
33 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
34 in responder mode now supports the new "-multi" option, which
35 spawns the specified number of child processes to handle OCSP
36 requests. The "-timeout" option now also limits the OCSP
37 responder's patience to wait to receive the full client request
38 on a newly accepted connection. Child processes are respawned
39 as needed, and the CA index file is automatically reloaded
40 when changed. This makes it possible to run the "ocsp" responder
41 as a long-running service, making the OpenSSL CA somewhat more
42 feature-complete. In this mode, most diagnostic messages logged
43 after entering the event loop are logged via syslog(3) rather than
47 *) Added support for X448 and Ed448. Heavily based on original work by
51 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
52 objects loaded. This adds the functions OSSL_STORE_expect() and
53 OSSL_STORE_find() as well as needed tools to construct searches and
54 get the search data out of them.
57 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
58 version of OpenSSL should review their configuration settings to ensure
59 that they are still appropriate for TLSv1.3. For further information see:
60 https://www.openssl.org/blog/blog/2018/02/08/tlsv1.3/
62 NOTE: In this pre-release of OpenSSL a draft version of the
63 TLSv1.3 standard has been implemented. Implementations of different draft
64 versions of the standard do not inter-operate, and this version will not
65 inter-operate with an implementation of the final standard when it is
66 eventually published. Different pre-release versions may implement
67 different versions of the draft. The final version of OpenSSL 1.1.1 will
68 implement the final version of the standard.
69 TODO(TLS1.3): Remove the above note before final release
72 *) Grand redesign of the OpenSSL random generator
74 The default RAND method now utilizes an AES-CTR DRBG according to
75 NIST standard SP 800-90Ar1. The new random generator is essentially
76 a port of the default random generator from the OpenSSL FIPS 2.0
77 object module. It is a hybrid deterministic random bit generator
78 using an AES-CTR bit stream and which seeds and reseeds itself
79 automatically using trusted system entropy sources.
81 Some of its new features are:
82 o Support for multiple DRBG instances with seed chaining.
83 o Add a public DRBG instance for the default RAND method.
84 o Add a dedicated DRBG instance for generating long term private keys.
85 o Make the DRBG instances fork-safe.
86 o Keep all global DRBG instances on the secure heap if it is enabled.
87 o Add a DRBG instance to every SSL instance for lock free operation
88 and to increase unpredictability.
89 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
91 *) Changed Configure so it only says what it does and doesn't dump
92 so much data. Instead, ./configdata.pm should be used as a script
93 to display all sorts of configuration data.
96 *) Added processing of "make variables" to Configure.
99 *) Added SHA512/224 and SHA512/256 algorithm support.
102 *) The last traces of Netware support, first removed in 1.1.0, have
106 *) Get rid of Makefile.shared, and in the process, make the processing
107 of certain files (rc.obj, or the .def/.map/.opt files produced from
108 the ordinal files) more visible and hopefully easier to trace and
109 debug (or make silent).
112 *) Make it possible to have environment variable assignments as
113 arguments to config / Configure.
116 *) Add multi-prime RSA (RFC 8017) support.
119 *) Add SM3 implemented according to GB/T 32905-2016
120 [ Jack Lloyd <jack.lloyd@ribose.com>,
121 Ronald Tse <ronald.tse@ribose.com>,
122 Erick Borsboom <erick.borsboom@ribose.com> ]
124 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
125 as documented in RFC6066.
126 Based on a patch from Tomasz Moń
127 [Filipe Raimundo da Silva]
129 *) Add SM4 implemented according to GB/T 32907-2016.
130 [ Jack Lloyd <jack.lloyd@ribose.com>,
131 Ronald Tse <ronald.tse@ribose.com>,
132 Erick Borsboom <erick.borsboom@ribose.com> ]
134 *) Reimplement -newreq-nodes and ERR_error_string_n; the
135 original author does not agree with the license change.
138 *) Add ARIA AEAD TLS support.
141 *) Some macro definitions to support VS6 have been removed. Visual
142 Studio 6 has not worked since 1.1.0
145 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
146 without clearing the errors.
149 *) Add "atfork" functions. If building on a system that without
150 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
151 requirements. The RAND facility now uses/requires this.
157 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
158 not possible to disable entirely. However, it's still possible to
159 disable the console reading UI method, UI_OpenSSL() (use UI_null()
162 To disable, configure with 'no-ui-console'. 'no-ui' is still
163 possible to use as an alias. Check at compile time with the
164 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
165 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
168 *) Add a STORE module, which implements a uniform and URI based reader of
169 stores that can contain keys, certificates, CRLs and numerous other
170 objects. The main API is loosely based on a few stdio functions,
171 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
172 OSSL_STORE_error and OSSL_STORE_close.
173 The implementation uses backends called "loaders" to implement arbitrary
174 URI schemes. There is one built in "loader" for the 'file' scheme.
177 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
178 then adjusted to work on FreeBSD 8.4 as well.
179 Enable by configuring with 'enable-devcryptoeng'. This is done by default
180 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
183 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
184 util/mkerr.pl, which is adapted to allow those prefixes, leading to
185 error code calls like this:
187 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
189 With this change, we claim the namespaces OSSL and OPENSSL in a manner
190 that can be encoded in C. For the foreseeable future, this will only
192 [Richard Levitte and Tim Hudson]
194 *) Removed BSD cryptodev engine.
197 *) Add a build target 'build_all_generated', to build all generated files
198 and only that. This can be used to prepare everything that requires
199 things like perl for a system that lacks perl and then move everything
200 to that system and do the rest of the build there.
203 *) In the UI interface, make it possible to duplicate the user data. This
204 can be used by engines that need to retain the data for a longer time
205 than just the call where this user data is passed.
208 *) Ignore the '-named_curve auto' value for compatibility of applications
210 [Tomas Mraz <tmraz@fedoraproject.org>]
212 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
213 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
214 alerts across multiple records (some of which could be empty). In practice
215 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
216 prohibts this altogether and other libraries (BoringSSL, NSS) do not
217 support this at all. Supporting it adds significant complexity to the
218 record layer, and its removal is unlikely to cause inter-operability
222 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
223 with Z. These are meant to replace LONG and ZLONG and to be size safe.
224 The use of LONG and ZLONG is discouraged and scheduled for deprecation
228 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
229 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
230 [Richard Levitte, Andy Polyakov]
232 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
236 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
237 platform rather than 'mingw'.
240 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
241 success if they are asked to add an object which already exists
242 in the store. This change cascades to other functions which load
243 certificates and CRLs.
246 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
247 facilitate stack unwinding even from assembly subroutines.
250 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
251 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
254 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
255 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
256 which is the minimum version we support.
259 *) Certificate time validation (X509_cmp_time) enforces stricter
260 compliance with RFC 5280. Fractional seconds and timezone offsets
261 are no longer allowed.
264 *) Add support for ARIA
267 *) s_client will now send the Server Name Indication (SNI) extension by
268 default unless the new "-noservername" option is used. The server name is
269 based on the host provided to the "-connect" option unless overridden by
273 *) Add support for SipHash
276 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
277 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
278 prevent issues where no progress is being made and the peer continually
279 sends unrecognised record types, using up resources processing them.
282 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
283 using the algorithm defined in
284 https://www.akkadia.org/drepper/SHA-crypt.txt
287 *) Heartbeat support has been removed; the ABI is changed for now.
288 [Richard Levitte, Rich Salz]
290 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
293 *) The RSA "null" method, which was partially supported to avoid patent
294 issues, has been replaced to always returns NULL.
297 Changes between 1.1.0g and 1.1.0h [xx XXX xxxx]
299 *) Constructed ASN.1 types with a recursive definition could exceed the stack
301 Constructed ASN.1 types with a recursive definition (such as can be found
302 in PKCS7) could eventually exceed the stack given malicious input with
303 excessive recursion. This could result in a Denial Of Service attack. There
304 are no such structures used within SSL/TLS that come from untrusted sources
305 so this is considered safe.
307 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
312 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
314 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
315 effectively reduced to only comparing the least significant bit of each
316 byte. This allows an attacker to forge messages that would be considered as
317 authenticated in an amount of tries lower than that guaranteed by the
318 security claims of the scheme. The module can only be compiled by the
319 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
321 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
326 *) Add a build target 'build_all_generated', to build all generated files
327 and only that. This can be used to prepare everything that requires
328 things like perl for a system that lacks perl and then move everything
329 to that system and do the rest of the build there.
332 *) Backport SSL_OP_NO_RENGOTIATION
334 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
335 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
336 changes this is no longer possible in 1.1.0. Therefore the new
337 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
338 1.1.0 to provide equivalent functionality.
340 Note that if an application built against 1.1.0h headers (or above) is run
341 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
342 accepted but nothing will happen, i.e. renegotiation will not be prevented.
345 *) Removed the OS390-Unix config target. It relied on a script that doesn't
349 *) rsaz_1024_mul_avx2 overflow bug on x86_64
351 There is an overflow bug in the AVX2 Montgomery multiplication procedure
352 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
353 Analysis suggests that attacks against RSA and DSA as a result of this
354 defect would be very difficult to perform and are not believed likely.
355 Attacks against DH1024 are considered just feasible, because most of the
356 work necessary to deduce information about a private key may be performed
357 offline. The amount of resources required for such an attack would be
358 significant. However, for an attack on TLS to be meaningful, the server
359 would have to share the DH1024 private key among multiple clients, which is
360 no longer an option since CVE-2016-0701.
362 This only affects processors that support the AVX2 but not ADX extensions
363 like Intel Haswell (4th generation).
365 This issue was reported to OpenSSL by David Benjamin (Google). The issue
366 was originally found via the OSS-Fuzz project.
370 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
372 *) bn_sqrx8x_internal carry bug on x86_64
374 There is a carry propagating bug in the x86_64 Montgomery squaring
375 procedure. No EC algorithms are affected. Analysis suggests that attacks
376 against RSA and DSA as a result of this defect would be very difficult to
377 perform and are not believed likely. Attacks against DH are considered just
378 feasible (although very difficult) because most of the work necessary to
379 deduce information about a private key may be performed offline. The amount
380 of resources required for such an attack would be very significant and
381 likely only accessible to a limited number of attackers. An attacker would
382 additionally need online access to an unpatched system using the target
383 private key in a scenario with persistent DH parameters and a private
384 key that is shared between multiple clients.
386 This only affects processors that support the BMI1, BMI2 and ADX extensions
387 like Intel Broadwell (5th generation) and later or AMD Ryzen.
389 This issue was reported to OpenSSL by the OSS-Fuzz project.
393 *) Malformed X.509 IPAddressFamily could cause OOB read
395 If an X.509 certificate has a malformed IPAddressFamily extension,
396 OpenSSL could do a one-byte buffer overread. The most likely result
397 would be an erroneous display of the certificate in text format.
399 This issue was reported to OpenSSL by the OSS-Fuzz project.
403 Changes between 1.1.0e and 1.1.0f [25 May 2017]
405 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
406 platform rather than 'mingw'.
409 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
410 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
411 which is the minimum version we support.
414 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
416 *) Encrypt-Then-Mac renegotiation crash
418 During a renegotiation handshake if the Encrypt-Then-Mac extension is
419 negotiated where it was not in the original handshake (or vice-versa) then
420 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
421 and servers are affected.
423 This issue was reported to OpenSSL by Joe Orton (Red Hat).
427 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
429 *) Truncated packet could crash via OOB read
431 If one side of an SSL/TLS path is running on a 32-bit host and a specific
432 cipher is being used, then a truncated packet can cause that host to
433 perform an out-of-bounds read, usually resulting in a crash.
435 This issue was reported to OpenSSL by Robert Święcki of Google.
439 *) Bad (EC)DHE parameters cause a client crash
441 If a malicious server supplies bad parameters for a DHE or ECDHE key
442 exchange then this can result in the client attempting to dereference a
443 NULL pointer leading to a client crash. This could be exploited in a Denial
446 This issue was reported to OpenSSL by Guido Vranken.
450 *) BN_mod_exp may produce incorrect results on x86_64
452 There is a carry propagating bug in the x86_64 Montgomery squaring
453 procedure. No EC algorithms are affected. Analysis suggests that attacks
454 against RSA and DSA as a result of this defect would be very difficult to
455 perform and are not believed likely. Attacks against DH are considered just
456 feasible (although very difficult) because most of the work necessary to
457 deduce information about a private key may be performed offline. The amount
458 of resources required for such an attack would be very significant and
459 likely only accessible to a limited number of attackers. An attacker would
460 additionally need online access to an unpatched system using the target
461 private key in a scenario with persistent DH parameters and a private
462 key that is shared between multiple clients. For example this can occur by
463 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
464 similar to CVE-2015-3193 but must be treated as a separate problem.
466 This issue was reported to OpenSSL by the OSS-Fuzz project.
470 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
472 *) ChaCha20/Poly1305 heap-buffer-overflow
474 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
475 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
476 crash. This issue is not considered to be exploitable beyond a DoS.
478 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
482 *) CMS Null dereference
484 Applications parsing invalid CMS structures can crash with a NULL pointer
485 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
486 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
487 structure callback if an attempt is made to free certain invalid encodings.
488 Only CHOICE structures using a callback which do not handle NULL value are
491 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
495 *) Montgomery multiplication may produce incorrect results
497 There is a carry propagating bug in the Broadwell-specific Montgomery
498 multiplication procedure that handles input lengths divisible by, but
499 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
500 and DH private keys are impossible. This is because the subroutine in
501 question is not used in operations with the private key itself and an input
502 of the attacker's direct choice. Otherwise the bug can manifest itself as
503 transient authentication and key negotiation failures or reproducible
504 erroneous outcome of public-key operations with specially crafted input.
505 Among EC algorithms only Brainpool P-512 curves are affected and one
506 presumably can attack ECDH key negotiation. Impact was not analyzed in
507 detail, because pre-requisites for attack are considered unlikely. Namely
508 multiple clients have to choose the curve in question and the server has to
509 share the private key among them, neither of which is default behaviour.
510 Even then only clients that chose the curve will be affected.
512 This issue was publicly reported as transient failures and was not
513 initially recognized as a security issue. Thanks to Richard Morgan for
514 providing reproducible case.
518 *) Removed automatic addition of RPATH in shared libraries and executables,
519 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
522 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
524 *) Fix Use After Free for large message sizes
526 The patch applied to address CVE-2016-6307 resulted in an issue where if a
527 message larger than approx 16k is received then the underlying buffer to
528 store the incoming message is reallocated and moved. Unfortunately a
529 dangling pointer to the old location is left which results in an attempt to
530 write to the previously freed location. This is likely to result in a
531 crash, however it could potentially lead to execution of arbitrary code.
533 This issue only affects OpenSSL 1.1.0a.
535 This issue was reported to OpenSSL by Robert Święcki.
539 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
541 *) OCSP Status Request extension unbounded memory growth
543 A malicious client can send an excessively large OCSP Status Request
544 extension. If that client continually requests renegotiation, sending a
545 large OCSP Status Request extension each time, then there will be unbounded
546 memory growth on the server. This will eventually lead to a Denial Of
547 Service attack through memory exhaustion. Servers with a default
548 configuration are vulnerable even if they do not support OCSP. Builds using
549 the "no-ocsp" build time option are not affected.
551 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
555 *) SSL_peek() hang on empty record
557 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
558 sends an empty record. This could be exploited by a malicious peer in a
559 Denial Of Service attack.
561 This issue was reported to OpenSSL by Alex Gaynor.
565 *) Excessive allocation of memory in tls_get_message_header() and
566 dtls1_preprocess_fragment()
568 A (D)TLS message includes 3 bytes for its length in the header for the
569 message. This would allow for messages up to 16Mb in length. Messages of
570 this length are excessive and OpenSSL includes a check to ensure that a
571 peer is sending reasonably sized messages in order to avoid too much memory
572 being consumed to service a connection. A flaw in the logic of version
573 1.1.0 means that memory for the message is allocated too early, prior to
574 the excessive message length check. Due to way memory is allocated in
575 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
576 to service a connection. This could lead to a Denial of Service through
577 memory exhaustion. However, the excessive message length check still takes
578 place, and this would cause the connection to immediately fail. Assuming
579 that the application calls SSL_free() on the failed connection in a timely
580 manner then the 21Mb of allocated memory will then be immediately freed
581 again. Therefore the excessive memory allocation will be transitory in
582 nature. This then means that there is only a security impact if:
584 1) The application does not call SSL_free() in a timely manner in the event
585 that the connection fails
587 2) The application is working in a constrained environment where there is
588 very little free memory
590 3) The attacker initiates multiple connection attempts such that there are
591 multiple connections in a state where memory has been allocated for the
592 connection; SSL_free() has not yet been called; and there is insufficient
593 memory to service the multiple requests.
595 Except in the instance of (1) above any Denial Of Service is likely to be
596 transitory because as soon as the connection fails the memory is
597 subsequently freed again in the SSL_free() call. However there is an
598 increased risk during this period of application crashes due to the lack of
599 memory - which would then mean a more serious Denial of Service.
601 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
602 (CVE-2016-6307 and CVE-2016-6308)
605 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
606 had to be removed. Primary reason is that vendor assembler can't
607 assemble our modules with -KPIC flag. As result it, assembly
608 support, was not even available as option. But its lack means
609 lack of side-channel resistant code, which is incompatible with
610 security by todays standards. Fortunately gcc is readily available
611 prepackaged option, which we firmly point at...
614 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
616 *) Windows command-line tool supports UTF-8 opt-in option for arguments
617 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
618 (to any value) allows Windows user to access PKCS#12 file generated
619 with Windows CryptoAPI and protected with non-ASCII password, as well
620 as files generated under UTF-8 locale on Linux also protected with
624 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
625 have been disabled by default and removed from DEFAULT, just like RC4.
626 See the RC4 item below to re-enable both.
629 *) The method for finding the storage location for the Windows RAND seed file
630 has changed. First we check %RANDFILE%. If that is not set then we check
631 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
632 all else fails we fall back to C:\.
635 *) The EVP_EncryptUpdate() function has had its return type changed from void
636 to int. A return of 0 indicates and error while a return of 1 indicates
640 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
641 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
642 off the constant time implementation for RSA, DSA and DH have been made
643 no-ops and deprecated.
646 *) Windows RAND implementation was simplified to only get entropy by
647 calling CryptGenRandom(). Various other RAND-related tickets
649 [Joseph Wylie Yandle, Rich Salz]
651 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
652 and OPENSSL_LH_, respectively. The old names are available
653 with API compatibility. They new names are now completely documented.
656 *) Unify TYPE_up_ref(obj) methods signature.
657 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
658 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
659 int (instead of void) like all others TYPE_up_ref() methods.
660 So now these methods also check the return value of CRYPTO_atomic_add(),
661 and the validity of object reference counter.
662 [fdasilvayy@gmail.com]
664 *) With Windows Visual Studio builds, the .pdb files are installed
665 alongside the installed libraries and executables. For a static
666 library installation, ossl_static.pdb is the associate compiler
667 generated .pdb file to be used when linking programs.
670 *) Remove openssl.spec. Packaging files belong with the packagers.
673 *) Automatic Darwin/OSX configuration has had a refresh, it will now
674 recognise x86_64 architectures automatically. You can still decide
675 to build for a different bitness with the environment variable
676 KERNEL_BITS (can be 32 or 64), for example:
678 KERNEL_BITS=32 ./config
682 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
683 256 bit AES and HMAC with SHA256.
686 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
689 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
692 *) To enable users to have their own config files and build file templates,
693 Configure looks in the directory indicated by the environment variable
694 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
695 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
696 name and is used as is.
699 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
700 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
701 X509_CERT_FILE_CTX was removed.
704 *) "shared" builds are now the default. To create only static libraries use
705 the "no-shared" Configure option.
708 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
709 All of these option have not worked for some while and are fundamental
713 *) Make various cleanup routines no-ops and mark them as deprecated. Most
714 global cleanup functions are no longer required because they are handled
715 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
716 Explicitly de-initing can cause problems (e.g. where a library that uses
717 OpenSSL de-inits, but an application is still using it). The affected
718 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
719 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
720 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
724 *) --strict-warnings no longer enables runtime debugging options
725 such as REF_DEBUG. Instead, debug options are automatically
726 enabled with '--debug' builds.
727 [Andy Polyakov, Emilia Käsper]
729 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
730 have been moved out of the public header files. New functions for managing
731 these have been added.
734 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
735 objects have been moved out of the public header files. New
736 functions for managing these have been added.
739 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
740 have been moved out of the public header files. New functions for managing
741 these have been added.
744 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
745 moved out of the public header files. New functions for managing these
749 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
752 *) Removed the mk1mf build scripts.
755 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
756 it is always safe to #include a header now.
759 *) Removed the aged BC-32 config and all its supporting scripts
762 *) Removed support for Ultrix, Netware, and OS/2.
765 *) Add support for HKDF.
768 *) Add support for blake2b and blake2s
771 *) Added support for "pipelining". Ciphers that have the
772 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
773 encryptions/decryptions simultaneously. There are currently no built-in
774 ciphers with this property but the expectation is that engines will be able
775 to offer it to significantly improve throughput. Support has been extended
776 into libssl so that multiple records for a single connection can be
777 processed in one go (for >=TLS 1.1).
780 *) Added the AFALG engine. This is an async capable engine which is able to
781 offload work to the Linux kernel. In this initial version it only supports
782 AES128-CBC. The kernel must be version 4.1.0 or greater.
785 *) OpenSSL now uses a new threading API. It is no longer necessary to
786 set locking callbacks to use OpenSSL in a multi-threaded environment. There
787 are two supported threading models: pthreads and windows threads. It is
788 also possible to configure OpenSSL at compile time for "no-threads". The
789 old threading API should no longer be used. The functions have been
790 replaced with "no-op" compatibility macros.
791 [Alessandro Ghedini, Matt Caswell]
793 *) Modify behavior of ALPN to invoke callback after SNI/servername
794 callback, such that updates to the SSL_CTX affect ALPN.
797 *) Add SSL_CIPHER queries for authentication and key-exchange.
800 *) Changes to the DEFAULT cipherlist:
801 - Prefer (EC)DHE handshakes over plain RSA.
802 - Prefer AEAD ciphers over legacy ciphers.
803 - Prefer ECDSA over RSA when both certificates are available.
804 - Prefer TLSv1.2 ciphers/PRF.
805 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
809 *) Change the ECC default curve list to be this, in order: x25519,
810 secp256r1, secp521r1, secp384r1.
813 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
814 disabled by default. They can be re-enabled using the
815 enable-weak-ssl-ciphers option to Configure.
818 *) If the server has ALPN configured, but supports no protocols that the
819 client advertises, send a fatal "no_application_protocol" alert.
820 This behaviour is SHALL in RFC 7301, though it isn't universally
821 implemented by other servers.
824 *) Add X25519 support.
825 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
826 for public and private key encoding using the format documented in
827 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
828 key generation and key derivation.
830 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
834 *) Deprecate SRP_VBASE_get_by_user.
835 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
836 In order to fix an unavoidable memory leak (CVE-2016-0798),
837 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
838 seed, even if the seed is configured.
840 Users should use SRP_VBASE_get1_by_user instead. Note that in
841 SRP_VBASE_get1_by_user, caller must free the returned value. Note
842 also that even though configuring the SRP seed attempts to hide
843 invalid usernames by continuing the handshake with fake
844 credentials, this behaviour is not constant time and no strong
845 guarantees are made that the handshake is indistinguishable from
846 that of a valid user.
849 *) Configuration change; it's now possible to build dynamic engines
850 without having to build shared libraries and vice versa. This
851 only applies to the engines in engines/, those in crypto/engine/
852 will always be built into libcrypto (i.e. "static").
854 Building dynamic engines is enabled by default; to disable, use
855 the configuration option "disable-dynamic-engine".
857 The only requirements for building dynamic engines are the
858 presence of the DSO module and building with position independent
859 code, so they will also automatically be disabled if configuring
860 with "disable-dso" or "disable-pic".
862 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
863 are also taken away from openssl/opensslconf.h, as they are
867 *) Configuration change; if there is a known flag to compile
868 position independent code, it will always be applied on the
869 libcrypto and libssl object files, and never on the application
870 object files. This means other libraries that use routines from
871 libcrypto / libssl can be made into shared libraries regardless
872 of how OpenSSL was configured.
874 If this isn't desirable, the configuration options "disable-pic"
875 or "no-pic" can be used to disable the use of PIC. This will
876 also disable building shared libraries and dynamic engines.
879 *) Removed JPAKE code. It was experimental and has no wide use.
882 *) The INSTALL_PREFIX Makefile variable has been renamed to
883 DESTDIR. That makes for less confusion on what this variable
884 is for. Also, the configuration option --install_prefix is
888 *) Heartbeat for TLS has been removed and is disabled by default
889 for DTLS; configure with enable-heartbeats. Code that uses the
890 old #define's might need to be updated.
891 [Emilia Käsper, Rich Salz]
893 *) Rename REF_CHECK to REF_DEBUG.
896 *) New "unified" build system
898 The "unified" build system is aimed to be a common system for all
899 platforms we support. With it comes new support for VMS.
901 This system builds supports building in a different directory tree
902 than the source tree. It produces one Makefile (for unix family
903 or lookalikes), or one descrip.mms (for VMS).
905 The source of information to make the Makefile / descrip.mms is
906 small files called 'build.info', holding the necessary
907 information for each directory with source to compile, and a
908 template in Configurations, like unix-Makefile.tmpl or
911 With this change, the library names were also renamed on Windows
912 and on VMS. They now have names that are closer to the standard
913 on Unix, and include the major version number, and in certain
914 cases, the architecture they are built for. See "Notes on shared
915 libraries" in INSTALL.
917 We rely heavily on the perl module Text::Template.
920 *) Added support for auto-initialisation and de-initialisation of the library.
921 OpenSSL no longer requires explicit init or deinit routines to be called,
922 except in certain circumstances. See the OPENSSL_init_crypto() and
923 OPENSSL_init_ssl() man pages for further information.
926 *) The arguments to the DTLSv1_listen function have changed. Specifically the
927 "peer" argument is now expected to be a BIO_ADDR object.
929 *) Rewrite of BIO networking library. The BIO library lacked consistent
930 support of IPv6, and adding it required some more extensive
931 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
932 which hold all types of addresses and chains of address information.
933 It also introduces a new API, with functions like BIO_socket,
934 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
935 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
936 have been adapted accordingly.
939 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
943 *) CRIME protection: disable compression by default, even if OpenSSL is
944 compiled with zlib enabled. Applications can still enable compression
945 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
946 using the SSL_CONF library to configure compression.
949 *) The signature of the session callback configured with
950 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
951 was explicitly marked as 'const unsigned char*' instead of
955 *) Always DPURIFY. Remove the use of uninitialized memory in the
956 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
959 *) Removed many obsolete configuration items, including
960 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
961 MD2_CHAR, MD2_INT, MD2_LONG
963 IDEA_SHORT, IDEA_LONG
964 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
965 [Rich Salz, with advice from Andy Polyakov]
967 *) Many BN internals have been moved to an internal header file.
968 [Rich Salz with help from Andy Polyakov]
970 *) Configuration and writing out the results from it has changed.
971 Files such as Makefile include/openssl/opensslconf.h and are now
972 produced through general templates, such as Makefile.in and
973 crypto/opensslconf.h.in and some help from the perl module
976 Also, the center of configuration information is no longer
977 Makefile. Instead, Configure produces a perl module in
978 configdata.pm which holds most of the config data (in the hash
979 table %config), the target data that comes from the target
980 configuration in one of the Configurations/*.conf files (in
984 *) To clarify their intended purposes, the Configure options
985 --prefix and --openssldir change their semantics, and become more
986 straightforward and less interdependent.
988 --prefix shall be used exclusively to give the location INSTALLTOP
989 where programs, scripts, libraries, include files and manuals are
990 going to be installed. The default is now /usr/local.
992 --openssldir shall be used exclusively to give the default
993 location OPENSSLDIR where certificates, private keys, CRLs are
994 managed. This is also where the default openssl.cnf gets
996 If the directory given with this option is a relative path, the
997 values of both the --prefix value and the --openssldir value will
998 be combined to become OPENSSLDIR.
999 The default for --openssldir is INSTALLTOP/ssl.
1001 Anyone who uses --openssldir to specify where OpenSSL is to be
1002 installed MUST change to use --prefix instead.
1005 *) The GOST engine was out of date and therefore it has been removed. An up
1006 to date GOST engine is now being maintained in an external repository.
1007 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1008 support for GOST ciphersuites (these are only activated if a GOST engine
1012 *) EGD is no longer supported by default; use enable-egd when
1014 [Ben Kaduk and Rich Salz]
1016 *) The distribution now has Makefile.in files, which are used to
1017 create Makefile's when Configure is run. *Configure must be run
1018 before trying to build now.*
1021 *) The return value for SSL_CIPHER_description() for error conditions
1025 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1027 Obtaining and performing DNSSEC validation of TLSA records is
1028 the application's responsibility. The application provides
1029 the TLSA records of its choice to OpenSSL, and these are then
1030 used to authenticate the peer.
1032 The TLSA records need not even come from DNS. They can, for
1033 example, be used to implement local end-entity certificate or
1034 trust-anchor "pinning", where the "pin" data takes the form
1035 of TLSA records, which can augment or replace verification
1036 based on the usual WebPKI public certification authorities.
1039 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1040 continues to support deprecated interfaces in default builds.
1041 However, applications are strongly advised to compile their
1042 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1043 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1044 or the 1.1.0 releases.
1046 In environments in which all applications have been ported to
1047 not use any deprecated interfaces OpenSSL's Configure script
1048 should be used with the --api=1.1.0 option to entirely remove
1049 support for the deprecated features from the library and
1050 unconditionally disable them in the installed headers.
1051 Essentially the same effect can be achieved with the "no-deprecated"
1052 argument to Configure, except that this will always restrict
1053 the build to just the latest API, rather than a fixed API
1056 As applications are ported to future revisions of the API,
1057 they should update their compile-time OPENSSL_API_COMPAT define
1058 accordingly, but in most cases should be able to continue to
1059 compile with later releases.
1061 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1062 0x10000000L and 0x00908000L, respectively. However those
1063 versions did not support the OPENSSL_API_COMPAT feature, and
1064 so applications are not typically tested for explicit support
1065 of just the undeprecated features of either release.
1068 *) Add support for setting the minimum and maximum supported protocol.
1069 It can bet set via the SSL_set_min_proto_version() and
1070 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1071 MaxProtocol. It's recommended to use the new APIs to disable
1072 protocols instead of disabling individual protocols using
1073 SSL_set_options() or SSL_CONF's Protocol. This change also
1074 removes support for disabling TLS 1.2 in the OpenSSL TLS
1075 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1078 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1081 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1082 and integrates ECDSA and ECDH functionality into EC. Implementations can
1083 now redirect key generation and no longer need to convert to or from
1086 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1087 include the ec.h header file instead.
1090 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1091 ciphers who are no longer supported and drops support the ephemeral RSA key
1092 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1095 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1096 opaque. For HMAC_CTX, the following constructors and destructors
1099 HMAC_CTX *HMAC_CTX_new(void);
1100 void HMAC_CTX_free(HMAC_CTX *ctx);
1102 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1103 destroy such methods has been added. See EVP_MD_meth_new(3) and
1104 EVP_CIPHER_meth_new(3) for documentation.
1107 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1108 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1109 EVP_MD_CTX_reset() should be called instead to reinitialise
1110 an already created structure.
1111 2) For consistency with the majority of our object creators and
1112 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1113 EVP_MD_CTX_(new|free). The old names are retained as macros
1114 for deprecated builds.
1117 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1118 cryptographic operations to be performed asynchronously as long as an
1119 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1120 further details. Libssl has also had this capability integrated with the
1121 introduction of the new mode SSL_MODE_ASYNC and associated error
1122 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1123 pages. This work was developed in partnership with Intel Corp.
1126 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1127 always enabled now. If you want to disable the support you should
1128 exclude it using the list of supported ciphers. This also means that the
1129 "-no_ecdhe" option has been removed from s_server.
1132 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1133 SSL_{CTX_}set1_curves() which can set a list.
1136 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1137 curve you want to support using SSL_{CTX_}set1_curves().
1140 *) State machine rewrite. The state machine code has been significantly
1141 refactored in order to remove much duplication of code and solve issues
1142 with the old code (see ssl/statem/README for further details). This change
1143 does have some associated API changes. Notably the SSL_state() function
1144 has been removed and replaced by SSL_get_state which now returns an
1145 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1146 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1150 *) All instances of the string "ssleay" in the public API were replaced
1151 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1152 Some error codes related to internal RSA_eay API's were renamed.
1155 *) The demo files in crypto/threads were moved to demo/threads.
1158 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1160 [Matt Caswell, Rich Salz]
1162 *) New ASN.1 embed macro.
1164 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1165 structure is not allocated: it is part of the parent. That is instead of
1173 This reduces memory fragmentation and make it impossible to accidentally
1174 set a mandatory field to NULL.
1176 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1177 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1178 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1182 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1185 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1186 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1187 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1188 DES and RC4 ciphersuites.
1191 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1192 This changes the decoding behaviour for some invalid messages,
1193 though the change is mostly in the more lenient direction, and
1194 legacy behaviour is preserved as much as possible.
1197 *) Fix no-stdio build.
1198 [ David Woodhouse <David.Woodhouse@intel.com> and also
1199 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1201 *) New testing framework
1202 The testing framework has been largely rewritten and is now using
1203 perl and the perl modules Test::Harness and an extended variant of
1204 Test::More called OpenSSL::Test to do its work. All test scripts in
1205 test/ have been rewritten into test recipes, and all direct calls to
1206 executables in test/Makefile have become individual recipes using the
1207 simplified testing OpenSSL::Test::Simple.
1209 For documentation on our testing modules, do:
1211 perldoc test/testlib/OpenSSL/Test/Simple.pm
1212 perldoc test/testlib/OpenSSL/Test.pm
1216 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1217 are used; the latter aborts on memory leaks (usually checked on exit).
1218 Some undocumented "set malloc, etc., hooks" functions were removed
1219 and others were changed. All are now documented.
1222 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1224 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1226 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1227 from RFC4279, RFC4785, RFC5487, RFC5489.
1229 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1230 original RSA_PSK patch.
1233 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1234 era flag was never set throughout the codebase (only read). Also removed
1235 SSL3_FLAGS_POP_BUFFER which was only used if
1236 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1239 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1240 to be "oneline" instead of "compat".
1243 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1244 not aware of clients that still exhibit this bug, and the workaround
1245 hasn't been working properly for a while.
1248 *) The return type of BIO_number_read() and BIO_number_written() as well as
1249 the corresponding num_read and num_write members in the BIO structure has
1250 changed from unsigned long to uint64_t. On platforms where an unsigned
1251 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1255 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1256 OpenSSL without support for them. It also means that maintaining
1257 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1258 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1261 *) Removed support for the two export grade static DH ciphersuites
1262 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1263 were newly added (along with a number of other static DH ciphersuites) to
1264 1.0.2. However the two export ones have *never* worked since they were
1265 introduced. It seems strange in any case to be adding new export
1266 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1269 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1270 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1271 and turned into macros which simply call the new preferred function names
1272 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1273 should use the new names instead. Also as part of this change the ssl23.h
1274 header file has been removed.
1277 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1278 code and the associated standard is no longer considered fit-for-purpose.
1281 *) RT2547 was closed. When generating a private key, try to make the
1282 output file readable only by the owner. This behavior change might
1283 be noticeable when interacting with other software.
1285 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1289 *) Added HTTP GET support to the ocsp command.
1292 *) Changed default digest for the dgst and enc commands from MD5 to
1296 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1299 *) Added support for TLS extended master secret from
1300 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1301 initial patch which was a great help during development.
1304 *) All libssl internal structures have been removed from the public header
1305 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1306 now redundant). Users should not attempt to access internal structures
1307 directly. Instead they should use the provided API functions.
1310 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1311 Access to deprecated functions can be re-enabled by running config with
1312 "enable-deprecated". In addition applications wishing to use deprecated
1313 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1314 will, by default, disable some transitive includes that previously existed
1315 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1318 *) Added support for OCB mode. OpenSSL has been granted a patent license
1319 compatible with the OpenSSL license for use of OCB. Details are available
1320 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1321 for OCB can be removed by calling config with no-ocb.
1324 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1325 compatible client hello.
1328 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1329 done while fixing the error code for the key-too-small case.
1330 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1332 *) CA.sh has been removed; use CA.pl instead.
1335 *) Removed old DES API.
1338 *) Remove various unsupported platforms:
1344 Sinix/ReliantUNIX RM400
1349 16-bit platforms such as WIN16
1352 *) Clean up OPENSSL_NO_xxx #define's
1353 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1354 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1355 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1356 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1357 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1358 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1359 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1360 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1361 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1362 Remove MS_STATIC; it's a relic from platforms <32 bits.
1365 *) Cleaned up dead code
1366 Remove all but one '#ifdef undef' which is to be looked at.
1369 *) Clean up calling of xxx_free routines.
1370 Just like free(), fix most of the xxx_free routines to accept
1371 NULL. Remove the non-null checks from callers. Save much code.
1374 *) Add secure heap for storage of private keys (when possible).
1375 Add BIO_s_secmem(), CBIGNUM, etc.
1376 Contributed by Akamai Technologies under our Corporate CLA.
1379 *) Experimental support for a new, fast, unbiased prime candidate generator,
1380 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1381 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1383 *) New output format NSS in the sess_id command line tool. This allows
1384 exporting the session id and the master key in NSS keylog format.
1385 [Martin Kaiser <martin@kaiser.cx>]
1387 *) Harmonize version and its documentation. -f flag is used to display
1389 [mancha <mancha1@zoho.com>]
1391 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1392 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1393 [mancha <mancha1@zoho.com>]
1395 *) Fix some double frees. These are not thought to be exploitable.
1396 [mancha <mancha1@zoho.com>]
1398 *) A missing bounds check in the handling of the TLS heartbeat extension
1399 can be used to reveal up to 64k of memory to a connected client or
1402 Thanks for Neel Mehta of Google Security for discovering this bug and to
1403 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1404 preparing the fix (CVE-2014-0160)
1405 [Adam Langley, Bodo Moeller]
1407 *) Fix for the attack described in the paper "Recovering OpenSSL
1408 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1409 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1410 http://eprint.iacr.org/2014/140
1412 Thanks to Yuval Yarom and Naomi Benger for discovering this
1413 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1414 [Yuval Yarom and Naomi Benger]
1416 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1417 this fixes a limitation in previous versions of OpenSSL.
1420 *) Experimental encrypt-then-mac support.
1422 Experimental support for encrypt then mac from
1423 draft-gutmann-tls-encrypt-then-mac-02.txt
1425 To enable it set the appropriate extension number (0x42 for the test
1426 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1428 For non-compliant peers (i.e. just about everything) this should have no
1431 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1435 *) Add EVP support for key wrapping algorithms, to avoid problems with
1436 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1437 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1438 algorithms and include tests cases.
1441 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1445 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1446 MGF1 digest and OAEP label.
1449 *) Make openssl verify return errors.
1450 [Chris Palmer <palmer@google.com> and Ben Laurie]
1452 *) New function ASN1_TIME_diff to calculate the difference between two
1453 ASN1_TIME structures or one structure and the current time.
1456 *) Update fips_test_suite to support multiple command line options. New
1457 test to induce all self test errors in sequence and check expected
1461 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1462 sign or verify all in one operation.
1465 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1466 test programs and fips_test_suite. Includes functionality to parse
1467 the minimal script output of fipsalgest.pl directly.
1470 *) Add authorisation parameter to FIPS_module_mode_set().
1473 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1476 *) Use separate DRBG fields for internal and external flags. New function
1477 FIPS_drbg_health_check() to perform on demand health checking. Add
1478 generation tests to fips_test_suite with reduced health check interval to
1479 demonstrate periodic health checking. Add "nodh" option to
1480 fips_test_suite to skip very slow DH test.
1483 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1487 *) More extensive health check for DRBG checking many more failure modes.
1488 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1489 combination: call this in fips_test_suite.
1492 *) Add support for canonical generation of DSA parameter 'g'. See
1495 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1496 POST to handle HMAC cases.
1499 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1500 to return numerical and string versions of the FIPS module number.
1503 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1504 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1505 outside the validated module in the FIPS capable OpenSSL.
1508 *) Minor change to DRBG entropy callback semantics. In some cases
1509 there is no multiple of the block length between min_len and
1510 max_len. Allow the callback to return more than max_len bytes
1511 of entropy but discard any extra: it is the callback's responsibility
1512 to ensure that the extra data discarded does not impact the
1513 requested amount of entropy.
1516 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1517 information in FIPS186-3, SP800-57 and SP800-131A.
1520 *) CCM support via EVP. Interface is very similar to GCM case except we
1521 must supply all data in one chunk (i.e. no update, final) and the
1522 message length must be supplied if AAD is used. Add algorithm test
1526 *) Initial version of POST overhaul. Add POST callback to allow the status
1527 of POST to be monitored and/or failures induced. Modify fips_test_suite
1528 to use callback. Always run all selftests even if one fails.
1531 *) XTS support including algorithm test driver in the fips_gcmtest program.
1532 Note: this does increase the maximum key length from 32 to 64 bytes but
1533 there should be no binary compatibility issues as existing applications
1534 will never use XTS mode.
1537 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1538 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1539 performs algorithm blocking for unapproved PRNG types. Also do not
1540 set PRNG type in FIPS_mode_set(): leave this to the application.
1541 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1542 the standard OpenSSL PRNG: set additional data to a date time vector.
1545 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1546 This shouldn't present any incompatibility problems because applications
1547 shouldn't be using these directly and any that are will need to rethink
1548 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1551 *) Extensive self tests and health checking required by SP800-90 DRBG.
1552 Remove strength parameter from FIPS_drbg_instantiate and always
1553 instantiate at maximum supported strength.
1556 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1559 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1562 *) New function DH_compute_key_padded() to compute a DH key and pad with
1563 leading zeroes if needed: this complies with SP800-56A et al.
1566 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1567 anything, incomplete, subject to change and largely untested at present.
1570 *) Modify fipscanisteronly build option to only build the necessary object
1571 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1574 *) Add experimental option FIPSSYMS to give all symbols in
1575 fipscanister.o and FIPS or fips prefix. This will avoid
1576 conflicts with future versions of OpenSSL. Add perl script
1577 util/fipsas.pl to preprocess assembly language source files
1578 and rename any affected symbols.
1581 *) Add selftest checks and algorithm block of non-fips algorithms in
1582 FIPS mode. Remove DES2 from selftests.
1585 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1586 return internal method without any ENGINE dependencies. Add new
1587 tiny fips sign and verify functions.
1590 *) New build option no-ec2m to disable characteristic 2 code.
1593 *) New build option "fipscanisteronly". This only builds fipscanister.o
1594 and (currently) associated fips utilities. Uses the file Makefile.fips
1595 instead of Makefile.org as the prototype.
1598 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1599 Update fips_gcmtest to use IV generator.
1602 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1603 setting output buffer to NULL. The *Final function must be
1604 called although it will not retrieve any additional data. The tag
1605 can be set or retrieved with a ctrl. The IV length is by default 12
1606 bytes (96 bits) but can be set to an alternative value. If the IV
1607 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1611 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1612 underlying do_cipher function handles all cipher semantics itself
1613 including padding and finalisation. This is useful if (for example)
1614 an ENGINE cipher handles block padding itself. The behaviour of
1615 do_cipher is subtly changed if this flag is set: the return value
1616 is the number of characters written to the output buffer (zero is
1617 no longer an error code) or a negative error code. Also if the
1618 input buffer is NULL and length 0 finalisation should be performed.
1621 *) If a candidate issuer certificate is already part of the constructed
1622 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1625 *) Improve forward-security support: add functions
1627 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1628 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1630 for use by SSL/TLS servers; the callback function will be called whenever a
1631 new session is created, and gets to decide whether the session may be
1632 cached to make it resumable (return 0) or not (return 1). (As by the
1633 SSL/TLS protocol specifications, the session_id sent by the server will be
1634 empty to indicate that the session is not resumable; also, the server will
1635 not generate RFC 4507 (RFC 5077) session tickets.)
1637 A simple reasonable callback implementation is to return is_forward_secure.
1638 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1639 by the SSL/TLS server library, indicating whether it can provide forward
1641 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1643 *) New -verify_name option in command line utilities to set verification
1647 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1648 Add CMAC pkey methods.
1651 *) Experimental renegotiation in s_server -www mode. If the client
1652 browses /reneg connection is renegotiated. If /renegcert it is
1653 renegotiated requesting a certificate.
1656 *) Add an "external" session cache for debugging purposes to s_server. This
1657 should help trace issues which normally are only apparent in deployed
1658 multi-process servers.
1661 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1662 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1663 BIO_set_cipher() and some obscure PEM functions were changed so they
1664 can now return an error. The RAND changes required a change to the
1665 RAND_METHOD structure.
1668 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1669 a gcc attribute to warn if the result of a function is ignored. This
1670 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1671 whose return value is often ignored.
1674 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1675 These allow SCTs (signed certificate timestamps) to be requested and
1676 validated when establishing a connection.
1677 [Rob Percival <robpercival@google.com>]
1679 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1681 *) Prevent padding oracle in AES-NI CBC MAC check
1683 A MITM attacker can use a padding oracle attack to decrypt traffic
1684 when the connection uses an AES CBC cipher and the server support
1687 This issue was introduced as part of the fix for Lucky 13 padding
1688 attack (CVE-2013-0169). The padding check was rewritten to be in
1689 constant time by making sure that always the same bytes are read and
1690 compared against either the MAC or padding bytes. But it no longer
1691 checked that there was enough data to have both the MAC and padding
1694 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1698 *) Fix EVP_EncodeUpdate overflow
1700 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1701 Base64 encoding of binary data. If an attacker is able to supply very large
1702 amounts of input data then a length check can overflow resulting in a heap
1705 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1706 the PEM_write_bio* family of functions. These are mainly used within the
1707 OpenSSL command line applications, so any application which processes data
1708 from an untrusted source and outputs it as a PEM file should be considered
1709 vulnerable to this issue. User applications that call these APIs directly
1710 with large amounts of untrusted data may also be vulnerable.
1712 This issue was reported by Guido Vranken.
1716 *) Fix EVP_EncryptUpdate overflow
1718 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1719 is able to supply very large amounts of input data after a previous call to
1720 EVP_EncryptUpdate() with a partial block then a length check can overflow
1721 resulting in a heap corruption. Following an analysis of all OpenSSL
1722 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1723 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1724 the first called function after an EVP_EncryptInit(), and therefore that
1725 specific call must be safe. The second form is where the length passed to
1726 EVP_EncryptUpdate() can be seen from the code to be some small value and
1727 therefore there is no possibility of an overflow. Since all instances are
1728 one of these two forms, it is believed that there can be no overflows in
1729 internal code due to this problem. It should be noted that
1730 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1731 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1732 of these calls have also been analysed too and it is believed there are no
1733 instances in internal usage where an overflow could occur.
1735 This issue was reported by Guido Vranken.
1739 *) Prevent ASN.1 BIO excessive memory allocation
1741 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1742 a short invalid encoding can cause allocation of large amounts of memory
1743 potentially consuming excessive resources or exhausting memory.
1745 Any application parsing untrusted data through d2i BIO functions is
1746 affected. The memory based functions such as d2i_X509() are *not* affected.
1747 Since the memory based functions are used by the TLS library, TLS
1748 applications are not affected.
1750 This issue was reported by Brian Carpenter.
1756 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1757 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1758 in arbitrary stack data being returned in the buffer.
1760 This issue was reported by Guido Vranken.
1764 *) Modify behavior of ALPN to invoke callback after SNI/servername
1765 callback, such that updates to the SSL_CTX affect ALPN.
1768 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1772 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1773 methods are enabled and ssl2 is disabled the methods return NULL.
1776 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1778 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1779 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1780 provide any "EXPORT" or "LOW" strength ciphers.
1783 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1784 is by default disabled at build-time. Builds that are not configured with
1785 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1786 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1787 will need to explicitly call either of:
1789 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1791 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1793 as appropriate. Even if either of those is used, or the application
1794 explicitly uses the version-specific SSLv2_method() or its client and
1795 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1796 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1797 ciphers, and SSLv2 56-bit DES are no longer available.
1801 *) Fix a double-free in DSA code
1803 A double free bug was discovered when OpenSSL parses malformed DSA private
1804 keys and could lead to a DoS attack or memory corruption for applications
1805 that receive DSA private keys from untrusted sources. This scenario is
1808 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1813 *) Disable SRP fake user seed to address a server memory leak.
1815 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1817 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1818 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1819 was changed to ignore the "fake user" SRP seed, even if the seed
1822 Users should use SRP_VBASE_get1_by_user instead. Note that in
1823 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1824 also that even though configuring the SRP seed attempts to hide
1825 invalid usernames by continuing the handshake with fake
1826 credentials, this behaviour is not constant time and no strong
1827 guarantees are made that the handshake is indistinguishable from
1828 that of a valid user.
1832 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1834 In the BN_hex2bn function the number of hex digits is calculated using an
1835 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1836 large values of |i| this can result in |bn_expand| not allocating any
1837 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1838 field as NULL leading to a subsequent NULL ptr deref. For very large values
1839 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1840 In this case memory is allocated to the internal BIGNUM data field, but it
1841 is insufficiently sized leading to heap corruption. A similar issue exists
1842 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1843 is ever called by user applications with very large untrusted hex/dec data.
1844 This is anticipated to be a rare occurrence.
1846 All OpenSSL internal usage of these functions use data that is not expected
1847 to be untrusted, e.g. config file data or application command line
1848 arguments. If user developed applications generate config file data based
1849 on untrusted data then it is possible that this could also lead to security
1850 consequences. This is also anticipated to be rare.
1852 This issue was reported to OpenSSL by Guido Vranken.
1856 *) Fix memory issues in BIO_*printf functions
1858 The internal |fmtstr| function used in processing a "%s" format string in
1859 the BIO_*printf functions could overflow while calculating the length of a
1860 string and cause an OOB read when printing very long strings.
1862 Additionally the internal |doapr_outch| function can attempt to write to an
1863 OOB memory location (at an offset from the NULL pointer) in the event of a
1864 memory allocation failure. In 1.0.2 and below this could be caused where
1865 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1866 could be in processing a very long "%s" format string. Memory leaks can
1869 The first issue may mask the second issue dependent on compiler behaviour.
1870 These problems could enable attacks where large amounts of untrusted data
1871 is passed to the BIO_*printf functions. If applications use these functions
1872 in this way then they could be vulnerable. OpenSSL itself uses these
1873 functions when printing out human-readable dumps of ASN.1 data. Therefore
1874 applications that print this data could be vulnerable if the data is from
1875 untrusted sources. OpenSSL command line applications could also be
1876 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1877 as command line arguments.
1879 Libssl is not considered directly vulnerable. Additionally certificates etc
1880 received via remote connections via libssl are also unlikely to be able to
1881 trigger these issues because of message size limits enforced within libssl.
1883 This issue was reported to OpenSSL Guido Vranken.
1887 *) Side channel attack on modular exponentiation
1889 A side-channel attack was found which makes use of cache-bank conflicts on
1890 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1891 of RSA keys. The ability to exploit this issue is limited as it relies on
1892 an attacker who has control of code in a thread running on the same
1893 hyper-threaded core as the victim thread which is performing decryptions.
1895 This issue was reported to OpenSSL by Yuval Yarom, The University of
1896 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1897 Nadia Heninger, University of Pennsylvania with more information at
1898 http://cachebleed.info.
1902 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1903 if no keysize is specified with default_bits. This fixes an
1904 omission in an earlier change that changed all RSA/DSA key generation
1905 apps to use 2048 bits by default.
1908 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1909 *) DH small subgroups
1911 Historically OpenSSL only ever generated DH parameters based on "safe"
1912 primes. More recently (in version 1.0.2) support was provided for
1913 generating X9.42 style parameter files such as those required for RFC 5114
1914 support. The primes used in such files may not be "safe". Where an
1915 application is using DH configured with parameters based on primes that are
1916 not "safe" then an attacker could use this fact to find a peer's private
1917 DH exponent. This attack requires that the attacker complete multiple
1918 handshakes in which the peer uses the same private DH exponent. For example
1919 this could be used to discover a TLS server's private DH exponent if it's
1920 reusing the private DH exponent or it's using a static DH ciphersuite.
1922 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1923 TLS. It is not on by default. If the option is not set then the server
1924 reuses the same private DH exponent for the life of the server process and
1925 would be vulnerable to this attack. It is believed that many popular
1926 applications do set this option and would therefore not be at risk.
1928 The fix for this issue adds an additional check where a "q" parameter is
1929 available (as is the case in X9.42 based parameters). This detects the
1930 only known attack, and is the only possible defense for static DH
1931 ciphersuites. This could have some performance impact.
1933 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1934 default and cannot be disabled. This could have some performance impact.
1936 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1940 *) SSLv2 doesn't block disabled ciphers
1942 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1943 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1944 been disabled, provided that the SSLv2 protocol was not also disabled via
1947 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1948 and Sebastian Schinzel.
1952 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1954 *) BN_mod_exp may produce incorrect results on x86_64
1956 There is a carry propagating bug in the x86_64 Montgomery squaring
1957 procedure. No EC algorithms are affected. Analysis suggests that attacks
1958 against RSA and DSA as a result of this defect would be very difficult to
1959 perform and are not believed likely. Attacks against DH are considered just
1960 feasible (although very difficult) because most of the work necessary to
1961 deduce information about a private key may be performed offline. The amount
1962 of resources required for such an attack would be very significant and
1963 likely only accessible to a limited number of attackers. An attacker would
1964 additionally need online access to an unpatched system using the target
1965 private key in a scenario with persistent DH parameters and a private
1966 key that is shared between multiple clients. For example this can occur by
1967 default in OpenSSL DHE based SSL/TLS ciphersuites.
1969 This issue was reported to OpenSSL by Hanno Böck.
1973 *) Certificate verify crash with missing PSS parameter
1975 The signature verification routines will crash with a NULL pointer
1976 dereference if presented with an ASN.1 signature using the RSA PSS
1977 algorithm and absent mask generation function parameter. Since these
1978 routines are used to verify certificate signature algorithms this can be
1979 used to crash any certificate verification operation and exploited in a
1980 DoS attack. Any application which performs certificate verification is
1981 vulnerable including OpenSSL clients and servers which enable client
1984 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1988 *) X509_ATTRIBUTE memory leak
1990 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1991 memory. This structure is used by the PKCS#7 and CMS routines so any
1992 application which reads PKCS#7 or CMS data from untrusted sources is
1993 affected. SSL/TLS is not affected.
1995 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2000 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2001 This changes the decoding behaviour for some invalid messages,
2002 though the change is mostly in the more lenient direction, and
2003 legacy behaviour is preserved as much as possible.
2006 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2008 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2010 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2012 *) Alternate chains certificate forgery
2014 During certificate verification, OpenSSL will attempt to find an
2015 alternative certificate chain if the first attempt to build such a chain
2016 fails. An error in the implementation of this logic can mean that an
2017 attacker could cause certain checks on untrusted certificates to be
2018 bypassed, such as the CA flag, enabling them to use a valid leaf
2019 certificate to act as a CA and "issue" an invalid certificate.
2021 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2025 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2027 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2028 incompatibility in the handling of HMAC. The previous ABI has now been
2032 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2034 *) Malformed ECParameters causes infinite loop
2036 When processing an ECParameters structure OpenSSL enters an infinite loop
2037 if the curve specified is over a specially malformed binary polynomial
2040 This can be used to perform denial of service against any
2041 system which processes public keys, certificate requests or
2042 certificates. This includes TLS clients and TLS servers with
2043 client authentication enabled.
2045 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2049 *) Exploitable out-of-bounds read in X509_cmp_time
2051 X509_cmp_time does not properly check the length of the ASN1_TIME
2052 string and can read a few bytes out of bounds. In addition,
2053 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2056 An attacker can use this to craft malformed certificates and CRLs of
2057 various sizes and potentially cause a segmentation fault, resulting in
2058 a DoS on applications that verify certificates or CRLs. TLS clients
2059 that verify CRLs are affected. TLS clients and servers with client
2060 authentication enabled may be affected if they use custom verification
2063 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2064 independently by Hanno Böck.
2068 *) PKCS7 crash with missing EnvelopedContent
2070 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2071 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2072 with missing content and trigger a NULL pointer dereference on parsing.
2074 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2075 structures from untrusted sources are affected. OpenSSL clients and
2076 servers are not affected.
2078 This issue was reported to OpenSSL by Michal Zalewski (Google).
2082 *) CMS verify infinite loop with unknown hash function
2084 When verifying a signedData message the CMS code can enter an infinite loop
2085 if presented with an unknown hash function OID. This can be used to perform
2086 denial of service against any system which verifies signedData messages using
2088 This issue was reported to OpenSSL by Johannes Bauer.
2092 *) Race condition handling NewSessionTicket
2094 If a NewSessionTicket is received by a multi-threaded client when attempting to
2095 reuse a previous ticket then a race condition can occur potentially leading to
2096 a double free of the ticket data.
2100 *) Only support 256-bit or stronger elliptic curves with the
2101 'ecdh_auto' setting (server) or by default (client). Of supported
2102 curves, prefer P-256 (both).
2105 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2107 *) ClientHello sigalgs DoS fix
2109 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2110 invalid signature algorithms extension a NULL pointer dereference will
2111 occur. This can be exploited in a DoS attack against the server.
2113 This issue was was reported to OpenSSL by David Ramos of Stanford
2116 [Stephen Henson and Matt Caswell]
2118 *) Multiblock corrupted pointer fix
2120 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2121 feature only applies on 64 bit x86 architecture platforms that support AES
2122 NI instructions. A defect in the implementation of "multiblock" can cause
2123 OpenSSL's internal write buffer to become incorrectly set to NULL when
2124 using non-blocking IO. Typically, when the user application is using a
2125 socket BIO for writing, this will only result in a failed connection.
2126 However if some other BIO is used then it is likely that a segmentation
2127 fault will be triggered, thus enabling a potential DoS attack.
2129 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2133 *) Segmentation fault in DTLSv1_listen fix
2135 The DTLSv1_listen function is intended to be stateless and processes the
2136 initial ClientHello from many peers. It is common for user code to loop
2137 over the call to DTLSv1_listen until a valid ClientHello is received with
2138 an associated cookie. A defect in the implementation of DTLSv1_listen means
2139 that state is preserved in the SSL object from one invocation to the next
2140 that can lead to a segmentation fault. Errors processing the initial
2141 ClientHello can trigger this scenario. An example of such an error could be
2142 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2145 This issue was reported to OpenSSL by Per Allansson.
2149 *) Segmentation fault in ASN1_TYPE_cmp fix
2151 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2152 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2153 certificate signature algorithm consistency this can be used to crash any
2154 certificate verification operation and exploited in a DoS attack. Any
2155 application which performs certificate verification is vulnerable including
2156 OpenSSL clients and servers which enable client authentication.
2160 *) Segmentation fault for invalid PSS parameters fix
2162 The signature verification routines will crash with a NULL pointer
2163 dereference if presented with an ASN.1 signature using the RSA PSS
2164 algorithm and invalid parameters. Since these routines are used to verify
2165 certificate signature algorithms this can be used to crash any
2166 certificate verification operation and exploited in a DoS attack. Any
2167 application which performs certificate verification is vulnerable including
2168 OpenSSL clients and servers which enable client authentication.
2170 This issue was was reported to OpenSSL by Brian Carpenter.
2174 *) ASN.1 structure reuse memory corruption fix
2176 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2177 memory corruption via an invalid write. Such reuse is and has been
2178 strongly discouraged and is believed to be rare.
2180 Applications that parse structures containing CHOICE or ANY DEFINED BY
2181 components may be affected. Certificate parsing (d2i_X509 and related
2182 functions) are however not affected. OpenSSL clients and servers are
2187 *) PKCS7 NULL pointer dereferences fix
2189 The PKCS#7 parsing code does not handle missing outer ContentInfo
2190 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2191 missing content and trigger a NULL pointer dereference on parsing.
2193 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2194 otherwise parse PKCS#7 structures from untrusted sources are
2195 affected. OpenSSL clients and servers are not affected.
2197 This issue was reported to OpenSSL by Michal Zalewski (Google).
2201 *) DoS via reachable assert in SSLv2 servers fix
2203 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2204 servers that both support SSLv2 and enable export cipher suites by sending
2205 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2207 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2208 (OpenSSL development team).
2212 *) Empty CKE with client auth and DHE fix
2214 If client auth is used then a server can seg fault in the event of a DHE
2215 ciphersuite being selected and a zero length ClientKeyExchange message
2216 being sent by the client. This could be exploited in a DoS attack.
2220 *) Handshake with unseeded PRNG fix
2222 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2223 with an unseeded PRNG. The conditions are:
2224 - The client is on a platform where the PRNG has not been seeded
2225 automatically, and the user has not seeded manually
2226 - A protocol specific client method version has been used (i.e. not
2227 SSL_client_methodv23)
2228 - A ciphersuite is used that does not require additional random data from
2229 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2231 If the handshake succeeds then the client random that has been used will
2232 have been generated from a PRNG with insufficient entropy and therefore the
2233 output may be predictable.
2235 For example using the following command with an unseeded openssl will
2236 succeed on an unpatched platform:
2238 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2242 *) Use After Free following d2i_ECPrivatekey error fix
2244 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2245 could cause a use after free condition. This, in turn, could cause a double
2246 free in several private key parsing functions (such as d2i_PrivateKey
2247 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2248 for applications that receive EC private keys from untrusted
2249 sources. This scenario is considered rare.
2251 This issue was discovered by the BoringSSL project and fixed in their
2256 *) X509_to_X509_REQ NULL pointer deref fix
2258 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2259 the certificate key is invalid. This function is rarely used in practice.
2261 This issue was discovered by Brian Carpenter.
2265 *) Removed the export ciphers from the DEFAULT ciphers
2268 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2270 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2271 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2272 So far those who have to target multiple platforms would compromise
2273 and argue that binary targeting say ARMv5 would still execute on
2274 ARMv8. "Universal" build resolves this compromise by providing
2275 near-optimal performance even on newer platforms.
2278 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2279 (other platforms pending).
2280 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2282 *) Add support for the SignedCertificateTimestampList certificate and
2283 OCSP response extensions from RFC6962.
2286 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2287 for corner cases. (Certain input points at infinity could lead to
2288 bogus results, with non-infinity inputs mapped to infinity too.)
2291 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2292 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2293 common cases are optimized and there still is room for further
2294 improvements. Vector Permutation AES for Altivec is also added.
2297 *) Add support for little-endian ppc64 Linux target.
2298 [Marcelo Cerri (IBM)]
2300 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2301 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2302 are optimized and there still is room for further improvements.
2303 Both 32- and 64-bit modes are supported.
2304 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2306 *) Improved ARMv7 NEON support.
2309 *) Support for SPARC Architecture 2011 crypto extensions, first
2310 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2311 SHA256/512, MD5, GHASH and modular exponentiation.
2312 [Andy Polyakov, David Miller]
2314 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2316 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2318 *) Support for new and upcoming Intel processors, including AVX2,
2319 BMI and SHA ISA extensions. This includes additional "stitched"
2320 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2323 This work was sponsored by Intel Corp.
2326 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2327 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2328 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2331 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2332 this fixes a limitation in previous versions of OpenSSL.
2335 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2336 MGF1 digest and OAEP label.
2339 *) Add EVP support for key wrapping algorithms, to avoid problems with
2340 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2341 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2342 algorithms and include tests cases.
2345 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2347 [Douglas E. Engert, Steve Henson]
2349 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2350 difference in days and seconds between two tm or ASN1_TIME structures.
2353 *) Add -rev test option to s_server to just reverse order of characters
2354 received by client and send back to server. Also prints an abbreviated
2355 summary of the connection parameters.
2358 *) New option -brief for s_client and s_server to print out a brief summary
2359 of connection parameters.
2362 *) Add callbacks for arbitrary TLS extensions.
2363 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2365 *) New option -crl_download in several openssl utilities to download CRLs
2366 from CRLDP extension in certificates.
2369 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2372 *) New function X509_CRL_diff to generate a delta CRL from the difference
2373 of two full CRLs. Add support to "crl" utility.
2376 *) New functions to set lookup_crls function and to retrieve
2377 X509_STORE from X509_STORE_CTX.
2380 *) Print out deprecated issuer and subject unique ID fields in
2384 *) Extend OCSP I/O functions so they can be used for simple general purpose
2385 HTTP as well as OCSP. New wrapper function which can be used to download
2386 CRLs using the OCSP API.
2389 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2392 *) SSL_CONF* functions. These provide a common framework for application
2393 configuration using configuration files or command lines.
2396 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2397 message callback and prints the results. Needs compile time option
2398 "enable-ssl-trace". New options to s_client and s_server to enable
2402 *) New ctrl and macro to retrieve supported points extensions.
2403 Print out extension in s_server and s_client.
2406 *) New functions to retrieve certificate signature and signature
2410 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2414 *) New Suite B modes for TLS code. These use and enforce the requirements
2415 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2416 only use Suite B curves. The Suite B modes can be set by using the
2417 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2420 *) New chain verification flags for Suite B levels of security. Check
2421 algorithms are acceptable when flags are set in X509_verify_cert.
2424 *) Make tls1_check_chain return a set of flags indicating checks passed
2425 by a certificate chain. Add additional tests to handle client
2426 certificates: checks for matching certificate type and issuer name
2430 *) If an attempt is made to use a signature algorithm not in the peer
2431 preference list abort the handshake. If client has no suitable
2432 signature algorithms in response to a certificate request do not
2433 use the certificate.
2436 *) If server EC tmp key is not in client preference list abort handshake.
2439 *) Add support for certificate stores in CERT structure. This makes it
2440 possible to have different stores per SSL structure or one store in
2441 the parent SSL_CTX. Include distinct stores for certificate chain
2442 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2443 to build and store a certificate chain in CERT structure: returning
2444 an error if the chain cannot be built: this will allow applications
2445 to test if a chain is correctly configured.
2447 Note: if the CERT based stores are not set then the parent SSL_CTX
2448 store is used to retain compatibility with existing behaviour.
2452 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2453 mask based on the current session, check mask when sending client
2454 hello and checking the requested ciphersuite.
2457 *) New ctrls to retrieve and set certificate types in a certificate
2458 request message. Print out received values in s_client. If certificate
2459 types is not set with custom values set sensible values based on
2460 supported signature algorithms.
2463 *) Support for distinct client and server supported signature algorithms.
2466 *) Add certificate callback. If set this is called whenever a certificate
2467 is required by client or server. An application can decide which
2468 certificate chain to present based on arbitrary criteria: for example
2469 supported signature algorithms. Add very simple example to s_server.
2470 This fixes many of the problems and restrictions of the existing client
2471 certificate callback: for example you can now clear an existing
2472 certificate and specify the whole chain.
2475 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2476 the certificate can be used for (if anything). Set valid_flags field
2477 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2478 to have similar checks in it.
2480 Add new "cert_flags" field to CERT structure and include a "strict mode".
2481 This enforces some TLS certificate requirements (such as only permitting
2482 certificate signature algorithms contained in the supported algorithms
2483 extension) which some implementations ignore: this option should be used
2484 with caution as it could cause interoperability issues.
2487 *) Update and tidy signature algorithm extension processing. Work out
2488 shared signature algorithms based on preferences and peer algorithms
2489 and print them out in s_client and s_server. Abort handshake if no
2490 shared signature algorithms.
2493 *) Add new functions to allow customised supported signature algorithms
2494 for SSL and SSL_CTX structures. Add options to s_client and s_server
2498 *) New function SSL_certs_clear() to delete all references to certificates
2499 from an SSL structure. Before this once a certificate had been added
2500 it couldn't be removed.
2503 *) Integrate hostname, email address and IP address checking with certificate
2504 verification. New verify options supporting checking in openssl utility.
2507 *) Fixes and wildcard matching support to hostname and email checking
2508 functions. Add manual page.
2509 [Florian Weimer (Red Hat Product Security Team)]
2511 *) New functions to check a hostname email or IP address against a
2512 certificate. Add options x509 utility to print results of checks against
2516 *) Fix OCSP checking.
2517 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2519 *) Initial experimental support for explicitly trusted non-root CAs.
2520 OpenSSL still tries to build a complete chain to a root but if an
2521 intermediate CA has a trust setting included that is used. The first
2522 setting is used: whether to trust (e.g., -addtrust option to the x509
2526 *) Add -trusted_first option which attempts to find certificates in the
2527 trusted store even if an untrusted chain is also supplied.
2530 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2531 platform support for Linux and Android.
2534 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2537 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2538 When in FIPS mode the approved implementations are used as normal,
2539 when not in FIPS mode the internal unapproved versions are used instead.
2540 This means that the FIPS capable OpenSSL isn't forced to use the
2541 (often lower performance) FIPS implementations outside FIPS mode.
2544 *) Transparently support X9.42 DH parameters when calling
2545 PEM_read_bio_DHparameters. This means existing applications can handle
2546 the new parameter format automatically.
2549 *) Initial experimental support for X9.42 DH parameter format: mainly
2550 to support use of 'q' parameter for RFC5114 parameters.
2553 *) Add DH parameters from RFC5114 including test data to dhtest.
2556 *) Support for automatic EC temporary key parameter selection. If enabled
2557 the most preferred EC parameters are automatically used instead of
2558 hardcoded fixed parameters. Now a server just has to call:
2559 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2560 support ECDH and use the most appropriate parameters.
2563 *) Enhance and tidy EC curve and point format TLS extension code. Use
2564 static structures instead of allocation if default values are used.
2565 New ctrls to set curves we wish to support and to retrieve shared curves.
2566 Print out shared curves in s_server. New options to s_server and s_client
2567 to set list of supported curves.
2570 *) New ctrls to retrieve supported signature algorithms and
2571 supported curve values as an array of NIDs. Extend openssl utility
2572 to print out received values.
2575 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2576 between NIDs and the more common NIST names such as "P-256". Enhance
2577 ecparam utility and ECC method to recognise the NIST names for curves.
2580 *) Enhance SSL/TLS certificate chain handling to support different
2581 chains for each certificate instead of one chain in the parent SSL_CTX.
2584 *) Support for fixed DH ciphersuite client authentication: where both
2585 server and client use DH certificates with common parameters.
2588 *) Support for fixed DH ciphersuites: those requiring DH server
2592 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2594 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2595 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2596 X509_CINF_get_signature were reverted post internal team review.
2598 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2600 *) Build fixes for the Windows and OpenVMS platforms
2601 [Matt Caswell and Richard Levitte]
2603 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2605 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2606 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2607 dereference. This could lead to a Denial Of Service attack. Thanks to
2608 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2612 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2613 dtls1_buffer_record function under certain conditions. In particular this
2614 could occur if an attacker sent repeated DTLS records with the same
2615 sequence number but for the next epoch. The memory leak could be exploited
2616 by an attacker in a Denial of Service attack through memory exhaustion.
2617 Thanks to Chris Mueller for reporting this issue.
2621 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2622 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2623 method would be set to NULL which could later result in a NULL pointer
2624 dereference. Thanks to Frank Schmirler for reporting this issue.
2628 *) Abort handshake if server key exchange message is omitted for ephemeral
2631 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2632 reporting this issue.
2636 *) Remove non-export ephemeral RSA code on client and server. This code
2637 violated the TLS standard by allowing the use of temporary RSA keys in
2638 non-export ciphersuites and could be used by a server to effectively
2639 downgrade the RSA key length used to a value smaller than the server
2640 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2641 INRIA or reporting this issue.
2645 *) Fixed issue where DH client certificates are accepted without verification.
2646 An OpenSSL server will accept a DH certificate for client authentication
2647 without the certificate verify message. This effectively allows a client to
2648 authenticate without the use of a private key. This only affects servers
2649 which trust a client certificate authority which issues certificates
2650 containing DH keys: these are extremely rare and hardly ever encountered.
2651 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2656 *) Ensure that the session ID context of an SSL is updated when its
2657 SSL_CTX is updated via SSL_set_SSL_CTX.
2659 The session ID context is typically set from the parent SSL_CTX,
2660 and can vary with the CTX.
2663 *) Fix various certificate fingerprint issues.
2665 By using non-DER or invalid encodings outside the signed portion of a
2666 certificate the fingerprint can be changed without breaking the signature.
2667 Although no details of the signed portion of the certificate can be changed
2668 this can cause problems with some applications: e.g. those using the
2669 certificate fingerprint for blacklists.
2671 1. Reject signatures with non zero unused bits.
2673 If the BIT STRING containing the signature has non zero unused bits reject
2674 the signature. All current signature algorithms require zero unused bits.
2676 2. Check certificate algorithm consistency.
2678 Check the AlgorithmIdentifier inside TBS matches the one in the
2679 certificate signature. NB: this will result in signature failure
2680 errors for some broken certificates.
2682 Thanks to Konrad Kraszewski from Google for reporting this issue.
2684 3. Check DSA/ECDSA signatures use DER.
2686 Re-encode DSA/ECDSA signatures and compare with the original received
2687 signature. Return an error if there is a mismatch.
2689 This will reject various cases including garbage after signature
2690 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2691 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2692 (negative or with leading zeroes).
2694 Further analysis was conducted and fixes were developed by Stephen Henson
2695 of the OpenSSL core team.
2700 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2701 results on some platforms, including x86_64. This bug occurs at random
2702 with a very low probability, and is not known to be exploitable in any
2703 way, though its exact impact is difficult to determine. Thanks to Pieter
2704 Wuille (Blockstream) who reported this issue and also suggested an initial
2705 fix. Further analysis was conducted by the OpenSSL development team and
2706 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2707 the OpenSSL core team.
2711 *) Do not resume sessions on the server if the negotiated protocol
2712 version does not match the session's version. Resuming with a different
2713 version, while not strictly forbidden by the RFC, is of questionable
2714 sanity and breaks all known clients.
2715 [David Benjamin, Emilia Käsper]
2717 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2718 early CCS messages during renegotiation. (Note that because
2719 renegotiation is encrypted, this early CCS was not exploitable.)
2722 *) Tighten client-side session ticket handling during renegotiation:
2723 ensure that the client only accepts a session ticket if the server sends
2724 the extension anew in the ServerHello. Previously, a TLS client would
2725 reuse the old extension state and thus accept a session ticket if one was
2726 announced in the initial ServerHello.
2728 Similarly, ensure that the client requires a session ticket if one
2729 was advertised in the ServerHello. Previously, a TLS client would
2730 ignore a missing NewSessionTicket message.
2733 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2735 *) SRTP Memory Leak.
2737 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2738 sends a carefully crafted handshake message, to cause OpenSSL to fail
2739 to free up to 64k of memory causing a memory leak. This could be
2740 exploited in a Denial Of Service attack. This issue affects OpenSSL
2741 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2742 whether SRTP is used or configured. Implementations of OpenSSL that
2743 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2745 The fix was developed by the OpenSSL team.
2749 *) Session Ticket Memory Leak.
2751 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2752 integrity of that ticket is first verified. In the event of a session
2753 ticket integrity check failing, OpenSSL will fail to free memory
2754 causing a memory leak. By sending a large number of invalid session
2755 tickets an attacker could exploit this issue in a Denial Of Service
2760 *) Build option no-ssl3 is incomplete.
2762 When OpenSSL is configured with "no-ssl3" as a build option, servers
2763 could accept and complete a SSL 3.0 handshake, and clients could be
2764 configured to send them.
2766 [Akamai and the OpenSSL team]
2768 *) Add support for TLS_FALLBACK_SCSV.
2769 Client applications doing fallback retries should call
2770 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2772 [Adam Langley, Bodo Moeller]
2774 *) Add additional DigestInfo checks.
2776 Re-encode DigestInto in DER and check against the original when
2777 verifying RSA signature: this will reject any improperly encoded
2778 DigestInfo structures.
2780 Note: this is a precautionary measure and no attacks are currently known.
2784 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2786 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2787 SRP code can be overrun an internal buffer. Add sanity check that
2788 g, A, B < N to SRP code.
2790 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2791 Group for discovering this issue.
2795 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2796 TLS 1.0 instead of higher protocol versions when the ClientHello message
2797 is badly fragmented. This allows a man-in-the-middle attacker to force a
2798 downgrade to TLS 1.0 even if both the server and the client support a
2799 higher protocol version, by modifying the client's TLS records.
2801 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2802 researching this issue.
2806 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2807 to a denial of service attack. A malicious server can crash the client
2808 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2809 ciphersuite and sending carefully crafted handshake messages.
2811 Thanks to Felix Gröbert (Google) for discovering and researching this
2816 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2817 to leak memory. This can be exploited through a Denial of Service attack.
2818 Thanks to Adam Langley for discovering and researching this issue.
2822 *) An attacker can force openssl to consume large amounts of memory whilst
2823 processing DTLS handshake messages. This can be exploited through a
2824 Denial of Service attack.
2825 Thanks to Adam Langley for discovering and researching this issue.
2829 *) An attacker can force an error condition which causes openssl to crash
2830 whilst processing DTLS packets due to memory being freed twice. This
2831 can be exploited through a Denial of Service attack.
2832 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2837 *) If a multithreaded client connects to a malicious server using a resumed
2838 session and the server sends an ec point format extension it could write
2839 up to 255 bytes to freed memory.
2841 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2846 *) A malicious server can crash an OpenSSL client with a null pointer
2847 dereference (read) by specifying an SRP ciphersuite even though it was not
2848 properly negotiated with the client. This can be exploited through a
2849 Denial of Service attack.
2851 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2852 discovering and researching this issue.
2856 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2857 X509_name_oneline, X509_name_print_ex et al. to leak some information
2858 from the stack. Applications may be affected if they echo pretty printing
2859 output to the attacker.
2861 Thanks to Ivan Fratric (Google) for discovering this issue.
2863 [Emilia Käsper, and Steve Henson]
2865 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2866 for corner cases. (Certain input points at infinity could lead to
2867 bogus results, with non-infinity inputs mapped to infinity too.)
2870 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2872 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2873 handshake can force the use of weak keying material in OpenSSL
2874 SSL/TLS clients and servers.
2876 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2877 researching this issue. (CVE-2014-0224)
2878 [KIKUCHI Masashi, Steve Henson]
2880 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2881 OpenSSL DTLS client the code can be made to recurse eventually crashing
2884 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2886 [Imre Rad, Steve Henson]
2888 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2889 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2890 client or server. This is potentially exploitable to run arbitrary
2891 code on a vulnerable client or server.
2893 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2894 [Jüri Aedla, Steve Henson]
2896 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2897 are subject to a denial of service attack.
2899 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2900 this issue. (CVE-2014-3470)
2901 [Felix Gröbert, Ivan Fratric, Steve Henson]
2903 *) Harmonize version and its documentation. -f flag is used to display
2905 [mancha <mancha1@zoho.com>]
2907 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2908 in i2d_ECPrivateKey.
2909 [mancha <mancha1@zoho.com>]
2911 *) Fix some double frees. These are not thought to be exploitable.
2912 [mancha <mancha1@zoho.com>]
2914 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2916 *) A missing bounds check in the handling of the TLS heartbeat extension
2917 can be used to reveal up to 64k of memory to a connected client or
2920 Thanks for Neel Mehta of Google Security for discovering this bug and to
2921 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2922 preparing the fix (CVE-2014-0160)
2923 [Adam Langley, Bodo Moeller]
2925 *) Fix for the attack described in the paper "Recovering OpenSSL
2926 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2927 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2928 http://eprint.iacr.org/2014/140
2930 Thanks to Yuval Yarom and Naomi Benger for discovering this
2931 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2932 [Yuval Yarom and Naomi Benger]
2934 *) TLS pad extension: draft-agl-tls-padding-03
2936 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2937 TLS client Hello record length value would otherwise be > 255 and
2938 less that 512 pad with a dummy extension containing zeroes so it
2939 is at least 512 bytes long.
2941 [Adam Langley, Steve Henson]
2943 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2945 *) Fix for TLS record tampering bug. A carefully crafted invalid
2946 handshake could crash OpenSSL with a NULL pointer exception.
2947 Thanks to Anton Johansson for reporting this issues.
2950 *) Keep original DTLS digest and encryption contexts in retransmission
2951 structures so we can use the previous session parameters if they need
2952 to be resent. (CVE-2013-6450)
2955 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2956 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2957 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2958 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2959 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2960 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2961 [Rob Stradling, Adam Langley]
2963 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2965 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2966 supporting platforms or when small records were transferred.
2967 [Andy Polyakov, Steve Henson]
2969 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2971 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2973 This addresses the flaw in CBC record processing discovered by
2974 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2975 at: http://www.isg.rhul.ac.uk/tls/
2977 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2978 Security Group at Royal Holloway, University of London
2979 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2980 Emilia Käsper for the initial patch.
2982 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2984 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2985 ciphersuites which can be exploited in a denial of service attack.
2986 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2987 and detecting this bug and to Wolfgang Ettlinger
2988 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2992 *) Return an error when checking OCSP signatures when key is NULL.
2993 This fixes a DoS attack. (CVE-2013-0166)
2996 *) Make openssl verify return errors.
2997 [Chris Palmer <palmer@google.com> and Ben Laurie]
2999 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3000 the right response is stapled. Also change SSL_get_certificate()
3001 so it returns the certificate actually sent.
3002 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3003 [Rob Stradling <rob.stradling@comodo.com>]
3005 *) Fix possible deadlock when decoding public keys.
3008 *) Don't use TLS 1.0 record version number in initial client hello
3012 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3014 *) Sanity check record length before skipping explicit IV in TLS
3015 1.2, 1.1 and DTLS to fix DoS attack.
3017 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3018 fuzzing as a service testing platform.
3022 *) Initialise tkeylen properly when encrypting CMS messages.
3023 Thanks to Solar Designer of Openwall for reporting this issue.
3026 *) In FIPS mode don't try to use composite ciphers as they are not
3030 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3032 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3033 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3034 mean any application compiled against OpenSSL 1.0.0 headers setting
3035 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3036 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3037 0x10000000L Any application which was previously compiled against
3038 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3039 will need to be recompiled as a result. Letting be results in
3040 inability to disable specifically TLS 1.1 and in client context,
3041 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3044 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3045 disable just protocol X, but all protocols above X *if* there are
3046 protocols *below* X still enabled. In more practical terms it means
3047 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3048 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3049 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3053 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3055 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3056 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3057 in CRYPTO_realloc_clean.
3059 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3060 issue and to Adam Langley <agl@chromium.org> for fixing it.
3062 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3064 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3067 *) Workarounds for some broken servers that "hang" if a client hello
3068 record length exceeds 255 bytes.
3070 1. Do not use record version number > TLS 1.0 in initial client
3071 hello: some (but not all) hanging servers will now work.
3072 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3073 the number of ciphers sent in the client hello. This should be
3074 set to an even number, such as 50, for example by passing:
3075 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3076 Most broken servers should now work.
3077 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3078 TLS 1.2 client support entirely.
3081 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3084 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3086 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3087 STRING form instead of a DigestInfo.
3090 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3091 and the RSA_sign/RSA_verify functions. This was made more apparent when
3092 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3093 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3094 the correct format in RSA_verify so both forms transparently work.
3097 *) Some servers which support TLS 1.0 can choke if we initially indicate
3098 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3099 encrypted premaster secret. As a workaround use the maximum permitted
3100 client version in client hello, this should keep such servers happy
3101 and still work with previous versions of OpenSSL.
3104 *) Add support for TLS/DTLS heartbeats.
3105 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3107 *) Add support for SCTP.
3108 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3110 *) Improved PRNG seeding for VOS.
3111 [Paul Green <Paul.Green@stratus.com>]
3113 *) Extensive assembler packs updates, most notably:
3115 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3116 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3117 - x86_64: bit-sliced AES implementation;
3118 - ARM: NEON support, contemporary platforms optimizations;
3119 - s390x: z196 support;
3120 - *: GHASH and GF(2^m) multiplication implementations;
3124 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3125 (removal of unnecessary code)
3126 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3128 *) Add TLS key material exporter from RFC 5705.
3131 *) Add DTLS-SRTP negotiation from RFC 5764.
3134 *) Add Next Protocol Negotiation,
3135 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3136 disabled with a no-npn flag to config or Configure. Code donated
3138 [Adam Langley <agl@google.com> and Ben Laurie]
3140 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3141 NIST-P256, NIST-P521, with constant-time single point multiplication on
3142 typical inputs. Compiler support for the nonstandard type __uint128_t is
3143 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3144 Code made available under Apache License version 2.0.
3146 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3147 line to include this in your build of OpenSSL, and run "make depend" (or
3148 "make update"). This enables the following EC_METHODs:
3150 EC_GFp_nistp224_method()
3151 EC_GFp_nistp256_method()
3152 EC_GFp_nistp521_method()
3154 EC_GROUP_new_by_curve_name() will automatically use these (while
3155 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3157 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3159 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3160 all platforms. Move ssize_t definition from e_os.h to the public
3161 header file e_os2.h as it now appears in public header file cms.h
3164 *) New -sigopt option to the ca, req and x509 utilities. Additional
3165 signature parameters can be passed using this option and in
3169 *) Add RSA PSS signing function. This will generate and set the
3170 appropriate AlgorithmIdentifiers for PSS based on those in the
3171 corresponding EVP_MD_CTX structure. No application support yet.
3174 *) Support for companion algorithm specific ASN1 signing routines.
3175 New function ASN1_item_sign_ctx() signs a pre-initialised
3176 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3177 the appropriate parameters.
3180 *) Add new algorithm specific ASN1 verification initialisation function
3181 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3182 handling will be the same no matter what EVP_PKEY_METHOD is used.
3183 Add a PSS handler to support verification of PSS signatures: checked
3184 against a number of sample certificates.
3187 *) Add signature printing for PSS. Add PSS OIDs.
3188 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3190 *) Add algorithm specific signature printing. An individual ASN1 method
3191 can now print out signatures instead of the standard hex dump.
3193 More complex signatures (e.g. PSS) can print out more meaningful
3194 information. Include DSA version that prints out the signature
3198 *) Password based recipient info support for CMS library: implementing
3202 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3203 neatly separates the code into cipher and PBE sections and is required
3204 for some algorithms that split PBES2 into separate pieces (such as
3205 password based CMS).
3208 *) Session-handling fixes:
3209 - Fix handling of connections that are resuming with a session ID,
3210 but also support Session Tickets.
3211 - Fix a bug that suppressed issuing of a new ticket if the client
3212 presented a ticket with an expired session.
3213 - Try to set the ticket lifetime hint to something reasonable.
3214 - Make tickets shorter by excluding irrelevant information.
3215 - On the client side, don't ignore renewed tickets.
3216 [Adam Langley, Bodo Moeller (Google)]
3218 *) Fix PSK session representation.
3221 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3223 This work was sponsored by Intel.
3226 *) Add GCM support to TLS library. Some custom code is needed to split
3227 the IV between the fixed (from PRF) and explicit (from TLS record)
3228 portions. This adds all GCM ciphersuites supported by RFC5288 and
3229 RFC5289. Generalise some AES* cipherstrings to include GCM and
3230 add a special AESGCM string for GCM only.
3233 *) Expand range of ctrls for AES GCM. Permit setting invocation
3234 field on decrypt and retrieval of invocation field only on encrypt.
3237 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3238 As required by RFC5289 these ciphersuites cannot be used if for
3239 versions of TLS earlier than 1.2.
3242 *) For FIPS capable OpenSSL interpret a NULL default public key method
3243 as unset and return the appropriate default but do *not* set the default.
3244 This means we can return the appropriate method in applications that
3245 switch between FIPS and non-FIPS modes.
3248 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3249 ENGINE is used then we cannot handle that in the FIPS module so we
3250 keep original code iff non-FIPS operations are allowed.
3253 *) Add -attime option to openssl utilities.
3254 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3256 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3259 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3260 FIPS EC methods unconditionally for now.
3263 *) New build option no-ec2m to disable characteristic 2 code.
3266 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3267 all cases can be covered as some introduce binary incompatibilities.
3270 *) Redirect RSA operations to FIPS module including keygen,
3271 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3274 *) Add similar low level API blocking to ciphers.
3277 *) Low level digest APIs are not approved in FIPS mode: any attempt
3278 to use these will cause a fatal error. Applications that *really* want
3279 to use them can use the private_* version instead.
3282 *) Redirect cipher operations to FIPS module for FIPS builds.
3285 *) Redirect digest operations to FIPS module for FIPS builds.
3288 *) Update build system to add "fips" flag which will link in fipscanister.o
3289 for static and shared library builds embedding a signature if needed.
3292 *) Output TLS supported curves in preference order instead of numerical
3293 order. This is currently hardcoded for the highest order curves first.
3294 This should be configurable so applications can judge speed vs strength.
3297 *) Add TLS v1.2 server support for client authentication.
3300 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3304 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3305 FIPS modules versions.
3308 *) Add TLS v1.2 client side support for client authentication. Keep cache
3309 of handshake records longer as we don't know the hash algorithm to use
3310 until after the certificate request message is received.
3313 *) Initial TLS v1.2 client support. Add a default signature algorithms
3314 extension including all the algorithms we support. Parse new signature
3315 format in client key exchange. Relax some ECC signing restrictions for
3316 TLS v1.2 as indicated in RFC5246.
3319 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3320 to new signature format when needed using client digest preference.
3321 All server ciphersuites should now work correctly in TLS v1.2. No client
3322 support yet and no support for client certificates.
3325 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3326 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3327 ciphersuites. At present only RSA key exchange ciphersuites work with
3328 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3329 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3330 and version checking.
3333 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3334 with this defined it will not be affected by any changes to ssl internal
3335 structures. Add several utility functions to allow openssl application
3336 to work with OPENSSL_NO_SSL_INTERN defined.
3339 *) A long standing patch to add support for SRP from EdelWeb (Peter
3340 Sylvester and Christophe Renou) was integrated.
3341 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3342 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3345 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3348 *) Permit abbreviated handshakes when renegotiating using the function
3349 SSL_renegotiate_abbreviated().
3350 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3352 *) Add call to ENGINE_register_all_complete() to
3353 ENGINE_load_builtin_engines(), so some implementations get used
3354 automatically instead of needing explicit application support.
3357 *) Add support for TLS key exporter as described in RFC5705.
3358 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3360 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3361 a few changes are required:
3363 Add SSL_OP_NO_TLSv1_1 flag.
3364 Add TLSv1_1 methods.
3365 Update version checking logic to handle version 1.1.
3366 Add explicit IV handling (ported from DTLS code).
3367 Add command line options to s_client/s_server.
3370 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3372 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3373 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3374 content decryption and always return the same error. Note: this attack
3375 needs on average 2^20 messages so it only affects automated senders. The
3376 old behaviour can be re-enabled in the CMS code by setting the
3377 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3378 an MMA defence is not necessary.
3379 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3380 this issue. (CVE-2012-0884)
3383 *) Fix CVE-2011-4619: make sure we really are receiving a
3384 client hello before rejecting multiple SGC restarts. Thanks to
3385 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3388 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3390 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3391 Thanks to Antonio Martin, Enterprise Secure Access Research and
3392 Development, Cisco Systems, Inc. for discovering this bug and
3393 preparing a fix. (CVE-2012-0050)
3396 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3398 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3399 of the Vaudenay padding oracle attack on CBC mode encryption
3400 which enables an efficient plaintext recovery attack against
3401 the OpenSSL implementation of DTLS. Their attack exploits timing
3402 differences arising during decryption processing. A research
3403 paper describing this attack can be found at:
3404 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3405 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3406 Security Group at Royal Holloway, University of London
3407 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3408 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3409 for preparing the fix. (CVE-2011-4108)
3410 [Robin Seggelmann, Michael Tuexen]
3412 *) Clear bytes used for block padding of SSL 3.0 records.
3414 [Adam Langley (Google)]
3416 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3417 Kadianakis <desnacked@gmail.com> for discovering this issue and
3418 Adam Langley for preparing the fix. (CVE-2011-4619)
3419 [Adam Langley (Google)]
3421 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3422 [Andrey Kulikov <amdeich@gmail.com>]
3424 *) Prevent malformed RFC3779 data triggering an assertion failure.
3425 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3426 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3427 [Rob Austein <sra@hactrn.net>]
3429 *) Improved PRNG seeding for VOS.
3430 [Paul Green <Paul.Green@stratus.com>]
3432 *) Fix ssl_ciph.c set-up race.
3433 [Adam Langley (Google)]
3435 *) Fix spurious failures in ecdsatest.c.
3436 [Emilia Käsper (Google)]
3438 *) Fix the BIO_f_buffer() implementation (which was mixing different
3439 interpretations of the '..._len' fields).
3440 [Adam Langley (Google)]
3442 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3443 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3444 threads won't reuse the same blinding coefficients.
3446 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3447 lock to call BN_BLINDING_invert_ex, and avoids one use of
3448 BN_BLINDING_update for each BN_BLINDING structure (previously,
3449 the last update always remained unused).
3450 [Emilia Käsper (Google)]
3452 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3453 [Bob Buckholz (Google)]
3455 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3457 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3458 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3459 [Kaspar Brand <ossl@velox.ch>]
3461 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3462 for multi-threaded use of ECDH. (CVE-2011-3210)
3463 [Adam Langley (Google)]
3465 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3468 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3469 signature public key algorithm by using OID xref utilities instead.
3470 Before this you could only use some ECC ciphersuites with SHA1 only.
3473 *) Add protection against ECDSA timing attacks as mentioned in the paper
3474 by Billy Bob Brumley and Nicola Tuveri, see:
3476 http://eprint.iacr.org/2011/232.pdf
3478 [Billy Bob Brumley and Nicola Tuveri]
3480 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3482 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3483 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3485 *) Fix bug in string printing code: if *any* escaping is enabled we must
3486 escape the escape character (backslash) or the resulting string is
3490 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3492 *) Disable code workaround for ancient and obsolete Netscape browsers
3493 and servers: an attacker can use it in a ciphersuite downgrade attack.
3494 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3497 *) Fixed J-PAKE implementation error, originally discovered by
3498 Sebastien Martini, further info and confirmation from Stefan
3499 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3502 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3504 *) Fix extension code to avoid race conditions which can result in a buffer
3505 overrun vulnerability: resumed sessions must not be modified as they can
3506 be shared by multiple threads. CVE-2010-3864
3509 *) Fix WIN32 build system to correctly link an ENGINE directory into
3513 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3515 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3517 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3519 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3521 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3522 context. The operation can be customised via the ctrl mechanism in
3523 case ENGINEs want to include additional functionality.
3526 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3529 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3530 output hashes compatible with older versions of OpenSSL.
3531 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3533 *) Fix compression algorithm handling: if resuming a session use the
3534 compression algorithm of the resumed session instead of determining
3535 it from client hello again. Don't allow server to change algorithm.
3538 *) Add load_crls() function to apps tidying load_certs() too. Add option
3539 to verify utility to allow additional CRLs to be included.
3542 *) Update OCSP request code to permit adding custom headers to the request:
3543 some responders need this.
3546 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3548 [Julia Lawall <julia@diku.dk>]
3550 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3551 needlessly dereferenced structures, used obsolete functions and
3552 didn't handle all updated verify codes correctly.
3555 *) Disable MD2 in the default configuration.
3558 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3559 indicate the initial BIO being pushed or popped. This makes it possible
3560 to determine whether the BIO is the one explicitly called or as a result
3561 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3562 it handles reference counts correctly and doesn't zero out the I/O bio
3563 when it is not being explicitly popped. WARNING: applications which
3564 included workarounds for the old buggy behaviour will need to be modified
3565 or they could free up already freed BIOs.
3568 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3569 renaming to all platforms (within the 0.9.8 branch, this was
3570 done conditionally on Netware platforms to avoid a name clash).
3571 [Guenter <lists@gknw.net>]
3573 *) Add ECDHE and PSK support to DTLS.
3574 [Michael Tuexen <tuexen@fh-muenster.de>]
3576 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3580 *) Add "missing" function EVP_MD_flags() (without this the only way to
3581 retrieve a digest flags is by accessing the structure directly. Update
3582 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3583 or cipher is registered as in the "from" argument. Print out all
3584 registered digests in the dgst usage message instead of manually
3585 attempting to work them out.
3588 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3589 this allows the use of compression and extensions. Change default cipher
3590 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3591 by default unless an application cipher string requests it.
3594 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3595 key ids to find matching certificates and keys but some PKCS#12 files
3596 don't follow the (somewhat unwritten) rules and this strategy fails.
3597 Now just gather all certificates together and the first private key
3598 then look for the first certificate that matches the key.
3601 *) Support use of registered digest and cipher names for dgst and cipher
3602 commands instead of having to add each one as a special case. So now
3609 openssl dgst -sha256 foo
3611 and this works for ENGINE based algorithms too.
3615 *) Update Gost ENGINE to support parameter files.
3616 [Victor B. Wagner <vitus@cryptocom.ru>]
3618 *) Support GeneralizedTime in ca utility.
3619 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3621 *) Enhance the hash format used for certificate directory links. The new
3622 form uses the canonical encoding (meaning equivalent names will work
3623 even if they aren't identical) and uses SHA1 instead of MD5. This form
3624 is incompatible with the older format and as a result c_rehash should
3625 be used to rebuild symbolic links.
3628 *) Make PKCS#8 the default write format for private keys, replacing the
3629 traditional format. This form is standardised, more secure and doesn't
3630 include an implicit MD5 dependency.
3633 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3634 committed to OpenSSL should pass this lot as a minimum.
3637 *) Add session ticket override functionality for use by EAP-FAST.
3638 [Jouni Malinen <j@w1.fi>]
3640 *) Modify HMAC functions to return a value. Since these can be implemented
3641 in an ENGINE errors can occur.
3644 *) Type-checked OBJ_bsearch_ex.
3647 *) Type-checked OBJ_bsearch. Also some constification necessitated
3648 by type-checking. Still to come: TXT_DB, bsearch(?),
3649 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3653 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3654 seconds to a tm structure directly, instead of going through OS
3655 specific date routines. This avoids any issues with OS routines such
3656 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3657 and X509_time_adj_ex() to cover the extended range. The existing
3658 X509_time_adj() is still usable and will no longer have any date issues.
3661 *) Delta CRL support. New use deltas option which will attempt to locate
3662 and search any appropriate delta CRLs available.
3664 This work was sponsored by Google.
3667 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3668 code and add additional score elements. Validate alternate CRL paths
3669 as part of the CRL checking and indicate a new error "CRL path validation
3670 error" in this case. Applications wanting additional details can use
3671 the verify callback and check the new "parent" field. If this is not
3672 NULL CRL path validation is taking place. Existing applications won't
3673 see this because it requires extended CRL support which is off by
3676 This work was sponsored by Google.
3679 *) Support for freshest CRL extension.
3681 This work was sponsored by Google.
3684 *) Initial indirect CRL support. Currently only supported in the CRLs
3685 passed directly and not via lookup. Process certificate issuer
3686 CRL entry extension and lookup CRL entries by bother issuer name
3687 and serial number. Check and process CRL issuer entry in IDP extension.
3689 This work was sponsored by Google.
3692 *) Add support for distinct certificate and CRL paths. The CRL issuer
3693 certificate is validated separately in this case. Only enabled if
3694 an extended CRL support flag is set: this flag will enable additional
3695 CRL functionality in future.
3697 This work was sponsored by Google.
3700 *) Add support for policy mappings extension.
3702 This work was sponsored by Google.
3705 *) Fixes to pathlength constraint, self issued certificate handling,
3706 policy processing to align with RFC3280 and PKITS tests.
3708 This work was sponsored by Google.
3711 *) Support for name constraints certificate extension. DN, email, DNS
3712 and URI types are currently supported.
3714 This work was sponsored by Google.
3717 *) To cater for systems that provide a pointer-based thread ID rather
3718 than numeric, deprecate the current numeric thread ID mechanism and
3719 replace it with a structure and associated callback type. This
3720 mechanism allows a numeric "hash" to be extracted from a thread ID in
3721 either case, and on platforms where pointers are larger than 'long',
3722 mixing is done to help ensure the numeric 'hash' is usable even if it
3723 can't be guaranteed unique. The default mechanism is to use "&errno"
3724 as a pointer-based thread ID to distinguish between threads.
3726 Applications that want to provide their own thread IDs should now use
3727 CRYPTO_THREADID_set_callback() to register a callback that will call
3728 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3730 Note that ERR_remove_state() is now deprecated, because it is tied
3731 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3732 to free the current thread's error state should be replaced by
3733 ERR_remove_thread_state(NULL).
3735 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3736 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3737 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3738 application was previously providing a numeric thread callback that
3739 was inappropriate for distinguishing threads, then uniqueness might
3740 have been obtained with &errno that happened immediately in the
3741 intermediate development versions of OpenSSL; this is no longer the
3742 case, the numeric thread callback will now override the automatic use
3744 [Geoff Thorpe, with help from Bodo Moeller]
3746 *) Initial support for different CRL issuing certificates. This covers a
3747 simple case where the self issued certificates in the chain exist and
3748 the real CRL issuer is higher in the existing chain.
3750 This work was sponsored by Google.
3753 *) Removed effectively defunct crypto/store from the build.
3756 *) Revamp of STACK to provide stronger type-checking. Still to come:
3757 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3758 ASN1_STRING, CONF_VALUE.
3761 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3762 RAM on SSL connections. This option can save about 34k per idle SSL.
3765 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3766 STACK, TXT_DB, bsearch, qsort.
3769 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3770 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3771 support for data, signedData, compressedData, digestedData and
3772 encryptedData, envelopedData types included. Scripts to check against
3773 RFC4134 examples draft and interop and consistency checks of many
3774 content types and variants.
3777 *) Add options to enc utility to support use of zlib compression BIO.
3780 *) Extend mk1mf to support importing of options and assembly language
3781 files from Configure script, currently only included in VC-WIN32.
3782 The assembly language rules can now optionally generate the source
3783 files from the associated perl scripts.
3786 *) Implement remaining functionality needed to support GOST ciphersuites.
3787 Interop testing has been performed using CryptoPro implementations.
3788 [Victor B. Wagner <vitus@cryptocom.ru>]
3790 *) s390x assembler pack.
3793 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3797 *) Implement Opaque PRF Input TLS extension as specified in
3798 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3799 official specification yet and no extension type assignment by
3800 IANA exists, this extension (for now) will have to be explicitly
3801 enabled when building OpenSSL by providing the extension number
3802 to use. For example, specify an option
3804 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3806 to the "config" or "Configure" script to enable the extension,
3807 assuming extension number 0x9527 (which is a completely arbitrary
3808 and unofficial assignment based on the MD5 hash of the Internet
3809 Draft). Note that by doing so, you potentially lose
3810 interoperability with other TLS implementations since these might
3811 be using the same extension number for other purposes.
3813 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3814 opaque PRF input value to use in the handshake. This will create
3815 an internal copy of the length-'len' string at 'src', and will
3816 return non-zero for success.
3818 To get more control and flexibility, provide a callback function
3821 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3822 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3826 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3829 Callback function 'cb' will be called in handshakes, and is
3830 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3831 Argument 'arg' is for application purposes (the value as given to
3832 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3833 be provided to the callback function). The callback function
3834 has to return non-zero to report success: usually 1 to use opaque
3835 PRF input just if possible, or 2 to enforce use of the opaque PRF
3836 input. In the latter case, the library will abort the handshake
3837 if opaque PRF input is not successfully negotiated.
3839 Arguments 'peerinput' and 'len' given to the callback function
3840 will always be NULL and 0 in the case of a client. A server will
3841 see the client's opaque PRF input through these variables if
3842 available (NULL and 0 otherwise). Note that if the server
3843 provides an opaque PRF input, the length must be the same as the
3844 length of the client's opaque PRF input.
3846 Note that the callback function will only be called when creating
3847 a new session (session resumption can resume whatever was
3848 previously negotiated), and will not be called in SSL 2.0
3849 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3850 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3851 for applications that need to enforce opaque PRF input.
3855 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3858 [Victor B. Wagner <vitus@cryptocom.ru>]
3860 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3861 RFC4507bis. The encrypted ticket format is an encrypted encoded
3862 SSL_SESSION structure, that way new session features are automatically
3865 If a client application caches session in an SSL_SESSION structure
3866 support is transparent because tickets are now stored in the encoded
3869 The SSL_CTX structure automatically generates keys for ticket
3870 protection in servers so again support should be possible
3871 with no application modification.
3873 If a client or server wishes to disable RFC4507 support then the option
3874 SSL_OP_NO_TICKET can be set.
3876 Add a TLS extension debugging callback to allow the contents of any client
3877 or server extensions to be examined.
3879 This work was sponsored by Google.
3882 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3883 OpenSSL should now compile cleanly on gcc 4.2
3884 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3886 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3887 support including streaming MAC support: this is required for GOST
3888 ciphersuite support.
3889 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3891 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3892 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3893 to output in BER and PEM format.
3896 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3897 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3898 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3899 ENGINE support for HMAC keys which are unextractable. New -mac and
3900 -macopt options to dgst utility.
3903 *) New option -sigopt to dgst utility. Update dgst to use
3904 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3905 alternative signing parameters such as X9.31 or PSS in the dgst
3909 *) Change ssl_cipher_apply_rule(), the internal function that does
3910 the work each time a ciphersuite string requests enabling
3911 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3912 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3913 the order of disabled ciphersuites such that those ciphersuites
3914 that most recently went from enabled to disabled not only stay
3915 in order with respect to each other, but also have higher priority
3916 than other disabled ciphersuites the next time ciphersuites are
3919 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3920 the same ciphersuites as with "HIGH" alone, but in a specific
3921 order where the PSK ciphersuites come first (since they are the
3922 most recently disabled ciphersuites when "HIGH" is parsed).
3924 Also, change ssl_create_cipher_list() (using this new
3925 functionality) such that between otherwise identical
3926 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
3930 *) Change ssl_create_cipher_list() so that it automatically
3931 arranges the ciphersuites in reasonable order before starting
3932 to process the rule string. Thus, the definition for "DEFAULT"
3933 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3934 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3935 This makes it much easier to arrive at a reasonable default order
3936 in applications for which anonymous ciphers are OK (meaning
3937 that you can't actually use DEFAULT).
3938 [Bodo Moeller; suggested by Victor Duchovni]
3940 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3941 processing) into multiple integers instead of setting
3942 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3943 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3944 (These masks as well as the individual bit definitions are hidden
3945 away into the non-exported interface ssl/ssl_locl.h, so this
3946 change to the definition of the SSL_CIPHER structure shouldn't
3947 affect applications.) This give us more bits for each of these
3948 categories, so there is no longer a need to coagulate AES128 and
3949 AES256 into a single algorithm bit, and to coagulate Camellia128
3950 and Camellia256 into a single algorithm bit, which has led to all
3953 Thus, among other things, the kludge introduced in 0.9.7m and
3954 0.9.8e for masking out AES256 independently of AES128 or masking
3955 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3957 With the change, we also introduce new ciphersuite aliases that
3958 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3962 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3963 Use the leftmost N bytes of the signature input if the input is
3964 larger than the prime q (with N being the size in bytes of q).
3967 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3968 it yet and it is largely untested.
3971 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3974 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3975 some compilers (gcc 4.2 and later) reject their use. Safestack is
3976 reimplemented. Update ASN1 to avoid use of legacy functions.
3979 *) Win32/64 targets are linked with Winsock2.
3982 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3983 to external functions. This can be used to increase CRL handling
3984 efficiency especially when CRLs are very large by (for example) storing
3985 the CRL revoked certificates in a database.
3988 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3989 new CRLs added to a directory can be used. New command line option
3990 -verify_return_error to s_client and s_server. This causes real errors
3991 to be returned by the verify callback instead of carrying on no matter
3992 what. This reflects the way a "real world" verify callback would behave.
3995 *) GOST engine, supporting several GOST algorithms and public key formats.
3996 Kindly donated by Cryptocom.
3999 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4000 partitioned by DP are handled but no indirect CRL or reason partitioning
4001 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4002 selected via a scoring technique which handles IDP and AKID in CRLs.
4005 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4006 will ultimately be used for all verify operations: this will remove the
4007 X509_STORE dependency on certificate verification and allow alternative
4008 lookup methods. X509_STORE based implementations of these two callbacks.
4011 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4012 Modify get_crl() to find a valid (unexpired) CRL if possible.
4015 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4016 this would be called X509_CRL_cmp() but that name is already used by
4017 a function that just compares CRL issuer names. Cache several CRL
4018 extensions in X509_CRL structure and cache CRLDP in X509.
4021 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4022 this maps equivalent X509_NAME structures into a consistent structure.
4023 Name comparison can then be performed rapidly using memcmp().
4026 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4030 *) Allow digests to supply their own micalg string for S/MIME type using
4031 the ctrl EVP_MD_CTRL_MICALG.
4034 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4035 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4036 ctrl. It can then customise the structure before and/or after signing
4040 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4041 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4042 to free up any added signature OIDs.
4045 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4046 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4047 digest and cipher tables. New options added to openssl utility:
4048 list-message-digest-algorithms and list-cipher-algorithms.
4051 *) Change the array representation of binary polynomials: the list
4052 of degrees of non-zero coefficients is now terminated with -1.
4053 Previously it was terminated with 0, which was also part of the
4054 value; thus, the array representation was not applicable to
4055 polynomials where t^0 has coefficient zero. This change makes
4056 the array representation useful in a more general context.
4059 *) Various modifications and fixes to SSL/TLS cipher string
4060 handling. For ECC, the code now distinguishes between fixed ECDH
4061 with RSA certificates on the one hand and with ECDSA certificates
4062 on the other hand, since these are separate ciphersuites. The
4063 unused code for Fortezza ciphersuites has been removed.
4065 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4066 (not "ECDHE"). For consistency with the code for DH
4067 certificates, use of ECDH certificates is now considered ECDH
4068 authentication, not RSA or ECDSA authentication (the latter is
4069 merely the CA's signing algorithm and not actively used in the
4072 The temporary ciphersuite alias "ECCdraft" is no longer
4073 available, and ECC ciphersuites are no longer excluded from "ALL"
4074 and "DEFAULT". The following aliases now exist for RFC 4492
4075 ciphersuites, most of these by analogy with the DH case:
4077 kECDHr - ECDH cert, signed with RSA
4078 kECDHe - ECDH cert, signed with ECDSA
4079 kECDH - ECDH cert (signed with either RSA or ECDSA)
4080 kEECDH - ephemeral ECDH
4081 ECDH - ECDH cert or ephemeral ECDH
4087 AECDH - anonymous ECDH
4088 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4092 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4093 Use correct micalg parameters depending on digest(s) in signed message.
4096 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4097 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4100 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4101 an engine to register a method. Add ENGINE lookups for methods and
4102 functional reference processing.
4105 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4106 EVP_{Sign,Verify}* which allow an application to customise the signature
4110 *) New -resign option to smime utility. This adds one or more signers
4111 to an existing PKCS#7 signedData structure. Also -md option to use an
4112 alternative message digest algorithm for signing.
4115 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4116 create PKCS7 structures containing multiple signers. Update smime
4117 application to support multiple signers.
4120 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4124 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4125 Reorganize PBE internals to lookup from a static table using NIDs,
4126 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4127 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4128 PRF which will be automatically used with PBES2.
4131 *) Replace the algorithm specific calls to generate keys in "req" with the
4135 *) Update PKCS#7 enveloped data routines to use new API. This is now
4136 supported by any public key method supporting the encrypt operation. A
4137 ctrl is added to allow the public key algorithm to examine or modify
4138 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4142 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4143 a default digest type to use. In most cases this will be SHA1 but some
4144 algorithms (such as GOST) need to specify an alternative digest. The
4145 return value indicates how strong the preference is 1 means optional and
4146 2 is mandatory (that is it is the only supported type). Modify
4147 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4148 use the default md. Update openssl utilities to use the default digest
4149 type for signing if it is not explicitly indicated.
4152 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4153 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4154 signing method from the key type. This effectively removes the link
4155 between digests and public key types.
4158 *) Add an OID cross reference table and utility functions. Its purpose is to
4159 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4160 rsaEncryption. This will allow some of the algorithm specific hackery
4161 needed to use the correct OID to be removed.
4164 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4165 structures for PKCS7_sign(). They are now set up by the relevant public
4169 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4172 *) Add support for key derivation (agreement) in the API, DH method and
4176 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4177 public and private key formats. As a side effect these add additional
4178 command line functionality not previously available: DSA signatures can be
4179 generated and verified using pkeyutl and DH key support and generation in
4184 [Oliver Tappe <zooey@hirschkaefer.de>]
4186 *) New make target "install_html_docs" installs HTML renditions of the
4188 [Oliver Tappe <zooey@hirschkaefer.de>]
4190 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4191 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4192 support key and parameter generation and add initial key generation
4193 functionality for RSA.
4196 *) Add functions for main EVP_PKEY_method operations. The undocumented
4197 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4198 EVP_PKEY_{encrypt,decrypt}_old.
4201 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4202 key API, doesn't do much yet.
4205 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4206 public key algorithms. New option to openssl utility:
4207 "list-public-key-algorithms" to print out info.
4210 *) Implement the Supported Elliptic Curves Extension for
4211 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4214 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4215 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4218 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4219 utilities such as rsa, dsa, dsaparam etc except they process any key
4223 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4224 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4225 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4229 *) Initial support for pluggable public key ASN1.
4230 De-spaghettify the public key ASN1 handling. Move public and private
4231 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4232 algorithm specific handling to a single module within the relevant
4233 algorithm directory. Add functions to allow (near) opaque processing
4234 of public and private key structures.
4237 *) Implement the Supported Point Formats Extension for
4238 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4241 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4242 for the psk identity [hint] and the psk callback functions to the
4243 SSL_SESSION, SSL and SSL_CTX structure.
4246 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4250 SSL_CTX_use_psk_identity_hint
4251 SSL_get_psk_identity_hint
4252 SSL_get_psk_identity
4253 SSL_use_psk_identity_hint
4255 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4257 *) Add RFC 3161 compliant time stamp request creation, response generation
4258 and response verification functionality.
4259 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4261 *) Add initial support for TLS extensions, specifically for the server_name
4262 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4263 have new members for a host name. The SSL data structure has an
4264 additional member SSL_CTX *initial_ctx so that new sessions can be
4265 stored in that context to allow for session resumption, even after the
4266 SSL has been switched to a new SSL_CTX in reaction to a client's
4267 server_name extension.
4269 New functions (subject to change):
4271 SSL_get_servername()
4272 SSL_get_servername_type()
4275 New CTRL codes and macros (subject to change):
4277 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4278 - SSL_CTX_set_tlsext_servername_callback()
4279 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4280 - SSL_CTX_set_tlsext_servername_arg()
4281 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4283 openssl s_client has a new '-servername ...' option.
4285 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4286 '-key2 ...', '-servername_fatal' (subject to change). This allows
4287 testing the HostName extension for a specific single host name ('-cert'
4288 and '-key' remain fallbacks for handshakes without HostName
4289 negotiation). If the unrecognized_name alert has to be sent, this by
4290 default is a warning; it becomes fatal with the '-servername_fatal'
4293 [Peter Sylvester, Remy Allais, Christophe Renou]
4295 *) Whirlpool hash implementation is added.
4298 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4299 bn(64,32). Because of instruction set limitations it doesn't have
4300 any negative impact on performance. This was done mostly in order
4301 to make it possible to share assembler modules, such as bn_mul_mont
4302 implementations, between 32- and 64-bit builds without hassle.
4305 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4306 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4310 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4311 dedicated Montgomery multiplication procedure, is introduced.
4312 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4313 "64-bit" performance on certain 32-bit targets.
4316 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4317 in SSL structures. New SSL ctrl to set maximum send fragment size.
4318 Save memory by setting the I/O buffer sizes dynamically instead of
4319 using the maximum available value.
4322 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4323 in addition to the text details.
4326 *) Very, very preliminary EXPERIMENTAL support for printing of general
4327 ASN1 structures. This currently produces rather ugly output and doesn't
4328 handle several customised structures at all.
4331 *) Integrated support for PVK file format and some related formats such
4332 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4333 these in the 'rsa' and 'dsa' utilities.
4336 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4339 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4340 place for the (very old) "NETSCAPE" format certificates which are now
4341 handled using new ASN1 code equivalents.
4344 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4345 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4346 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4349 *) Modify CRL distribution points extension code to print out previously
4350 unsupported fields. Enhance extension setting code to allow setting of
4354 *) Add print and set support for Issuing Distribution Point CRL extension.
4357 *) Change 'Configure' script to enable Camellia by default.
4360 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4362 *) When rejecting SSL/TLS records due to an incorrect version number, never
4363 update s->server with a new major version number. As of
4364 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4365 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4366 the previous behavior could result in a read attempt at NULL when
4367 receiving specific incorrect SSL/TLS records once record payload
4368 protection is active. (CVE-2010-0740)
4369 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4371 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4372 could be crashed if the relevant tables were not present (e.g. chrooted).
4373 [Tomas Hoger <thoger@redhat.com>]
4375 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4377 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4378 [Martin Olsson, Neel Mehta]
4380 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4381 accommodate for stack sorting, always a write lock!).
4384 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4385 excessive delays in the RAND_poll(): over a minute. As a workaround
4386 include a time check in the inner Heap32Next loop too.
4389 *) The code that handled flushing of data in SSL/TLS originally used the
4390 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4391 the problem outlined in PR#1949. The fix suggested there however can
4392 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4393 of Apache). So instead simplify the code to flush unconditionally.
4394 This should be fine since flushing with no data to flush is a no op.
4397 *) Handle TLS versions 2.0 and later properly and correctly use the
4398 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4399 off ancient servers have a habit of sticking around for a while...
4402 *) Modify compression code so it frees up structures without using the
4403 ex_data callbacks. This works around a problem where some applications
4404 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4405 restarting) then use compression (e.g. SSL with compression) later.
4406 This results in significant per-connection memory leaks and
4407 has caused some security issues including CVE-2008-1678 and
4411 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4412 change when encrypting or decrypting.
4415 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4416 connect and renegotiate with servers which do not support RI.
4417 Until RI is more widely deployed this option is enabled by default.
4420 *) Add "missing" ssl ctrls to clear options and mode.
4423 *) If client attempts to renegotiate and doesn't support RI respond with
4424 a no_renegotiation alert as required by RFC5746. Some renegotiating
4425 TLS clients will continue a connection gracefully when they receive
4426 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4427 waiting for a server hello which it will never receive. Now we treat a
4428 received no_renegotiation alert as a fatal error. This is because
4429 applications requesting a renegotiation might well expect it to succeed
4430 and would have no code in place to handle the server denying it so the
4431 only safe thing to do is to terminate the connection.
4434 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4435 peer supports secure renegotiation and 0 otherwise. Print out peer
4436 renegotiation support in s_client/s_server.
4439 *) Replace the highly broken and deprecated SPKAC certification method with
4440 the updated NID creation version. This should correctly handle UTF8.
4443 *) Implement RFC5746. Re-enable renegotiation but require the extension
4444 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4445 turns out to be a bad idea. It has been replaced by
4446 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4447 SSL_CTX_set_options(). This is really not recommended unless you
4448 know what you are doing.
4449 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4451 *) Fixes to stateless session resumption handling. Use initial_ctx when
4452 issuing and attempting to decrypt tickets in case it has changed during
4453 servername handling. Use a non-zero length session ID when attempting
4454 stateless session resumption: this makes it possible to determine if
4455 a resumption has occurred immediately after receiving server hello
4456 (several places in OpenSSL subtly assume this) instead of later in
4460 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4461 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4462 fixes for a few places where the return code is not checked
4464 [Julia Lawall <julia@diku.dk>]
4466 *) Add --strict-warnings option to Configure script to include devteam
4467 warnings in other configurations.
4470 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4471 makes it possible to install openssl libraries in locations which
4472 have names other than "lib", for example "/usr/lib64" which some
4474 [Steve Henson, based on patch from Jeremy Utley]
4476 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4477 X690 8.9.12 and can produce some misleading textual output of OIDs.
4478 [Steve Henson, reported by Dan Kaminsky]
4480 *) Delete MD2 from algorithm tables. This follows the recommendation in
4481 several standards that it is not used in new applications due to
4482 several cryptographic weaknesses. For binary compatibility reasons
4483 the MD2 API is still compiled in by default.
4486 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4490 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4491 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4493 [Guenter <lists@gknw.net>]
4495 *) Fix the server certificate chain building code to use X509_verify_cert(),
4496 it used to have an ad-hoc builder which was unable to cope with anything
4497 other than a simple chain.
4498 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4500 *) Don't check self signed certificate signatures in X509_verify_cert()
4501 by default (a flag can override this): it just wastes time without
4502 adding any security. As a useful side effect self signed root CAs
4503 with non-FIPS digests are now usable in FIPS mode.
4506 *) In dtls1_process_out_of_seq_message() the check if the current message
4507 is already buffered was missing. For every new message was memory
4508 allocated, allowing an attacker to perform an denial of service attack
4509 with sending out of seq handshake messages until there is no memory
4510 left. Additionally every future message was buffered, even if the
4511 sequence number made no sense and would be part of another handshake.
4512 So only messages with sequence numbers less than 10 in advance will be
4513 buffered. (CVE-2009-1378)
4514 [Robin Seggelmann, discovered by Daniel Mentz]
4516 *) Records are buffered if they arrive with a future epoch to be
4517 processed after finishing the corresponding handshake. There is
4518 currently no limitation to this buffer allowing an attacker to perform
4519 a DOS attack with sending records with future epochs until there is no
4520 memory left. This patch adds the pqueue_size() function to determine
4521 the size of a buffer and limits the record buffer to 100 entries.
4523 [Robin Seggelmann, discovered by Daniel Mentz]
4525 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4526 parent structure is freed. (CVE-2009-1379)
4529 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4530 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4533 [Ilya O. <vrghost@gmail.com>]
4535 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4537 *) Disable renegotiation completely - this fixes a severe security
4538 problem (CVE-2009-3555) at the cost of breaking all
4539 renegotiation. Renegotiation can be re-enabled by setting
4540 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4541 run-time. This is really not recommended unless you know what
4545 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4547 *) Don't set val to NULL when freeing up structures, it is freed up by
4548 underlying code. If sizeof(void *) > sizeof(long) this can result in
4549 zeroing past the valid field. (CVE-2009-0789)
4550 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4552 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4553 checked correctly. This would allow some invalid signed attributes to
4554 appear to verify correctly. (CVE-2009-0591)
4555 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4557 *) Reject UniversalString and BMPString types with invalid lengths. This
4558 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4559 a legal length. (CVE-2009-0590)
4562 *) Set S/MIME signing as the default purpose rather than setting it
4563 unconditionally. This allows applications to override it at the store
4567 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4568 to handle some structures.
4571 *) Improve efficiency of mem_gets: don't search whole buffer each time
4573 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4575 *) New -hex option for openssl rand.
4578 *) Print out UTF8String and NumericString when parsing ASN1.
4581 *) Support NumericString type for name components.
4584 *) Allow CC in the environment to override the automatically chosen
4585 compiler. Note that nothing is done to ensure flags work with the
4589 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4591 *) Properly check EVP_VerifyFinal() and similar return values
4593 [Ben Laurie, Bodo Moeller, Google Security Team]
4595 *) Enable TLS extensions by default.
4598 *) Allow the CHIL engine to be loaded, whether the application is
4599 multithreaded or not. (This does not release the developer from the
4600 obligation to set up the dynamic locking callbacks.)
4601 [Sander Temme <sander@temme.net>]
4603 *) Use correct exit code if there is an error in dgst command.
4604 [Steve Henson; problem pointed out by Roland Dirlewanger]
4606 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4607 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4610 *) Add experimental JPAKE support, including demo authentication in
4611 s_client and s_server.
4614 *) Set the comparison function in v3_addr_canonize().
4615 [Rob Austein <sra@hactrn.net>]
4617 *) Add support for XMPP STARTTLS in s_client.
4618 [Philip Paeps <philip@freebsd.org>]
4620 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4621 to ensure that even with this option, only ciphersuites in the
4622 server's preference list will be accepted. (Note that the option
4623 applies only when resuming a session, so the earlier behavior was
4624 just about the algorithm choice for symmetric cryptography.)
4627 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4629 *) Fix NULL pointer dereference if a DTLS server received
4630 ChangeCipherSpec as first record (CVE-2009-1386).
4633 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4634 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4637 *) The fix in 0.9.8c that supposedly got rid of unsafe
4638 double-checked locking was incomplete for RSA blinding,
4639 addressing just one layer of what turns out to have been
4640 doubly unsafe triple-checked locking.
4642 So now fix this for real by retiring the MONT_HELPER macro
4643 in crypto/rsa/rsa_eay.c.
4645 [Bodo Moeller; problem pointed out by Marius Schilder]
4647 *) Various precautionary measures:
4649 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4651 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4652 (NB: This would require knowledge of the secret session ticket key
4653 to exploit, in which case you'd be SOL either way.)
4655 - Change bn_nist.c so that it will properly handle input BIGNUMs
4656 outside the expected range.
4658 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4661 [Neel Mehta, Bodo Moeller]
4663 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4664 the load fails. Useful for distros.
4665 [Ben Laurie and the FreeBSD team]
4667 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4670 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4673 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4675 This work was sponsored by Logica.
4678 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4679 keystores. Support for SSL/TLS client authentication too.
4680 Not compiled unless enable-capieng specified to Configure.
4682 This work was sponsored by Logica.
4685 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4686 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4687 attribute creation routines such as certificate requests and PKCS#12
4691 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4693 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4694 handshake which could lead to a client crash as found using the
4695 Codenomicon TLS test suite (CVE-2008-1672)
4696 [Steve Henson, Mark Cox]
4698 *) Fix double free in TLS server name extensions which could lead to
4699 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4702 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4704 Clear the error queue to ensure that error entries left from
4705 older function calls do not interfere with the correct operation.
4706 [Lutz Jaenicke, Erik de Castro Lopo]
4708 *) Remove root CA certificates of commercial CAs:
4710 The OpenSSL project does not recommend any specific CA and does not
4711 have any policy with respect to including or excluding any CA.
4712 Therefore it does not make any sense to ship an arbitrary selection
4713 of root CA certificates with the OpenSSL software.
4716 *) RSA OAEP patches to fix two separate invalid memory reads.
4717 The first one involves inputs when 'lzero' is greater than
4718 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4719 before the beginning of from). The second one involves inputs where
4720 the 'db' section contains nothing but zeroes (there is a one-byte
4721 invalid read after the end of 'db').
4722 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4724 *) Partial backport from 0.9.9-dev:
4726 Introduce bn_mul_mont (dedicated Montgomery multiplication
4727 procedure) as a candidate for BIGNUM assembler implementation.
4728 While 0.9.9-dev uses assembler for various architectures, only
4729 x86_64 is available by default here in the 0.9.8 branch, and
4730 32-bit x86 is available through a compile-time setting.
4732 To try the 32-bit x86 assembler implementation, use Configure
4733 option "enable-montasm" (which exists only for this backport).
4735 As "enable-montasm" for 32-bit x86 disclaims code stability
4736 anyway, in this constellation we activate additional code
4737 backported from 0.9.9-dev for further performance improvements,
4738 namely BN_from_montgomery_word. (To enable this otherwise,
4739 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4741 [Andy Polyakov (backport partially by Bodo Moeller)]
4743 *) Add TLS session ticket callback. This allows an application to set
4744 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4745 values. This is useful for key rollover for example where several key
4746 sets may exist with different names.
4749 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4750 This was broken until now in 0.9.8 releases, such that the only way
4751 a registered ENGINE could be used (assuming it initialises
4752 successfully on the host) was to explicitly set it as the default
4753 for the relevant algorithms. This is in contradiction with 0.9.7
4754 behaviour and the documentation. With this fix, when an ENGINE is
4755 registered into a given algorithm's table of implementations, the
4756 'uptodate' flag is reset so that auto-discovery will be used next
4757 time a new context for that algorithm attempts to select an
4759 [Ian Lister (tweaked by Geoff Thorpe)]
4761 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4762 implementation in the following ways:
4764 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4767 Lack of BER streaming support means one pass streaming processing is
4768 only supported if data is detached: setting the streaming flag is
4769 ignored for embedded content.
4771 CMS support is disabled by default and must be explicitly enabled
4772 with the enable-cms configuration option.
4775 *) Update the GMP engine glue to do direct copies between BIGNUM and
4776 mpz_t when openssl and GMP use the same limb size. Otherwise the
4777 existing "conversion via a text string export" trick is still used.
4778 [Paul Sheer <paulsheer@gmail.com>]
4780 *) Zlib compression BIO. This is a filter BIO which compressed and
4781 uncompresses any data passed through it.
4784 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4785 RFC3394 compatible AES key wrapping.
4788 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4789 sets string data without copying. X509_ALGOR_set0() and
4790 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4791 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4792 from an X509_ATTRIBUTE structure optionally checking it occurs only
4793 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4797 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4798 to get the expected BN_FLG_CONSTTIME behavior.
4799 [Bodo Moeller (Google)]
4803 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4804 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4805 - added some more tests to do_tests.pl
4806 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4807 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4808 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4809 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4810 - various changes to netware.pl to enable gcc-cross builds on Win32
4812 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4813 - various changes to fix missing prototype warnings
4814 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4815 - added AES, WHIRLPOOL and CPUID assembler code to build files
4816 - added missing AES assembler make rules to mk1mf.pl
4817 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4818 [Guenter Knauf <eflash@gmx.net>]
4820 *) Implement certificate status request TLS extension defined in RFC3546.
4821 A client can set the appropriate parameters and receive the encoded
4822 OCSP response via a callback. A server can query the supplied parameters
4823 and set the encoded OCSP response in the callback. Add simplified examples
4824 to s_client and s_server.
4827 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4829 *) Fix various bugs:
4830 + Binary incompatibility of ssl_ctx_st structure
4831 + DTLS interoperation with non-compliant servers
4832 + Don't call get_session_cb() without proposed session
4833 + Fix ia64 assembler code
4834 [Andy Polyakov, Steve Henson]
4836 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4838 *) DTLS Handshake overhaul. There were longstanding issues with
4839 OpenSSL DTLS implementation, which were making it impossible for
4840 RFC 4347 compliant client to communicate with OpenSSL server.
4841 Unfortunately just fixing these incompatibilities would "cut off"
4842 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4843 server keeps tolerating non RFC compliant syntax. The opposite is
4844 not true, 0.9.8f client can not communicate with earlier server.
4845 This update even addresses CVE-2007-4995.
4848 *) Changes to avoid need for function casts in OpenSSL: some compilers
4849 (gcc 4.2 and later) reject their use.
4850 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4853 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4854 RFC4507bis. The encrypted ticket format is an encrypted encoded
4855 SSL_SESSION structure, that way new session features are automatically
4858 If a client application caches session in an SSL_SESSION structure
4859 support is transparent because tickets are now stored in the encoded
4862 The SSL_CTX structure automatically generates keys for ticket
4863 protection in servers so again support should be possible
4864 with no application modification.
4866 If a client or server wishes to disable RFC4507 support then the option
4867 SSL_OP_NO_TICKET can be set.
4869 Add a TLS extension debugging callback to allow the contents of any client
4870 or server extensions to be examined.
4872 This work was sponsored by Google.
4875 *) Add initial support for TLS extensions, specifically for the server_name
4876 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4877 have new members for a host name. The SSL data structure has an
4878 additional member SSL_CTX *initial_ctx so that new sessions can be
4879 stored in that context to allow for session resumption, even after the
4880 SSL has been switched to a new SSL_CTX in reaction to a client's
4881 server_name extension.
4883 New functions (subject to change):
4885 SSL_get_servername()
4886 SSL_get_servername_type()
4889 New CTRL codes and macros (subject to change):
4891 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4892 - SSL_CTX_set_tlsext_servername_callback()
4893 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4894 - SSL_CTX_set_tlsext_servername_arg()
4895 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4897 openssl s_client has a new '-servername ...' option.
4899 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4900 '-key2 ...', '-servername_fatal' (subject to change). This allows
4901 testing the HostName extension for a specific single host name ('-cert'
4902 and '-key' remain fallbacks for handshakes without HostName
4903 negotiation). If the unrecognized_name alert has to be sent, this by
4904 default is a warning; it becomes fatal with the '-servername_fatal'
4907 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4909 *) Add AES and SSE2 assembly language support to VC++ build.
4912 *) Mitigate attack on final subtraction in Montgomery reduction.
4915 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4916 (which previously caused an internal error).
4919 *) Squeeze another 10% out of IGE mode when in != out.
4922 *) AES IGE mode speedup.
4923 [Dean Gaudet (Google)]
4925 *) Add the Korean symmetric 128-bit cipher SEED (see
4926 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4927 add SEED ciphersuites from RFC 4162:
4929 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4930 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4931 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4932 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4934 To minimize changes between patchlevels in the OpenSSL 0.9.8
4935 series, SEED remains excluded from compilation unless OpenSSL
4936 is configured with 'enable-seed'.
4937 [KISA, Bodo Moeller]
4939 *) Mitigate branch prediction attacks, which can be practical if a
4940 single processor is shared, allowing a spy process to extract
4941 information. For detailed background information, see
4942 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4943 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4944 and Necessary Software Countermeasures"). The core of the change
4945 are new versions BN_div_no_branch() and
4946 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4947 respectively, which are slower, but avoid the security-relevant
4948 conditional branches. These are automatically called by BN_div()
4949 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4950 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4951 remove a conditional branch.
4953 BN_FLG_CONSTTIME is the new name for the previous
4954 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4955 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4956 in the exponent causes BN_mod_exp_mont() to use the alternative
4957 implementation in BN_mod_exp_mont_consttime().) The old name
4958 remains as a deprecated alias.
4960 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4961 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4962 constant-time implementations for more than just exponentiation.
4963 Here too the old name is kept as a deprecated alias.
4965 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4966 the BN_BLINDING structure gets an independent copy of the
4967 modulus. This means that the previous "BIGNUM *m" argument to
4968 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4969 essentially becomes "const BIGNUM *m", although we can't actually
4970 change this in the header file before 0.9.9. It allows
4971 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4972 enable BN_FLG_CONSTTIME.
4974 [Matthew D Wood (Intel Corp)]
4976 *) In the SSL/TLS server implementation, be strict about session ID
4977 context matching (which matters if an application uses a single
4978 external cache for different purposes). Previously,
4979 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4980 set. This did ensure strict client verification, but meant that,
4981 with applications using a single external cache for quite
4982 different requirements, clients could circumvent ciphersuite
4983 restrictions for a given session ID context by starting a session
4984 in a different context.
4987 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4988 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4989 authentication-only ciphersuites.
4992 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4993 not complete and could lead to a possible single byte overflow
4994 (CVE-2007-5135) [Ben Laurie]
4996 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4998 *) Since AES128 and AES256 (and similarly Camellia128 and
4999 Camellia256) share a single mask bit in the logic of
5000 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5001 kludge to work properly if AES128 is available and AES256 isn't
5002 (or if Camellia128 is available and Camellia256 isn't).
5005 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5006 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5007 When a point or a seed is encoded in a BIT STRING, we need to
5008 prevent the removal of trailing zero bits to get the proper DER
5009 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5010 of a NamedBitList, for which trailing 0 bits need to be removed.)
5013 *) Have SSL/TLS server implementation tolerate "mismatched" record
5014 protocol version while receiving ClientHello even if the
5015 ClientHello is fragmented. (The server can't insist on the
5016 particular protocol version it has chosen before the ServerHello
5017 message has informed the client about his choice.)
5020 *) Add RFC 3779 support.
5021 [Rob Austein for ARIN, Ben Laurie]
5023 *) Load error codes if they are not already present instead of using a
5024 static variable. This allows them to be cleanly unloaded and reloaded.
5025 Improve header file function name parsing.
5028 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5029 or CAPABILITY handshake as required by RFCs.
5032 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5034 *) Introduce limits to prevent malicious keys being able to
5035 cause a denial of service. (CVE-2006-2940)
5036 [Steve Henson, Bodo Moeller]
5038 *) Fix ASN.1 parsing of certain invalid structures that can result
5039 in a denial of service. (CVE-2006-2937) [Steve Henson]
5041 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5042 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5044 *) Fix SSL client code which could crash if connecting to a
5045 malicious SSLv2 server. (CVE-2006-4343)
5046 [Tavis Ormandy and Will Drewry, Google Security Team]
5048 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5049 match only those. Before that, "AES256-SHA" would be interpreted
5050 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5051 the same strength classification in 0.9.7h) as we currently only
5052 have a single AES bit in the ciphersuite description bitmap.
5053 That change, however, also applied to ciphersuite strings such as
5054 "RC4-MD5" that intentionally matched multiple ciphersuites --
5055 namely, SSL 2.0 ciphersuites in addition to the more common ones
5056 from SSL 3.0/TLS 1.0.
5058 So we change the selection algorithm again: Naming an explicit
5059 ciphersuite selects this one ciphersuite, and any other similar
5060 ciphersuite (same bitmap) from *other* protocol versions.
5061 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5062 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5064 Since SSL 2.0 does not have any ciphersuites for which the
5065 128/256 bit distinction would be relevant, this works for now.
5066 The proper fix will be to use different bits for AES128 and
5067 AES256, which would have avoided the problems from the beginning;
5068 however, bits are scarce, so we can only do this in a new release
5069 (not just a patchlevel) when we can change the SSL_CIPHER
5070 definition to split the single 'unsigned long mask' bitmap into
5071 multiple values to extend the available space.
5075 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5077 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5078 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5080 *) Add AES IGE and biIGE modes.
5083 *) Change the Unix randomness entropy gathering to use poll() when
5084 possible instead of select(), since the latter has some
5085 undesirable limitations.
5086 [Darryl Miles via Richard Levitte and Bodo Moeller]
5088 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5089 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5090 cannot be implicitly activated as part of, e.g., the "AES" alias.
5091 However, please upgrade to OpenSSL 0.9.9[-dev] for
5092 non-experimental use of the ECC ciphersuites to get TLS extension
5093 support, which is required for curve and point format negotiation
5094 to avoid potential handshake problems.
5097 *) Disable rogue ciphersuites:
5099 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5100 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5101 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5103 The latter two were purportedly from
5104 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5107 Also deactivate the remaining ciphersuites from
5108 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5109 unofficial, and the ID has long expired.
5112 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5113 dual-core machines) and other potential thread-safety issues.
5116 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5117 versions), which is now available for royalty-free use
5118 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5119 Also, add Camellia TLS ciphersuites from RFC 4132.
5121 To minimize changes between patchlevels in the OpenSSL 0.9.8
5122 series, Camellia remains excluded from compilation unless OpenSSL
5123 is configured with 'enable-camellia'.
5126 *) Disable the padding bug check when compression is in use. The padding
5127 bug check assumes the first packet is of even length, this is not
5128 necessarily true if compression is enabled and can result in false
5129 positives causing handshake failure. The actual bug test is ancient
5130 code so it is hoped that implementations will either have fixed it by
5131 now or any which still have the bug do not support compression.
5134 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5136 *) When applying a cipher rule check to see if string match is an explicit
5137 cipher suite and only match that one cipher suite if it is.
5140 *) Link in manifests for VC++ if needed.
5141 [Austin Ziegler <halostatue@gmail.com>]
5143 *) Update support for ECC-based TLS ciphersuites according to
5144 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5145 TLS extensions, which are supported starting with the 0.9.9
5146 branch, not in the OpenSSL 0.9.8 branch).
5149 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5150 opaque EVP_CIPHER_CTX handling.
5153 *) Fixes and enhancements to zlib compression code. We now only use
5154 "zlib1.dll" and use the default __cdecl calling convention on Win32
5155 to conform with the standards mentioned here:
5156 http://www.zlib.net/DLL_FAQ.txt
5157 Static zlib linking now works on Windows and the new --with-zlib-include
5158 --with-zlib-lib options to Configure can be used to supply the location
5159 of the headers and library. Gracefully handle case where zlib library
5163 *) Several fixes and enhancements to the OID generation code. The old code
5164 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5165 handle numbers larger than ULONG_MAX, truncated printing and had a
5166 non standard OBJ_obj2txt() behaviour.
5169 *) Add support for building of engines under engine/ as shared libraries
5170 under VC++ build system.
5173 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5174 Hopefully, we will not see any false combination of paths any more.
5177 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5179 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5180 (part of SSL_OP_ALL). This option used to disable the
5181 countermeasure against man-in-the-middle protocol-version
5182 rollback in the SSL 2.0 server implementation, which is a bad
5183 idea. (CVE-2005-2969)
5185 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5186 for Information Security, National Institute of Advanced Industrial
5187 Science and Technology [AIST], Japan)]
5189 *) Add two function to clear and return the verify parameter flags.
5192 *) Keep cipherlists sorted in the source instead of sorting them at
5193 runtime, thus removing the need for a lock.
5196 *) Avoid some small subgroup attacks in Diffie-Hellman.
5197 [Nick Mathewson and Ben Laurie]
5199 *) Add functions for well-known primes.
5202 *) Extended Windows CE support.
5203 [Satoshi Nakamura and Andy Polyakov]
5205 *) Initialize SSL_METHOD structures at compile time instead of during
5206 runtime, thus removing the need for a lock.
5209 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5210 attempting to decrypt each encrypted key in turn. Add support to
5214 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5216 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5219 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5222 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5223 key into the same file any more.
5226 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5229 *) Add -utf8 command line and config file option to 'ca'.
5230 [Stefan <stf@udoma.org]
5232 *) Removed the macro des_crypt(), as it seems to conflict with some
5233 libraries. Use DES_crypt().
5236 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5237 involves renaming the source and generated shared-libs for
5238 both. The engines will accept the corrected or legacy ids
5239 ('ncipher' and '4758_cca' respectively) when binding. NB,
5240 this only applies when building 'shared'.
5241 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5243 *) Add attribute functions to EVP_PKEY structure. Modify
5244 PKCS12_create() to recognize a CSP name attribute and
5245 use it. Make -CSP option work again in pkcs12 utility.
5248 *) Add new functionality to the bn blinding code:
5249 - automatic re-creation of the BN_BLINDING parameters after
5250 a fixed number of uses (currently 32)
5251 - add new function for parameter creation
5252 - introduce flags to control the update behaviour of the
5253 BN_BLINDING parameters
5254 - hide BN_BLINDING structure
5255 Add a second BN_BLINDING slot to the RSA structure to improve
5256 performance when a single RSA object is shared among several
5260 *) Add support for DTLS.
5261 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5263 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5264 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5267 *) Remove buggy and incomplete DH cert support from
5268 ssl/ssl_rsa.c and ssl/s3_both.c
5271 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5272 the apps/openssl applications.
5275 *) Compile clean with "-Wall -Wmissing-prototypes
5276 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5277 DEBUG_SAFESTACK must also be set.
5280 *) Change ./Configure so that certain algorithms can be disabled by default.
5281 The new counterpiece to "no-xxx" is "enable-xxx".
5283 The patented RC5 and MDC2 algorithms will now be disabled unless
5284 "enable-rc5" and "enable-mdc2", respectively, are specified.
5286 (IDEA remains enabled despite being patented. This is because IDEA
5287 is frequently required for interoperability, and there is no license
5288 fee for non-commercial use. As before, "no-idea" can be used to
5289 avoid this algorithm.)
5293 *) Add processing of proxy certificates (see RFC 3820). This work was
5294 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5295 EGEE (Enabling Grids for E-science in Europe).
5298 *) RC4 performance overhaul on modern architectures/implementations, such
5299 as Intel P4, IA-64 and AMD64.
5302 *) New utility extract-section.pl. This can be used specify an alternative
5303 section number in a pod file instead of having to treat each file as
5304 a separate case in Makefile. This can be done by adding two lines to the
5307 =for comment openssl_section:XXX
5309 The blank line is mandatory.
5313 *) New arguments -certform, -keyform and -pass for s_client and s_server
5314 to allow alternative format key and certificate files and passphrase
5318 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5319 update associated structures and add various utility functions.
5321 Add new policy related verify parameters, include policy checking in
5322 standard verify code. Enhance 'smime' application with extra parameters
5323 to support policy checking and print out.
5326 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5327 Nehemiah processors. These extensions support AES encryption in hardware
5328 as well as RNG (though RNG support is currently disabled).
5329 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5331 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5334 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5335 [Andy Polyakov and a number of other people]
5337 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5338 implementation contributed by IBM.
5339 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5341 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5342 exponent rather than 'unsigned long'. There is a corresponding change to
5343 the new 'rsa_keygen' element of the RSA_METHOD structure.
5344 [Jelte Jansen, Geoff Thorpe]
5346 *) Functionality for creating the initial serial number file is now
5347 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5349 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5350 number file to 1, which is bound to cause problems. To avoid
5351 the problems while respecting compatibility between different 0.9.7
5352 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5353 CA.pl for serial number initialization. With the new release 0.9.8,
5354 we can fix the problem directly in the 'ca' utility.)
5357 *) Reduced header interdependencies by declaring more opaque objects in
5358 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5359 give fewer recursive includes, which could break lazy source code - so
5360 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5361 developers should define this symbol when building and using openssl to
5362 ensure they track the recommended behaviour, interfaces, [etc], but
5363 backwards-compatible behaviour prevails when this isn't defined.
5366 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5369 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5370 This will generate a random key of the appropriate length based on the
5371 cipher context. The EVP_CIPHER can provide its own random key generation
5372 routine to support keys of a specific form. This is used in the des and
5373 3des routines to generate a key of the correct parity. Update S/MIME
5374 code to use new functions and hence generate correct parity DES keys.
5375 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5376 valid (weak or incorrect parity).
5379 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5380 as looking them up. This is useful when the verified structure may contain
5381 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5382 present unless the new PKCS7_NO_CRL flag is asserted.
5385 *) Extend ASN1 oid configuration module. It now additionally accepts the
5388 shortName = some long name, 1.2.3.4
5391 *) Reimplemented the BN_CTX implementation. There is now no more static
5392 limitation on the number of variables it can handle nor the depth of the
5393 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5394 information can now expand as required, and rather than having a single
5395 static array of bignums, BN_CTX now uses a linked-list of such arrays
5396 allowing it to expand on demand whilst maintaining the usefulness of
5397 BN_CTX's "bundling".
5400 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5401 to allow all RSA operations to function using a single BN_CTX.
5404 *) Preliminary support for certificate policy evaluation and checking. This
5405 is initially intended to pass the tests outlined in "Conformance Testing
5406 of Relying Party Client Certificate Path Processing Logic" v1.07.
5409 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5410 remained unused and not that useful. A variety of other little bignum
5411 tweaks and fixes have also been made continuing on from the audit (see
5415 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5416 associated ASN1, EVP and SSL functions and old ASN1 macros.
5419 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5420 and this should never fail. So the return value from the use of
5421 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5422 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5425 *) BN_CTX_get() should return zero-valued bignums, providing the same
5426 initialised value as BN_new().
5427 [Geoff Thorpe, suggested by Ulf Möller]
5429 *) Support for inhibitAnyPolicy certificate extension.
5432 *) An audit of the BIGNUM code is underway, for which debugging code is
5433 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5434 is considered valid when processing BIGNUMs, and causes execution to
5435 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5436 further steps are taken to deliberately pollute unused data in BIGNUM
5437 structures to try and expose faulty code further on. For now, openssl will
5438 (in its default mode of operation) continue to tolerate the inconsistent
5439 forms that it has tolerated in the past, but authors and packagers should
5440 consider trying openssl and their own applications when compiled with
5441 these debugging symbols defined. It will help highlight potential bugs in
5442 their own code, and will improve the test coverage for OpenSSL itself. At
5443 some point, these tighter rules will become openssl's default to improve
5444 maintainability, though the assert()s and other overheads will remain only
5445 in debugging configurations. See bn.h for more details.
5446 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5448 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5449 that can only be obtained through BN_CTX_new() (which implicitly
5450 initialises it). The presence of this function only made it possible
5451 to overwrite an existing structure (and cause memory leaks).
5454 *) Because of the callback-based approach for implementing LHASH as a
5455 template type, lh_insert() adds opaque objects to hash-tables and
5456 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5457 to clean up those corresponding objects before destroying the hash table
5458 (and losing the object pointers). So some over-zealous constifications in
5459 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5460 objects as "const" and the lh_doall[_arg] callback wrappers are not
5461 prototyped to have "const" restrictions on the object pointers they are
5462 given (and so aren't required to cast them away any more).
5465 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5466 (speed) prefers to use its own implementation. The two implementations
5467 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5468 its object type properly exposed (MS_TM) instead of casting to/from "char
5469 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5470 aren't necessarily the greatest nomenclatures - but this is what was used
5471 internally to the implementation so I've used that for now.
5474 *) Ensure that deprecated functions do not get compiled when
5475 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5476 the self-tests were still using deprecated key-generation functions so
5477 these have been updated also.
5480 *) Reorganise PKCS#7 code to separate the digest location functionality
5481 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5482 New function PKCS7_set_digest() to set the digest type for PKCS#7
5483 digestedData type. Add additional code to correctly generate the
5484 digestedData type and add support for this type in PKCS7 initialization
5488 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5489 structure of type "other".
5492 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5493 sure the loop does correctly stop and breaking ("division by zero")
5494 modulus operations are not performed. The (pre-generated) prime
5495 table crypto/bn/bn_prime.h was already correct, but it could not be
5496 re-generated on some platforms because of the "division by zero"
5497 situation in the script.
5498 [Ralf S. Engelschall]
5500 *) Update support for ECC-based TLS ciphersuites according to
5501 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5502 SHA-1 now is only used for "small" curves (where the
5503 representation of a field element takes up to 24 bytes); for
5504 larger curves, the field element resulting from ECDH is directly
5505 used as premaster secret.
5506 [Douglas Stebila (Sun Microsystems Laboratories)]
5508 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5509 curve secp160r1 to the tests.
5510 [Douglas Stebila (Sun Microsystems Laboratories)]
5512 *) Add the possibility to load symbols globally with DSO.
5513 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5515 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5516 control of the error stack.
5519 *) Add support for STORE in ENGINE.
5522 *) Add the STORE type. The intention is to provide a common interface
5523 to certificate and key stores, be they simple file-based stores, or
5524 HSM-type store, or LDAP stores, or...
5525 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5528 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5529 pass a list of arguments to any function as well as provide a way
5530 for a function to pass data back to the caller.
5533 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5534 works like BUF_strdup() but can be used to duplicate a portion of
5535 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5539 *) Add the function sk_find_ex() which works like sk_find(), but will
5540 return an index to an element even if an exact match couldn't be
5541 found. The index is guaranteed to point at the element where the
5542 searched-for key would be inserted to preserve sorting order.
5545 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5546 takes an extra flags argument for optional functionality. Currently,
5547 the following flags are defined:
5549 OBJ_BSEARCH_VALUE_ON_NOMATCH
5550 This one gets OBJ_bsearch_ex() to return a pointer to the first
5551 element where the comparing function returns a negative or zero
5554 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5555 This one gets OBJ_bsearch_ex() to return a pointer to the first
5556 element where the comparing function returns zero. This is useful
5557 if there are more than one element where the comparing function
5561 *) Make it possible to create self-signed certificates with 'openssl ca'
5562 in such a way that the self-signed certificate becomes part of the
5563 CA database and uses the same mechanisms for serial number generation
5564 as all other certificate signing. The new flag '-selfsign' enables
5565 this functionality. Adapt CA.sh and CA.pl.in.
5568 *) Add functionality to check the public key of a certificate request
5569 against a given private. This is useful to check that a certificate
5570 request can be signed by that key (self-signing).
5573 *) Make it possible to have multiple active certificates with the same
5574 subject in the CA index file. This is done only if the keyword
5575 'unique_subject' is set to 'no' in the main CA section (default
5576 if 'CA_default') of the configuration file. The value is saved
5577 with the database itself in a separate index attribute file,
5578 named like the index file with '.attr' appended to the name.
5581 *) Generate multi-valued AVAs using '+' notation in config files for
5585 *) Support for nameConstraints certificate extension.
5588 *) Support for policyConstraints certificate extension.
5591 *) Support for policyMappings certificate extension.
5594 *) Make sure the default DSA_METHOD implementation only uses its
5595 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5596 and change its own handlers to be NULL so as to remove unnecessary
5597 indirection. This lets alternative implementations fallback to the
5598 default implementation more easily.
5601 *) Support for directoryName in GeneralName related extensions
5605 *) Make it possible to link applications using Makefile.shared.
5606 Make that possible even when linking against static libraries!
5609 *) Support for single pass processing for S/MIME signing. This now
5610 means that S/MIME signing can be done from a pipe, in addition
5611 cleartext signing (multipart/signed type) is effectively streaming
5612 and the signed data does not need to be all held in memory.
5614 This is done with a new flag PKCS7_STREAM. When this flag is set
5615 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5616 is done after the data is output (and digests calculated) in
5617 SMIME_write_PKCS7().
5620 *) Add full support for -rpath/-R, both in shared libraries and
5621 applications, at least on the platforms where it's known how
5625 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5626 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5627 will now compute a table of multiples of the generator that
5628 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5629 faster (notably in the case of a single point multiplication,
5630 scalar * generator).
5631 [Nils Larsch, Bodo Moeller]
5633 *) IPv6 support for certificate extensions. The various extensions
5634 which use the IP:a.b.c.d can now take IPv6 addresses using the
5635 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5639 *) Added an ENGINE that implements RSA by performing private key
5640 exponentiations with the GMP library. The conversions to and from
5641 GMP's mpz_t format aren't optimised nor are any montgomery forms
5642 cached, and on x86 it appears OpenSSL's own performance has caught up.
5643 However there are likely to be other architectures where GMP could
5644 provide a boost. This ENGINE is not built in by default, but it can be
5645 specified at Configure time and should be accompanied by the necessary
5646 linker additions, eg;
5647 ./config -DOPENSSL_USE_GMP -lgmp
5650 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5651 testing availability of engines with "-t" - the old behaviour is
5652 produced by increasing the feature's verbosity with "-tt".
5655 *) ECDSA routines: under certain error conditions uninitialized BN objects
5656 could be freed. Solution: make sure initialization is performed early
5657 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5661 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5662 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5663 software implementations. For DSA and DH, parameter generation can
5664 also be overridden by providing the appropriate method callbacks.
5667 *) Change the "progress" mechanism used in key-generation and
5668 primality testing to functions that take a new BN_GENCB pointer in
5669 place of callback/argument pairs. The new API functions have "_ex"
5670 postfixes and the older functions are reimplemented as wrappers for
5671 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5672 declarations of the old functions to help (graceful) attempts to
5673 migrate to the new functions. Also, the new key-generation API
5674 functions operate on a caller-supplied key-structure and return
5675 success/failure rather than returning a key or NULL - this is to
5676 help make "keygen" another member function of RSA_METHOD etc.
5678 Example for using the new callback interface:
5680 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5684 BN_GENCB_set(&my_cb, my_callback, my_arg);
5686 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5687 /* For the meaning of a, b in calls to my_callback(), see the
5688 * documentation of the function that calls the callback.
5689 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5690 * my_callback should return 1 if it wants BN_is_prime_ex()
5691 * to continue, or 0 to stop.
5696 *) Change the ZLIB compression method to be stateful, and make it
5697 available to TLS with the number defined in
5698 draft-ietf-tls-compression-04.txt.
5701 *) Add the ASN.1 structures and functions for CertificatePair, which
5702 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5704 CertificatePair ::= SEQUENCE {
5705 forward [0] Certificate OPTIONAL,
5706 reverse [1] Certificate OPTIONAL,
5707 -- at least one of the pair shall be present -- }
5709 Also implement the PEM functions to read and write certificate
5710 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5712 This needed to be defined, mostly for the sake of the LDAP
5713 attribute crossCertificatePair, but may prove useful elsewhere as
5717 *) Make it possible to inhibit symlinking of shared libraries in
5718 Makefile.shared, for Cygwin's sake.
5721 *) Extend the BIGNUM API by creating a function
5722 void BN_set_negative(BIGNUM *a, int neg);
5723 and a macro that behave like
5724 int BN_is_negative(const BIGNUM *a);
5726 to avoid the need to access 'a->neg' directly in applications.
5729 *) Implement fast modular reduction for pseudo-Mersenne primes
5730 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5731 EC_GROUP_new_curve_GFp() will now automatically use this
5733 [Nils Larsch <nla@trustcenter.de>]
5735 *) Add new lock type (CRYPTO_LOCK_BN).
5738 *) Change the ENGINE framework to automatically load engines
5739 dynamically from specific directories unless they could be
5740 found to already be built in or loaded. Move all the
5741 current engines except for the cryptodev one to a new
5743 The engines in engines/ are built as shared libraries if
5744 the "shared" options was given to ./Configure or ./config.
5745 Otherwise, they are inserted in libcrypto.a.
5746 /usr/local/ssl/engines is the default directory for dynamic
5747 engines, but that can be overridden at configure time through
5748 the usual use of --prefix and/or --openssldir, and at run
5749 time with the environment variable OPENSSL_ENGINES.
5750 [Geoff Thorpe and Richard Levitte]
5752 *) Add Makefile.shared, a helper makefile to build shared
5753 libraries. Adapt Makefile.org.
5756 *) Add version info to Win32 DLLs.
5757 [Peter 'Luna' Runestig" <peter@runestig.com>]
5759 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5760 can be added using this API to created arbitrary PKCS#12
5761 files while avoiding the low level API.
5763 New options to PKCS12_create(), key or cert can be NULL and
5764 will then be omitted from the output file. The encryption
5765 algorithm NIDs can be set to -1 for no encryption, the mac
5766 iteration count can be set to 0 to omit the mac.
5768 Enhance pkcs12 utility by making the -nokeys and -nocerts
5769 options work when creating a PKCS#12 file. New option -nomac
5770 to omit the mac, NONE can be set for an encryption algorithm.
5771 New code is modified to use the enhanced PKCS12_create()
5772 instead of the low level API.
5775 *) Extend ASN1 encoder to support indefinite length constructed
5776 encoding. This can output sequences tags and octet strings in
5777 this form. Modify pk7_asn1.c to support indefinite length
5778 encoding. This is experimental and needs additional code to
5779 be useful, such as an ASN1 bio and some enhanced streaming
5782 Extend template encode functionality so that tagging is passed
5783 down to the template encoder.
5786 *) Let 'openssl req' fail if an argument to '-newkey' is not
5787 recognized instead of using RSA as a default.
5790 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5791 As these are not official, they are not included in "ALL";
5792 the "ECCdraft" ciphersuite group alias can be used to select them.
5793 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5795 *) Add ECDH engine support.
5796 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5798 *) Add ECDH in new directory crypto/ecdh/.
5799 [Douglas Stebila (Sun Microsystems Laboratories)]
5801 *) Let BN_rand_range() abort with an error after 100 iterations
5802 without success (which indicates a broken PRNG).
5805 *) Change BN_mod_sqrt() so that it verifies that the input value
5806 is really the square of the return value. (Previously,
5807 BN_mod_sqrt would show GIGO behaviour.)
5810 *) Add named elliptic curves over binary fields from X9.62, SECG,
5811 and WAP/WTLS; add OIDs that were still missing.
5813 [Sheueling Chang Shantz and Douglas Stebila
5814 (Sun Microsystems Laboratories)]
5816 *) Extend the EC library for elliptic curves over binary fields
5817 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5820 EC_GF2m_simple_method
5824 EC_GROUP_new_curve_GF2m
5825 EC_GROUP_set_curve_GF2m
5826 EC_GROUP_get_curve_GF2m
5827 EC_POINT_set_affine_coordinates_GF2m
5828 EC_POINT_get_affine_coordinates_GF2m
5829 EC_POINT_set_compressed_coordinates_GF2m
5831 Point compression for binary fields is disabled by default for
5832 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5835 As binary polynomials are represented as BIGNUMs, various members
5836 of the EC_GROUP and EC_POINT data structures can be shared
5837 between the implementations for prime fields and binary fields;
5838 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5839 are essentially identical to their ..._GFp counterparts.
5840 (For simplicity, the '..._GFp' prefix has been dropped from
5841 various internal method names.)
5843 An internal 'field_div' method (similar to 'field_mul' and
5844 'field_sqr') has been added; this is used only for binary fields.
5846 [Sheueling Chang Shantz and Douglas Stebila
5847 (Sun Microsystems Laboratories)]
5849 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5850 through methods ('mul', 'precompute_mult').
5852 The generic implementations (now internally called 'ec_wNAF_mul'
5853 and 'ec_wNAF_precomputed_mult') remain the default if these
5854 methods are undefined.
5856 [Sheueling Chang Shantz and Douglas Stebila
5857 (Sun Microsystems Laboratories)]
5859 *) New function EC_GROUP_get_degree, which is defined through
5860 EC_METHOD. For curves over prime fields, this returns the bit
5861 length of the modulus.
5863 [Sheueling Chang Shantz and Douglas Stebila
5864 (Sun Microsystems Laboratories)]
5866 *) New functions EC_GROUP_dup, EC_POINT_dup.
5867 (These simply call ..._new and ..._copy).
5869 [Sheueling Chang Shantz and Douglas Stebila
5870 (Sun Microsystems Laboratories)]
5872 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5873 Polynomials are represented as BIGNUMs (where the sign bit is not
5874 used) in the following functions [macros]:
5877 BN_GF2m_sub [= BN_GF2m_add]
5878 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5879 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5880 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5882 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5883 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5884 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5885 BN_GF2m_cmp [= BN_ucmp]
5887 (Note that only the 'mod' functions are actually for fields GF(2^m).
5888 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5890 For some functions, an the irreducible polynomial defining a
5891 field can be given as an 'unsigned int[]' with strictly
5892 decreasing elements giving the indices of those bits that are set;
5893 i.e., p[] represents the polynomial
5894 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5896 p[0] > p[1] > ... > p[k] = 0.
5897 This applies to the following functions:
5902 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5903 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5905 BN_GF2m_mod_sqrt_arr
5906 BN_GF2m_mod_solve_quad_arr
5910 Conversion can be performed by the following functions:
5915 bntest.c has additional tests for binary polynomial arithmetic.
5917 Two implementations for BN_GF2m_mod_div() are available.
5918 The default algorithm simply uses BN_GF2m_mod_inv() and
5919 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5920 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5921 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5923 [Sheueling Chang Shantz and Douglas Stebila
5924 (Sun Microsystems Laboratories)]
5926 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5927 functionality is disabled at compile-time.
5928 [Douglas Stebila <douglas.stebila@sun.com>]
5930 *) Change default behaviour of 'openssl asn1parse' so that more
5931 information is visible when viewing, e.g., a certificate:
5933 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5934 mode the content of non-printable OCTET STRINGs is output in a
5935 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5936 avoid the appearance of a printable string.
5937 [Nils Larsch <nla@trustcenter.de>]
5939 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5941 EC_GROUP_set_asn1_flag()
5942 EC_GROUP_get_asn1_flag()
5943 EC_GROUP_set_point_conversion_form()
5944 EC_GROUP_get_point_conversion_form()
5945 These control ASN1 encoding details:
5946 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5947 has been set to OPENSSL_EC_NAMED_CURVE.
5948 - Points are encoded in uncompressed form by default; options for
5949 asn1_for are as for point2oct, namely
5950 POINT_CONVERSION_COMPRESSED
5951 POINT_CONVERSION_UNCOMPRESSED
5952 POINT_CONVERSION_HYBRID
5954 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5957 EC_GROUP_get0_seed()
5958 EC_GROUP_get_seed_len()
5959 This is used only for ASN1 purposes (so far).
5960 [Nils Larsch <nla@trustcenter.de>]
5962 *) Add 'field_type' member to EC_METHOD, which holds the NID
5963 of the appropriate field type OID. The new function
5964 EC_METHOD_get_field_type() returns this value.
5965 [Nils Larsch <nla@trustcenter.de>]
5970 EC_POINT_point2hex()
5971 EC_POINT_hex2point()
5972 providing useful interfaces to EC_POINT_point2oct() and
5973 EC_POINT_oct2point().
5974 [Nils Larsch <nla@trustcenter.de>]
5976 *) Change internals of the EC library so that the functions
5977 EC_GROUP_set_generator()
5978 EC_GROUP_get_generator()
5979 EC_GROUP_get_order()
5980 EC_GROUP_get_cofactor()
5981 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5982 to methods, which would lead to unnecessary code duplication when
5983 adding different types of curves.
5984 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5986 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5987 arithmetic, and such that modified wNAFs are generated
5988 (which avoid length expansion in many cases).
5991 *) Add a function EC_GROUP_check_discriminant() (defined via
5992 EC_METHOD) that verifies that the curve discriminant is non-zero.
5994 Add a function EC_GROUP_check() that makes some sanity tests
5995 on a EC_GROUP, its generator and order. This includes
5996 EC_GROUP_check_discriminant().
5997 [Nils Larsch <nla@trustcenter.de>]
5999 *) Add ECDSA in new directory crypto/ecdsa/.
6001 Add applications 'openssl ecparam' and 'openssl ecdsa'
6002 (these are based on 'openssl dsaparam' and 'openssl dsa').
6004 ECDSA support is also included in various other files across the
6005 library. Most notably,
6006 - 'openssl req' now has a '-newkey ecdsa:file' option;
6007 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6008 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6009 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6010 them suitable for ECDSA where domain parameters must be
6011 extracted before the specific public key;
6012 - ECDSA engine support has been added.
6013 [Nils Larsch <nla@trustcenter.de>]
6015 *) Include some named elliptic curves, and add OIDs from X9.62,
6016 SECG, and WAP/WTLS. Each curve can be obtained from the new
6018 EC_GROUP_new_by_curve_name(),
6019 and the list of available named curves can be obtained with
6020 EC_get_builtin_curves().
6021 Also add a 'curve_name' member to EC_GROUP objects, which can be
6023 EC_GROUP_set_curve_name()
6024 EC_GROUP_get_curve_name()
6025 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6027 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6028 was actually never needed) and in BN_mul(). The removal in BN_mul()
6029 required a small change in bn_mul_part_recursive() and the addition
6030 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6031 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6032 bn_sub_words() and bn_add_words() except they take arrays with
6036 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6038 *) Cleanse PEM buffers before freeing them since they may contain
6040 [Benjamin Bennett <ben@psc.edu>]
6042 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6043 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6044 authentication-only ciphersuites.
6047 *) Since AES128 and AES256 share a single mask bit in the logic of
6048 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6049 kludge to work properly if AES128 is available and AES256 isn't.
6052 *) Expand security boundary to match 1.1.1 module.
6055 *) Remove redundant features: hash file source, editing of test vectors
6056 modify fipsld to use external fips_premain.c signature.
6059 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6060 run algorithm test programs.
6063 *) Make algorithm test programs more tolerant of whitespace.
6066 *) Have SSL/TLS server implementation tolerate "mismatched" record
6067 protocol version while receiving ClientHello even if the
6068 ClientHello is fragmented. (The server can't insist on the
6069 particular protocol version it has chosen before the ServerHello
6070 message has informed the client about his choice.)
6073 *) Load error codes if they are not already present instead of using a
6074 static variable. This allows them to be cleanly unloaded and reloaded.
6077 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6079 *) Introduce limits to prevent malicious keys being able to
6080 cause a denial of service. (CVE-2006-2940)
6081 [Steve Henson, Bodo Moeller]
6083 *) Fix ASN.1 parsing of certain invalid structures that can result
6084 in a denial of service. (CVE-2006-2937) [Steve Henson]
6086 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6087 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6089 *) Fix SSL client code which could crash if connecting to a
6090 malicious SSLv2 server. (CVE-2006-4343)
6091 [Tavis Ormandy and Will Drewry, Google Security Team]
6093 *) Change ciphersuite string processing so that an explicit
6094 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6095 will no longer include "AES128-SHA"), and any other similar
6096 ciphersuite (same bitmap) from *other* protocol versions (so that
6097 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6098 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6099 changes from 0.9.8b and 0.9.8d.
6102 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6104 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6105 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6107 *) Change the Unix randomness entropy gathering to use poll() when
6108 possible instead of select(), since the latter has some
6109 undesirable limitations.
6110 [Darryl Miles via Richard Levitte and Bodo Moeller]
6112 *) Disable rogue ciphersuites:
6114 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6115 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6116 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6118 The latter two were purportedly from
6119 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6122 Also deactivate the remaining ciphersuites from
6123 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6124 unofficial, and the ID has long expired.
6127 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6128 dual-core machines) and other potential thread-safety issues.
6131 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6133 *) Adapt fipsld and the build system to link against the validated FIPS
6134 module in FIPS mode.
6137 *) Fixes for VC++ 2005 build under Windows.
6140 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6141 from a Windows bash shell such as MSYS. It is autodetected from the
6142 "config" script when run from a VC++ environment. Modify standard VC++
6143 build to use fipscanister.o from the GNU make build.
6146 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6148 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6149 The value now differs depending on if you build for FIPS or not.
6150 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6151 safely run with a non-FIPSed libcrypto, as it may crash because of
6152 the difference induced by this change.
6155 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6157 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6158 (part of SSL_OP_ALL). This option used to disable the
6159 countermeasure against man-in-the-middle protocol-version
6160 rollback in the SSL 2.0 server implementation, which is a bad
6161 idea. (CVE-2005-2969)
6163 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6164 for Information Security, National Institute of Advanced Industrial
6165 Science and Technology [AIST], Japan)]
6167 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6168 mainly for FIPS compliance and not fully integrated at this stage.
6171 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6172 the exponentiation using a fixed-length exponent. (Otherwise,
6173 the information leaked through timing could expose the secret key
6174 after many signatures; cf. Bleichenbacher's attack on DSA with
6178 *) Make a new fixed-window mod_exp implementation the default for
6179 RSA, DSA, and DH private-key operations so that the sequence of
6180 squares and multiplies and the memory access pattern are
6181 independent of the particular secret key. This will mitigate
6182 cache-timing and potential related attacks.
6184 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6185 and this is automatically used by BN_mod_exp_mont() if the new flag
6186 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6187 will use this BN flag for private exponents unless the flag
6188 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6189 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6191 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6193 *) Change the client implementation for SSLv23_method() and
6194 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6195 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6196 (Previously, the SSL 2.0 backwards compatible Client Hello
6197 message format would be used even with SSL_OP_NO_SSLv2.)
6200 *) Add support for smime-type MIME parameter in S/MIME messages which some
6204 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6205 a threadsafe manner. Modify rsa code to use new function and add calls
6206 to dsa and dh code (which had race conditions before).
6209 *) Include the fixed error library code in the C error file definitions
6210 instead of fixing them up at runtime. This keeps the error code
6211 structures constant.
6214 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6216 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6219 *) Fixes for newer kerberos headers. NB: the casts are needed because
6220 the 'length' field is signed on one version and unsigned on another
6221 with no (?) obvious way to tell the difference, without these VC++
6222 complains. Also the "definition" of FAR (blank) is no longer included
6223 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6224 some needed definitions.
6227 *) Undo Cygwin change.
6230 *) Added support for proxy certificates according to RFC 3820.
6231 Because they may be a security thread to unaware applications,
6232 they must be explicitly allowed in run-time. See
6233 docs/HOWTO/proxy_certificates.txt for further information.
6236 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6238 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6239 server and client random values. Previously
6240 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6241 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6243 This change has negligible security impact because:
6245 1. Server and client random values still have 24 bytes of pseudo random
6248 2. Server and client random values are sent in the clear in the initial
6251 3. The master secret is derived using the premaster secret (48 bytes in
6252 size for static RSA ciphersuites) as well as client server and random
6255 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6258 [Stephen Henson, reported by UK NISCC]
6260 *) Use Windows randomness collection on Cygwin.
6263 *) Fix hang in EGD/PRNGD query when communication socket is closed
6264 prematurely by EGD/PRNGD.
6265 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6267 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6270 *) Back-port of selected performance improvements from development
6271 branch, as well as improved support for PowerPC platforms.
6274 *) Add lots of checks for memory allocation failure, error codes to indicate
6275 failure and freeing up memory if a failure occurs.
6276 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6278 *) Add new -passin argument to dgst.
6281 *) Perform some character comparisons of different types in X509_NAME_cmp:
6282 this is needed for some certificates that re-encode DNs into UTF8Strings
6283 (in violation of RFC3280) and can't or won't issue name rollover
6287 *) Make an explicit check during certificate validation to see that
6288 the CA setting in each certificate on the chain is correct. As a
6289 side effect always do the following basic checks on extensions,
6290 not just when there's an associated purpose to the check:
6292 - if there is an unhandled critical extension (unless the user
6293 has chosen to ignore this fault)
6294 - if the path length has been exceeded (if one is set at all)
6295 - that certain extensions fit the associated purpose (if one has
6299 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6301 *) Avoid a race condition when CRLs are checked in a multi threaded
6302 environment. This would happen due to the reordering of the revoked
6303 entries during signature checking and serial number lookup. Now the
6304 encoding is cached and the serial number sort performed under a lock.
6305 Add new STACK function sk_is_sorted().
6308 *) Add Delta CRL to the extension code.
6311 *) Various fixes to s3_pkt.c so alerts are sent properly.
6312 [David Holmes <d.holmes@f5.com>]
6314 *) Reduce the chances of duplicate issuer name and serial numbers (in
6315 violation of RFC3280) using the OpenSSL certificate creation utilities.
6316 This is done by creating a random 64 bit value for the initial serial
6317 number when a serial number file is created or when a self signed
6318 certificate is created using 'openssl req -x509'. The initial serial
6319 number file is created using 'openssl x509 -next_serial' in CA.pl
6320 rather than being initialized to 1.
6323 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6325 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6326 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6327 [Joe Orton, Steve Henson]
6329 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6331 [Joe Orton, Steve Henson]
6333 *) Make it possible to have multiple active certificates with the same
6334 subject in the CA index file. This is done only if the keyword
6335 'unique_subject' is set to 'no' in the main CA section (default
6336 if 'CA_default') of the configuration file. The value is saved
6337 with the database itself in a separate index attribute file,
6338 named like the index file with '.attr' appended to the name.
6341 *) X509 verify fixes. Disable broken certificate workarounds when
6342 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6343 keyUsage extension present. Don't accept CRLs with unhandled critical
6344 extensions: since verify currently doesn't process CRL extensions this
6345 rejects a CRL with *any* critical extensions. Add new verify error codes
6349 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6350 A clarification of RFC2560 will require the use of OCTET STRINGs and
6351 some implementations cannot handle the current raw format. Since OpenSSL
6352 copies and compares OCSP nonces as opaque blobs without any attempt at
6353 parsing them this should not create any compatibility issues.
6356 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6357 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6358 this HMAC (and other) operations are several times slower than OpenSSL
6362 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6363 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6365 *) Use the correct content when signing type "other".
6368 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6370 *) Fix various bugs revealed by running the NISCC test suite:
6372 Stop out of bounds reads in the ASN1 code when presented with
6373 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6375 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6377 If verify callback ignores invalid public key errors don't try to check
6378 certificate signature with the NULL public key.
6382 *) New -ignore_err option in ocsp application to stop the server
6383 exiting on the first error in a request.
6386 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6387 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6391 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6392 extra data after the compression methods not only for TLS 1.0
6393 but also for SSL 3.0 (as required by the specification).
6394 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6396 *) Change X509_certificate_type() to mark the key as exported/exportable
6397 when it's 512 *bits* long, not 512 bytes.
6400 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6401 blocks during encryption.
6404 *) Various fixes to base64 BIO and non blocking I/O. On write
6405 flushes were not handled properly if the BIO retried. On read
6406 data was not being buffered properly and had various logic bugs.
6407 This also affects blocking I/O when the data being decoded is a
6411 *) Various S/MIME bugfixes and compatibility changes:
6412 output correct application/pkcs7 MIME type if
6413 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6414 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6415 of files as .eml work). Correctly handle very long lines in MIME
6419 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6421 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6422 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6423 a protocol version number mismatch like a decryption error
6424 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6427 *) Turn on RSA blinding by default in the default implementation
6428 to avoid a timing attack. Applications that don't want it can call
6429 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6430 They would be ill-advised to do so in most cases.
6431 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6433 *) Change RSA blinding code so that it works when the PRNG is not
6434 seeded (in this case, the secret RSA exponent is abused as
6435 an unpredictable seed -- if it is not unpredictable, there
6436 is no point in blinding anyway). Make RSA blinding thread-safe
6437 by remembering the creator's thread ID in rsa->blinding and
6438 having all other threads use local one-time blinding factors
6439 (this requires more computation than sharing rsa->blinding, but
6440 avoids excessive locking; and if an RSA object is not shared
6441 between threads, blinding will still be very fast).
6444 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6445 ENGINE as defaults for all supported algorithms irrespective of
6446 the 'flags' parameter. 'flags' is now honoured, so applications
6447 should make sure they are passing it correctly.
6450 *) Target "mingw" now allows native Windows code to be generated in
6451 the Cygwin environment as well as with the MinGW compiler.
6454 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6456 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6457 via timing by performing a MAC computation even if incorrect
6458 block cipher padding has been found. This is a countermeasure
6459 against active attacks where the attacker has to distinguish
6460 between bad padding and a MAC verification error. (CVE-2003-0078)
6462 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6463 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6464 Martin Vuagnoux (EPFL, Ilion)]
6466 *) Make the no-err option work as intended. The intention with no-err
6467 is not to have the whole error stack handling routines removed from
6468 libcrypto, it's only intended to remove all the function name and
6469 reason texts, thereby removing some of the footprint that may not
6470 be interesting if those errors aren't displayed anyway.
6472 NOTE: it's still possible for any application or module to have it's
6473 own set of error texts inserted. The routines are there, just not
6474 used by default when no-err is given.
6477 *) Add support for FreeBSD on IA64.
6478 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6480 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6481 Kerberos function mit_des_cbc_cksum(). Before this change,
6482 the value returned by DES_cbc_cksum() was like the one from
6483 mit_des_cbc_cksum(), except the bytes were swapped.
6484 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6486 *) Allow an application to disable the automatic SSL chain building.
6487 Before this a rather primitive chain build was always performed in
6488 ssl3_output_cert_chain(): an application had no way to send the
6489 correct chain if the automatic operation produced an incorrect result.
6491 Now the chain builder is disabled if either:
6493 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6495 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6497 The reasoning behind this is that an application would not want the
6498 auto chain building to take place if extra chain certificates are
6499 present and it might also want a means of sending no additional
6500 certificates (for example the chain has two certificates and the
6504 *) Add the possibility to build without the ENGINE framework.
6505 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6507 *) Under Win32 gmtime() can return NULL: check return value in
6508 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6511 *) DSA routines: under certain error conditions uninitialized BN objects
6512 could be freed. Solution: make sure initialization is performed early
6513 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6514 Nils Larsch <nla@trustcenter.de> via PR#459)
6517 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6518 checked on reconnect on the client side, therefore session resumption
6519 could still fail with a "ssl session id is different" error. This
6520 behaviour is masked when SSL_OP_ALL is used due to
6521 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6522 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6523 followup to PR #377.
6526 *) IA-32 assembler support enhancements: unified ELF targets, support
6527 for SCO/Caldera platforms, fix for Cygwin shared build.
6530 *) Add support for FreeBSD on sparc64. As a consequence, support for
6531 FreeBSD on non-x86 processors is separate from x86 processors on
6532 the config script, much like the NetBSD support.
6533 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6535 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6537 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6540 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6541 code (06) was taken as the first octet of the session ID and the last
6542 octet was ignored consequently. As a result SSLv2 client side session
6543 caching could not have worked due to the session ID mismatch between
6545 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6549 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6550 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6554 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6555 seems that in spite of existing for more than a year, many application
6556 author have done nothing to provide the necessary callbacks, which
6557 means that this particular engine will not work properly anywhere.
6558 This is a very unfortunate situation which forces us, in the name
6559 of usability, to give the hw_ncipher.c a static lock, which is part
6561 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6562 appear in 0.9.8 or later. We EXPECT application authors to have
6563 dealt properly with this when 0.9.8 is released (unless we actually
6564 make such changes in the libcrypto locking code that changes will
6565 have to be made anyway).
6568 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6569 octets have been read, EOF or an error occurs. Without this change
6570 some truncated ASN1 structures will not produce an error.
6573 *) Disable Heimdal support, since it hasn't been fully implemented.
6574 Still give the possibility to force the use of Heimdal, but with
6575 warnings and a request that patches get sent to openssl-dev.
6578 *) Add the VC-CE target, introduce the WINCE sysname, and add
6579 INSTALL.WCE and appropriate conditionals to make it build.
6580 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6582 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6583 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6584 edit numbers of the version.
6585 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6587 *) Introduce safe string copy and catenation functions
6588 (BUF_strlcpy() and BUF_strlcat()).
6589 [Ben Laurie (CHATS) and Richard Levitte]
6591 *) Avoid using fixed-size buffers for one-line DNs.
6592 [Ben Laurie (CHATS)]
6594 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6595 resizing buffers containing secrets, and use where appropriate.
6596 [Ben Laurie (CHATS)]
6598 *) Avoid using fixed size buffers for configuration file location.
6599 [Ben Laurie (CHATS)]
6601 *) Avoid filename truncation for various CA files.
6602 [Ben Laurie (CHATS)]
6604 *) Use sizeof in preference to magic numbers.
6605 [Ben Laurie (CHATS)]
6607 *) Avoid filename truncation in cert requests.
6608 [Ben Laurie (CHATS)]
6610 *) Add assertions to check for (supposedly impossible) buffer
6612 [Ben Laurie (CHATS)]
6614 *) Don't cache truncated DNS entries in the local cache (this could
6615 potentially lead to a spoofing attack).
6616 [Ben Laurie (CHATS)]
6618 *) Fix various buffers to be large enough for hex/decimal
6619 representations in a platform independent manner.
6620 [Ben Laurie (CHATS)]
6622 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6623 resizing buffers containing secrets, and use where appropriate.
6624 [Ben Laurie (CHATS)]
6626 *) Add BIO_indent() to avoid much slightly worrying code to do
6628 [Ben Laurie (CHATS)]
6630 *) Convert sprintf()/BIO_puts() to BIO_printf().
6631 [Ben Laurie (CHATS)]
6633 *) buffer_gets() could terminate with the buffer only half
6635 [Ben Laurie (CHATS)]
6637 *) Add assertions to prevent user-supplied crypto functions from
6638 overflowing internal buffers by having large block sizes, etc.
6639 [Ben Laurie (CHATS)]
6641 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6643 [Ben Laurie (CHATS)]
6645 *) Eliminate unused copy of key in RC4.
6646 [Ben Laurie (CHATS)]
6648 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6649 [Ben Laurie (CHATS)]
6651 *) Fix off-by-one error in EGD path.
6652 [Ben Laurie (CHATS)]
6654 *) If RANDFILE path is too long, ignore instead of truncating.
6655 [Ben Laurie (CHATS)]
6657 *) Eliminate unused and incorrectly sized X.509 structure
6659 [Ben Laurie (CHATS)]
6661 *) Eliminate unused and dangerous function knumber().
6662 [Ben Laurie (CHATS)]
6664 *) Eliminate unused and dangerous structure, KSSL_ERR.
6665 [Ben Laurie (CHATS)]
6667 *) Protect against overlong session ID context length in an encoded
6668 session object. Since these are local, this does not appear to be
6670 [Ben Laurie (CHATS)]
6672 *) Change from security patch (see 0.9.6e below) that did not affect
6673 the 0.9.6 release series:
6675 Remote buffer overflow in SSL3 protocol - an attacker could
6676 supply an oversized master key in Kerberos-enabled versions.
6678 [Ben Laurie (CHATS)]
6680 *) Change the SSL kerb5 codes to match RFC 2712.
6683 *) Make -nameopt work fully for req and add -reqopt switch.
6684 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6686 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6687 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6689 *) Make sure tests can be performed even if the corresponding algorithms
6690 have been removed entirely. This was also the last step to make
6691 OpenSSL compilable with DJGPP under all reasonable conditions.
6692 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6694 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6695 to allow version independent disabling of normally unselected ciphers,
6696 which may be activated as a side-effect of selecting a single cipher.
6698 (E.g., cipher list string "RSA" enables ciphersuites that are left
6699 out of "ALL" because they do not provide symmetric encryption.
6700 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6701 [Lutz Jaenicke, Bodo Moeller]
6703 *) Add appropriate support for separate platform-dependent build
6704 directories. The recommended way to make a platform-dependent
6705 build directory is the following (tested on Linux), maybe with
6708 # Place yourself outside of the OpenSSL source tree. In
6709 # this example, the environment variable OPENSSL_SOURCE
6710 # is assumed to contain the absolute OpenSSL source directory.
6711 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6712 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6713 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6714 mkdir -p `dirname $F`
6715 ln -s $OPENSSL_SOURCE/$F $F
6718 To be absolutely sure not to disturb the source tree, a "make clean"
6719 is a good thing. If it isn't successful, don't worry about it,
6720 it probably means the source directory is very clean.
6723 *) Make sure any ENGINE control commands make local copies of string
6724 pointers passed to them whenever necessary. Otherwise it is possible
6725 the caller may have overwritten (or deallocated) the original string
6726 data when a later ENGINE operation tries to use the stored values.
6727 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6729 *) Improve diagnostics in file reading and command-line digests.
6730 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6732 *) Add AES modes CFB and OFB to the object database. Correct an
6733 error in AES-CFB decryption.
6736 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6737 allows existing EVP_CIPHER_CTX structures to be reused after
6738 calling EVP_*Final(). This behaviour is used by encryption
6739 BIOs and some applications. This has the side effect that
6740 applications must explicitly clean up cipher contexts with
6741 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6744 *) Check the values of dna and dnb in bn_mul_recursive before calling
6745 bn_mul_comba (a non zero value means the a or b arrays do not contain
6746 n2 elements) and fallback to bn_mul_normal if either is not zero.
6749 *) Fix escaping of non-ASCII characters when using the -subj option
6750 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6753 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6754 form for "surname", serialNumber has no short form.
6755 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6756 therefore remove "mail" short name for "internet 7".
6757 The OID for unique identifiers in X509 certificates is
6758 x500UniqueIdentifier, not uniqueIdentifier.
6759 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6762 *) Add an "init" command to the ENGINE config module and auto initialize
6763 ENGINEs. Without any "init" command the ENGINE will be initialized
6764 after all ctrl commands have been executed on it. If init=1 the
6765 ENGINE is initialized at that point (ctrls before that point are run
6766 on the uninitialized ENGINE and after on the initialized one). If
6767 init=0 then the ENGINE will not be initialized at all.
6770 *) Fix the 'app_verify_callback' interface so that the user-defined
6771 argument is actually passed to the callback: In the
6772 SSL_CTX_set_cert_verify_callback() prototype, the callback
6773 declaration has been changed from
6776 int (*cb)(X509_STORE_CTX *,void *);
6777 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6778 i=s->ctx->app_verify_callback(&ctx)
6779 has been changed into
6780 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6782 To update applications using SSL_CTX_set_cert_verify_callback(),
6783 a dummy argument can be added to their callback functions.
6784 [D. K. Smetters <smetters@parc.xerox.com>]
6786 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6787 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6789 *) Add and OPENSSL_LOAD_CONF define which will cause
6790 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6791 This allows older applications to transparently support certain
6792 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6793 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6794 load the config file and OPENSSL_add_all_algorithms_conf() which will
6795 always load it have also been added.
6798 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6799 Adjust NIDs and EVP layer.
6800 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6802 *) Config modules support in openssl utility.
6804 Most commands now load modules from the config file,
6805 though in a few (such as version) this isn't done
6806 because it couldn't be used for anything.
6808 In the case of ca and req the config file used is
6809 the same as the utility itself: that is the -config
6810 command line option can be used to specify an
6814 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6815 use "openssl_conf" if filename is NULL use default openssl config file.
6818 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6819 config section name. Add a new flag to tolerate a missing config file
6820 and move code to CONF_modules_load_file().
6823 *) Support for crypto accelerator cards from Accelerated Encryption
6824 Processing, www.aep.ie. (Use engine 'aep')
6825 The support was copied from 0.9.6c [engine] and adapted/corrected
6826 to work with the new engine framework.
6827 [AEP Inc. and Richard Levitte]
6829 *) Support for SureWare crypto accelerator cards from Baltimore
6830 Technologies. (Use engine 'sureware')
6831 The support was copied from 0.9.6c [engine] and adapted
6832 to work with the new engine framework.
6835 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6836 make the newer ENGINE framework commands for the CHIL engine work.
6837 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6839 *) Make it possible to produce shared libraries on ReliantUNIX.
6840 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6842 *) Add the configuration target debug-linux-ppro.
6843 Make 'openssl rsa' use the general key loading routines
6844 implemented in apps.c, and make those routines able to
6845 handle the key format FORMAT_NETSCAPE and the variant
6847 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6849 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6850 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6852 *) Add -keyform to rsautl, and document -engine.
6853 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6855 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6856 BIO_R_NO_SUCH_FILE error code rather than the generic
6857 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6860 *) Add new functions
6862 ERR_peek_last_error_line
6863 ERR_peek_last_error_line_data.
6864 These are similar to
6867 ERR_peek_error_line_data,
6868 but report on the latest error recorded rather than the first one
6869 still in the error queue.
6870 [Ben Laurie, Bodo Moeller]
6872 *) default_algorithms option in ENGINE config module. This allows things
6874 default_algorithms = ALL
6875 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6878 *) Preliminary ENGINE config module.
6881 *) New experimental application configuration code.
6884 *) Change the AES code to follow the same name structure as all other
6885 symmetric ciphers, and behave the same way. Move everything to
6886 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6887 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6889 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6890 [Ben Laurie and Theo de Raadt]
6892 *) Add option to output public keys in req command.
6893 [Massimiliano Pala madwolf@openca.org]
6895 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6896 (up to about 10% better than before for P-192 and P-224).
6899 *) New functions/macros
6901 SSL_CTX_set_msg_callback(ctx, cb)
6902 SSL_CTX_set_msg_callback_arg(ctx, arg)
6903 SSL_set_msg_callback(ssl, cb)
6904 SSL_set_msg_callback_arg(ssl, arg)
6906 to request calling a callback function
6908 void cb(int write_p, int version, int content_type,
6909 const void *buf, size_t len, SSL *ssl, void *arg)
6911 whenever a protocol message has been completely received
6912 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6913 protocol version according to which the SSL library interprets
6914 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6915 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6916 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6917 specification (change_cipher_spec(20), alert(21), handshake(22)).
6918 'buf' and 'len' point to the actual message, 'ssl' to the
6919 SSL object, and 'arg' is the application-defined value set by
6920 SSL[_CTX]_set_msg_callback_arg().
6922 'openssl s_client' and 'openssl s_server' have new '-msg' options
6923 to enable a callback that displays all protocol messages.
6926 *) Change the shared library support so shared libraries are built as
6927 soon as the corresponding static library is finished, and thereby get
6928 openssl and the test programs linked against the shared library.
6929 This still only happens when the keyword "shard" has been given to
6930 the configuration scripts.
6932 NOTE: shared library support is still an experimental thing, and
6933 backward binary compatibility is still not guaranteed.
6934 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6936 *) Add support for Subject Information Access extension.
6937 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6939 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6940 additional bytes when new memory had to be allocated, not just
6941 when reusing an existing buffer.
6944 *) New command line and configuration option 'utf8' for the req command.
6945 This allows field values to be specified as UTF8 strings.
6948 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6949 runs for the former and machine-readable output for the latter.
6952 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6953 of the e-mail address in the DN (i.e., it will go into a certificate
6954 extension only). The new configuration file option 'email_in_dn = no'
6955 has the same effect.
6956 [Massimiliano Pala madwolf@openca.org]
6958 *) Change all functions with names starting with des_ to be starting
6959 with DES_ instead. Add wrappers that are compatible with libdes,
6960 but are named _ossl_old_des_*. Finally, add macros that map the
6961 des_* symbols to the corresponding _ossl_old_des_* if libdes
6962 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6963 desired, the des_* symbols will be mapped to DES_*, with one
6966 Since we provide two compatibility mappings, the user needs to
6967 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6968 compatibility is desired. The default (i.e., when that macro
6969 isn't defined) is OpenSSL 0.9.6c compatibility.
6971 There are also macros that enable and disable the support of old
6972 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6973 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6974 are defined, the default will apply: to support the old des routines.
6976 In either case, one must include openssl/des.h to get the correct
6977 definitions. Do not try to just include openssl/des_old.h, that
6980 NOTE: This is a major break of an old API into a new one. Software
6981 authors are encouraged to switch to the DES_ style functions. Some
6982 time in the future, des_old.h and the libdes compatibility functions
6983 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6984 default), and then completely removed.
6987 *) Test for certificates which contain unsupported critical extensions.
6988 If such a certificate is found during a verify operation it is
6989 rejected by default: this behaviour can be overridden by either
6990 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6991 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6992 X509_supported_extension() has also been added which returns 1 if a
6993 particular extension is supported.
6996 *) Modify the behaviour of EVP cipher functions in similar way to digests
6997 to retain compatibility with existing code.
7000 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7001 compatibility with existing code. In particular the 'ctx' parameter does
7002 not have to be to be initialized before the call to EVP_DigestInit() and
7003 it is tidied up after a call to EVP_DigestFinal(). New function
7004 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7005 EVP_MD_CTX_copy() changed to not require the destination to be
7006 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7007 requires the destination to be valid.
7009 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7010 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7013 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7014 so that complete 'Handshake' protocol structures are kept in memory
7015 instead of overwriting 'msg_type' and 'length' with 'body' data.
7018 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7019 [Massimo Santin via Richard Levitte]
7021 *) Major restructuring to the underlying ENGINE code. This includes
7022 reduction of linker bloat, separation of pure "ENGINE" manipulation
7023 (initialisation, etc) from functionality dealing with implementations
7024 of specific crypto interfaces. This change also introduces integrated
7025 support for symmetric ciphers and digest implementations - so ENGINEs
7026 can now accelerate these by providing EVP_CIPHER and EVP_MD
7027 implementations of their own. This is detailed in crypto/engine/README
7028 as it couldn't be adequately described here. However, there are a few
7029 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7030 were changed in the original introduction of ENGINE code have now
7031 reverted back - the hooking from this code to ENGINE is now a good
7032 deal more passive and at run-time, operations deal directly with
7033 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7034 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7035 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7036 they were not being used by the framework as there is no concept of a
7037 BIGNUM_METHOD and they could not be generalised to the new
7038 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7039 ENGINE_cpy() has been removed as it cannot be consistently defined in
7043 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7046 *) Change mkdef.pl to sort symbols that get the same entry number,
7047 and make sure the automatically generated functions ERR_load_*
7048 become part of libeay.num as well.
7051 *) New function SSL_renegotiate_pending(). This returns true once
7052 renegotiation has been requested (either SSL_renegotiate() call
7053 or HelloRequest/ClientHello received from the peer) and becomes
7054 false once a handshake has been completed.
7055 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7056 sends a HelloRequest, but does not ensure that a handshake takes
7057 place. SSL_renegotiate_pending() is useful for checking if the
7058 client has followed the request.)
7061 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7062 By default, clients may request session resumption even during
7063 renegotiation (if session ID contexts permit); with this option,
7064 session resumption is possible only in the first handshake.
7066 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7067 more bits available for options that should not be part of
7068 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7071 *) Add some demos for certificate and certificate request creation.
7074 *) Make maximum certificate chain size accepted from the peer application
7075 settable (SSL*_get/set_max_cert_list()), as proposed by
7076 "Douglas E. Engert" <deengert@anl.gov>.
7079 *) Add support for shared libraries for Unixware-7
7080 (Boyd Lynn Gerber <gerberb@zenez.com>).
7083 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7084 be done prior to destruction. Use this to unload error strings from
7085 ENGINEs that load their own error strings. NB: This adds two new API
7086 functions to "get" and "set" this destroy handler in an ENGINE.
7089 *) Alter all existing ENGINE implementations (except "openssl" and
7090 "openbsd") to dynamically instantiate their own error strings. This
7091 makes them more flexible to be built both as statically-linked ENGINEs
7092 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7093 Also, add stub code to each that makes building them as self-contained
7094 shared-libraries easier (see README.ENGINE).
7097 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7098 implementations into applications that are completely implemented in
7099 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7100 commands that can be used to configure what shared-library to load and
7101 to control aspects of the way it is handled. Also, made an update to
7102 the README.ENGINE file that brings its information up-to-date and
7103 provides some information and instructions on the "dynamic" ENGINE
7104 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7107 *) Make it possible to unload ranges of ERR strings with a new
7108 "ERR_unload_strings" function.
7111 *) Add a copy() function to EVP_MD.
7114 *) Make EVP_MD routines take a context pointer instead of just the
7115 md_data void pointer.
7118 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7119 that the digest can only process a single chunk of data
7120 (typically because it is provided by a piece of
7121 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7122 is only going to provide a single chunk of data, and hence the
7123 framework needn't accumulate the data for oneshot drivers.
7126 *) As with "ERR", make it possible to replace the underlying "ex_data"
7127 functions. This change also alters the storage and management of global
7128 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7129 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7130 index counters. The API functions that use this state have been changed
7131 to take a "class_index" rather than pointers to the class's local STACK
7132 and counter, and there is now an API function to dynamically create new
7133 classes. This centralisation allows us to (a) plug a lot of the
7134 thread-safety problems that existed, and (b) makes it possible to clean
7135 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7136 such data would previously have always leaked in application code and
7137 workarounds were in place to make the memory debugging turn a blind eye
7138 to it. Application code that doesn't use this new function will still
7139 leak as before, but their memory debugging output will announce it now
7140 rather than letting it slide.
7142 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7143 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7144 has a return value to indicate success or failure.
7147 *) Make it possible to replace the underlying "ERR" functions such that the
7148 global state (2 LHASH tables and 2 locks) is only used by the "default"
7149 implementation. This change also adds two functions to "get" and "set"
7150 the implementation prior to it being automatically set the first time
7151 any other ERR function takes place. Ie. an application can call "get",
7152 pass the return value to a module it has just loaded, and that module
7153 can call its own "set" function using that value. This means the
7154 module's "ERR" operations will use (and modify) the error state in the
7155 application and not in its own statically linked copy of OpenSSL code.
7158 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7159 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7160 the operation, and provides a more encapsulated way for external code
7161 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7162 to use these functions rather than manually incrementing the counts.
7164 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7167 *) Add EVP test program.
7170 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7173 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7174 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7175 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7176 These allow a CRL to be built without having to access X509_CRL fields
7177 directly. Modify 'ca' application to use new functions.
7180 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7181 bug workarounds. Rollback attack detection is a security feature.
7182 The problem will only arise on OpenSSL servers when TLSv1 is not
7183 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7184 Software authors not wanting to support TLSv1 will have special reasons
7185 for their choice and can explicitly enable this option.
7186 [Bodo Moeller, Lutz Jaenicke]
7188 *) Rationalise EVP so it can be extended: don't include a union of
7189 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7190 (similar to those existing for EVP_CIPHER_CTX).
7195 EVP_MD_CTX_init(&md); /* new function call */
7196 EVP_DigestInit(&md, EVP_sha1());
7197 EVP_DigestUpdate(&md, in, len);
7198 EVP_DigestFinal(&md, out, NULL);
7199 EVP_MD_CTX_cleanup(&md); /* new function call */
7203 *) Make DES key schedule conform to the usual scheme, as well as
7204 correcting its structure. This means that calls to DES functions
7205 now have to pass a pointer to a des_key_schedule instead of a
7206 plain des_key_schedule (which was actually always a pointer
7209 des_key_schedule ks;
7211 des_set_key_checked(..., &ks);
7212 des_ncbc_encrypt(..., &ks, ...);
7214 (Note that a later change renames 'des_...' into 'DES_...'.)
7217 *) Initial reduction of linker bloat: the use of some functions, such as
7218 PEM causes large amounts of unused functions to be linked in due to
7219 poor organisation. For example pem_all.c contains every PEM function
7220 which has a knock on effect of linking in large amounts of (unused)
7221 ASN1 code. Grouping together similar functions and splitting unrelated
7222 functions prevents this.
7225 *) Cleanup of EVP macros.
7228 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7229 correct _ecb suffix.
7232 *) Add initial OCSP responder support to ocsp application. The
7233 revocation information is handled using the text based index
7234 use by the ca application. The responder can either handle
7235 requests generated internally, supplied in files (for example
7236 via a CGI script) or using an internal minimal server.
7239 *) Add configuration choices to get zlib compression for TLS.
7242 *) Changes to Kerberos SSL for RFC 2712 compliance:
7243 1. Implemented real KerberosWrapper, instead of just using
7244 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7245 2. Implemented optional authenticator field of KerberosWrapper.
7247 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7248 and authenticator structs; see crypto/krb5/.
7250 Generalized Kerberos calls to support multiple Kerberos libraries.
7251 [Vern Staats <staatsvr@asc.hpc.mil>,
7252 Jeffrey Altman <jaltman@columbia.edu>
7253 via Richard Levitte]
7255 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7256 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7257 values for each of the key sizes rather than having just
7258 parameters (and 'speed' generating keys each time).
7261 *) Speed up EVP routines.
7264 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7265 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7266 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7267 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7269 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7270 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7271 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7274 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7276 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7279 *) Added the OS2-EMX target.
7280 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7282 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7283 to support NCONF routines in extension code. New function CONF_set_nconf()
7284 to allow functions which take an NCONF to also handle the old LHASH
7285 structure: this means that the old CONF compatible routines can be
7286 retained (in particular wrt extensions) without having to duplicate the
7287 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7290 *) Enhance the general user interface with mechanisms for inner control
7291 and with possibilities to have yes/no kind of prompts.
7294 *) Change all calls to low level digest routines in the library and
7295 applications to use EVP. Add missing calls to HMAC_cleanup() and
7296 don't assume HMAC_CTX can be copied using memcpy().
7297 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7299 *) Add the possibility to control engines through control names but with
7300 arbitrary arguments instead of just a string.
7301 Change the key loaders to take a UI_METHOD instead of a callback
7302 function pointer. NOTE: this breaks binary compatibility with earlier
7303 versions of OpenSSL [engine].
7304 Adapt the nCipher code for these new conditions and add a card insertion
7308 *) Enhance the general user interface with mechanisms to better support
7309 dialog box interfaces, application-defined prompts, the possibility
7310 to use defaults (for example default passwords from somewhere else)
7311 and interrupts/cancellations.
7314 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7315 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7318 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7319 tidy up some unnecessarily weird code in 'sk_new()').
7320 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7322 *) Change the key loading routines for ENGINEs to use the same kind
7323 callback (pem_password_cb) as all other routines that need this
7327 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7328 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7329 than this minimum value is recommended.
7332 *) New random seeder for OpenVMS, using the system process statistics
7333 that are easily reachable.
7336 *) Windows apparently can't transparently handle global
7337 variables defined in DLLs. Initialisations such as:
7339 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7341 won't compile. This is used by the any applications that need to
7342 declare their own ASN1 modules. This was fixed by adding the option
7343 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7344 needed for static libraries under Win32.
7347 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7348 setting of purpose and trust fields. New X509_STORE trust and
7349 purpose functions and tidy up setting in other SSL functions.
7352 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7353 structure. These are inherited by X509_STORE_CTX when it is
7354 initialised. This allows various defaults to be set in the
7355 X509_STORE structure (such as flags for CRL checking and custom
7356 purpose or trust settings) for functions which only use X509_STORE_CTX
7357 internally such as S/MIME.
7359 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7360 trust settings if they are not set in X509_STORE. This allows X509_STORE
7361 purposes and trust (in S/MIME for example) to override any set by default.
7363 Add command line options for CRL checking to smime, s_client and s_server
7367 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7368 are set then the CRL is looked up in the X509_STORE structure and
7369 its validity and signature checked, then if the certificate is found
7370 in the CRL the verify fails with a revoked error.
7372 Various new CRL related callbacks added to X509_STORE_CTX structure.
7374 Command line options added to 'verify' application to support this.
7376 This needs some additional work, such as being able to handle multiple
7377 CRLs with different times, extension based lookup (rather than just
7378 by subject name) and ultimately more complete V2 CRL extension
7382 *) Add a general user interface API (crypto/ui/). This is designed
7383 to replace things like des_read_password and friends (backward
7384 compatibility functions using this new API are provided).
7385 The purpose is to remove prompting functions from the DES code
7386 section as well as provide for prompting through dialog boxes in
7387 a window system and the like.
7390 *) Add "ex_data" support to ENGINE so implementations can add state at a
7391 per-structure level rather than having to store it globally.
7394 *) Make it possible for ENGINE structures to be copied when retrieved by
7395 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7396 This causes the "original" ENGINE structure to act like a template,
7397 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7398 operational state can be localised to each ENGINE structure, despite the
7399 fact they all share the same "methods". New ENGINE structures returned in
7400 this case have no functional references and the return value is the single
7401 structural reference. This matches the single structural reference returned
7402 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7406 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7407 needs to match any other type at all we need to manually clear the
7411 *) Changes to the "openssl engine" utility to include;
7412 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7413 about an ENGINE's available control commands.
7414 - executing control commands from command line arguments using the
7415 '-pre' and '-post' switches. '-post' is only used if '-t' is
7416 specified and the ENGINE is successfully initialised. The syntax for
7417 the individual commands are colon-separated, for example;
7418 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7421 *) New dynamic control command support for ENGINEs. ENGINEs can now
7422 declare their own commands (numbers), names (strings), descriptions,
7423 and input types for run-time discovery by calling applications. A
7424 subset of these commands are implicitly classed as "executable"
7425 depending on their input type, and only these can be invoked through
7426 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7427 can be based on user input, config files, etc). The distinction is
7428 that "executable" commands cannot return anything other than a boolean
7429 result and can only support numeric or string input, whereas some
7430 discoverable commands may only be for direct use through
7431 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7432 pointers, or other custom uses. The "executable" commands are to
7433 support parameterisations of ENGINE behaviour that can be
7434 unambiguously defined by ENGINEs and used consistently across any
7435 OpenSSL-based application. Commands have been added to all the
7436 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7437 control over shared-library paths without source code alterations.
7440 *) Changed all ENGINE implementations to dynamically allocate their
7441 ENGINEs rather than declaring them statically. Apart from this being
7442 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7443 this also allows the implementations to compile without using the
7444 internal engine_int.h header.
7447 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7448 'const' value. Any code that should be able to modify a RAND_METHOD
7449 should already have non-const pointers to it (ie. they should only
7450 modify their own ones).
7453 *) Made a variety of little tweaks to the ENGINE code.
7454 - "atalla" and "ubsec" string definitions were moved from header files
7455 to C code. "nuron" string definitions were placed in variables
7456 rather than hard-coded - allowing parameterisation of these values
7457 later on via ctrl() commands.
7458 - Removed unused "#if 0"'d code.
7459 - Fixed engine list iteration code so it uses ENGINE_free() to release
7460 structural references.
7461 - Constified the RAND_METHOD element of ENGINE structures.
7462 - Constified various get/set functions as appropriate and added
7463 missing functions (including a catch-all ENGINE_cpy that duplicates
7464 all ENGINE values onto a new ENGINE except reference counts/state).
7465 - Removed NULL parameter checks in get/set functions. Setting a method
7466 or function to NULL is a way of cancelling out a previously set
7467 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7468 and doesn't justify the extra error symbols and code.
7469 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7470 flags from engine_int.h to engine.h.
7471 - Changed prototypes for ENGINE handler functions (init(), finish(),
7472 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7475 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7476 to the algorithm using long division. The binary algorithm can be
7477 used only if the modulus is odd. On 32-bit systems, it is faster
7478 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7479 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7480 up to 450 bits. In 64-bit environments, the binary algorithm
7481 appears to be advantageous for much longer moduli; here we use it
7482 for moduli up to 2048 bits.
7485 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7486 could not support the combine flag in choice fields.
7489 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7490 extensions from a certificate request to the certificate.
7493 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7494 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7495 file: this allows the display of the certificate about to be
7496 signed to be customised, to allow certain fields to be included
7497 or excluded and extension details. The old system didn't display
7498 multicharacter strings properly, omitted fields not in the policy
7499 and couldn't display additional details such as extensions.
7502 *) Function EC_POINTs_mul for multiple scalar multiplication
7503 of an arbitrary number of elliptic curve points
7504 \sum scalars[i]*points[i],
7505 optionally including the generator defined for the EC_GROUP:
7506 scalar*generator + \sum scalars[i]*points[i].
7508 EC_POINT_mul is a simple wrapper function for the typical case
7509 that the point list has just one item (besides the optional
7513 *) First EC_METHODs for curves over GF(p):
7515 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7516 operations and provides various method functions that can also
7517 operate with faster implementations of modular arithmetic.
7519 EC_GFp_mont_method() reuses most functions that are part of
7520 EC_GFp_simple_method, but uses Montgomery arithmetic.
7522 [Bodo Moeller; point addition and point doubling
7523 implementation directly derived from source code provided by
7524 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7526 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7527 crypto/ec/ec_lib.c):
7529 Curves are EC_GROUP objects (with an optional group generator)
7530 based on EC_METHODs that are built into the library.
7532 Points are EC_POINT objects based on EC_GROUP objects.
7534 Most of the framework would be able to handle curves over arbitrary
7535 finite fields, but as there are no obvious types for fields other
7536 than GF(p), some functions are limited to that for now.
7539 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7540 that the file contains a complete HTTP response.
7543 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7544 change the def and num file printf format specifier from "%-40sXXX"
7545 to "%-39s XXX". The latter will always guarantee a space after the
7546 field while the former will cause them to run together if the field
7547 is 40 of more characters long.
7550 *) Constify the cipher and digest 'method' functions and structures
7551 and modify related functions to take constant EVP_MD and EVP_CIPHER
7555 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7556 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7559 *) Modify EVP_Digest*() routines so they now return values. Although the
7560 internal software routines can never fail additional hardware versions
7564 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7566 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7567 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7570 ERR_R_NESTED_ASN1_ERROR
7572 ERR_R_MISSING_ASN1_EOS
7573 were 4 .. 9, conflicting with
7574 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7576 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7577 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7579 Add new error code 'ERR_R_INTERNAL_ERROR'.
7582 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7586 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7587 sets the subject name for a new request or supersedes the
7588 subject name in a given request. Formats that can be parsed are
7589 'CN=Some Name, OU=myOU, C=IT'
7591 'CN=Some Name/OU=myOU/C=IT'.
7593 Add options '-batch' and '-verbose' to 'openssl req'.
7594 [Massimiliano Pala <madwolf@hackmasters.net>]
7596 *) Introduce the possibility to access global variables through
7597 functions on platform were that's the best way to handle exporting
7598 global variables in shared libraries. To enable this functionality,
7599 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7600 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7601 is normally done by Configure or something similar).
7603 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7604 in the source file (foo.c) like this:
7606 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7607 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7609 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7610 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7612 OPENSSL_DECLARE_GLOBAL(int,foo);
7613 #define foo OPENSSL_GLOBAL_REF(foo)
7614 OPENSSL_DECLARE_GLOBAL(double,bar);
7615 #define bar OPENSSL_GLOBAL_REF(bar)
7617 The #defines are very important, and therefore so is including the
7618 header file everywhere where the defined globals are used.
7620 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7621 of ASN.1 items, but that structure is a bit different.
7623 The largest change is in util/mkdef.pl which has been enhanced with
7624 better and easier to understand logic to choose which symbols should
7625 go into the Windows .def files as well as a number of fixes and code
7626 cleanup (among others, algorithm keywords are now sorted
7627 lexicographically to avoid constant rewrites).
7630 *) In BN_div() keep a copy of the sign of 'num' before writing the
7631 result to 'rm' because if rm==num the value will be overwritten
7632 and produce the wrong result if 'num' is negative: this caused
7633 problems with BN_mod() and BN_nnmod().
7636 *) Function OCSP_request_verify(). This checks the signature on an
7637 OCSP request and verifies the signer certificate. The signer
7638 certificate is just checked for a generic purpose and OCSP request
7642 *) Add OCSP_check_validity() function to check the validity of OCSP
7643 responses. OCSP responses are prepared in real time and may only
7644 be a few seconds old. Simply checking that the current time lies
7645 between thisUpdate and nextUpdate max reject otherwise valid responses
7646 caused by either OCSP responder or client clock inaccuracy. Instead
7647 we allow thisUpdate and nextUpdate to fall within a certain period of
7648 the current time. The age of the response can also optionally be
7649 checked. Two new options -validity_period and -status_age added to
7653 *) If signature or public key algorithm is unrecognized print out its
7654 OID rather that just UNKNOWN.
7657 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7658 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7659 ID to be generated from the issuer certificate alone which can then be
7660 passed to OCSP_id_issuer_cmp().
7663 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7664 ASN1 modules to export functions returning ASN1_ITEM pointers
7665 instead of the ASN1_ITEM structures themselves. This adds several
7666 new macros which allow the underlying ASN1 function/structure to
7667 be accessed transparently. As a result code should not use ASN1_ITEM
7668 references directly (such as &X509_it) but instead use the relevant
7669 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7670 use of the new ASN1 code on platforms where exporting structures
7671 is problematical (for example in shared libraries) but exporting
7672 functions returning pointers to structures is not.
7675 *) Add support for overriding the generation of SSL/TLS session IDs.
7676 These callbacks can be registered either in an SSL_CTX or per SSL.
7677 The purpose of this is to allow applications to control, if they wish,
7678 the arbitrary values chosen for use as session IDs, particularly as it
7679 can be useful for session caching in multiple-server environments. A
7680 command-line switch for testing this (and any client code that wishes
7681 to use such a feature) has been added to "s_server".
7682 [Geoff Thorpe, Lutz Jaenicke]
7684 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7685 of the form '#if defined(...) || defined(...) || ...' and
7686 '#if !defined(...) && !defined(...) && ...'. This also avoids
7687 the growing number of special cases it was previously handling.
7690 *) Make all configuration macros available for application by making
7691 sure they are available in opensslconf.h, by giving them names starting
7692 with "OPENSSL_" to avoid conflicts with other packages and by making
7693 sure e_os2.h will cover all platform-specific cases together with
7695 Additionally, it is now possible to define configuration/platform-
7696 specific names (called "system identities"). In the C code, these
7697 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7698 macro with the name beginning with "OPENSSL_SYS_", which is determined
7699 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7703 *) New option -set_serial to 'req' and 'x509' this allows the serial
7704 number to use to be specified on the command line. Previously self
7705 signed certificates were hard coded with serial number 0 and the
7706 CA options of 'x509' had to use a serial number in a file which was
7710 *) New options to 'ca' utility to support V2 CRL entry extensions.
7711 Currently CRL reason, invalidity date and hold instruction are
7712 supported. Add new CRL extensions to V3 code and some new objects.
7715 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7716 disable standard block padding (aka PKCS#5 padding) in the EVP
7717 API, which was previously mandatory. This means that the data is
7718 not padded in any way and so the total length much be a multiple
7719 of the block size, otherwise an error occurs.
7722 *) Initial (incomplete) OCSP SSL support.
7725 *) New function OCSP_parse_url(). This splits up a URL into its host,
7726 port and path components: primarily to parse OCSP URLs. New -url
7727 option to ocsp utility.
7730 *) New nonce behavior. The return value of OCSP_check_nonce() now
7731 reflects the various checks performed. Applications can decide
7732 whether to tolerate certain situations such as an absent nonce
7733 in a response when one was present in a request: the ocsp application
7734 just prints out a warning. New function OCSP_add1_basic_nonce()
7735 this is to allow responders to include a nonce in a response even if
7736 the request is nonce-less.
7739 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7740 skipped when using openssl x509 multiple times on a single input file,
7741 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7744 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7745 set string type: to handle setting ASN1_TIME structures. Fix ca
7746 utility to correctly initialize revocation date of CRLs.
7749 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7750 the clients preferred ciphersuites and rather use its own preferences.
7751 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7752 Internet Explorer by ensuring unchanged hash method during stepup.
7753 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7756 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7757 to aes and add a new 'exist' option to print out symbols that don't
7761 *) Additional options to ocsp utility to allow flags to be set and
7762 additional certificates supplied.
7765 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7766 OCSP client a number of certificate to only verify the response
7770 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7771 handle the new API. Currently only ECB, CBC modes supported. Add new
7774 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7775 Encryption Standard (AES) Ciphersuites for Transport Layer
7776 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7777 not enabled by default and were not part of the "ALL" ciphersuite
7778 alias because they were not yet official; they could be
7779 explicitly requested by specifying the "AESdraft" ciphersuite
7780 group alias. In the final release of OpenSSL 0.9.7, the group
7781 alias is called "AES" and is part of "ALL".)
7782 [Ben Laurie, Steve Henson, Bodo Moeller]
7784 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7785 request to response.
7788 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7789 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7790 extract information from a certificate request. OCSP_response_create()
7791 creates a response and optionally adds a basic response structure.
7792 OCSP_basic_add1_status() adds a complete single response to a basic
7793 response and returns the OCSP_SINGLERESP structure just added (to allow
7794 extensions to be included for example). OCSP_basic_add1_cert() adds a
7795 certificate to a basic response and OCSP_basic_sign() signs a basic
7796 response with various flags. New helper functions ASN1_TIME_check()
7797 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7798 (converts ASN1_TIME to GeneralizedTime).
7801 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7802 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7803 structure from a certificate. X509_pubkey_digest() digests the public_key
7804 contents: this is used in various key identifiers.
7807 *) Make sk_sort() tolerate a NULL argument.
7808 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7810 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7811 passed by the function are trusted implicitly. If any of them signed the
7812 response then it is assumed to be valid and is not verified.
7815 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7816 to data. This was previously part of the PKCS7 ASN1 code. This
7817 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7818 [Steve Henson, reported by Kenneth R. Robinette
7819 <support@securenetterm.com>]
7821 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7822 routines: without these tracing memory leaks is very painful.
7823 Fix leaks in PKCS12 and PKCS7 routines.
7826 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7827 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7828 effectively meant GeneralizedTime would never be used. Now it
7829 is initialised to -1 but X509_time_adj() now has to check the value
7830 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7831 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7832 [Steve Henson, reported by Kenneth R. Robinette
7833 <support@securenetterm.com>]
7835 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7836 result in a zero length in the ASN1_INTEGER structure which was
7837 not consistent with the structure when d2i_ASN1_INTEGER() was used
7838 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7839 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7840 where it did not print out a minus for negative ASN1_INTEGER.
7843 *) Add summary printout to ocsp utility. The various functions which
7844 convert status values to strings have been renamed to:
7845 OCSP_response_status_str(), OCSP_cert_status_str() and
7846 OCSP_crl_reason_str() and are no longer static. New options
7847 to verify nonce values and to disable verification. OCSP response
7848 printout format cleaned up.
7851 *) Add additional OCSP certificate checks. These are those specified
7852 in RFC2560. This consists of two separate checks: the CA of the
7853 certificate being checked must either be the OCSP signer certificate
7854 or the issuer of the OCSP signer certificate. In the latter case the
7855 OCSP signer certificate must contain the OCSP signing extended key
7856 usage. This check is performed by attempting to match the OCSP
7857 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7858 in the OCSP_CERTID structures of the response.
7861 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7862 and related routines. This uses the standard OpenSSL certificate
7863 verify routines to perform initial checks (just CA validity) and
7864 to obtain the certificate chain. Then additional checks will be
7865 performed on the chain. Currently the root CA is checked to see
7866 if it is explicitly trusted for OCSP signing. This is used to set
7867 a root CA as a global signing root: that is any certificate that
7868 chains to that CA is an acceptable OCSP signing certificate.
7871 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7872 extensions from a separate configuration file.
7873 As when reading extensions from the main configuration file,
7874 the '-extensions ...' option may be used for specifying the
7876 [Massimiliano Pala <madwolf@comune.modena.it>]
7878 *) New OCSP utility. Allows OCSP requests to be generated or
7879 read. The request can be sent to a responder and the output
7880 parsed, outputed or printed in text form. Not complete yet:
7881 still needs to check the OCSP response validity.
7884 *) New subcommands for 'openssl ca':
7885 'openssl ca -status <serial>' prints the status of the cert with
7886 the given serial number (according to the index file).
7887 'openssl ca -updatedb' updates the expiry status of certificates
7889 [Massimiliano Pala <madwolf@comune.modena.it>]
7891 *) New '-newreq-nodes' command option to CA.pl. This is like
7892 '-newreq', but calls 'openssl req' with the '-nodes' option
7893 so that the resulting key is not encrypted.
7894 [Damien Miller <djm@mindrot.org>]
7896 *) New configuration for the GNU Hurd.
7897 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7899 *) Initial code to implement OCSP basic response verify. This
7900 is currently incomplete. Currently just finds the signer's
7901 certificate and verifies the signature on the response.
7904 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7905 value of OPENSSLDIR. This is available via the new '-d' option
7906 to 'openssl version', and is also included in 'openssl version -a'.
7909 *) Allowing defining memory allocation callbacks that will be given
7910 file name and line number information in additional arguments
7911 (a const char* and an int). The basic functionality remains, as
7912 well as the original possibility to just replace malloc(),
7913 realloc() and free() by functions that do not know about these
7914 additional arguments. To register and find out the current
7915 settings for extended allocation functions, the following
7916 functions are provided:
7918 CRYPTO_set_mem_ex_functions
7919 CRYPTO_set_locked_mem_ex_functions
7920 CRYPTO_get_mem_ex_functions
7921 CRYPTO_get_locked_mem_ex_functions
7923 These work the same way as CRYPTO_set_mem_functions and friends.
7924 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7925 extended allocation function is enabled.
7926 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7927 a conventional allocation function is enabled.
7928 [Richard Levitte, Bodo Moeller]
7930 *) Finish off removing the remaining LHASH function pointer casts.
7931 There should no longer be any prototype-casting required when using
7932 the LHASH abstraction, and any casts that remain are "bugs". See
7933 the callback types and macros at the head of lhash.h for details
7934 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7937 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7938 If /dev/[u]random devices are not available or do not return enough
7939 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7941 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7942 /etc/entropy will be queried once each in this sequence, querying stops
7943 when enough entropy was collected without querying more sockets.
7946 *) Change the Unix RAND_poll() variant to be able to poll several
7947 random devices, as specified by DEVRANDOM, until a sufficient amount
7948 of data has been collected. We spend at most 10 ms on each file
7949 (select timeout) and read in non-blocking mode. DEVRANDOM now
7950 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7951 (previously it was just the string "/dev/urandom"), so on typical
7952 platforms the 10 ms delay will never occur.
7953 Also separate out the Unix variant to its own file, rand_unix.c.
7954 For VMS, there's a currently-empty rand_vms.c.
7957 *) Move OCSP client related routines to ocsp_cl.c. These
7958 provide utility functions which an application needing
7959 to issue a request to an OCSP responder and analyse the
7960 response will typically need: as opposed to those which an
7961 OCSP responder itself would need which will be added later.
7963 OCSP_request_sign() signs an OCSP request with an API similar
7964 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7965 response. OCSP_response_get1_basic() extracts basic response
7966 from response. OCSP_resp_find_status(): finds and extracts status
7967 information from an OCSP_CERTID structure (which will be created
7968 when the request structure is built). These are built from lower
7969 level functions which work on OCSP_SINGLERESP structures but
7970 won't normally be used unless the application wishes to examine
7971 extensions in the OCSP response for example.
7973 Replace nonce routines with a pair of functions.
7974 OCSP_request_add1_nonce() adds a nonce value and optionally
7975 generates a random value. OCSP_check_nonce() checks the
7976 validity of the nonce in an OCSP response.
7979 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7980 This doesn't copy the supplied OCSP_CERTID and avoids the
7981 need to free up the newly created id. Change return type
7982 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7983 This can then be used to add extensions to the request.
7984 Deleted OCSP_request_new(), since most of its functionality
7985 is now in OCSP_REQUEST_new() (and the case insensitive name
7986 clash) apart from the ability to set the request name which
7987 will be added elsewhere.
7990 *) Update OCSP API. Remove obsolete extensions argument from
7991 various functions. Extensions are now handled using the new
7992 OCSP extension code. New simple OCSP HTTP function which
7993 can be used to send requests and parse the response.
7996 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7997 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7998 uses the special reorder version of SET OF to sort the attributes
7999 and reorder them to match the encoded order. This resolves a long
8000 standing problem: a verify on a PKCS7 structure just after signing
8001 it used to fail because the attribute order did not match the
8002 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8003 it uses the received order. This is necessary to tolerate some broken
8004 software that does not order SET OF. This is handled by encoding
8005 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8006 to produce the required SET OF.
8009 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8010 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8011 files to get correct declarations of the ASN.1 item variables.
8014 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8015 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8016 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8017 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8018 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8019 ASN1_ITEM and no wrapper functions.
8022 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8023 replace the old function pointer based I/O routines. Change most of
8024 the *_d2i_bio() and *_d2i_fp() functions to use these.
8027 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8028 lines, recognize more "algorithms" that can be deselected, and make
8029 it complain about algorithm deselection that isn't recognised.
8032 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8033 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8034 to use new functions. Add NO_ASN1_OLD which can be set to remove
8035 some old style ASN1 functions: this can be used to determine if old
8036 code will still work when these eventually go away.
8039 *) New extension functions for OCSP structures, these follow the
8040 same conventions as certificates and CRLs.
8043 *) New function X509V3_add1_i2d(). This automatically encodes and
8044 adds an extension. Its behaviour can be customised with various
8045 flags to append, replace or delete. Various wrappers added for
8046 certificates and CRLs.
8049 *) Fix to avoid calling the underlying ASN1 print routine when
8050 an extension cannot be parsed. Correct a typo in the
8051 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8054 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8055 entries for variables.
8058 *) Add functionality to apps/openssl.c for detecting locking
8059 problems: As the program is single-threaded, all we have
8060 to do is register a locking callback using an array for
8061 storing which locks are currently held by the program.
8064 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8065 SSL_get_ex_data_X509_STORE_idx(), which is used in
8066 ssl_verify_cert_chain() and thus can be called at any time
8067 during TLS/SSL handshakes so that thread-safety is essential.
8068 Unfortunately, the ex_data design is not at all suited
8069 for multi-threaded use, so it probably should be abolished.
8072 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8073 [Broadcom, tweaked and integrated by Geoff Thorpe]
8075 *) Move common extension printing code to new function
8076 X509V3_print_extensions(). Reorganise OCSP print routines and
8077 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8080 *) New function X509_signature_print() to remove duplication in some
8084 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8085 set (this was treated exactly the same as SET OF previously). This
8086 is used to reorder the STACK representing the structure to match the
8087 encoding. This will be used to get round a problem where a PKCS7
8088 structure which was signed could not be verified because the STACK
8089 order did not reflect the encoded order.
8092 *) Reimplement the OCSP ASN1 module using the new code.
8095 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8096 for its ASN1 operations. The old style function pointers still exist
8097 for now but they will eventually go away.
8100 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8101 completely replaces the old ASN1 functionality with a table driven
8102 encoder and decoder which interprets an ASN1_ITEM structure describing
8103 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8104 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8105 has also been converted to the new form.
8108 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8109 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8110 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8111 for negative moduli.
8114 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8115 of not touching the result's sign bit.
8118 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8122 *) Changed the LHASH code to use prototypes for callbacks, and created
8123 macros to declare and implement thin (optionally static) functions
8124 that provide type-safety and avoid function pointer casting for the
8125 type-specific callbacks.
8128 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8130 [Veers Staats <staatsvr@asc.hpc.mil>,
8131 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8133 *) Reformat the FAQ so the different questions and answers can be divided
8134 in sections depending on the subject.
8137 *) Have the zlib compression code load ZLIB.DLL dynamically under
8141 *) New function BN_mod_sqrt for computing square roots modulo a prime
8142 (using the probabilistic Tonelli-Shanks algorithm unless
8143 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8144 be handled deterministically).
8145 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8147 *) Make BN_mod_inverse faster by explicitly handling small quotients
8148 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8149 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8152 *) New function BN_kronecker.
8155 *) Fix BN_gcd so that it works on negative inputs; the result is
8156 positive unless both parameters are zero.
8157 Previously something reasonably close to an infinite loop was
8158 possible because numbers could be growing instead of shrinking
8159 in the implementation of Euclid's algorithm.
8162 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8163 sign of the number in question.
8165 Fix BN_is_word(a,w) to work correctly for w == 0.
8167 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8168 because its test if the absolute value of 'a' equals 'w'.
8169 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8170 it exists mostly for use in the implementations of BN_is_zero(),
8171 BN_is_one(), and BN_is_word().
8174 *) New function BN_swap.
8177 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8178 the exponentiation functions are more likely to produce reasonable
8179 results on negative inputs.
8182 *) Change BN_mod_mul so that the result is always non-negative.
8183 Previously, it could be negative if one of the factors was negative;
8184 I don't think anyone really wanted that behaviour.
8187 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8188 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8189 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8190 and add new functions:
8199 BN_mod_lshift1_quick
8203 These functions always generate non-negative results.
8205 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8206 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8208 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8209 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8210 be reduced modulo m.
8211 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8214 The following entry accidentally appeared in the CHANGES file
8215 distributed with OpenSSL 0.9.7. The modifications described in
8216 it do *not* apply to OpenSSL 0.9.7.
8218 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8219 was actually never needed) and in BN_mul(). The removal in BN_mul()
8220 required a small change in bn_mul_part_recursive() and the addition
8221 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8222 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8223 bn_sub_words() and bn_add_words() except they take arrays with
8228 *) In 'openssl passwd', verify passwords read from the terminal
8229 unless the '-salt' option is used (which usually means that
8230 verification would just waste user's time since the resulting
8231 hash is going to be compared with some given password hash)
8232 or the new '-noverify' option is used.
8234 This is an incompatible change, but it does not affect
8235 non-interactive use of 'openssl passwd' (passwords on the command
8236 line, '-stdin' option, '-in ...' option) and thus should not
8240 *) Remove all references to RSAref, since there's no more need for it.
8243 *) Make DSO load along a path given through an environment variable
8244 (SHLIB_PATH) with shl_load().
8247 *) Constify the ENGINE code as a result of BIGNUM constification.
8248 Also constify the RSA code and most things related to it. In a
8249 few places, most notable in the depth of the ASN.1 code, ugly
8250 casts back to non-const were required (to be solved at a later
8254 *) Make it so the openssl application has all engines loaded by default.
8257 *) Constify the BIGNUM routines a little more.
8260 *) Add the following functions:
8262 ENGINE_load_cswift()
8264 ENGINE_load_atalla()
8266 ENGINE_load_builtin_engines()
8268 That way, an application can itself choose if external engines that
8269 are built-in in OpenSSL shall ever be used or not. The benefit is
8270 that applications won't have to be linked with libdl or other dso
8271 libraries unless it's really needed.
8273 Changed 'openssl engine' to load all engines on demand.
8274 Changed the engine header files to avoid the duplication of some
8275 declarations (they differed!).
8278 *) 'openssl engine' can now list capabilities.
8281 *) Better error reporting in 'openssl engine'.
8284 *) Never call load_dh_param(NULL) in s_server.
8287 *) Add engine application. It can currently list engines by name and
8288 identity, and test if they are actually available.
8291 *) Improve RPM specification file by forcing symbolic linking and making
8292 sure the installed documentation is also owned by root.root.
8293 [Damien Miller <djm@mindrot.org>]
8295 *) Give the OpenSSL applications more possibilities to make use of
8296 keys (public as well as private) handled by engines.
8299 *) Add OCSP code that comes from CertCo.
8302 *) Add VMS support for the Rijndael code.
8305 *) Added untested support for Nuron crypto accelerator.
8308 *) Add support for external cryptographic devices. This code was
8309 previously distributed separately as the "engine" branch.
8310 [Geoff Thorpe, Richard Levitte]
8312 *) Rework the filename-translation in the DSO code. It is now possible to
8313 have far greater control over how a "name" is turned into a filename
8314 depending on the operating environment and any oddities about the
8315 different shared library filenames on each system.
8318 *) Support threads on FreeBSD-elf in Configure.
8321 *) Fix for SHA1 assembly problem with MASM: it produces
8322 warnings about corrupt line number information when assembling
8323 with debugging information. This is caused by the overlapping
8325 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8328 NCONF_get_number() has no error checking at all. As a replacement,
8329 NCONF_get_number_e() is defined (_e for "error checking") and is
8330 promoted strongly. The old NCONF_get_number is kept around for
8331 binary backward compatibility.
8332 Make it possible for methods to load from something other than a BIO,
8333 by providing a function pointer that is given a name instead of a BIO.
8334 For example, this could be used to load configuration data from an
8338 *) Fix for non blocking accept BIOs. Added new I/O special reason
8339 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8340 with non blocking I/O was not possible because no retry code was
8341 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8345 *) Added the beginnings of Rijndael support.
8348 *) Fix for bug in DirectoryString mask setting. Add support for
8349 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8350 to allow certificate printing to more controllable, additional
8351 'certopt' option to 'x509' to allow new printing options to be
8355 *) Clean old EAY MD5 hack from e_os.h.
8358 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8360 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8361 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8362 [Joe Orton, Steve Henson]
8364 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8366 *) Fix additional bug revealed by the NISCC test suite:
8368 Stop bug triggering large recursion when presented with
8369 certain ASN.1 tags (CVE-2003-0851)
8372 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8374 *) Fix various bugs revealed by running the NISCC test suite:
8376 Stop out of bounds reads in the ASN1 code when presented with
8377 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8379 If verify callback ignores invalid public key errors don't try to check
8380 certificate signature with the NULL public key.
8384 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8385 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8389 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8390 extra data after the compression methods not only for TLS 1.0
8391 but also for SSL 3.0 (as required by the specification).
8392 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8394 *) Change X509_certificate_type() to mark the key as exported/exportable
8395 when it's 512 *bits* long, not 512 bytes.
8398 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8400 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8401 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8402 a protocol version number mismatch like a decryption error
8403 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8406 *) Turn on RSA blinding by default in the default implementation
8407 to avoid a timing attack. Applications that don't want it can call
8408 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8409 They would be ill-advised to do so in most cases.
8410 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8412 *) Change RSA blinding code so that it works when the PRNG is not
8413 seeded (in this case, the secret RSA exponent is abused as
8414 an unpredictable seed -- if it is not unpredictable, there
8415 is no point in blinding anyway). Make RSA blinding thread-safe
8416 by remembering the creator's thread ID in rsa->blinding and
8417 having all other threads use local one-time blinding factors
8418 (this requires more computation than sharing rsa->blinding, but
8419 avoids excessive locking; and if an RSA object is not shared
8420 between threads, blinding will still be very fast).
8423 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8425 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8426 via timing by performing a MAC computation even if incorrect
8427 block cipher padding has been found. This is a countermeasure
8428 against active attacks where the attacker has to distinguish
8429 between bad padding and a MAC verification error. (CVE-2003-0078)
8431 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8432 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8433 Martin Vuagnoux (EPFL, Ilion)]
8435 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8437 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8438 memory from it's contents. This is done with a counter that will
8439 place alternating values in each byte. This can be used to solve
8440 two issues: 1) the removal of calls to memset() by highly optimizing
8441 compilers, and 2) cleansing with other values than 0, since those can
8442 be read through on certain media, for example a swap space on disk.
8445 *) Bugfix: client side session caching did not work with external caching,
8446 because the session->cipher setting was not restored when reloading
8447 from the external cache. This problem was masked, when
8448 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8449 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8452 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8453 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8454 [Zeev Lieber <zeev-l@yahoo.com>]
8456 *) Undo an undocumented change introduced in 0.9.6e which caused
8457 repeated calls to OpenSSL_add_all_ciphers() and
8458 OpenSSL_add_all_digests() to be ignored, even after calling
8462 *) Change the default configuration reader to deal with last line not
8463 being properly terminated.
8466 *) Change X509_NAME_cmp() so it applies the special rules on handling
8467 DN values that are of type PrintableString, as well as RDNs of type
8468 emailAddress where the value has the type ia5String.
8469 [stefank@valicert.com via Richard Levitte]
8471 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8472 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8473 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8474 the bitwise-OR of the two for use by the majority of applications
8475 wanting this behaviour, and update the docs. The documented
8476 behaviour and actual behaviour were inconsistent and had been
8477 changing anyway, so this is more a bug-fix than a behavioural
8479 [Geoff Thorpe, diagnosed by Nadav Har'El]
8481 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8482 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8485 *) Fix initialization code race conditions in
8486 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8487 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8488 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8489 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8490 ssl2_get_cipher_by_char(),
8491 ssl3_get_cipher_by_char().
8492 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8494 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8495 the cached sessions are flushed, as the remove_cb() might use ex_data
8496 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8497 (see [openssl.org #212]).
8498 [Geoff Thorpe, Lutz Jaenicke]
8500 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8501 length, instead of the encoding length to d2i_ASN1_OBJECT.
8504 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8506 *) [In 0.9.6g-engine release:]
8507 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8508 [Lynn Gazis <lgazis@rainbow.com>]
8510 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8512 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8513 and get fix the header length calculation.
8514 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8515 Alon Kantor <alonk@checkpoint.com> (and others),
8518 *) Use proper error handling instead of 'assertions' in buffer
8519 overflow checks added in 0.9.6e. This prevents DoS (the
8520 assertions could call abort()).
8521 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8523 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8525 *) Add various sanity checks to asn1_get_length() to reject
8526 the ASN1 length bytes if they exceed sizeof(long), will appear
8527 negative or the content length exceeds the length of the
8529 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8531 *) Fix cipher selection routines: ciphers without encryption had no flags
8532 for the cipher strength set and where therefore not handled correctly
8533 by the selection routines (PR #130).
8536 *) Fix EVP_dsa_sha macro.
8540 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8541 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8542 that was added in OpenSSL 0.9.6d.
8544 As the countermeasure turned out to be incompatible with some
8545 broken SSL implementations, the new option is part of SSL_OP_ALL.
8546 SSL_OP_ALL is usually employed when compatibility with weird SSL
8547 implementations is desired (e.g. '-bugs' option to 's_client' and
8548 's_server'), so the new option is automatically set in many
8552 *) Changes in security patch:
8554 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8555 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8556 Air Force Materiel Command, USAF, under agreement number
8559 *) Add various sanity checks to asn1_get_length() to reject
8560 the ASN1 length bytes if they exceed sizeof(long), will appear
8561 negative or the content length exceeds the length of the
8562 supplied buffer. (CVE-2002-0659)
8563 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8565 *) Assertions for various potential buffer overflows, not known to
8567 [Ben Laurie (CHATS)]
8569 *) Various temporary buffers to hold ASCII versions of integers were
8570 too small for 64 bit platforms. (CVE-2002-0655)
8571 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8573 *) Remote buffer overflow in SSL3 protocol - an attacker could
8574 supply an oversized session ID to a client. (CVE-2002-0656)
8575 [Ben Laurie (CHATS)]
8577 *) Remote buffer overflow in SSL2 protocol - an attacker could
8578 supply an oversized client master key. (CVE-2002-0656)
8579 [Ben Laurie (CHATS)]
8581 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8583 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8584 encoded as NULL) with id-dsa-with-sha1.
8585 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8587 *) Check various X509_...() return values in apps/req.c.
8588 [Nils Larsch <nla@trustcenter.de>]
8590 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8591 an end-of-file condition would erroneously be flagged, when the CRLF
8592 was just at the end of a processed block. The bug was discovered when
8593 processing data through a buffering memory BIO handing the data to a
8594 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8595 <ptsekov@syntrex.com> and Nedelcho Stanev.
8598 *) Implement a countermeasure against a vulnerability recently found
8599 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8600 before application data chunks to avoid the use of known IVs
8601 with data potentially chosen by the attacker.
8604 *) Fix length checks in ssl3_get_client_hello().
8607 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8608 to prevent ssl3_read_internal() from incorrectly assuming that
8609 ssl3_read_bytes() found application data while handshake
8610 processing was enabled when in fact s->s3->in_read_app_data was
8611 merely automatically cleared during the initial handshake.
8612 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8614 *) Fix object definitions for Private and Enterprise: they were not
8615 recognized in their shortname (=lowercase) representation. Extend
8616 obj_dat.pl to issue an error when using undefined keywords instead
8617 of silently ignoring the problem (Svenning Sorensen
8618 <sss@sss.dnsalias.net>).
8621 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8622 generators, i.e. generators other than 2 and 5. (Previously, the
8623 code did not properly initialise the 'add' and 'rem' values to
8624 BN_generate_prime().)
8626 In the new general case, we do not insist that 'generator' is
8627 actually a primitive root: This requirement is rather pointless;
8628 a generator of the order-q subgroup is just as good, if not
8632 *) Map new X509 verification errors to alerts. Discovered and submitted by
8633 Tom Wu <tom@arcot.com>.
8636 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8637 returning non-zero before the data has been completely received
8638 when using non-blocking I/O.
8639 [Bodo Moeller; problem pointed out by John Hughes]
8641 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8642 [Ben Laurie, Lutz Jaenicke]
8644 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8645 Yoram Zahavi <YoramZ@gilian.com>).
8648 *) Add information about CygWin 1.3 and on, and preserve proper
8649 configuration for the versions before that.
8650 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8652 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8653 check whether we deal with a copy of a session and do not delete from
8654 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8655 <izhar@checkpoint.com>.
8658 *) Do not store session data into the internal session cache, if it
8659 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8660 flag is set). Proposed by Aslam <aslam@funk.com>.
8663 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8667 *) [In 0.9.6d-engine release:]
8668 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8669 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8671 *) Add the configuration target linux-s390x.
8672 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8674 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8675 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8676 variable as an indication that a ClientHello message has been
8677 received. As the flag value will be lost between multiple
8678 invocations of ssl3_accept when using non-blocking I/O, the
8679 function may not be aware that a handshake has actually taken
8680 place, thus preventing a new session from being added to the
8683 To avoid this problem, we now set s->new_session to 2 instead of
8684 using a local variable.
8685 [Lutz Jaenicke, Bodo Moeller]
8687 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8688 if the SSL_R_LENGTH_MISMATCH error is detected.
8689 [Geoff Thorpe, Bodo Moeller]
8691 *) New 'shared_ldflag' column in Configure platform table.
8694 *) Fix EVP_CIPHER_mode macro.
8695 ["Dan S. Camper" <dan@bti.net>]
8697 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8698 type, we must throw them away by setting rr->length to 0.
8699 [D P Chang <dpc@qualys.com>]
8701 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8703 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8704 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8705 worked incorrectly for those cases where range = 10..._2 and
8706 3*range is two bits longer than range.)
8709 *) Only add signing time to PKCS7 structures if it is not already
8713 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8714 OBJ_ld_ce should be OBJ_id_ce.
8715 Also some ip-pda OIDs in crypto/objects/objects.txt were
8716 incorrect (cf. RFC 3039).
8717 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8719 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8720 returns early because it has nothing to do.
8721 [Andy Schneider <andy.schneider@bjss.co.uk>]
8723 *) [In 0.9.6c-engine release:]
8724 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8725 [Andy Schneider <andy.schneider@bjss.co.uk>]
8727 *) [In 0.9.6c-engine release:]
8728 Add support for Cryptographic Appliance's keyserver technology.
8729 (Use engine 'keyclient')
8730 [Cryptographic Appliances and Geoff Thorpe]
8732 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8733 is called via tools/c89.sh because arguments have to be
8734 rearranged (all '-L' options must appear before the first object
8736 [Richard Shapiro <rshapiro@abinitio.com>]
8738 *) [In 0.9.6c-engine release:]
8739 Add support for Broadcom crypto accelerator cards, backported
8741 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8743 *) [In 0.9.6c-engine release:]
8744 Add support for SureWare crypto accelerator cards from
8745 Baltimore Technologies. (Use engine 'sureware')
8746 [Baltimore Technologies and Mark Cox]
8748 *) [In 0.9.6c-engine release:]
8749 Add support for crypto accelerator cards from Accelerated
8750 Encryption Processing, www.aep.ie. (Use engine 'aep')
8751 [AEP Inc. and Mark Cox]
8753 *) Add a configuration entry for gcc on UnixWare.
8754 [Gary Benson <gbenson@redhat.com>]
8756 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8757 messages are stored in a single piece (fixed-length part and
8758 variable-length part combined) and fix various bugs found on the way.
8761 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8762 instead. BIO_gethostbyname() does not know what timeouts are
8763 appropriate, so entries would stay in cache even when they have
8765 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8767 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8768 faced with a pathologically small ClientHello fragment that does
8769 not contain client_version: Instead of aborting with an error,
8770 simply choose the highest available protocol version (i.e.,
8771 TLS 1.0 unless it is disabled). In practice, ClientHello
8772 messages are never sent like this, but this change gives us
8773 strictly correct behaviour at least for TLS.
8776 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8777 never resets s->method to s->ctx->method when called from within
8778 one of the SSL handshake functions.
8779 [Bodo Moeller; problem pointed out by Niko Baric]
8781 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8782 (sent using the client's version number) if client_version is
8783 smaller than the protocol version in use. Also change
8784 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8785 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8786 the client will at least see that alert.
8789 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8793 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8794 client receives HelloRequest while in a handshake.
8795 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8797 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8798 should end in 'break', not 'goto end' which circumvents various
8799 cleanups done in state SSL_ST_OK. But session related stuff
8800 must be disabled for SSL_ST_OK in the case that we just sent a
8803 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8804 before just sending a HelloRequest.
8805 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8807 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8808 reveal whether illegal block cipher padding was found or a MAC
8809 verification error occurred. (Neither SSLerr() codes nor alerts
8810 are directly visible to potential attackers, but the information
8811 may leak via logfiles.)
8813 Similar changes are not required for the SSL 2.0 implementation
8814 because the number of padding bytes is sent in clear for SSL 2.0,
8815 and the extra bytes are just ignored. However ssl/s2_pkt.c
8816 failed to verify that the purported number of padding bytes is in
8820 *) Add OpenUNIX-8 support including shared libraries
8821 (Boyd Lynn Gerber <gerberb@zenez.com>).
8824 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8825 'wristwatch attack' using huge encoding parameters (cf.
8826 James H. Manger's CRYPTO 2001 paper). Note that the
8827 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8828 encoding parameters and hence was not vulnerable.
8831 *) BN_sqr() bug fix.
8832 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8834 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8835 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8836 followed by modular reduction.
8837 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8839 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8840 equivalent based on BN_pseudo_rand() instead of BN_rand().
8843 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8844 This function was broken, as the check for a new client hello message
8845 to handle SGC did not allow these large messages.
8846 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8849 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8852 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8853 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8856 *) Rework the configuration and shared library support for Tru64 Unix.
8857 The configuration part makes use of modern compiler features and
8858 still retains old compiler behavior for those that run older versions
8859 of the OS. The shared library support part includes a variant that
8860 uses the RPATH feature, and is available through the special
8861 configuration target "alpha-cc-rpath", which will never be selected
8863 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8865 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8866 with the same message size as in ssl3_get_certificate_request().
8867 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8868 messages might inadvertently be reject as too long.
8869 [Petr Lampa <lampa@fee.vutbr.cz>]
8871 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8874 *) Modified SSL library such that the verify_callback that has been set
8875 specificly for an SSL object with SSL_set_verify() is actually being
8876 used. Before the change, a verify_callback set with this function was
8877 ignored and the verify_callback() set in the SSL_CTX at the time of
8878 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8879 to allow the necessary settings.
8882 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8883 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8884 done automatically (in contradiction to the requirements of the C
8885 standard). This made problems when used from OpenSSH.
8888 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8889 dh->length and always used
8891 BN_rand_range(priv_key, dh->p).
8893 BN_rand_range() is not necessary for Diffie-Hellman, and this
8894 specific range makes Diffie-Hellman unnecessarily inefficient if
8895 dh->length (recommended exponent length) is much smaller than the
8896 length of dh->p. We could use BN_rand_range() if the order of
8897 the subgroup was stored in the DH structure, but we only have
8902 BN_rand(priv_key, l, ...)
8904 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8910 RSA_eay_public_encrypt
8911 RSA_eay_private_decrypt
8912 RSA_eay_private_encrypt (signing)
8913 RSA_eay_public_decrypt (signature verification)
8915 (default implementations for RSA_public_encrypt,
8916 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8917 always reject numbers >= n.
8920 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8921 to synchronize access to 'locking_thread'. This is necessary on
8922 systems where access to 'locking_thread' (an 'unsigned long'
8923 variable) is not atomic.
8926 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8927 *before* setting the 'crypto_lock_rand' flag. The previous code had
8928 a race condition if 0 is a valid thread ID.
8929 [Travis Vitek <vitek@roguewave.com>]
8931 *) Add support for shared libraries under Irix.
8932 [Albert Chin-A-Young <china@thewrittenword.com>]
8934 *) Add configuration option to build on Linux on both big-endian and
8936 [Ralf Baechle <ralf@uni-koblenz.de>]
8938 *) Add the possibility to create shared libraries on HP-UX.
8941 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8943 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8944 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8945 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8946 PRNG state recovery was possible based on the output of
8947 one PRNG request appropriately sized to gain knowledge on
8948 'md' followed by enough consecutive 1-byte PRNG requests
8949 to traverse all of 'state'.
8951 1. When updating 'md_local' (the current thread's copy of 'md')
8952 during PRNG output generation, hash all of the previous
8953 'md_local' value, not just the half used for PRNG output.
8955 2. Make the number of bytes from 'state' included into the hash
8956 independent from the number of PRNG bytes requested.
8958 The first measure alone would be sufficient to avoid
8959 Markku-Juhani's attack. (Actually it had never occurred
8960 to me that the half of 'md_local' used for chaining was the
8961 half from which PRNG output bytes were taken -- I had always
8962 assumed that the secret half would be used.) The second
8963 measure makes sure that additional data from 'state' is never
8964 mixed into 'md_local' in small portions; this heuristically
8965 further strengthens the PRNG.
8968 *) Fix crypto/bn/asm/mips3.s.
8971 *) When only the key is given to "enc", the IV is undefined. Print out
8972 an error message in this case.
8975 *) Handle special case when X509_NAME is empty in X509 printing routines.
8978 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8979 positive and less than q.
8982 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8983 used: it isn't thread safe and the add_lock_callback should handle
8985 [Paul Rose <Paul.Rose@bridge.com>]
8987 *) Verify that incoming data obeys the block size in
8988 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8992 [Ulf Möller, Bodo Möller]
8994 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8995 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8996 when fixing the server behaviour for backwards-compatible 'client
8997 hello' messages. (Note that the attack is impractical against
8998 SSL 3.0 and TLS 1.0 anyway because length and version checking
8999 means that the probability of guessing a valid ciphertext is
9000 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9003 Before 0.9.5, the countermeasure (hide the error by generating a
9004 random 'decryption result') did not work properly because
9005 ERR_clear_error() was missing, meaning that SSL_get_error() would
9006 detect the supposedly ignored error.
9008 Both problems are now fixed.
9011 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9012 (previously it was 1024).
9015 *) Fix for compatibility mode trust settings: ignore trust settings
9016 unless some valid trust or reject settings are present.
9019 *) Fix for blowfish EVP: its a variable length cipher.
9022 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9023 parameters in DSA public key structures and return an error in the
9024 DSA routines if parameters are absent.
9027 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9028 in the current directory if neither $RANDFILE nor $HOME was set.
9029 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9030 caused some confusion to Windows users who haven't defined $HOME.
9031 Thus RAND_file_name() is changed again: e_os.h can define a
9032 DEFAULT_HOME, which will be used if $HOME is not set.
9033 For Windows, we use "C:"; on other platforms, we still require
9034 environment variables.
9036 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9037 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9038 having multiple threads call RAND_poll() concurrently.
9041 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9042 combination of a flag and a thread ID variable.
9043 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9044 flag), *other* threads can enter ssleay_add_bytes without obeying
9045 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9046 that they do not hold after the first thread unsets add_do_not_lock).
9049 *) Change bctest again: '-x' expressions are not available in all
9053 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9055 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9056 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9058 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9059 the default extension for executables, if any. Also, make the perl
9060 scripts that use symlink() to test if it really exists and use "cp"
9061 if it doesn't. All this made OpenSSL compilable and installable in
9065 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9066 If SEQUENCE is length is indefinite just set c->slen to the total
9067 amount of data available.
9068 [Steve Henson, reported by shige@FreeBSD.org]
9069 [This change does not apply to 0.9.7.]
9071 *) Change bctest to avoid here-documents inside command substitution
9072 (workaround for FreeBSD /bin/sh bug).
9073 For compatibility with Ultrix, avoid shell functions (introduced
9074 in the bctest version that searches along $PATH).
9077 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9078 with des_encrypt() defined on some operating systems, like Solaris
9082 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9083 On the Importance of Eliminating Errors in Cryptographic
9084 Computations, J. Cryptology 14 (2001) 2, 101-119,
9085 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9088 *) MIPS assembler BIGNUM division bug fix.
9091 *) Disabled incorrect Alpha assembler code.
9094 *) Fix PKCS#7 decode routines so they correctly update the length
9095 after reading an EOC for the EXPLICIT tag.
9097 [This change does not apply to 0.9.7.]
9099 *) Fix bug in PKCS#12 key generation routines. This was triggered
9100 if a 3DES key was generated with a 0 initial byte. Include
9101 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9102 (but broken) behaviour.
9105 *) Enhance bctest to search for a working bc along $PATH and print
9107 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9109 *) Fix memory leaks in err.c: free err_data string if necessary;
9110 don't write to the wrong index in ERR_set_error_data.
9113 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9117 *) Replace rdtsc with _emit statements for VC++ version 5.
9118 [Jeremy Cooper <jeremy@baymoo.org>]
9120 *) Make it possible to reuse SSLv2 sessions.
9123 *) In copy_email() check for >= 0 as a return value for
9124 X509_NAME_get_index_by_NID() since 0 is a valid index.
9125 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9127 *) Avoid coredump with unsupported or invalid public keys by checking if
9128 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9129 PKCS7_verify() fails with non detached data.
9132 *) Don't use getenv in library functions when run as setuid/setgid.
9133 New function OPENSSL_issetugid().
9136 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9137 due to incorrect handling of multi-threading:
9139 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9141 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9143 3. Count how many times MemCheck_off() has been called so that
9144 nested use can be treated correctly. This also avoids
9145 inband-signalling in the previous code (which relied on the
9146 assumption that thread ID 0 is impossible).
9149 *) Add "-rand" option also to s_client and s_server.
9152 *) Fix CPU detection on Irix 6.x.
9153 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9154 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9156 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9159 [This change does not apply to 0.9.7.]
9161 *) Use the cached encoding of an X509_NAME structure rather than
9162 copying it. This is apparently the reason for the libsafe "errors"
9163 but the code is actually correct.
9166 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9167 Bleichenbacher's DSA attack.
9168 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9169 to be set and top=0 forces the highest bit to be set; top=-1 is new
9170 and leaves the highest bit random.
9171 [Ulf Moeller, Bodo Moeller]
9173 *) In the NCONF_...-based implementations for CONF_... queries
9174 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9175 a temporary CONF structure with the data component set to NULL
9176 (which gives segmentation faults in lh_retrieve).
9177 Instead, use NULL for the CONF pointer in CONF_get_string and
9178 CONF_get_number (which may use environment variables) and directly
9179 return NULL from CONF_get_section.
9182 *) Fix potential buffer overrun for EBCDIC.
9185 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9186 keyUsage if basicConstraints absent for a CA.
9189 *) Make SMIME_write_PKCS7() write mail header values with a format that
9190 is more generally accepted (no spaces before the semicolon), since
9191 some programs can't parse those values properly otherwise. Also make
9192 sure BIO's that break lines after each write do not create invalid
9196 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9197 macros previously used would not encode an empty SEQUENCE OF
9198 and break the signature.
9200 [This change does not apply to 0.9.7.]
9202 *) Zero the premaster secret after deriving the master secret in
9206 *) Add some EVP_add_digest_alias registrations (as found in
9207 OpenSSL_add_all_digests()) to SSL_library_init()
9208 aka OpenSSL_add_ssl_algorithms(). This provides improved
9209 compatibility with peers using X.509 certificates
9210 with unconventional AlgorithmIdentifier OIDs.
9213 *) Fix for Irix with NO_ASM.
9214 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9216 *) ./config script fixes.
9217 [Ulf Moeller, Richard Levitte]
9219 *) Fix 'openssl passwd -1'.
9222 *) Change PKCS12_key_gen_asc() so it can cope with non null
9223 terminated strings whose length is passed in the passlen
9224 parameter, for example from PEM callbacks. This was done
9225 by adding an extra length parameter to asc2uni().
9226 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9228 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9229 call failed, free the DSA structure.
9232 *) Fix to uni2asc() to cope with zero length Unicode strings.
9233 These are present in some PKCS#12 files.
9236 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9237 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9238 when writing a 32767 byte record.
9239 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9241 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9242 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9244 (RSA objects have a reference count access to which is protected
9245 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9246 so they are meant to be shared between threads.)
9247 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9248 "Reddie, Steven" <Steven.Reddie@ca.com>]
9250 *) Fix a deadlock in CRYPTO_mem_leaks().
9253 *) Use better test patterns in bntest.
9256 *) rand_win.c fix for Borland C.
9259 *) BN_rshift bugfix for n == 0.
9262 *) Add a 'bctest' script that checks for some known 'bc' bugs
9263 so that 'make test' does not abort just because 'bc' is broken.
9266 *) Store verify_result within SSL_SESSION also for client side to
9267 avoid potential security hole. (Re-used sessions on the client side
9268 always resulted in verify_result==X509_V_OK, not using the original
9269 result of the server certificate verification.)
9272 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9273 SSL3_RT_APPLICATION_DATA, return 0.
9274 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9278 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9279 releases, have been re-implemented by renaming the previous
9280 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9281 and ssl3_read_internal, respectively, and adding 'peek' parameters
9282 to them. The new ssl[23]_{read,peek} functions are calls to
9283 ssl[23]_read_internal with the 'peek' flag set appropriately.
9284 A 'peek' parameter has also been added to ssl3_read_bytes, which
9285 does the actual work for ssl3_read_internal.
9288 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9289 the method-specific "init()" handler. Also clean up ex_data after
9290 calling the method-specific "finish()" handler. Previously, this was
9291 happening the other way round.
9294 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9295 The previous value, 12, was not always sufficient for BN_mod_exp().
9298 *) Make sure that shared libraries get the internal name engine with
9299 the full version number and not just 0. This should mark the
9300 shared libraries as not backward compatible. Of course, this should
9301 be changed again when we can guarantee backward binary compatibility.
9304 *) Fix typo in get_cert_by_subject() in by_dir.c
9305 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9307 *) Rework the system to generate shared libraries:
9309 - Make note of the expected extension for the shared libraries and
9310 if there is a need for symbolic links from for example libcrypto.so.0
9311 to libcrypto.so.0.9.7. There is extended info in Configure for
9314 - Make as few rebuilds of the shared libraries as possible.
9316 - Still avoid linking the OpenSSL programs with the shared libraries.
9318 - When installing, install the shared libraries separately from the
9322 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9324 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9325 and not in SSL_clear because the latter is also used by the
9326 accept/connect functions; previously, the settings made by
9327 SSL_set_read_ahead would be lost during the handshake.
9328 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9330 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9331 Previously, it would create entries for disabled algorithms no
9335 *) Added several new manual pages for SSL_* function.
9338 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9340 *) In ssl23_get_client_hello, generate an error message when faced
9341 with an initial SSL 3.0/TLS record that is too small to contain the
9342 first two bytes of the ClientHello message, i.e. client_version.
9343 (Note that this is a pathologic case that probably has never happened
9344 in real life.) The previous approach was to use the version number
9345 from the record header as a substitute; but our protocol choice
9346 should not depend on that one because it is not authenticated
9347 by the Finished messages.
9350 *) More robust randomness gathering functions for Windows.
9351 [Jeffrey Altman <jaltman@columbia.edu>]
9353 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9354 not set then we don't setup the error code for issuer check errors
9355 to avoid possibly overwriting other errors which the callback does
9356 handle. If an application does set the flag then we assume it knows
9357 what it is doing and can handle the new informational codes
9361 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9362 a general "ANY" type, as such it should be able to decode anything
9363 including tagged types. However it didn't check the class so it would
9364 wrongly interpret tagged types in the same way as their universal
9365 counterpart and unknown types were just rejected. Changed so that the
9366 tagged and unknown types are handled in the same way as a SEQUENCE:
9367 that is the encoding is stored intact. There is also a new type
9368 "V_ASN1_OTHER" which is used when the class is not universal, in this
9369 case we have no idea what the actual type is so we just lump them all
9373 *) On VMS, stdout may very well lead to a file that is written to
9374 in a record-oriented fashion. That means that every write() will
9375 write a separate record, which will be read separately by the
9376 programs trying to read from it. This can be very confusing.
9378 The solution is to put a BIO filter in the way that will buffer
9379 text until a linefeed is reached, and then write everything a
9380 line at a time, so every record written will be an actual line,
9381 not chunks of lines and not (usually doesn't happen, but I've
9382 seen it once) several lines in one record. BIO_f_linebuffer() is
9385 Currently, it's a VMS-only method, because that's where it has
9386 been tested well enough.
9389 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9390 it can return incorrect results.
9391 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9392 but it was in 0.9.6-beta[12].)
9395 *) Disable the check for content being present when verifying detached
9396 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9397 include zero length content when signing messages.
9400 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9401 BIO_ctrl (for BIO pairs).
9404 *) Add DSO method for VMS.
9407 *) Bug fix: Montgomery multiplication could produce results with the
9411 *) Add RPM specification openssl.spec and modify it to build three
9412 packages. The default package contains applications, application
9413 documentation and run-time libraries. The devel package contains
9414 include files, static libraries and function documentation. The
9415 doc package contains the contents of the doc directory. The original
9416 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9419 *) Add a large number of documentation files for many SSL routines.
9420 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9422 *) Add a configuration entry for Sony News 4.
9423 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9425 *) Don't set the two most significant bits to one when generating a
9426 random number < q in the DSA library.
9429 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9430 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9431 the underlying transport is blocking) if a handshake took place.
9432 (The default behaviour is needed by applications such as s_client
9433 and s_server that use select() to determine when to use SSL_read;
9434 but for applications that know in advance when to expect data, it
9435 just makes things more complicated.)
9438 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9442 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9443 work better on such systems.
9444 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9446 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9447 Update PKCS12_parse() so it copies the friendlyName and the
9448 keyid to the certificates aux info.
9451 *) Fix bug in PKCS7_verify() which caused an infinite loop
9452 if there was more than one signature.
9453 [Sven Uszpelkat <su@celocom.de>]
9455 *) Major change in util/mkdef.pl to include extra information
9456 about each symbol, as well as presenting variables as well
9457 as functions. This change means that there's n more need
9458 to rebuild the .num files when some algorithms are excluded.
9461 *) Allow the verify time to be set by an application,
9462 rather than always using the current time.
9465 *) Phase 2 verify code reorganisation. The certificate
9466 verify code now looks up an issuer certificate by a
9467 number of criteria: subject name, authority key id
9468 and key usage. It also verifies self signed certificates
9469 by the same criteria. The main comparison function is
9470 X509_check_issued() which performs these checks.
9472 Lot of changes were necessary in order to support this
9473 without completely rewriting the lookup code.
9475 Authority and subject key identifier are now cached.
9477 The LHASH 'certs' is X509_STORE has now been replaced
9478 by a STACK_OF(X509_OBJECT). This is mainly because an
9479 LHASH can't store or retrieve multiple objects with
9480 the same hash value.
9482 As a result various functions (which were all internal
9483 use only) have changed to handle the new X509_STORE
9484 structure. This will break anything that messed round
9485 with X509_STORE internally.
9487 The functions X509_STORE_add_cert() now checks for an
9488 exact match, rather than just subject name.
9490 The X509_STORE API doesn't directly support the retrieval
9491 of multiple certificates matching a given criteria, however
9492 this can be worked round by performing a lookup first
9493 (which will fill the cache with candidate certificates)
9494 and then examining the cache for matches. This is probably
9495 the best we can do without throwing out X509_LOOKUP
9496 entirely (maybe later...).
9498 The X509_VERIFY_CTX structure has been enhanced considerably.
9500 All certificate lookup operations now go via a get_issuer()
9501 callback. Although this currently uses an X509_STORE it
9502 can be replaced by custom lookups. This is a simple way
9503 to bypass the X509_STORE hackery necessary to make this
9504 work and makes it possible to use more efficient techniques
9505 in future. A very simple version which uses a simple
9506 STACK for its trusted certificate store is also provided
9507 using X509_STORE_CTX_trusted_stack().
9509 The verify_cb() and verify() callbacks now have equivalents
9510 in the X509_STORE_CTX structure.
9512 X509_STORE_CTX also has a 'flags' field which can be used
9513 to customise the verify behaviour.
9516 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9517 excludes S/MIME capabilities.
9520 *) When a certificate request is read in keep a copy of the
9521 original encoding of the signed data and use it when outputting
9522 again. Signatures then use the original encoding rather than
9523 a decoded, encoded version which may cause problems if the
9524 request is improperly encoded.
9527 *) For consistency with other BIO_puts implementations, call
9528 buffer_write(b, ...) directly in buffer_puts instead of calling
9531 In BIO_puts, increment b->num_write as in BIO_write.
9532 [Peter.Sylvester@EdelWeb.fr]
9534 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9535 BN_zero, we may not return a BIGNUM with an array consisting of
9539 *) Avoid calling abort() from within the library when problems are
9540 detected, except if preprocessor symbols have been defined
9541 (such as REF_CHECK, BN_DEBUG etc.).
9544 *) New openssl application 'rsautl'. This utility can be
9545 used for low level RSA operations. DER public key
9546 BIO/fp routines also added.
9549 *) New Configure entry and patches for compiling on QNX 4.
9550 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9552 *) A demo state-machine implementation was sponsored by
9553 Nuron (http://www.nuron.com/) and is now available in
9554 demos/state_machine.
9557 *) New options added to the 'dgst' utility for signature
9558 generation and verification.
9561 *) Unrecognized PKCS#7 content types are now handled via a
9562 catch all ASN1_TYPE structure. This allows unsupported
9563 types to be stored as a "blob" and an application can
9564 encode and decode it manually.
9567 *) Fix various signed/unsigned issues to make a_strex.c
9569 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9571 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9572 length if passed a buffer. ASN1_INTEGER_to_BN failed
9573 if passed a NULL BN and its argument was negative.
9574 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9576 *) Modification to PKCS#7 encoding routines to output definite
9577 length encoding. Since currently the whole structures are in
9578 memory there's not real point in using indefinite length
9579 constructed encoding. However if OpenSSL is compiled with
9580 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9583 *) Added BIO_vprintf() and BIO_vsnprintf().
9586 *) Added more prefixes to parse for in the strings written
9587 through a logging bio, to cover all the levels that are available
9588 through syslog. The prefixes are now:
9590 PANIC, EMERG, EMR => LOG_EMERG
9591 ALERT, ALR => LOG_ALERT
9592 CRIT, CRI => LOG_CRIT
9593 ERROR, ERR => LOG_ERR
9594 WARNING, WARN, WAR => LOG_WARNING
9595 NOTICE, NOTE, NOT => LOG_NOTICE
9596 INFO, INF => LOG_INFO
9597 DEBUG, DBG => LOG_DEBUG
9599 and as before, if none of those prefixes are present at the
9600 beginning of the string, LOG_ERR is chosen.
9602 On Win32, the LOG_* levels are mapped according to this:
9604 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9605 LOG_WARNING => EVENTLOG_WARNING_TYPE
9606 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9610 *) Made it possible to reconfigure with just the configuration
9611 argument "reconf" or "reconfigure". The command line arguments
9612 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9613 and are retrieved from there when reconfiguring.
9617 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9619 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9622 *) The obj_dat.pl script was messing up the sorting of object
9623 names. The reason was that it compared the quoted version
9624 of strings as a result "OCSP" > "OCSP Signing" because
9625 " > SPACE. Changed script to store unquoted versions of
9626 names and add quotes on output. It was also omitting some
9627 names from the lookup table if they were given a default
9628 value (that is if SN is missing it is given the same
9629 value as LN and vice versa), these are now added on the
9630 grounds that if an object has a name we should be able to
9631 look it up. Finally added warning output when duplicate
9632 short or long names are found.
9635 *) Changes needed for Tandem NSK.
9636 [Scott Uroff <scott@xypro.com>]
9638 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9639 RSA_padding_check_SSLv23(), special padding was never detected
9640 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9641 version rollback attacks was not effective.
9643 In s23_clnt.c, don't use special rollback-attack detection padding
9644 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9645 client; similarly, in s23_srvr.c, don't do the rollback check if
9646 SSL 2.0 is the only protocol enabled in the server.
9649 *) Make it possible to get hexdumps of unprintable data with 'openssl
9650 asn1parse'. By implication, the functions ASN1_parse_dump() and
9651 BIO_dump_indent() are added.
9654 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9655 these print out strings and name structures based on various
9656 flags including RFC2253 support and proper handling of
9657 multibyte characters. Added options to the 'x509' utility
9658 to allow the various flags to be set.
9661 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9662 Also change the functions X509_cmp_current_time() and
9663 X509_gmtime_adj() work with an ASN1_TIME structure,
9664 this will enable certificates using GeneralizedTime in validity
9665 dates to be checked.
9668 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9669 negative public key encodings) on by default,
9670 NO_NEG_PUBKEY_BUG can be set to disable it.
9673 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9674 content octets. An i2c_ASN1_OBJECT is unnecessary because
9675 the encoding can be trivially obtained from the structure.
9678 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9679 not read locks (CRYPTO_r_[un]lock).
9682 *) A first attempt at creating official support for shared
9683 libraries through configuration. I've kept it so the
9684 default is static libraries only, and the OpenSSL programs
9685 are always statically linked for now, but there are
9686 preparations for dynamic linking in place.
9687 This has been tested on Linux and Tru64.
9690 *) Randomness polling function for Win9x, as described in:
9691 Peter Gutmann, Software Generation of Practically Strong
9695 *) Fix so PRNG is seeded in req if using an already existing
9699 *) New options to smime application. -inform and -outform
9700 allow alternative formats for the S/MIME message including
9701 PEM and DER. The -content option allows the content to be
9702 specified separately. This should allow things like Netscape
9703 form signing output easier to verify.
9706 *) Fix the ASN1 encoding of tags using the 'long form'.
9709 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9710 STRING types. These convert content octets to and from the
9711 underlying type. The actual tag and length octets are
9712 already assumed to have been read in and checked. These
9713 are needed because all other string types have virtually
9714 identical handling apart from the tag. By having versions
9715 of the ASN1 functions that just operate on content octets
9716 IMPLICIT tagging can be handled properly. It also allows
9717 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9718 and ASN1_INTEGER are identical apart from the tag.
9721 *) Change the handling of OID objects as follows:
9723 - New object identifiers are inserted in objects.txt, following
9724 the syntax given in objects.README.
9725 - objects.pl is used to process obj_mac.num and create a new
9727 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9730 This is currently kind of a hack, and the perl code in objects.pl
9731 isn't very elegant, but it works as I intended. The simplest way
9732 to check that it worked correctly is to look in obj_dat.h and
9733 check the array nid_objs and make sure the objects haven't moved
9734 around (this is important!). Additions are OK, as well as
9735 consistent name changes.
9738 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9741 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9742 The given file adds to whatever has already been seeded into the
9743 random pool through the RANDFILE configuration file option or
9744 environment variable, or the default random state file.
9747 *) mkstack.pl now sorts each macro group into lexical order.
9748 Previously the output order depended on the order the files
9749 appeared in the directory, resulting in needless rewriting
9753 *) Patches to make OpenSSL compile under Win32 again. Mostly
9754 work arounds for the VC++ problem that it treats func() as
9755 func(void). Also stripped out the parts of mkdef.pl that
9756 added extra typesafe functions: these no longer exist.
9759 *) Reorganisation of the stack code. The macros are now all
9760 collected in safestack.h . Each macro is defined in terms of
9761 a "stack macro" of the form SKM_<name>(type, a, b). The
9762 DEBUG_SAFESTACK is now handled in terms of function casts,
9763 this has the advantage of retaining type safety without the
9764 use of additional functions. If DEBUG_SAFESTACK is not defined
9765 then the non typesafe macros are used instead. Also modified the
9766 mkstack.pl script to handle the new form. Needs testing to see
9767 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9768 the default if no major problems. Similar behaviour for ASN1_SET_OF
9769 and PKCS12_STACK_OF.
9772 *) When some versions of IIS use the 'NET' form of private key the
9773 key derivation algorithm is different. Normally MD5(password) is
9774 used as a 128 bit RC4 key. In the modified case
9775 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9776 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9777 as the old Netscape_RSA functions except they have an additional
9778 'sgckey' parameter which uses the modified algorithm. Also added
9779 an -sgckey command line option to the rsa utility. Thanks to
9780 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9781 algorithm to openssl-dev.
9784 *) The evp_local.h macros were using 'c.##kname' which resulted in
9785 invalid expansion on some systems (SCO 5.0.5 for example).
9786 Corrected to 'c.kname'.
9787 [Phillip Porch <root@theporch.com>]
9789 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9790 a STACK of email addresses from a certificate or request, these look
9791 in the subject name and the subject alternative name extensions and
9792 omit any duplicate addresses.
9795 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9796 This makes DSA verification about 2 % faster.
9799 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9800 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9801 plus overhead for 1024 bit moduli).
9802 This makes exponentiations about 0.5 % faster for 1024 bit
9803 exponents (as measured by "openssl speed rsa2048").
9806 *) Rename memory handling macros to avoid conflicts with other
9808 Malloc => OPENSSL_malloc
9809 Malloc_locked => OPENSSL_malloc_locked
9810 Realloc => OPENSSL_realloc
9811 Free => OPENSSL_free
9814 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9815 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9818 *) CygWin32 support.
9819 [John Jarvie <jjarvie@newsguy.com>]
9821 *) The type-safe stack code has been rejigged. It is now only compiled
9822 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9823 by default all type-specific stack functions are "#define"d back to
9824 standard stack functions. This results in more streamlined output
9825 but retains the type-safety checking possibilities of the original
9829 *) The STACK code has been cleaned up, and certain type declarations
9830 that didn't make a lot of sense have been brought in line. This has
9831 also involved a cleanup of sorts in safestack.h to more correctly
9832 map type-safe stack functions onto their plain stack counterparts.
9833 This work has also resulted in a variety of "const"ifications of
9834 lots of the code, especially "_cmp" operations which should normally
9835 be prototyped with "const" parameters anyway.
9838 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9839 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9840 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9841 where all of 'md' is used each time the PRNG is used, but 'state'
9842 is used only indexed by a cyclic counter. As entropy may not be
9843 well distributed from the beginning, 'md' is important as a
9844 chaining variable. However, the output function chains only half
9845 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9846 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9847 in all of 'state' being rewritten, with the new values depending
9848 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9851 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9852 the handshake is continued after ssl_verify_cert_chain();
9853 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9854 can lead to 'unexplainable' connection aborts later.
9855 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9857 *) Major EVP API cipher revision.
9858 Add hooks for extra EVP features. This allows various cipher
9859 parameters to be set in the EVP interface. Support added for variable
9860 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9861 setting of RC2 and RC5 parameters.
9863 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9866 Remove lots of duplicated code from the EVP library. For example *every*
9867 cipher init() function handles the 'iv' in the same way according to the
9868 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9869 for CFB and OFB modes they zero ctx->num.
9871 New functionality allows removal of S/MIME code RC2 hack.
9873 Most of the routines have the same form and so can be declared in terms
9876 By shifting this to the top level EVP_CipherInit() it can be removed from
9877 all individual ciphers. If the cipher wants to handle IVs or keys
9878 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9881 Change lots of functions like EVP_EncryptUpdate() to now return a
9882 value: although software versions of the algorithms cannot fail
9883 any installed hardware versions can.
9886 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9887 this option is set, tolerate broken clients that send the negotiated
9888 protocol version number instead of the requested protocol version
9892 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9893 i.e. non-zero for export ciphersuites, zero otherwise.
9894 Previous versions had this flag inverted, inconsistent with
9895 rsa_tmp_cb (..._TMP_RSA_CB).
9896 [Bodo Moeller; problem reported by Amit Chopra]
9898 *) Add missing DSA library text string. Work around for some IIS
9899 key files with invalid SEQUENCE encoding.
9902 *) Add a document (doc/standards.txt) that list all kinds of standards
9903 and so on that are implemented in OpenSSL.
9906 *) Enhance c_rehash script. Old version would mishandle certificates
9907 with the same subject name hash and wouldn't handle CRLs at all.
9908 Added -fingerprint option to crl utility, to support new c_rehash
9912 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9915 *) Fix for SSL server purpose checking. Server checking was
9916 rejecting certificates which had extended key usage present
9917 but no ssl client purpose.
9918 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9920 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9921 is a little unclear about how a blank password is handled.
9922 Since the password in encoded as a BMPString with terminating
9923 double NULL a zero length password would end up as just the
9924 double NULL. However no password at all is different and is
9925 handled differently in the PKCS#12 key generation code. NS
9926 treats a blank password as zero length. MSIE treats it as no
9927 password on export: but it will try both on import. We now do
9928 the same: PKCS12_parse() tries zero length and no password if
9929 the password is set to "" or NULL (NULL is now a valid password:
9930 it wasn't before) as does the pkcs12 application.
9933 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9934 perror when PEM_read_bio_X509_REQ fails, the error message must
9935 be obtained from the error queue.
9938 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9939 it in ERR_remove_state if appropriate, and change ERR_get_state
9940 accordingly to avoid race conditions (this is necessary because
9941 thread_hash is no longer constant once set).
9944 *) Bugfix for linux-elf makefile.one.
9947 *) RSA_get_default_method() will now cause a default
9948 RSA_METHOD to be chosen if one doesn't exist already.
9949 Previously this was only set during a call to RSA_new()
9950 or RSA_new_method(NULL) meaning it was possible for
9951 RSA_get_default_method() to return NULL.
9954 *) Added native name translation to the existing DSO code
9955 that will convert (if the flag to do so is set) filenames
9956 that are sufficiently small and have no path information
9957 into a canonical native form. Eg. "blah" converted to
9958 "libblah.so" or "blah.dll" etc.
9961 *) New function ERR_error_string_n(e, buf, len) which is like
9962 ERR_error_string(e, buf), but writes at most 'len' bytes
9963 including the 0 terminator. For ERR_error_string_n, 'buf'
9965 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9967 *) CONF library reworked to become more general. A new CONF
9968 configuration file reader "class" is implemented as well as a
9969 new functions (NCONF_*, for "New CONF") to handle it. The now
9970 old CONF_* functions are still there, but are reimplemented to
9971 work in terms of the new functions. Also, a set of functions
9972 to handle the internal storage of the configuration data is
9973 provided to make it easier to write new configuration file
9974 reader "classes" (I can definitely see something reading a
9975 configuration file in XML format, for example), called _CONF_*,
9976 or "the configuration storage API"...
9978 The new configuration file reading functions are:
9980 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9981 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9983 NCONF_default, NCONF_WIN32
9985 NCONF_dump_fp, NCONF_dump_bio
9987 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9988 NCONF_new creates a new CONF object. This works in the same way
9989 as other interfaces in OpenSSL, like the BIO interface.
9990 NCONF_dump_* dump the internal storage of the configuration file,
9991 which is useful for debugging. All other functions take the same
9992 arguments as the old CONF_* functions wth the exception of the
9993 first that must be a `CONF *' instead of a `LHASH *'.
9995 To make it easer to use the new classes with the old CONF_* functions,
9996 the function CONF_set_default_method is provided.
9999 *) Add '-tls1' option to 'openssl ciphers', which was already
10000 mentioned in the documentation but had not been implemented.
10001 (This option is not yet really useful because even the additional
10002 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10005 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10006 OpenSSL-based applications) load shared libraries and bind to
10007 them in a portable way.
10008 [Geoff Thorpe, with contributions from Richard Levitte]
10010 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10012 *) Make sure _lrotl and _lrotr are only used with MSVC.
10014 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10015 (the default implementation of RAND_status).
10017 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10018 to '-clrext' (= clear extensions), as intended and documented.
10019 [Bodo Moeller; inconsistency pointed out by Michael Attili
10020 <attili@amaxo.com>]
10022 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10023 was larger than the MD block size.
10024 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10026 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10027 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10028 using the passed key: if the passed key was a private key the result
10029 of X509_print(), for example, would be to print out all the private key
10033 *) des_quad_cksum() byte order bug fix.
10034 [Ulf Möller, using the problem description in krb4-0.9.7, where
10035 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10037 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10039 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10041 *) For easily testing in shell scripts whether some command
10042 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10043 returns with exit code 0 iff no command of the given name is available.
10044 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10045 the output goes to stdout and nothing is printed to stderr.
10046 Additional arguments are always ignored.
10048 Since for each cipher there is a command of the same name,
10049 the 'no-cipher' compilation switches can be tested this way.
10051 ('openssl no-XXX' is not able to detect pseudo-commands such
10052 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10055 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10058 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10059 is set; it will be thrown away anyway because each handshake creates
10061 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10062 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10063 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10064 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10067 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10068 'Q' and 'R' lose their special meanings (quit/renegotiate).
10069 This is part of what -quiet does; unlike -quiet, -ign_eof
10070 does not suppress any output.
10073 *) Add compatibility options to the purpose and trust code. The
10074 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10075 accepts a certificate or CA, this was the previous behaviour,
10076 with all the associated security issues.
10078 X509_TRUST_COMPAT is the old trust behaviour: only and
10079 automatically trust self signed roots in certificate store. A
10080 new trust setting X509_TRUST_DEFAULT is used to specify that
10081 a purpose has no associated trust setting and it should instead
10082 use the value in the default purpose.
10085 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10086 and fix a memory leak.
10089 *) In util/mkerr.pl (which implements 'make errors'), preserve
10090 reason strings from the previous version of the .c file, as
10091 the default to have only downcase letters (and digits) in
10092 automatically generated reasons codes is not always appropriate.
10095 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10096 using strerror. Previously, ERR_reason_error_string() returned
10097 library names as reason strings for SYSerr; but SYSerr is a special
10098 case where small numbers are errno values, not library numbers.
10101 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10102 converts DSA parameters into DH parameters. (When creating parameters,
10103 DSA_generate_parameters is used.)
10106 *) Include 'length' (recommended exponent length) in C code generated
10107 by 'openssl dhparam -C'.
10110 *) The second argument to set_label in perlasm was already being used
10111 so couldn't be used as a "file scope" flag. Moved to third argument
10115 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10116 instead of RAND_bytes for encryption IVs and salts.
10119 *) Include RAND_status() into RAND_METHOD instead of implementing
10120 it only for md_rand.c Otherwise replacing the PRNG by calling
10121 RAND_set_rand_method would be impossible.
10124 *) Don't let DSA_generate_key() enter an infinite loop if the random
10125 number generation fails.
10128 *) New 'rand' application for creating pseudo-random output.
10131 *) Added configuration support for Linux/IA64
10132 [Rolf Haberrecker <rolf@suse.de>]
10134 *) Assembler module support for Mingw32.
10137 *) Shared library support for HPUX (in shlib/).
10138 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10140 *) Shared library support for Solaris gcc.
10141 [Lutz Behnke <behnke@trustcenter.de>]
10143 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10145 *) PKCS7_encrypt() was adding text MIME headers twice because they
10146 were added manually and by SMIME_crlf_copy().
10149 *) In bntest.c don't call BN_rand with zero bits argument.
10150 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10152 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10153 case was implemented. This caused BN_div_recp() to fail occasionally.
10156 *) Add an optional second argument to the set_label() in the perl
10157 assembly language builder. If this argument exists and is set
10158 to 1 it signals that the assembler should use a symbol whose
10159 scope is the entire file, not just the current function. This
10160 is needed with MASM which uses the format label:: for this scope.
10161 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10163 *) Change the ASN1 types so they are typedefs by default. Before
10164 almost all types were #define'd to ASN1_STRING which was causing
10165 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10169 *) Change names of new functions to the new get1/get0 naming
10170 convention: After 'get1', the caller owns a reference count
10171 and has to call ..._free; 'get0' returns a pointer to some
10172 data structure without incrementing reference counters.
10173 (Some of the existing 'get' functions increment a reference
10174 counter, some don't.)
10175 Similarly, 'set1' and 'add1' functions increase reference
10176 counters or duplicate objects.
10179 *) Allow for the possibility of temp RSA key generation failure:
10180 the code used to assume it always worked and crashed on failure.
10183 *) Fix potential buffer overrun problem in BIO_printf().
10184 [Ulf Möller, using public domain code by Patrick Powell; problem
10185 pointed out by David Sacerdote <das33@cornell.edu>]
10187 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10188 RAND_egd() and RAND_status(). In the command line application,
10189 the EGD socket can be specified like a seed file using RANDFILE
10193 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10194 Some CAs (e.g. Verisign) distribute certificates in this form.
10197 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10198 list to exclude them. This means that no special compilation option
10199 is needed to use anonymous DH: it just needs to be included in the
10203 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10204 EVP_MD_type. The old functionality is available in a new macro called
10205 EVP_MD_md(). Change code that uses it and update docs.
10208 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10209 where the 'void *' argument is replaced by a function pointer argument.
10210 Previously 'void *' was abused to point to functions, which works on
10211 many platforms, but is not correct. As these functions are usually
10212 called by macros defined in OpenSSL header files, most source code
10213 should work without changes.
10216 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10217 sections with information on -D... compiler switches used for
10218 compiling the library so that applications can see them. To enable
10219 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10220 must be defined. E.g.,
10221 #define OPENSSL_ALGORITHM_DEFINES
10222 #include <openssl/opensslconf.h>
10223 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10224 [Richard Levitte, Ulf and Bodo Möller]
10226 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10230 *) Change the 'other' type in certificate aux info to a STACK_OF
10231 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10232 the required ASN1 format: arbitrary types determined by an OID.
10235 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10236 argument to 'req'. This is not because the function is newer or
10237 better than others it just uses the work 'NEW' in the certificate
10238 request header lines. Some software needs this.
10241 *) Reorganise password command line arguments: now passwords can be
10242 obtained from various sources. Delete the PEM_cb function and make
10243 it the default behaviour: i.e. if the callback is NULL and the
10244 usrdata argument is not NULL interpret it as a null terminated pass
10245 phrase. If usrdata and the callback are NULL then the pass phrase
10246 is prompted for as usual.
10249 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10250 the support is automatically enabled. The resulting binaries will
10251 autodetect the card and use it if present.
10252 [Ben Laurie and Compaq Inc.]
10254 *) Work around for Netscape hang bug. This sends certificate request
10255 and server done in one record. Since this is perfectly legal in the
10256 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10257 the bugs/SSLv3 entry for more info.
10260 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10263 *) Add -rand argument to smime and pkcs12 applications and read/write
10267 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10270 *) Add command line password options to the remaining applications.
10273 *) Bug fix for BN_div_recp() for numerators with an even number of
10277 *) More tests in bntest.c, and changed test_bn output.
10280 *) ./config recognizes MacOS X now.
10283 *) Bug fix for BN_div() when the first words of num and divisor are
10284 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10287 *) Add support for various broken PKCS#8 formats, and command line
10288 options to produce them.
10291 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10292 get temporary BIGNUMs from a BN_CTX.
10295 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10299 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10300 include a #define from the old name to the new. The original intent
10301 was that statically linked binaries could for example just call
10302 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10303 link with digests. This never worked because SSLeay_add_all_digests()
10304 and SSLeay_add_all_ciphers() were in the same source file so calling
10305 one would link with the other. They are now in separate source files.
10308 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10311 *) Use a less unusual form of the Miller-Rabin primality test (it used
10312 a binary algorithm for exponentiation integrated into the Miller-Rabin
10313 loop, our standard modexp algorithms are faster).
10316 *) Support for the EBCDIC character set completed.
10317 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10319 *) Source code cleanups: use const where appropriate, eliminate casts,
10320 use void * instead of char * in lhash.
10323 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10324 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10325 this the server could overwrite ephemeral keys that the client
10329 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10330 using 50 iterations of the Rabin-Miller test.
10332 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10333 iterations of the Rabin-Miller test as required by the appendix
10334 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10335 As BN_is_prime_fasttest includes trial division, DSA parameter
10336 generation becomes much faster.
10338 This implies a change for the callback functions in DSA_is_prime
10339 and DSA_generate_parameters: The callback function is called once
10340 for each positive witness in the Rabin-Miller test, not just
10341 occasionally in the inner loop; and the parameters to the
10342 callback function now provide an iteration count for the outer
10343 loop rather than for the current invocation of the inner loop.
10344 DSA_generate_parameters additionally can call the callback
10345 function with an 'iteration count' of -1, meaning that a
10346 candidate has passed the trial division test (when q is generated
10347 from an application-provided seed, trial division is skipped).
10350 *) New function BN_is_prime_fasttest that optionally does trial
10351 division before starting the Rabin-Miller test and has
10352 an additional BN_CTX * argument (whereas BN_is_prime always
10353 has to allocate at least one BN_CTX).
10354 'callback(1, -1, cb_arg)' is called when a number has passed the
10355 trial division stage.
10358 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10362 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10365 *) New function BN_pseudo_rand().
10368 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10369 bignum version of BN_from_montgomery() with the working code from
10370 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10374 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10375 made it impossible to use the same SSL_SESSION data structure in
10376 SSL2 clients in multiple threads.
10379 *) The return value of RAND_load_file() no longer counts bytes obtained
10380 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10381 to seed the PRNG (previously an explicit byte count was required).
10382 [Ulf Möller, Bodo Möller]
10384 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10385 used (char *) instead of (void *) and had casts all over the place.
10388 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10391 *) Retain source code compatibility for BN_prime_checks macro:
10392 BN_is_prime(..., BN_prime_checks, ...) now uses
10393 BN_prime_checks_for_size to determine the appropriate number of
10394 Rabin-Miller iterations.
10397 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10398 DH_CHECK_P_NOT_SAFE_PRIME.
10399 (Check if this is true? OpenPGP calls them "strong".)
10402 *) Merge the functionality of "dh" and "gendh" programs into a new program
10403 "dhparam". The old programs are retained for now but will handle DH keys
10404 (instead of parameters) in future.
10407 *) Make the ciphers, s_server and s_client programs check the return values
10408 when a new cipher list is set.
10411 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10412 ciphers. Before when the 56bit ciphers were enabled the sorting was
10415 The syntax for the cipher sorting has been extended to support sorting by
10416 cipher-strength (using the strength_bits hard coded in the tables).
10417 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10419 Fix a bug in the cipher-command parser: when supplying a cipher command
10420 string with an "undefined" symbol (neither command nor alphanumeric
10421 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10422 an error is flagged.
10424 Due to the strength-sorting extension, the code of the
10425 ssl_create_cipher_list() function was completely rearranged. I hope that
10426 the readability was also increased :-)
10427 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10429 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10430 for the first serial number and places 2 in the serial number file. This
10431 avoids problems when the root CA is created with serial number zero and
10432 the first user certificate has the same issuer name and serial number
10436 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10437 the new code. Add documentation for this stuff.
10440 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10441 X509_*() to X509at_*() on the grounds that they don't handle X509
10442 structures and behave in an analogous way to the X509v3 functions:
10443 they shouldn't be called directly but wrapper functions should be used
10446 So we also now have some wrapper functions that call the X509at functions
10447 when passed certificate requests. (TO DO: similar things can be done with
10448 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10449 things. Some of these need some d2i or i2d and print functionality
10450 because they handle more complex structures.)
10453 *) Add missing #ifndefs that caused missing symbols when building libssl
10454 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10455 NO_RSA in ssl/s2*.c.
10456 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10458 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10459 has a return value which indicates the quality of the random data
10460 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10461 error queue. New function RAND_pseudo_bytes() generates output that is
10462 guaranteed to be unique but not unpredictable. RAND_add is like
10463 RAND_seed, but takes an extra argument for an entropy estimate
10464 (RAND_seed always assumes full entropy).
10467 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10468 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10469 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10470 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10471 false-positive rate of at most 2^-80 for random input.
10474 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10477 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10478 in the 0.9.5 release), this returns the chain
10479 from an X509_CTX structure with a dup of the stack and all
10480 the X509 reference counts upped: so the stack will exist
10481 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10484 Also make SSL_SESSION_print() print out the verify return
10488 *) Add manpage for the pkcs12 command. Also change the default
10489 behaviour so MAC iteration counts are used unless the new
10490 -nomaciter option is used. This improves file security and
10491 only older versions of MSIE (4.0 for example) need it.
10494 *) Honor the no-xxx Configure options when creating .DEF files.
10497 *) Add PKCS#10 attributes to field table: challengePassword,
10498 unstructuredName and unstructuredAddress. These are taken from
10499 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10500 international characters are used.
10502 More changes to X509_ATTRIBUTE code: allow the setting of types
10503 based on strings. Remove the 'loc' parameter when adding
10504 attributes because these will be a SET OF encoding which is sorted
10508 *) Initial changes to the 'req' utility to allow request generation
10509 automation. This will allow an application to just generate a template
10510 file containing all the field values and have req construct the
10513 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10514 used all over the place including certificate requests and PKCS#7
10515 structures. They are currently handled manually where necessary with
10516 some primitive wrappers for PKCS#7. The new functions behave in a
10517 manner analogous to the X509 extension functions: they allow
10518 attributes to be looked up by NID and added.
10520 Later something similar to the X509V3 code would be desirable to
10521 automatically handle the encoding, decoding and printing of the
10522 more complex types. The string types like challengePassword can
10523 be handled by the string table functions.
10525 Also modified the multi byte string table handling. Now there is
10526 a 'global mask' which masks out certain types. The table itself
10527 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10528 is useful when for example there is only one permissible type
10529 (as in countryName) and using the mask might result in no valid
10533 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10534 SSL_get_peer_finished to allow applications to obtain the latest
10535 Finished messages sent to the peer or expected from the peer,
10536 respectively. (SSL_get_peer_finished is usually the Finished message
10537 actually received from the peer, otherwise the protocol will be aborted.)
10539 As the Finished message are message digests of the complete handshake
10540 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10541 be used for external authentication procedures when the authentication
10542 provided by SSL/TLS is not desired or is not enough.
10545 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10546 the host supports BWX extension and if Compaq C is present on the
10547 $PATH. Just exploiting of the BWX extension results in 20-30%
10548 performance kick for some algorithms, e.g. DES and RC4 to mention
10549 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10553 *) Add support for MS "fast SGC". This is arguably a violation of the
10554 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10555 weak crypto and after checking the certificate is SGC a second one
10556 with strong crypto. MS SGC stops the first handshake after receiving
10557 the server certificate message and sends a second client hello. Since
10558 a server will typically do all the time consuming operations before
10559 expecting any further messages from the client (server key exchange
10560 is the most expensive) there is little difference between the two.
10562 To get OpenSSL to support MS SGC we have to permit a second client
10563 hello message after we have sent server done. In addition we have to
10564 reset the MAC if we do get this second client hello.
10567 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10568 if a DER encoded private key is RSA or DSA traditional format. Changed
10569 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10570 format DER encoded private key. Newer code should use PKCS#8 format which
10571 has the key type encoded in the ASN1 structure. Added DER private key
10572 support to pkcs8 application.
10575 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10576 ciphersuites has been selected (as required by the SSL 3/TLS 1
10577 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10578 is set, we interpret this as a request to violate the specification
10579 (the worst that can happen is a handshake failure, and 'correct'
10580 behaviour would result in a handshake failure anyway).
10583 *) In SSL_CTX_add_session, take into account that there might be multiple
10584 SSL_SESSION structures with the same session ID (e.g. when two threads
10585 concurrently obtain them from an external cache).
10586 The internal cache can handle only one SSL_SESSION with a given ID,
10587 so if there's a conflict, we now throw out the old one to achieve
10591 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10592 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10593 some routines that use cipher OIDs: some ciphers do not have OIDs
10594 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10598 *) Simplify the trust setting structure and code. Now we just have
10599 two sequences of OIDs for trusted and rejected settings. These will
10600 typically have values the same as the extended key usage extension
10601 and any application specific purposes.
10603 The trust checking code now has a default behaviour: it will just
10604 check for an object with the same NID as the passed id. Functions can
10605 be provided to override either the default behaviour or the behaviour
10606 for a given id. SSL client, server and email already have functions
10607 in place for compatibility: they check the NID and also return "trusted"
10608 if the certificate is self signed.
10611 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10612 traditional format into an EVP_PKEY structure.
10615 *) Add a password callback function PEM_cb() which either prompts for
10616 a password if usr_data is NULL or otherwise assumes it is a null
10617 terminated password. Allow passwords to be passed on command line
10618 environment or config files in a few more utilities.
10621 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10622 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10623 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10624 Update documentation.
10627 *) Support for ASN1 "NULL" type. This could be handled before by using
10628 ASN1_TYPE but there wasn't any function that would try to read a NULL
10629 and produce an error if it couldn't. For compatibility we also have
10630 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10631 don't allocate anything because they don't need to.
10634 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10636 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10638 *) Rebuild of the memory allocation routines used by OpenSSL code and
10639 possibly others as well. The purpose is to make an interface that
10640 provide hooks so anyone can build a separate set of allocation and
10641 deallocation routines to be used by OpenSSL, for example memory
10642 pool implementations, or something else, which was previously hard
10643 since Malloc(), Realloc() and Free() were defined as macros having
10644 the values malloc, realloc and free, respectively (except for Win32
10645 compilations). The same is provided for memory debugging code.
10646 OpenSSL already comes with functionality to find memory leaks, but
10647 this gives people a chance to debug other memory problems.
10649 With these changes, a new set of functions and macros have appeared:
10651 CRYPTO_set_mem_debug_functions() [F]
10652 CRYPTO_get_mem_debug_functions() [F]
10653 CRYPTO_dbg_set_options() [F]
10654 CRYPTO_dbg_get_options() [F]
10655 CRYPTO_malloc_debug_init() [M]
10657 The memory debug functions are NULL by default, unless the library
10658 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10659 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10660 gives the standard debugging functions that come with OpenSSL) or
10661 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10662 provided by the library user) must be used. When the standard
10663 debugging functions are used, CRYPTO_dbg_set_options can be used to
10664 request additional information:
10665 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10666 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10668 Also, things like CRYPTO_set_mem_functions will always give the
10669 expected result (the new set of functions is used for allocation
10670 and deallocation) at all times, regardless of platform and compiler
10673 To finish it up, some functions that were never use in any other
10674 way than through macros have a new API and new semantic:
10676 CRYPTO_dbg_malloc()
10677 CRYPTO_dbg_realloc()
10680 All macros of value have retained their old syntax.
10681 [Richard Levitte and Bodo Moeller]
10683 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10684 ordering of SMIMECapabilities wasn't in "strength order" and there
10685 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10689 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10690 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10691 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10693 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10694 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10695 functionality to handle multipart/signed properly) and a utility
10696 called 'smime' to call all this stuff. This is based on code I
10697 originally wrote for Celo who have kindly allowed it to be
10698 included in OpenSSL.
10701 *) Add variants des_set_key_checked and des_set_key_unchecked of
10702 des_set_key (aka des_key_sched). Global variable des_check_key
10703 decides which of these is called by des_set_key; this way
10704 des_check_key behaves as it always did, but applications and
10705 the library itself, which was buggy for des_check_key == 1,
10706 have a cleaner way to pick the version they need.
10709 *) New function PKCS12_newpass() which changes the password of a
10713 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10714 dynamic mix. In both cases the ids can be used as an index into the
10715 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10716 functions so they accept a list of the field values and the
10717 application doesn't need to directly manipulate the X509_TRUST
10721 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10725 *) Modify the way the V3 extension code looks up extensions. This now
10726 works in a similar way to the object code: we have some "standard"
10727 extensions in a static table which is searched with OBJ_bsearch()
10728 and the application can add dynamic ones if needed. The file
10729 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10730 updated whenever a new extension is added to the core code and kept
10731 in ext_nid order. There is a simple program 'tabtest.c' which checks
10732 this. New extensions are not added too often so this file can readily
10733 be maintained manually.
10735 There are two big advantages in doing things this way. The extensions
10736 can be looked up immediately and no longer need to be "added" using
10737 X509V3_add_standard_extensions(): this function now does nothing.
10738 [Side note: I get *lots* of email saying the extension code doesn't
10739 work because people forget to call this function]
10740 Also no dynamic allocation is done unless new extensions are added:
10741 so if we don't add custom extensions there is no need to call
10742 X509V3_EXT_cleanup().
10745 *) Modify enc utility's salting as follows: make salting the default. Add a
10746 magic header, so unsalted files fail gracefully instead of just decrypting
10747 to garbage. This is because not salting is a big security hole, so people
10748 should be discouraged from doing it.
10751 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10752 digest to be passed on the command line but it only used this
10753 parameter when signing a certificate. Modified so all relevant
10754 operations are affected by the digest parameter including the
10755 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10756 DSA key was used because it didn't fix the digest.
10759 *) Initial certificate chain verify code. Currently tests the untrusted
10760 certificates for consistency with the verify purpose (which is set
10761 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10763 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10764 this is because it will reject chains with invalid extensions whereas
10765 every previous version of OpenSSL and SSLeay made no checks at all.
10767 Trust code: checks the root CA for the relevant trust settings. Trust
10768 settings have an initial value consistent with the verify purpose: e.g.
10769 if the verify purpose is for SSL client use it expects the CA to be
10770 trusted for SSL client use. However the default value can be changed to
10771 permit custom trust settings: one example of this would be to only trust
10772 certificates from a specific "secure" set of CAs.
10774 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10775 which should be used for version portability: especially since the
10776 verify structure is likely to change more often now.
10778 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10779 to set them. If not set then assume SSL clients will verify SSL servers
10782 Two new options to the verify program: -untrusted allows a set of
10783 untrusted certificates to be passed in and -purpose which sets the
10784 intended purpose of the certificate. If a purpose is set then the
10785 new chain verify code is used to check extension consistency.
10788 *) Support for the authority information access extension.
10791 *) Modify RSA and DSA PEM read routines to transparently handle
10792 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10793 public keys in a format compatible with certificate
10794 SubjectPublicKeyInfo structures. Unfortunately there were already
10795 functions called *_PublicKey_* which used various odd formats so
10796 these are retained for compatibility: however the DSA variants were
10797 never in a public release so they have been deleted. Changed dsa/rsa
10798 utilities to handle the new format: note no releases ever handled public
10799 keys so we should be OK.
10801 The primary motivation for this change is to avoid the same fiasco
10802 that dogs private keys: there are several incompatible private key
10803 formats some of which are standard and some OpenSSL specific and
10804 require various evil hacks to allow partial transparent handling and
10805 even then it doesn't work with DER formats. Given the option anything
10806 other than PKCS#8 should be dumped: but the other formats have to
10807 stay in the name of compatibility.
10809 With public keys and the benefit of hindsight one standard format
10810 is used which works with EVP_PKEY, RSA or DSA structures: though
10811 it clearly returns an error if you try to read the wrong kind of key.
10813 Added a -pubkey option to the 'x509' utility to output the public key.
10814 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10815 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10816 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10817 that do the same as the EVP_PKEY_assign_*() except they up the
10818 reference count of the added key (they don't "swallow" the
10822 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10823 CRLs would fail if the file contained no certificates or no CRLs:
10824 added a new function to read in both types and return the number
10825 read: this means that if none are read it will be an error. The
10826 DER versions of the certificate and CRL reader would always fail
10827 because it isn't possible to mix certificates and CRLs in DER format
10828 without choking one or the other routine. Changed this to just read
10829 a certificate: this is the best we can do. Also modified the code
10830 in apps/verify.c to take notice of return codes: it was previously
10831 attempting to read in certificates from NULL pointers and ignoring
10832 any errors: this is one reason why the cert and CRL reader seemed
10833 to work. It doesn't check return codes from the default certificate
10834 routines: these may well fail if the certificates aren't installed.
10837 *) Code to support otherName option in GeneralName.
10840 *) First update to verify code. Change the verify utility
10841 so it warns if it is passed a self signed certificate:
10842 for consistency with the normal behaviour. X509_verify
10843 has been modified to it will now verify a self signed
10844 certificate if *exactly* the same certificate appears
10845 in the store: it was previously impossible to trust a
10846 single self signed certificate. This means that:
10847 openssl verify ss.pem
10848 now gives a warning about a self signed certificate but
10849 openssl verify -CAfile ss.pem ss.pem
10853 *) For servers, store verify_result in SSL_SESSION data structure
10854 (and add it to external session representation).
10855 This is needed when client certificate verifications fails,
10856 but an application-provided verification callback (set by
10857 SSL_CTX_set_cert_verify_callback) allows accepting the session
10858 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10859 but returns 1): When the session is reused, we have to set
10860 ssl->verify_result to the appropriate error code to avoid
10862 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10864 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10865 case in PKCS7_dataInit() where the signed PKCS7 structure
10866 didn't contain any existing data because it was being created.
10867 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10869 *) Add a salt to the key derivation routines in enc.c. This
10870 forms the first 8 bytes of the encrypted file. Also add a
10871 -S option to allow a salt to be input on the command line.
10874 *) New function X509_cmp(). Oddly enough there wasn't a function
10875 to compare two certificates. We do this by working out the SHA1
10876 hash and comparing that. X509_cmp() will be needed by the trust
10880 *) SSL_get1_session() is like SSL_get_session(), but increments
10881 the reference count in the SSL_SESSION returned.
10882 [Geoff Thorpe <geoff@eu.c2.net>]
10884 *) Fix for 'req': it was adding a null to request attributes.
10885 Also change the X509_LOOKUP and X509_INFO code to handle
10886 certificate auxiliary information.
10889 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10893 *) Add the possibility to add extra information to the memory leak
10894 detecting output, to form tracebacks, showing from where each
10895 allocation was originated: CRYPTO_push_info("constant string") adds
10896 the string plus current file name and line number to a per-thread
10897 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10898 is like calling CYRPTO_pop_info() until the stack is empty.
10899 Also updated memory leak detection code to be multi-thread-safe.
10902 *) Add options -text and -noout to pkcs7 utility and delete the
10903 encryption options which never did anything. Update docs.
10906 *) Add options to some of the utilities to allow the pass phrase
10907 to be included on either the command line (not recommended on
10908 OSes like Unix) or read from the environment. Update the
10909 manpages and fix a few bugs.
10912 *) Add a few manpages for some of the openssl commands.
10915 *) Fix the -revoke option in ca. It was freeing up memory twice,
10916 leaking and not finding already revoked certificates.
10919 *) Extensive changes to support certificate auxiliary information.
10920 This involves the use of X509_CERT_AUX structure and X509_AUX
10921 functions. An X509_AUX function such as PEM_read_X509_AUX()
10922 can still read in a certificate file in the usual way but it
10923 will also read in any additional "auxiliary information". By
10924 doing things this way a fair degree of compatibility can be
10925 retained: existing certificates can have this information added
10926 using the new 'x509' options.
10928 Current auxiliary information includes an "alias" and some trust
10929 settings. The trust settings will ultimately be used in enhanced
10930 certificate chain verification routines: currently a certificate
10931 can only be trusted if it is self signed and then it is trusted
10935 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10936 The problem was that one of the replacement routines had not been working
10937 since SSLeay releases. For now the offending routine has been replaced
10938 with non-optimised assembler. Even so, this now gives around 95%
10939 performance improvement for 1024 bit RSA signs.
10942 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10943 handling. Most clients have the effective key size in bits equal to
10944 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10945 A few however don't do this and instead use the size of the decrypted key
10946 to determine the RC2 key length and the AlgorithmIdentifier to determine
10947 the effective key length. In this case the effective key length can still
10948 be 40 bits but the key length can be 168 bits for example. This is fixed
10949 by manually forcing an RC2 key into the EVP_PKEY structure because the
10950 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10951 the key length and effective key length are equal.
10954 *) Add a bunch of functions that should simplify the creation of
10955 X509_NAME structures. Now you should be able to do:
10956 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10957 and have it automatically work out the correct field type and fill in
10958 the structures. The more adventurous can try:
10959 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10960 and it will (hopefully) work out the correct multibyte encoding.
10963 *) Change the 'req' utility to use the new field handling and multibyte
10964 copy routines. Before the DN field creation was handled in an ad hoc
10965 way in req, ca, and x509 which was rather broken and didn't support
10966 BMPStrings or UTF8Strings. Since some software doesn't implement
10967 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10968 using the dirstring_type option. See the new comment in the default
10969 openssl.cnf for more info.
10972 *) Make crypto/rand/md_rand.c more robust:
10973 - Assure unique random numbers after fork().
10974 - Make sure that concurrent threads access the global counter and
10975 md serializably so that we never lose entropy in them
10976 or use exactly the same state in multiple threads.
10977 Access to the large state is not always serializable because
10978 the additional locking could be a performance killer, and
10979 md should be large enough anyway.
10982 *) New file apps/app_rand.c with commonly needed functionality
10983 for handling the random seed file.
10985 Use the random seed file in some applications that previously did not:
10987 dsaparam -genkey (which also ignored its '-rand' option),
10990 x509 (when signing).
10991 Except on systems with /dev/urandom, it is crucial to have a random
10992 seed file at least for key creation, DSA signing, and for DH exchanges;
10993 for RSA signatures we could do without one.
10995 gendh and gendsa (unlike genrsa) used to read only the first byte
10996 of each file listed in the '-rand' option. The function as previously
10997 found in genrsa is now in app_rand.c and is used by all programs
10998 that support '-rand'.
11001 *) In RAND_write_file, use mode 0600 for creating files;
11002 don't just chmod when it may be too late.
11005 *) Report an error from X509_STORE_load_locations
11006 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11009 *) New function ASN1_mbstring_copy() this copies a string in either
11010 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11011 into an ASN1_STRING type. A mask of permissible types is passed
11012 and it chooses the "minimal" type to use or an error if not type
11016 *) Add function equivalents to the various macros in asn1.h. The old
11017 macros are retained with an M_ prefix. Code inside the library can
11018 use the M_ macros. External code (including the openssl utility)
11019 should *NOT* in order to be "shared library friendly".
11022 *) Add various functions that can check a certificate's extensions
11023 to see if it usable for various purposes such as SSL client,
11024 server or S/MIME and CAs of these types. This is currently
11025 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11026 verification. Also added a -purpose flag to x509 utility to
11027 print out all the purposes.
11030 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11034 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11035 for, obtain and decode and extension and obtain its critical flag.
11036 This allows all the necessary extension code to be handled in a
11037 single function call.
11040 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11041 platforms. See crypto/rc4/rc4_enc.c for further details.
11044 *) New -noout option to asn1parse. This causes no output to be produced
11045 its main use is when combined with -strparse and -out to extract data
11046 from a file (which may not be in ASN.1 format).
11049 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11050 when producing the local key id.
11051 [Richard Levitte <levitte@stacken.kth.se>]
11053 *) New option -dhparam in s_server. This allows a DH parameter file to be
11054 stated explicitly. If it is not stated then it tries the first server
11055 certificate file. The previous behaviour hard coded the filename
11059 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11060 a public key to be input or output. For example:
11061 openssl rsa -in key.pem -pubout -out pubkey.pem
11062 Also added necessary DSA public key functions to handle this.
11065 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11066 in the message. This was handled by allowing
11067 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11068 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11070 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11071 to the end of the strings whereas this didn't. This would cause problems
11072 if strings read with d2i_ASN1_bytes() were later modified.
11073 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11075 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11076 data and it contains EOF it will end up returning an error. This is
11077 caused by input 46 bytes long. The cause is due to the way base64
11078 BIOs find the start of base64 encoded data. They do this by trying a
11079 trial decode on each line until they find one that works. When they
11080 do a flag is set and it starts again knowing it can pass all the
11081 data directly through the decoder. Unfortunately it doesn't reset
11082 the context it uses. This means that if EOF is reached an attempt
11083 is made to pass two EOFs through the context and this causes the
11084 resulting error. This can also cause other problems as well. As is
11085 usual with these problems it takes *ages* to find and the fix is
11086 trivial: move one line.
11087 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11089 *) Ugly workaround to get s_client and s_server working under Windows. The
11090 old code wouldn't work because it needed to select() on sockets and the
11091 tty (for keypresses and to see if data could be written). Win32 only
11092 supports select() on sockets so we select() with a 1s timeout on the
11093 sockets and then see if any characters are waiting to be read, if none
11094 are present then we retry, we also assume we can always write data to
11095 the tty. This isn't nice because the code then blocks until we've
11096 received a complete line of data and it is effectively polling the
11097 keyboard at 1s intervals: however it's quite a bit better than not
11098 working at all :-) A dedicated Windows application might handle this
11099 with an event loop for example.
11102 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11103 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11104 will be called when RSA_sign() and RSA_verify() are used. This is useful
11105 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11106 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11107 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11108 This necessitated the support of an extra signature type NID_md5_sha1
11109 for SSL signatures and modifications to the SSL library to use it instead
11110 of calling RSA_public_decrypt() and RSA_private_encrypt().
11113 *) Add new -verify -CAfile and -CApath options to the crl program, these
11114 will lookup a CRL issuers certificate and verify the signature in a
11115 similar way to the verify program. Tidy up the crl program so it
11116 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11117 less strict. It will now permit CRL extensions even if it is not
11118 a V2 CRL: this will allow it to tolerate some broken CRLs.
11121 *) Initialize all non-automatic variables each time one of the openssl
11122 sub-programs is started (this is necessary as they may be started
11123 multiple times from the "OpenSSL>" prompt).
11124 [Lennart Bang, Bodo Moeller]
11126 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11127 removing all other RSA functionality (this is what NO_RSA does). This
11128 is so (for example) those in the US can disable those operations covered
11129 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11133 *) Non-copying interface to BIO pairs.
11134 (still largely untested)
11137 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11138 ASCII string. This was handled independently in various places before.
11141 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11142 UTF8 strings a character at a time.
11145 *) Use client_version from client hello to select the protocol
11146 (s23_srvr.c) and for RSA client key exchange verification
11147 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11150 *) Add various utility functions to handle SPKACs, these were previously
11151 handled by poking round in the structure internals. Added new function
11152 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11153 print, verify and generate SPKACs. Based on an original idea from
11154 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11157 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11160 *) Allow the config file extension section to be overwritten on the
11161 command line. Based on an original idea from Massimiliano Pala
11162 <madwolf@comune.modena.it>. The new option is called -extensions
11163 and can be applied to ca, req and x509. Also -reqexts to override
11164 the request extensions in req and -crlexts to override the crl extensions
11168 *) Add new feature to the SPKAC handling in ca. Now you can include
11169 the same field multiple times by preceding it by "XXXX." for example:
11172 this is the same syntax as used in the req config file.
11175 *) Allow certificate extensions to be added to certificate requests. These
11176 are specified in a 'req_extensions' option of the req section of the
11177 config file. They can be printed out with the -text option to req but
11178 are otherwise ignored at present.
11181 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11182 data read consists of only the final block it would not decrypted because
11183 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11184 A misplaced 'break' also meant the decrypted final block might not be
11185 copied until the next read.
11188 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11189 a few extra parameters to the DH structure: these will be useful if
11190 for example we want the value of 'q' or implement X9.42 DH.
11193 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11194 provides hooks that allow the default DSA functions or functions on a
11195 "per key" basis to be replaced. This allows hardware acceleration and
11196 hardware key storage to be handled without major modification to the
11197 library. Also added low level modexp hooks and CRYPTO_EX structure and
11198 associated functions.
11201 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11202 as "read only": it can't be written to and the buffer it points to will
11203 not be freed. Reading from a read only BIO is much more efficient than
11204 a normal memory BIO. This was added because there are several times when
11205 an area of memory needs to be read from a BIO. The previous method was
11206 to create a memory BIO and write the data to it, this results in two
11207 copies of the data and an O(n^2) reading algorithm. There is a new
11208 function BIO_new_mem_buf() which creates a read only memory BIO from
11209 an area of memory. Also modified the PKCS#7 routines to use read only
11213 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11214 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11215 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11216 but a retry condition occurred while trying to read the rest.
11219 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11220 NID_pkcs7_encrypted by default: this was wrong since this should almost
11221 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11222 the encrypted data type: this is a more sensible place to put it and it
11223 allows the PKCS#12 code to be tidied up that duplicated this
11227 *) Changed obj_dat.pl script so it takes its input and output files on
11228 the command line. This should avoid shell escape redirection problems
11232 *) Initial support for certificate extension requests, these are included
11233 in things like Xenroll certificate requests. Included functions to allow
11234 extensions to be obtained and added.
11237 *) -crlf option to s_client and s_server for sending newlines as
11238 CRLF (as required by many protocols).
11241 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11243 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11244 [Ralf S. Engelschall]
11246 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11247 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11249 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11253 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11254 DH parameters/keys (q is lost during that conversion, but the resulting
11255 DH parameters contain its length).
11257 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11258 much faster than DH_generate_parameters (which creates parameters
11259 where p = 2*q + 1), and also the smaller q makes DH computations
11260 much more efficient (160-bit exponentiation instead of 1024-bit
11261 exponentiation); so this provides a convenient way to support DHE
11262 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11263 utter importance to use
11264 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11266 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11267 when such DH parameters are used, because otherwise small subgroup
11268 attacks may become possible!
11271 *) Avoid memory leak in i2d_DHparams.
11274 *) Allow the -k option to be used more than once in the enc program:
11275 this allows the same encrypted message to be read by multiple recipients.
11278 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11279 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11280 it will always use the numerical form of the OID, even if it has a short
11284 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11285 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11286 otherwise bn_mod_exp was called. In the case of hardware keys for example
11287 no private key components need be present and it might store extra data
11288 in the RSA structure, which cannot be accessed from bn_mod_exp.
11289 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11290 private key operations.
11293 *) Added support for SPARC Linux.
11296 *) pem_password_cb function type incompatibly changed from
11297 typedef int pem_password_cb(char *buf, int size, int rwflag);
11299 ....(char *buf, int size, int rwflag, void *userdata);
11300 so that applications can pass data to their callbacks:
11301 The PEM[_ASN1]_{read,write}... functions and macros now take an
11302 additional void * argument, which is just handed through whenever
11303 the password callback is called.
11304 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11306 New function SSL_CTX_set_default_passwd_cb_userdata.
11308 Compatibility note: As many C implementations push function arguments
11309 onto the stack in reverse order, the new library version is likely to
11310 interoperate with programs that have been compiled with the old
11311 pem_password_cb definition (PEM_whatever takes some data that
11312 happens to be on the stack as its last argument, and the callback
11313 just ignores this garbage); but there is no guarantee whatsoever that
11316 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11317 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11318 problems not only on Windows, but also on some Unix platforms.
11319 To avoid problematic command lines, these definitions are now in an
11320 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11321 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11324 *) MIPS III/IV assembler module is reimplemented.
11327 *) More DES library cleanups: remove references to srand/rand and
11328 delete an unused file.
11331 *) Add support for the free Netwide assembler (NASM) under Win32,
11332 since not many people have MASM (ml) and it can be hard to obtain.
11333 This is currently experimental but it seems to work OK and pass all
11334 the tests. Check out INSTALL.W32 for info.
11337 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11338 without temporary keys kept an extra copy of the server key,
11339 and connections with temporary keys did not free everything in case
11343 *) New function RSA_check_key and new openssl rsa option -check
11344 for verifying the consistency of RSA keys.
11345 [Ulf Moeller, Bodo Moeller]
11347 *) Various changes to make Win32 compile work:
11348 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11349 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11350 comparison" warnings.
11351 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11354 *) Add a debugging option to PKCS#5 v2 key generation function: when
11355 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11356 derived keys are printed to stderr.
11359 *) Copy the flags in ASN1_STRING_dup().
11360 [Roman E. Pavlov <pre@mo.msk.ru>]
11362 *) The x509 application mishandled signing requests containing DSA
11363 keys when the signing key was also DSA and the parameters didn't match.
11365 It was supposed to omit the parameters when they matched the signing key:
11366 the verifying software was then supposed to automatically use the CA's
11367 parameters if they were absent from the end user certificate.
11369 Omitting parameters is no longer recommended. The test was also
11370 the wrong way round! This was probably due to unusual behaviour in
11371 EVP_cmp_parameters() which returns 1 if the parameters match.
11372 This meant that parameters were omitted when they *didn't* match and
11373 the certificate was useless. Certificates signed with 'ca' didn't have
11375 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11377 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11378 The interface is as follows:
11379 Applications can use
11380 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11381 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11382 "off" is now the default.
11383 The library internally uses
11384 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11385 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11386 to disable memory-checking temporarily.
11388 Some inconsistent states that previously were possible (and were
11389 even the default) are now avoided.
11391 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11392 with each memory chunk allocated; this is occasionally more helpful
11393 than just having a counter.
11395 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11397 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11401 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11402 which largely parallels "options", but is for changing API behaviour,
11403 whereas "options" are about protocol behaviour.
11404 Initial "mode" flags are:
11406 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11407 a single record has been written.
11408 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11409 retries use the same buffer location.
11410 (But all of the contents must be
11414 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11417 *) Fix problems with no-hmac etc.
11418 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11420 *) New functions RSA_get_default_method(), RSA_set_method() and
11421 RSA_get_method(). These allows replacement of RSA_METHODs without having
11422 to mess around with the internals of an RSA structure.
11425 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11426 Also really enable memory leak checks in openssl.c and in some
11428 [Chad C. Mulligan, Bodo Moeller]
11430 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11431 up the length of negative integers. This has now been simplified to just
11432 store the length when it is first determined and use it later, rather
11433 than trying to keep track of where data is copied and updating it to
11435 [Steve Henson, reported by Brien Wheeler
11436 <bwheeler@authentica-security.com>]
11438 *) Add a new function PKCS7_signatureVerify. This allows the verification
11439 of a PKCS#7 signature but with the signing certificate passed to the
11440 function itself. This contrasts with PKCS7_dataVerify which assumes the
11441 certificate is present in the PKCS#7 structure. This isn't always the
11442 case: certificates can be omitted from a PKCS#7 structure and be
11443 distributed by "out of band" means (such as a certificate database).
11446 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11447 function prototypes in pem.h, also change util/mkdef.pl to add the
11448 necessary function names.
11451 *) mk1mf.pl (used by Windows builds) did not properly read the
11452 options set by Configure in the top level Makefile, and Configure
11453 was not even able to write more than one option correctly.
11454 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11457 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11458 file to be loaded from a BIO or FILE pointer. The BIO version will
11459 for example allow memory BIOs to contain config info.
11462 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11463 Whoever hopes to achieve shared-library compatibility across versions
11464 must use this, not the compile-time macro.
11465 (Exercise 0.9.4: Which is the minimum library version required by
11467 Note: All this applies only to multi-threaded programs, others don't
11471 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11472 through a BIO pair triggered the default case, i.e.
11473 SSLerr(...,SSL_R_UNKNOWN_STATE).
11476 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11477 can use the SSL library even if none of the specific BIOs is
11481 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11482 for the encoded length.
11483 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11485 *) Add initial documentation of the X509V3 functions.
11488 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11489 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11490 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11491 secure PKCS#8 private key format with a high iteration count.
11494 *) Fix determination of Perl interpreter: A perl or perl5
11495 _directory_ in $PATH was also accepted as the interpreter.
11496 [Ralf S. Engelschall]
11498 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11499 wrong with it but it was very old and did things like calling
11500 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11501 unusual formatting.
11504 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11505 to use the new extension code.
11508 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11509 with macros. This should make it easier to change their form, add extra
11510 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11514 *) Add to configuration table a new entry that can specify an alternative
11515 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11516 according to Mark Crispin <MRC@Panda.COM>.
11520 *) DES CBC did not update the IV. Weird.
11523 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11524 Changing the behaviour of the former might break existing programs --
11525 where IV updating is needed, des_ncbc_encrypt can be used.
11528 *) When bntest is run from "make test" it drives bc to check its
11529 calculations, as well as internally checking them. If an internal check
11530 fails, it needs to cause bc to give a non-zero result or make test carries
11531 on without noticing the failure. Fixed.
11534 *) DES library cleanups.
11537 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11538 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11539 ciphers. NOTE: although the key derivation function has been verified
11540 against some published test vectors it has not been extensively tested
11541 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11545 *) Instead of "mkdir -p", which is not fully portable, use new
11546 Perl script "util/mkdir-p.pl".
11549 *) Rewrite the way password based encryption (PBE) is handled. It used to
11550 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11551 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11552 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11553 the 'parameter' field of the AlgorithmIdentifier is passed to the
11554 underlying key generation function so it must do its own ASN1 parsing.
11555 This has also changed the EVP_PBE_CipherInit() function which now has a
11556 'parameter' argument instead of literal salt and iteration count values
11557 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11560 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11561 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11562 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11563 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11564 value was just used as a "magic string" and not used directly its
11565 value doesn't matter.
11568 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11572 *) "linux-sparc64" configuration (ultrapenguin).
11573 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11574 "linux-sparc" configuration.
11575 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11577 *) config now generates no-xxx options for missing ciphers.
11580 *) Support the EBCDIC character set (work in progress).
11581 File ebcdic.c not yet included because it has a different license.
11582 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11584 *) Support BS2000/OSD-POSIX.
11585 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11587 *) Make callbacks for key generation use void * instead of char *.
11590 *) Make S/MIME samples compile (not yet tested).
11593 *) Additional typesafe stacks.
11596 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11600 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11602 *) New configuration variant "sco5-gcc".
11604 *) Updated some demos.
11605 [Sean O Riordain, Wade Scholine]
11607 *) Add missing BIO_free at exit of pkcs12 application.
11610 *) Fix memory leak in conf.c.
11613 *) Updates for Win32 to assembler version of MD5.
11616 *) Set #! path to perl in apps/der_chop to where we found it
11617 instead of using a fixed path.
11620 *) SHA library changes for irix64-mips4-cc.
11623 *) Improvements for VMS support.
11627 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11629 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11630 This also avoids the problems with SC4.2 and unpatched SC5.
11631 [Andy Polyakov <appro@fy.chalmers.se>]
11633 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11634 These are required because of the typesafe stack would otherwise break
11635 existing code. If old code used a structure member which used to be STACK
11636 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11637 sk_num or sk_value it would produce an error because the num, data members
11638 are not present in STACK_OF. Now it just produces a warning. sk_set
11639 replaces the old method of assigning a value to sk_value
11640 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11641 that does this will no longer work (and should use sk_set instead) but
11642 this could be regarded as a "questionable" behaviour anyway.
11645 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11646 correctly handle encrypted S/MIME data.
11649 *) Change type of various DES function arguments from des_cblock
11650 (which means, in function argument declarations, pointer to char)
11651 to des_cblock * (meaning pointer to array with 8 char elements),
11652 which allows the compiler to do more typechecking; it was like
11653 that back in SSLeay, but with lots of ugly casts.
11655 Introduce new type const_des_cblock.
11658 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11659 problems: find RecipientInfo structure that matches recipient certificate
11660 and initialise the ASN1 structures properly based on passed cipher.
11663 *) Belatedly make the BN tests actually check the results.
11666 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11667 to and from BNs: it was completely broken. New compilation option
11668 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11669 key elements as negative integers.
11672 *) Reorganize and speed up MD5.
11673 [Andy Polyakov <appro@fy.chalmers.se>]
11676 [Richard Levitte <richard@levitte.org>]
11678 *) New option -out to asn1parse to allow the parsed structure to be
11679 output to a file. This is most useful when combined with the -strparse
11680 option to examine the output of things like OCTET STRINGS.
11683 *) Make SSL library a little more fool-proof by not requiring any longer
11684 that SSL_set_{accept,connect}_state be called before
11685 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11686 in many applications because usually everything *appeared* to work as
11687 intended anyway -- now it really works as intended).
11690 *) Move openssl.cnf out of lib/.
11693 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11694 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11695 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11696 [Ralf S. Engelschall]
11698 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11699 handle PKCS#7 enveloped data properly.
11700 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11702 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11703 copying pointers. The cert_st handling is changed by this in
11704 various ways (and thus what used to be known as ctx->default_cert
11705 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11706 any longer when s->cert does not give us what we need).
11707 ssl_cert_instantiate becomes obsolete by this change.
11708 As soon as we've got the new code right (possibly it already is?),
11709 we have solved a couple of bugs of the earlier code where s->cert
11710 was used as if it could not have been shared with other SSL structures.
11712 Note that using the SSL API in certain dirty ways now will result
11713 in different behaviour than observed with earlier library versions:
11714 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11715 does not influence s as it used to.
11717 In order to clean up things more thoroughly, inside SSL_SESSION
11718 we don't use CERT any longer, but a new structure SESS_CERT
11719 that holds per-session data (if available); currently, this is
11720 the peer's certificate chain and, for clients, the server's certificate
11721 and temporary key. CERT holds only those values that can have
11722 meaningful defaults in an SSL_CTX.
11725 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11726 from the internal representation. Various PKCS#7 fixes: remove some
11727 evil casts and set the enc_dig_alg field properly based on the signing
11731 *) Allow PKCS#12 password to be set from the command line or the
11732 environment. Let 'ca' get its config file name from the environment
11733 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11737 *) Allow certificate policies extension to use an IA5STRING for the
11738 organization field. This is contrary to the PKIX definition but
11739 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11743 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11744 without disallowing inline assembler and the like for non-pedantic builds.
11747 *) Support Borland C++ builder.
11748 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11750 *) Support Mingw32.
11753 *) SHA-1 cleanups and performance enhancements.
11754 [Andy Polyakov <appro@fy.chalmers.se>]
11756 *) Sparc v8plus assembler for the bignum library.
11757 [Andy Polyakov <appro@fy.chalmers.se>]
11759 *) Accept any -xxx and +xxx compiler options in Configure.
11762 *) Update HPUX configuration.
11765 *) Add missing sk_<type>_unshift() function to safestack.h
11766 [Ralf S. Engelschall]
11768 *) New function SSL_CTX_use_certificate_chain_file that sets the
11769 "extra_cert"s in addition to the certificate. (This makes sense
11770 only for "PEM" format files, as chains as a whole are not
11774 *) Support verify_depth from the SSL API.
11775 x509_vfy.c had what can be considered an off-by-one-error:
11776 Its depth (which was not part of the external interface)
11777 was actually counting the number of certificates in a chain;
11778 now it really counts the depth.
11781 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11782 instead of X509err, which often resulted in confusing error
11783 messages since the error codes are not globally unique
11784 (e.g. an alleged error in ssl3_accept when a certificate
11785 didn't match the private key).
11787 *) New function SSL_CTX_set_session_id_context that allows to set a default
11788 value (so that you don't need SSL_set_session_id_context for each
11789 connection using the SSL_CTX).
11792 *) OAEP decoding bug fix.
11795 *) Support INSTALL_PREFIX for package builders, as proposed by
11799 *) New Configure options "threads" and "no-threads". For systems
11800 where the proper compiler options are known (currently Solaris
11801 and Linux), "threads" is the default.
11804 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11807 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11808 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11809 such as /usr/local/bin.
11812 *) "make linux-shared" to build shared libraries.
11813 [Niels Poppe <niels@netbox.org>]
11815 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11818 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11819 extension adding in x509 utility.
11822 *) Remove NOPROTO sections and error code comments.
11825 *) Partial rewrite of the DEF file generator to now parse the ANSI
11829 *) New Configure options --prefix=DIR and --openssldir=DIR.
11832 *) Complete rewrite of the error code script(s). It is all now handled
11833 by one script at the top level which handles error code gathering,
11834 header rewriting and C source file generation. It should be much better
11835 than the old method: it now uses a modified version of Ulf's parser to
11836 read the ANSI prototypes in all header files (thus the old K&R definitions
11837 aren't needed for error creation any more) and do a better job of
11838 translating function codes into names. The old 'ASN1 error code imbedded
11839 in a comment' is no longer necessary and it doesn't use .err files which
11840 have now been deleted. Also the error code call doesn't have to appear all
11841 on one line (which resulted in some large lines...).
11844 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11847 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11848 0 (which usually indicates a closed connection), but continue reading.
11851 *) Fix some race conditions.
11854 *) Add support for CRL distribution points extension. Add Certificate
11855 Policies and CRL distribution points documentation.
11858 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11861 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11862 8 of keying material. Merlin has also confirmed interop with this fix
11863 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11864 [Merlin Hughes <merlin@baltimore.ie>]
11866 *) Fix lots of warnings.
11867 [Richard Levitte <levitte@stacken.kth.se>]
11869 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11870 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11871 [Richard Levitte <levitte@stacken.kth.se>]
11873 *) Fix problems with sizeof(long) == 8.
11874 [Andy Polyakov <appro@fy.chalmers.se>]
11876 *) Change functions to ANSI C.
11879 *) Fix typos in error codes.
11880 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11882 *) Remove defunct assembler files from Configure.
11885 *) SPARC v8 assembler BIGNUM implementation.
11886 [Andy Polyakov <appro@fy.chalmers.se>]
11888 *) Support for Certificate Policies extension: both print and set.
11889 Various additions to support the r2i method this uses.
11892 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11893 return a const string when you are expecting an allocated buffer.
11896 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11897 types DirectoryString and DisplayText.
11900 *) Add code to allow r2i extensions to access the configuration database,
11901 add an LHASH database driver and add several ctx helper functions.
11904 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11905 fail when they extended the size of a BIGNUM.
11908 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11909 support typesafe stack.
11912 *) Fix typo in SSL_[gs]et_options().
11913 [Nils Frostberg <nils@medcom.se>]
11915 *) Delete various functions and files that belonged to the (now obsolete)
11916 old X509V3 handling code.
11919 *) New Configure option "rsaref".
11922 *) Don't auto-generate pem.h.
11925 *) Introduce type-safe ASN.1 SETs.
11928 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11929 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11931 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11932 that links with OpenSSL (well at least cause lots of warnings), but fear
11933 not: the conversion is trivial, and it eliminates loads of evil casts. A
11934 few STACKed things have been converted already. Feel free to convert more.
11935 In the fullness of time, I'll do away with the STACK type altogether.
11938 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11939 specified in <certfile> by updating the entry in the index.txt file.
11940 This way one no longer has to edit the index.txt file manually for
11941 revoking a certificate. The -revoke option does the gory details now.
11942 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11944 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11945 `-text' option at all and this way the `-noout -text' combination was
11946 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11947 [Ralf S. Engelschall]
11949 *) Make sure a corresponding plain text error message exists for the
11950 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11951 verify callback function determined that a certificate was revoked.
11952 [Ralf S. Engelschall]
11954 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11955 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11956 all available ciphers including rc5, which was forgotten until now.
11957 In order to let the testing shell script know which algorithms
11958 are available, a new (up to now undocumented) command
11959 "openssl list-cipher-commands" is used.
11962 *) Bugfix: s_client occasionally would sleep in select() when
11963 it should have checked SSL_pending() first.
11966 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11967 the raw DSA values prior to ASN.1 encoding.
11970 *) Tweaks to Configure
11971 [Niels Poppe <niels@netbox.org>]
11973 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11977 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11980 *) New config option to avoid instructions that are illegal on the 80386.
11981 The default code is faster, but requires at least a 486.
11984 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11985 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11986 same as SSL2_VERSION anyway.
11989 *) New "-showcerts" option for s_client.
11992 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11993 application. Various cleanups and fixes.
11996 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11997 modify error routines to work internally. Add error codes and PBE init
11998 to library startup routines.
12001 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12002 packing functions to asn1 and evp. Changed function names and error
12003 codes along the way.
12006 *) PKCS12 integration: and so it begins... First of several patches to
12007 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12008 objects to objects.h
12011 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12012 and display support for Thawte strong extranet extension.
12015 *) Add LinuxPPC support.
12016 [Jeff Dubrule <igor@pobox.org>]
12018 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12019 bn_div_words in alpha.s.
12020 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12022 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12023 OAEP isn't supported when OpenSSL is built with RSAref.
12024 [Ulf Moeller <ulf@fitug.de>]
12026 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12027 so they no longer are missing under -DNOPROTO.
12028 [Soren S. Jorvang <soren@t.dk>]
12031 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12033 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12034 doesn't work when the session is reused. Coming soon!
12037 *) Fix a security hole, that allows sessions to be reused in the wrong
12038 context thus bypassing client cert protection! All software that uses
12039 client certs and session caches in multiple contexts NEEDS PATCHING to
12040 allow session reuse! A fuller solution is in the works.
12041 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12043 *) Some more source tree cleanups (removed obsolete files
12044 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12045 permission on "config" script to be executable) and a fix for the INSTALL
12047 [Ulf Moeller <ulf@fitug.de>]
12049 *) Remove some legacy and erroneous uses of malloc, free instead of
12051 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12053 *) Make rsa_oaep_test return non-zero on error.
12054 [Ulf Moeller <ulf@fitug.de>]
12056 *) Add support for native Solaris shared libraries. Configure
12057 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12058 if someone would make that last step automatic.
12059 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12061 *) ctx_size was not built with the right compiler during "make links". Fixed.
12064 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12065 except NULL ciphers". This means the default cipher list will no longer
12066 enable NULL ciphers. They need to be specifically enabled e.g. with
12067 the string "DEFAULT:eNULL".
12070 *) Fix to RSA private encryption routines: if p < q then it would
12071 occasionally produce an invalid result. This will only happen with
12072 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12075 *) Be less restrictive and allow also `perl util/perlpath.pl
12076 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12077 because this way one can also use an interpreter named `perl5' (which is
12078 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12079 installed as `perl').
12080 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12082 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12083 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12085 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12086 advapi32.lib to Win32 build and change the pem test comparison
12087 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12088 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12089 and crypto/des/ede_cbcm_enc.c.
12092 *) DES quad checksum was broken on big-endian architectures. Fixed.
12095 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12096 Win32 test batch file so it (might) work again. The Win32 test batch file
12097 is horrible: I feel ill....
12100 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12101 in e_os.h. Audit of header files to check ANSI and non ANSI
12102 sections: 10 functions were absent from non ANSI section and not exported
12103 from Windows DLLs. Fixed up libeay.num for new functions.
12106 *) Make `openssl version' output lines consistent.
12107 [Ralf S. Engelschall]
12109 *) Fix Win32 symbol export lists for BIO functions: Added
12110 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12111 to ms/libeay{16,32}.def.
12112 [Ralf S. Engelschall]
12114 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12115 fine under Unix and passes some trivial tests I've now added. But the
12116 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12117 added to make sure no one expects that this stuff really works in the
12118 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12119 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12121 [Ralf S. Engelschall]
12123 *) Fix the generation of two part addresses in perl.
12124 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12126 *) Add config entry for Linux on MIPS.
12127 [John Tobey <jtobey@channel1.com>]
12129 *) Make links whenever Configure is run, unless we are on Windoze.
12132 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12133 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12137 *) Add a useful kludge to allow package maintainers to specify compiler and
12138 other platforms details on the command line without having to patch the
12139 Configure script everytime: One now can use ``perl Configure
12140 <id>:<details>'', i.e. platform ids are allowed to have details appended
12141 to them (separated by colons). This is treated as there would be a static
12142 pre-configured entry in Configure's %table under key <id> with value
12143 <details> and ``perl Configure <id>'' is called. So, when you want to
12144 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12145 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12146 now, which overrides the FreeBSD-elf entry on-the-fly.
12147 [Ralf S. Engelschall]
12149 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12152 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12153 on the `perl Configure ...' command line. This way one can compile
12154 OpenSSL libraries with Position Independent Code (PIC) which is needed
12155 for linking it into DSOs.
12156 [Ralf S. Engelschall]
12158 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12162 *) Cleaned up the LICENSE document: The official contact for any license
12163 questions now is the OpenSSL core team under openssl-core@openssl.org.
12164 And add a paragraph about the dual-license situation to make sure people
12165 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12166 to the OpenSSL toolkit.
12167 [Ralf S. Engelschall]
12169 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12170 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12171 Additionally cleaned up the `make links' target: Remove unnecessary
12172 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12173 to speed processing and no longer clutter the display with confusing
12174 stuff. Instead only the actually done links are displayed.
12175 [Ralf S. Engelschall]
12177 *) Permit null encryption ciphersuites, used for authentication only. It used
12178 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12179 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12183 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12184 signed attributes when verifying signatures (this would break them),
12185 the detached data encoding was wrong and public keys obtained using
12186 X509_get_pubkey() weren't freed.
12189 *) Add text documentation for the BUFFER functions. Also added a work around
12190 to a Win95 console bug. This was triggered by the password read stuff: the
12191 last character typed gets carried over to the next fread(). If you were
12192 generating a new cert request using 'req' for example then the last
12193 character of the passphrase would be CR which would then enter the first
12197 *) Added the new `Includes OpenSSL Cryptography Software' button as
12198 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12199 button and can be used by applications based on OpenSSL to show the
12200 relationship to the OpenSSL project.
12201 [Ralf S. Engelschall]
12203 *) Remove confusing variables in function signatures in files
12204 ssl/ssl_lib.c and ssl/ssl.h.
12205 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12207 *) Don't install bss_file.c under PREFIX/include/
12208 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12210 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12211 functions that return function pointers and has support for NT specific
12212 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12213 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12214 unsigned to signed types: this was killing the Win32 compile.
12217 *) Add new certificate file to stack functions,
12218 SSL_add_dir_cert_subjects_to_stack() and
12219 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12220 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12221 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12222 This means that Apache-SSL and similar packages don't have to mess around
12223 to add as many CAs as they want to the preferred list.
12226 *) Experiment with doxygen documentation. Currently only partially applied to
12228 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12229 openssl.doxy as the configuration file.
12232 *) Get rid of remaining C++-style comments which strict C compilers hate.
12233 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12235 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12236 compiled in by default: it has problems with large keys.
12239 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12240 DH private keys and/or callback functions which directly correspond to
12241 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12242 is needed for applications which have to configure certificates on a
12243 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12245 For the RSA certificate situation is makes no difference, but
12246 for the DSA certificate situation this fixes the "no shared cipher"
12247 problem where the OpenSSL cipher selection procedure failed because the
12248 temporary keys were not overtaken from the context and the API provided
12249 no way to reconfigure them.
12250 The new functions now let applications reconfigure the stuff and they
12251 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12252 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12253 non-public-API function ssl_cert_instantiate() is used as a helper
12254 function and also to reduce code redundancy inside ssl_rsa.c.
12255 [Ralf S. Engelschall]
12257 *) Move s_server -dcert and -dkey options out of the undocumented feature
12258 area because they are useful for the DSA situation and should be
12259 recognized by the users.
12260 [Ralf S. Engelschall]
12262 *) Fix the cipher decision scheme for export ciphers: the export bits are
12263 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12264 SSL_EXP_MASK. So, the original variable has to be used instead of the
12265 already masked variable.
12266 [Richard Levitte <levitte@stacken.kth.se>]
12268 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12269 [Richard Levitte <levitte@stacken.kth.se>]
12271 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12272 from `int' to `unsigned int' because it's a length and initialized by
12273 EVP_DigestFinal() which expects an `unsigned int *'.
12274 [Richard Levitte <levitte@stacken.kth.se>]
12276 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12277 script. Instead use the usual Shell->Perl transition trick.
12278 [Ralf S. Engelschall]
12280 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12281 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12282 -noout -modulus' as it's already the case for `openssl rsa -noout
12283 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12284 currently the public key is printed (a decision which was already done by
12285 `openssl dsa -modulus' in the past) which serves a similar purpose.
12286 Additionally the NO_RSA no longer completely removes the whole -modulus
12287 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12289 [Ralf S. Engelschall]
12291 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12292 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12293 [Arne Ansper <arne@ats.cyber.ee>]
12295 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12296 to be added. Now both 'req' and 'ca' can use new objects defined in the
12300 *) Add cool BIO that does syslog (or event log on NT).
12301 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12303 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12304 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12305 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12306 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12309 *) Add preliminary config info for new extension code.
12312 *) Make RSA_NO_PADDING really use no padding.
12313 [Ulf Moeller <ulf@fitug.de>]
12315 *) Generate errors when private/public key check is done.
12318 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12319 for some CRL extensions and new objects added.
12322 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12323 key usage extension and fuller support for authority key id.
12326 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12327 padding method for RSA, which is recommended for new applications in PKCS
12328 #1 v2.0 (RFC 2437, October 1998).
12329 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12330 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12331 against Bleichbacher's attack on RSA.
12332 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12335 *) Updates to the new SSL compression code
12336 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12338 *) Fix so that the version number in the master secret, when passed
12339 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12340 (because the server will not accept higher), that the version number
12341 is 0x03,0x01, not 0x03,0x00
12342 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12344 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12345 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12346 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12349 *) Support for RAW extensions where an arbitrary extension can be
12350 created by including its DER encoding. See apps/openssl.cnf for
12354 *) Make sure latest Perl versions don't interpret some generated C array
12355 code as Perl array code in the crypto/err/err_genc.pl script.
12356 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12358 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12359 not many people have the assembler. Various Win32 compilation fixes and
12360 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12361 build instructions.
12364 *) Modify configure script 'Configure' to automatically create crypto/date.h
12365 file under Win32 and also build pem.h from pem.org. New script
12366 util/mkfiles.pl to create the MINFO file on environments that can't do a
12367 'make files': perl util/mkfiles.pl >MINFO should work.
12370 *) Major rework of DES function declarations, in the pursuit of correctness
12371 and purity. As a result, many evil casts evaporated, and some weirdness,
12372 too. You may find this causes warnings in your code. Zapping your evil
12373 casts will probably fix them. Mostly.
12376 *) Fix for a typo in asn1.h. Bug fix to object creation script
12377 obj_dat.pl. It considered a zero in an object definition to mean
12378 "end of object": none of the objects in objects.h have any zeros
12379 so it wasn't spotted.
12380 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12382 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12383 Masking (CBCM). In the absence of test vectors, the best I have been able
12384 to do is check that the decrypt undoes the encrypt, so far. Send me test
12385 vectors if you have them.
12388 *) Correct calculation of key length for export ciphers (too much space was
12389 allocated for null ciphers). This has not been tested!
12392 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12393 message is now correct (it understands "crypto" and "ssl" on its
12394 command line). There is also now an "update" option. This will update
12395 the util/ssleay.num and util/libeay.num files with any new functions.
12397 perl util/mkdef.pl crypto ssl update
12398 it will update them.
12401 *) Overhauled the Perl interface (perl/*):
12402 - ported BN stuff to OpenSSL's different BN library
12403 - made the perl/ source tree CVS-aware
12404 - renamed the package from SSLeay to OpenSSL (the files still contain
12405 their history because I've copied them in the repository)
12406 - removed obsolete files (the test scripts will be replaced
12407 by better Test::Harness variants in the future)
12408 [Ralf S. Engelschall]
12410 *) First cut for a very conservative source tree cleanup:
12411 1. merge various obsolete readme texts into doc/ssleay.txt
12412 where we collect the old documents and readme texts.
12413 2. remove the first part of files where I'm already sure that we no
12414 longer need them because of three reasons: either they are just temporary
12415 files which were left by Eric or they are preserved original files where
12416 I've verified that the diff is also available in the CVS via "cvs diff
12417 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12418 the crypto/md/ stuff).
12419 [Ralf S. Engelschall]
12421 *) More extension code. Incomplete support for subject and issuer alt
12422 name, issuer and authority key id. Change the i2v function parameters
12423 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12424 what that's for :-) Fix to ASN1 macro which messed up
12425 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12428 *) Preliminary support for ENUMERATED type. This is largely copied from the
12432 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12433 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12435 *) Make sure `make rehash' target really finds the `openssl' program.
12436 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12438 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12439 like to hear about it if this slows down other processors.
12442 *) Add CygWin32 platform information to Configure script.
12443 [Alan Batie <batie@aahz.jf.intel.com>]
12445 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12446 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12448 *) New program nseq to manipulate netscape certificate sequences
12451 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12455 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12456 but the BN code had some problems that would cause failures when
12457 doing certificate verification and some other functions.
12458 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12460 *) Add ASN1 and PEM code to support netscape certificate sequences.
12463 *) Add ASN1 and PEM code to support netscape certificate sequences.
12466 *) Add several PKIX and private extended key usage OIDs.
12469 *) Modify the 'ca' program to handle the new extension code. Modify
12470 openssl.cnf for new extension format, add comments.
12473 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12474 and add a sample to openssl.cnf so req -x509 now adds appropriate
12478 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12479 error code, add initial support to X509_print() and x509 application.
12482 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12483 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12484 stuff is currently isolated and isn't even compiled yet.
12487 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12488 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12489 Removed the versions check from X509 routines when loading extensions:
12490 this allows certain broken certificates that don't set the version
12491 properly to be processed.
12494 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12495 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12496 can still be regenerated with "make depend".
12499 *) Spelling mistake in C version of CAST-128.
12500 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12502 *) Changes to the error generation code. The perl script err-code.pl
12503 now reads in the old error codes and retains the old numbers, only
12504 adding new ones if necessary. It also only changes the .err files if new
12505 codes are added. The makefiles have been modified to only insert errors
12506 when needed (to avoid needlessly modifying header files). This is done
12507 by only inserting errors if the .err file is newer than the auto generated
12508 C file. To rebuild all the error codes from scratch (the old behaviour)
12509 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12510 or delete all the .err files.
12513 *) CAST-128 was incorrectly implemented for short keys. The C version has
12514 been fixed, but is untested. The assembler versions are also fixed, but
12515 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12516 to regenerate it if needed.
12517 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12518 Hagino <itojun@kame.net>]
12520 *) File was opened incorrectly in randfile.c.
12521 [Ulf Möller <ulf@fitug.de>]
12523 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12524 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12525 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12526 al: it's just almost always a UTCTime. Note this patch adds new error
12527 codes so do a "make errors" if there are problems.
12530 *) Correct Linux 1 recognition in config.
12531 [Ulf Möller <ulf@fitug.de>]
12533 *) Remove pointless MD5 hash when using DSA keys in ca.
12534 [Anonymous <nobody@replay.com>]
12536 *) Generate an error if given an empty string as a cert directory. Also
12537 generate an error if handed NULL (previously returned 0 to indicate an
12538 error, but didn't set one).
12539 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12541 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12544 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12545 parameters. This was causing a warning which killed off the Win32 compile.
12548 *) Remove C++ style comments from crypto/bn/bn_local.h.
12549 [Neil Costigan <neil.costigan@celocom.com>]
12551 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12552 based on a text string, looking up short and long names and finally
12553 "dot" format. The "dot" format stuff didn't work. Added new function
12554 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12555 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12556 OID is not part of the table.
12559 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12560 X509_LOOKUP_by_alias().
12563 *) Sort openssl functions by name.
12566 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12567 encryption from sample DSA keys (in case anyone is interested the password
12571 *) Make _all_ *_free functions accept a NULL pointer.
12572 [Frans Heymans <fheymans@isaserver.be>]
12574 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12576 [Anonymous <nobody@replay.com>]
12578 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12579 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12581 *) Don't blow it for numeric -newkey arguments to apps/req.
12582 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12584 *) Temp key "for export" tests were wrong in s3_srvr.c.
12585 [Anonymous <nobody@replay.com>]
12587 *) Add prototype for temp key callback functions
12588 SSL_CTX_set_tmp_{rsa,dh}_callback().
12591 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12592 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12595 *) X509_name_add_entry() freed the wrong thing after an error.
12596 [Arne Ansper <arne@ats.cyber.ee>]
12598 *) rsa_eay.c would attempt to free a NULL context.
12599 [Arne Ansper <arne@ats.cyber.ee>]
12601 *) BIO_s_socket() had a broken should_retry() on Windoze.
12602 [Arne Ansper <arne@ats.cyber.ee>]
12604 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12605 [Arne Ansper <arne@ats.cyber.ee>]
12607 *) Make sure the already existing X509_STORE->depth variable is initialized
12608 in X509_STORE_new(), but document the fact that this variable is still
12609 unused in the certificate verification process.
12610 [Ralf S. Engelschall]
12612 *) Fix the various library and apps files to free up pkeys obtained from
12613 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12616 *) Fix reference counting in X509_PUBKEY_get(). This makes
12617 demos/maurice/example2.c work, amongst others, probably.
12618 [Steve Henson and Ben Laurie]
12620 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12621 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12622 are no longer created. This way we have a single and consistent command
12623 line interface `openssl <command>', similar to `cvs <command>'.
12624 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12626 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12627 BIT STRING wrapper always have zero unused bits.
12630 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12633 *) Make the top-level INSTALL documentation easier to understand.
12636 *) Makefiles updated to exit if an error occurs in a sub-directory
12637 make (including if user presses ^C) [Paul Sutton]
12639 *) Make Montgomery context stuff explicit in RSA data structure.
12642 *) Fix build order of pem and err to allow for generated pem.h.
12645 *) Fix renumbering bug in X509_NAME_delete_entry().
12648 *) Enhanced the err-ins.pl script so it makes the error library number
12649 global and can add a library name. This is needed for external ASN1 and
12650 other error libraries.
12653 *) Fixed sk_insert which never worked properly.
12656 *) Fix ASN1 macros so they can handle indefinite length constructed
12657 EXPLICIT tags. Some non standard certificates use these: they can now
12661 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12662 into a single doc/ssleay.txt bundle. This way the information is still
12663 preserved but no longer messes up this directory. Now it's new room for
12664 the new set of documentation files.
12665 [Ralf S. Engelschall]
12667 *) SETs were incorrectly DER encoded. This was a major pain, because they
12668 shared code with SEQUENCEs, which aren't coded the same. This means that
12669 almost everything to do with SETs or SEQUENCEs has either changed name or
12670 number of arguments.
12671 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12673 *) Fix test data to work with the above.
12676 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12677 was already fixed by Eric for 0.9.1 it seems.
12678 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12680 *) Autodetect FreeBSD3.
12683 *) Fix various bugs in Configure. This affects the following platforms:
12687 unixware-2.0-pentium
12691 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12692 before they are needed.
12695 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12699 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12701 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12702 changed SSLeay to OpenSSL in version strings.
12703 [Ralf S. Engelschall]
12705 *) Some fixups to the top-level documents.
12708 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12709 because the symlink to include/ was missing.
12710 [Ralf S. Engelschall]
12712 *) Incorporated the popular no-RSA/DSA-only patches
12713 which allow to compile a RSA-free SSLeay.
12714 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12716 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12717 when "ssleay" is still not found.
12718 [Ralf S. Engelschall]
12720 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12721 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12723 *) Updated the README file.
12724 [Ralf S. Engelschall]
12726 *) Added various .cvsignore files in the CVS repository subdirs
12727 to make a "cvs update" really silent.
12728 [Ralf S. Engelschall]
12730 *) Recompiled the error-definition header files and added
12731 missing symbols to the Win32 linker tables.
12732 [Ralf S. Engelschall]
12734 *) Cleaned up the top-level documents;
12735 o new files: CHANGES and LICENSE
12736 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12737 o merged COPYRIGHT into LICENSE
12738 o removed obsolete TODO file
12739 o renamed MICROSOFT to INSTALL.W32
12740 [Ralf S. Engelschall]
12742 *) Removed dummy files from the 0.9.1b source tree:
12743 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12744 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12745 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12746 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12747 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12748 [Ralf S. Engelschall]
12750 *) Added various platform portability fixes.
12753 *) The Genesis of the OpenSSL rpject:
12754 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12755 Young and Tim J. Hudson created while they were working for C2Net until
12757 [The OpenSSL Project]
12760 Changes between 0.9.0b and 0.9.1b [not released]
12762 *) Updated a few CA certificates under certs/
12765 *) Changed some BIGNUM api stuff.
12768 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12769 DGUX x86, Linux Alpha, etc.
12772 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12773 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12777 *) Add -strparse option to asn1pars program which parses nested
12779 [Dr Stephen Henson <shenson@bigfoot.com>]
12781 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12784 *) DSA fix for "ca" program.
12787 *) Added "-genkey" option to "dsaparam" program.
12790 *) Added RIPE MD160 (rmd160) message digest.
12793 *) Added -a (all) option to "ssleay version" command.
12796 *) Added PLATFORM define which is the id given to Configure.
12799 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12802 *) Extended the ASN.1 parser routines.
12805 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12808 *) Added a BN_CTX to the BN library.
12811 *) Fixed the weak key values in DES library
12814 *) Changed API in EVP library for cipher aliases.
12817 *) Added support for RC2/64bit cipher.
12820 *) Converted the lhash library to the crypto/mem.c functions.
12823 *) Added more recognized ASN.1 object ids.
12826 *) Added more RSA padding checks for SSL/TLS.
12829 *) Added BIO proxy/filter functionality.
12832 *) Added extra_certs to SSL_CTX which can be used
12833 send extra CA certificates to the client in the CA cert chain sending
12834 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12837 *) Now Fortezza is denied in the authentication phase because
12838 this is key exchange mechanism is not supported by SSLeay at all.
12841 *) Additional PKCS1 checks.
12844 *) Support the string "TLSv1" for all TLS v1 ciphers.
12847 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12848 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12851 *) Fixed a few memory leaks.
12854 *) Fixed various code and comment typos.
12857 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12858 bytes sent in the client random.
12859 [Edward Bishop <ebishop@spyglass.com>]
projects / oweals / openssl.git / blob