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 1.1.0l and 1.1.0m [xx XXX xxxx]
14 Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
16 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
17 used even when parsing explicit parameters, when loading a serialized key
18 or calling `EC_GROUP_new_from_ecpkparameters()`/
19 `EC_GROUP_new_from_ecparameters()`.
20 This prevents bypass of security hardening and performance gains,
21 especially for curves with specialized EC_METHODs.
22 By default, if a key encoded with explicit parameters is loaded and later
23 serialized, the output is still encoded with explicit parameters, even if
24 internally a "named" EC_GROUP is used for computation.
27 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
28 this change, EC_GROUP_set_generator would accept order and/or cofactor as
29 NULL. After this change, only the cofactor parameter can be NULL. It also
30 does some minimal sanity checks on the passed order.
34 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
35 An attack is simple, if the first CMS_recipientInfo is valid but the
36 second CMS_recipientInfo is chosen ciphertext. If the second
37 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
38 encryption key will be replaced by garbage, and the message cannot be
39 decoded, but if the RSA decryption fails, the correct encryption key is
40 used and the recipient will not notice the attack.
41 As a work around for this potential attack the length of the decrypted
42 key must be equal to the cipher default key length, in case the
43 certifiate is not given and all recipientInfo are tried out.
44 The old behaviour can be re-enabled in the CMS code by setting the
45 CMS_DEBUG_DECRYPT flag.
49 *) Use Windows installation paths in the mingw builds
51 Mingw isn't a POSIX environment per se, which means that Windows
52 paths should be used for installation.
56 Changes between 1.1.0j and 1.1.0k [28 May 2019]
58 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
59 This changes the size when using the genpkey app when no size is given. It
60 fixes an omission in earlier changes that changed all RSA, DSA and DH
61 generation apps to use 2048 bits by default.
64 *) Prevent over long nonces in ChaCha20-Poly1305.
66 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
67 for every encryption operation. RFC 7539 specifies that the nonce value
68 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
69 and front pads the nonce with 0 bytes if it is less than 12
70 bytes. However it also incorrectly allows a nonce to be set of up to 16
71 bytes. In this case only the last 12 bytes are significant and any
72 additional leading bytes are ignored.
74 It is a requirement of using this cipher that nonce values are
75 unique. Messages encrypted using a reused nonce value are susceptible to
76 serious confidentiality and integrity attacks. If an application changes
77 the default nonce length to be longer than 12 bytes and then makes a
78 change to the leading bytes of the nonce expecting the new value to be a
79 new unique nonce then such an application could inadvertently encrypt
80 messages with a reused nonce.
82 Additionally the ignored bytes in a long nonce are not covered by the
83 integrity guarantee of this cipher. Any application that relies on the
84 integrity of these ignored leading bytes of a long nonce may be further
85 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
86 is safe because no such use sets such a long nonce value. However user
87 applications that use this cipher directly and set a non-default nonce
88 length to be longer than 12 bytes may be vulnerable.
90 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
95 *) Added SCA hardening for modular field inversion in EC_GROUP through
96 a new dedicated field_inv() pointer in EC_METHOD.
97 This also addresses a leakage affecting conversions from projective
98 to affine coordinates.
99 [Billy Bob Brumley, Nicola Tuveri]
101 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
102 re-used X509_PUBKEY object if the second PUBKEY is malformed.
105 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
108 *) Remove the 'dist' target and add a tarball building script. The
109 'dist' target has fallen out of use, and it shouldn't be
110 necessary to configure just to create a source distribution.
113 Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
115 *) Timing vulnerability in DSA signature generation
117 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
118 timing side channel attack. An attacker could use variations in the signing
119 algorithm to recover the private key.
121 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
125 *) Timing vulnerability in ECDSA signature generation
127 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
128 timing side channel attack. An attacker could use variations in the signing
129 algorithm to recover the private key.
131 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
135 *) Add coordinate blinding for EC_POINT and implement projective
136 coordinate blinding for generic prime curves as a countermeasure to
137 chosen point SCA attacks.
138 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
140 Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
142 *) Client DoS due to large DH parameter
144 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
145 malicious server can send a very large prime value to the client. This will
146 cause the client to spend an unreasonably long period of time generating a
147 key for this prime resulting in a hang until the client has finished. This
148 could be exploited in a Denial Of Service attack.
150 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
154 *) Cache timing vulnerability in RSA Key Generation
156 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
157 a cache timing side channel attack. An attacker with sufficient access to
158 mount cache timing attacks during the RSA key generation process could
159 recover the private key.
161 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
162 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
166 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
167 parameter is no longer accepted, as it leads to a corrupt table. NULL
168 pem_str is reserved for alias entries only.
171 *) Revert blinding in ECDSA sign and instead make problematic addition
172 length-invariant. Switch even to fixed-length Montgomery multiplication.
175 *) Change generating and checking of primes so that the error rate of not
176 being prime depends on the intended use based on the size of the input.
177 For larger primes this will result in more rounds of Miller-Rabin.
178 The maximal error rate for primes with more than 1080 bits is lowered
180 [Kurt Roeckx, Annie Yousar]
182 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
185 *) Add blinding to ECDSA and DSA signatures to protect against side channel
186 attacks discovered by Keegan Ryan (NCC Group).
189 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
190 now allow empty (zero character) pass phrases.
193 *) Certificate time validation (X509_cmp_time) enforces stricter
194 compliance with RFC 5280. Fractional seconds and timezone offsets
195 are no longer allowed.
198 *) Fixed a text canonicalisation bug in CMS
200 Where a CMS detached signature is used with text content the text goes
201 through a canonicalisation process first prior to signing or verifying a
202 signature. This process strips trailing space at the end of lines, converts
203 line terminators to CRLF and removes additional trailing line terminators
204 at the end of a file. A bug in the canonicalisation process meant that
205 some characters, such as form-feed, were incorrectly treated as whitespace
206 and removed. This is contrary to the specification (RFC5485). This fix
207 could mean that detached text data signed with an earlier version of
208 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
209 signed with a fixed OpenSSL may fail to verify with an earlier version of
210 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
211 and use the "-binary" flag (for the "cms" command line application) or set
212 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
215 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
217 *) Constructed ASN.1 types with a recursive definition could exceed the stack
219 Constructed ASN.1 types with a recursive definition (such as can be found
220 in PKCS7) could eventually exceed the stack given malicious input with
221 excessive recursion. This could result in a Denial Of Service attack. There
222 are no such structures used within SSL/TLS that come from untrusted sources
223 so this is considered safe.
225 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
230 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
232 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
233 effectively reduced to only comparing the least significant bit of each
234 byte. This allows an attacker to forge messages that would be considered as
235 authenticated in an amount of tries lower than that guaranteed by the
236 security claims of the scheme. The module can only be compiled by the
237 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
239 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
244 *) Add a build target 'build_all_generated', to build all generated files
245 and only that. This can be used to prepare everything that requires
246 things like perl for a system that lacks perl and then move everything
247 to that system and do the rest of the build there.
250 *) Backport SSL_OP_NO_RENGOTIATION
252 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
253 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
254 changes this is no longer possible in 1.1.0. Therefore the new
255 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
256 1.1.0 to provide equivalent functionality.
258 Note that if an application built against 1.1.0h headers (or above) is run
259 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
260 accepted but nothing will happen, i.e. renegotiation will not be prevented.
263 *) Removed the OS390-Unix config target. It relied on a script that doesn't
267 *) rsaz_1024_mul_avx2 overflow bug on x86_64
269 There is an overflow bug in the AVX2 Montgomery multiplication procedure
270 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
271 Analysis suggests that attacks against RSA and DSA as a result of this
272 defect would be very difficult to perform and are not believed likely.
273 Attacks against DH1024 are considered just feasible, because most of the
274 work necessary to deduce information about a private key may be performed
275 offline. The amount of resources required for such an attack would be
276 significant. However, for an attack on TLS to be meaningful, the server
277 would have to share the DH1024 private key among multiple clients, which is
278 no longer an option since CVE-2016-0701.
280 This only affects processors that support the AVX2 but not ADX extensions
281 like Intel Haswell (4th generation).
283 This issue was reported to OpenSSL by David Benjamin (Google). The issue
284 was originally found via the OSS-Fuzz project.
288 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
290 *) bn_sqrx8x_internal carry bug on x86_64
292 There is a carry propagating bug in the x86_64 Montgomery squaring
293 procedure. No EC algorithms are affected. Analysis suggests that attacks
294 against RSA and DSA as a result of this defect would be very difficult to
295 perform and are not believed likely. Attacks against DH are considered just
296 feasible (although very difficult) because most of the work necessary to
297 deduce information about a private key may be performed offline. The amount
298 of resources required for such an attack would be very significant and
299 likely only accessible to a limited number of attackers. An attacker would
300 additionally need online access to an unpatched system using the target
301 private key in a scenario with persistent DH parameters and a private
302 key that is shared between multiple clients.
304 This only affects processors that support the BMI1, BMI2 and ADX extensions
305 like Intel Broadwell (5th generation) and later or AMD Ryzen.
307 This issue was reported to OpenSSL by the OSS-Fuzz project.
311 *) Malformed X.509 IPAddressFamily could cause OOB read
313 If an X.509 certificate has a malformed IPAddressFamily extension,
314 OpenSSL could do a one-byte buffer overread. The most likely result
315 would be an erroneous display of the certificate in text format.
317 This issue was reported to OpenSSL by the OSS-Fuzz project.
321 *) Ignore the '-named_curve auto' value for compatibility of applications
323 [Tomas Mraz <tmraz@fedoraproject.org>]
325 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
328 Changes between 1.1.0e and 1.1.0f [25 May 2017]
330 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
331 platform rather than 'mingw'.
334 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
335 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
336 which is the minimum version we support.
339 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
341 *) Encrypt-Then-Mac renegotiation crash
343 During a renegotiation handshake if the Encrypt-Then-Mac extension is
344 negotiated where it was not in the original handshake (or vice-versa) then
345 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
346 and servers are affected.
348 This issue was reported to OpenSSL by Joe Orton (Red Hat).
352 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
354 *) Truncated packet could crash via OOB read
356 If one side of an SSL/TLS path is running on a 32-bit host and a specific
357 cipher is being used, then a truncated packet can cause that host to
358 perform an out-of-bounds read, usually resulting in a crash.
360 This issue was reported to OpenSSL by Robert Święcki of Google.
364 *) Bad (EC)DHE parameters cause a client crash
366 If a malicious server supplies bad parameters for a DHE or ECDHE key
367 exchange then this can result in the client attempting to dereference a
368 NULL pointer leading to a client crash. This could be exploited in a Denial
371 This issue was reported to OpenSSL by Guido Vranken.
375 *) BN_mod_exp may produce incorrect results on x86_64
377 There is a carry propagating bug in the x86_64 Montgomery squaring
378 procedure. No EC algorithms are affected. Analysis suggests that attacks
379 against RSA and DSA as a result of this defect would be very difficult to
380 perform and are not believed likely. Attacks against DH are considered just
381 feasible (although very difficult) because most of the work necessary to
382 deduce information about a private key may be performed offline. The amount
383 of resources required for such an attack would be very significant and
384 likely only accessible to a limited number of attackers. An attacker would
385 additionally need online access to an unpatched system using the target
386 private key in a scenario with persistent DH parameters and a private
387 key that is shared between multiple clients. For example this can occur by
388 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
389 similar to CVE-2015-3193 but must be treated as a separate problem.
391 This issue was reported to OpenSSL by the OSS-Fuzz project.
395 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
397 *) ChaCha20/Poly1305 heap-buffer-overflow
399 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
400 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
401 crash. This issue is not considered to be exploitable beyond a DoS.
403 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
407 *) CMS Null dereference
409 Applications parsing invalid CMS structures can crash with a NULL pointer
410 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
411 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
412 structure callback if an attempt is made to free certain invalid encodings.
413 Only CHOICE structures using a callback which do not handle NULL value are
416 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
420 *) Montgomery multiplication may produce incorrect results
422 There is a carry propagating bug in the Broadwell-specific Montgomery
423 multiplication procedure that handles input lengths divisible by, but
424 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
425 and DH private keys are impossible. This is because the subroutine in
426 question is not used in operations with the private key itself and an input
427 of the attacker's direct choice. Otherwise the bug can manifest itself as
428 transient authentication and key negotiation failures or reproducible
429 erroneous outcome of public-key operations with specially crafted input.
430 Among EC algorithms only Brainpool P-512 curves are affected and one
431 presumably can attack ECDH key negotiation. Impact was not analyzed in
432 detail, because pre-requisites for attack are considered unlikely. Namely
433 multiple clients have to choose the curve in question and the server has to
434 share the private key among them, neither of which is default behaviour.
435 Even then only clients that chose the curve will be affected.
437 This issue was publicly reported as transient failures and was not
438 initially recognized as a security issue. Thanks to Richard Morgan for
439 providing reproducible case.
443 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
444 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
445 prevent issues where no progress is being made and the peer continually
446 sends unrecognised record types, using up resources processing them.
449 *) Removed automatic addition of RPATH in shared libraries and executables,
450 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
453 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
455 *) Fix Use After Free for large message sizes
457 The patch applied to address CVE-2016-6307 resulted in an issue where if a
458 message larger than approx 16k is received then the underlying buffer to
459 store the incoming message is reallocated and moved. Unfortunately a
460 dangling pointer to the old location is left which results in an attempt to
461 write to the previously freed location. This is likely to result in a
462 crash, however it could potentially lead to execution of arbitrary code.
464 This issue only affects OpenSSL 1.1.0a.
466 This issue was reported to OpenSSL by Robert Święcki.
470 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
472 *) OCSP Status Request extension unbounded memory growth
474 A malicious client can send an excessively large OCSP Status Request
475 extension. If that client continually requests renegotiation, sending a
476 large OCSP Status Request extension each time, then there will be unbounded
477 memory growth on the server. This will eventually lead to a Denial Of
478 Service attack through memory exhaustion. Servers with a default
479 configuration are vulnerable even if they do not support OCSP. Builds using
480 the "no-ocsp" build time option are not affected.
482 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
486 *) SSL_peek() hang on empty record
488 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
489 sends an empty record. This could be exploited by a malicious peer in a
490 Denial Of Service attack.
492 This issue was reported to OpenSSL by Alex Gaynor.
496 *) Excessive allocation of memory in tls_get_message_header() and
497 dtls1_preprocess_fragment()
499 A (D)TLS message includes 3 bytes for its length in the header for the
500 message. This would allow for messages up to 16Mb in length. Messages of
501 this length are excessive and OpenSSL includes a check to ensure that a
502 peer is sending reasonably sized messages in order to avoid too much memory
503 being consumed to service a connection. A flaw in the logic of version
504 1.1.0 means that memory for the message is allocated too early, prior to
505 the excessive message length check. Due to way memory is allocated in
506 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
507 to service a connection. This could lead to a Denial of Service through
508 memory exhaustion. However, the excessive message length check still takes
509 place, and this would cause the connection to immediately fail. Assuming
510 that the application calls SSL_free() on the failed connection in a timely
511 manner then the 21Mb of allocated memory will then be immediately freed
512 again. Therefore the excessive memory allocation will be transitory in
513 nature. This then means that there is only a security impact if:
515 1) The application does not call SSL_free() in a timely manner in the event
516 that the connection fails
518 2) The application is working in a constrained environment where there is
519 very little free memory
521 3) The attacker initiates multiple connection attempts such that there are
522 multiple connections in a state where memory has been allocated for the
523 connection; SSL_free() has not yet been called; and there is insufficient
524 memory to service the multiple requests.
526 Except in the instance of (1) above any Denial Of Service is likely to be
527 transitory because as soon as the connection fails the memory is
528 subsequently freed again in the SSL_free() call. However there is an
529 increased risk during this period of application crashes due to the lack of
530 memory - which would then mean a more serious Denial of Service.
532 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
533 (CVE-2016-6307 and CVE-2016-6308)
536 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
537 had to be removed. Primary reason is that vendor assembler can't
538 assemble our modules with -KPIC flag. As result it, assembly
539 support, was not even available as option. But its lack means
540 lack of side-channel resistant code, which is incompatible with
541 security by todays standards. Fortunately gcc is readily available
542 prepackaged option, which we firmly point at...
545 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
547 *) Windows command-line tool supports UTF-8 opt-in option for arguments
548 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
549 (to any value) allows Windows user to access PKCS#12 file generated
550 with Windows CryptoAPI and protected with non-ASCII password, as well
551 as files generated under UTF-8 locale on Linux also protected with
555 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
556 have been disabled by default and removed from DEFAULT, just like RC4.
557 See the RC4 item below to re-enable both.
560 *) The method for finding the storage location for the Windows RAND seed file
561 has changed. First we check %RANDFILE%. If that is not set then we check
562 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
563 all else fails we fall back to C:\.
566 *) The EVP_EncryptUpdate() function has had its return type changed from void
567 to int. A return of 0 indicates and error while a return of 1 indicates
571 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
572 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
573 off the constant time implementation for RSA, DSA and DH have been made
574 no-ops and deprecated.
577 *) Windows RAND implementation was simplified to only get entropy by
578 calling CryptGenRandom(). Various other RAND-related tickets
580 [Joseph Wylie Yandle, Rich Salz]
582 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
583 and OPENSSL_LH_, respectively. The old names are available
584 with API compatibility. They new names are now completely documented.
587 *) Unify TYPE_up_ref(obj) methods signature.
588 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
589 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
590 int (instead of void) like all others TYPE_up_ref() methods.
591 So now these methods also check the return value of CRYPTO_atomic_add(),
592 and the validity of object reference counter.
593 [fdasilvayy@gmail.com]
595 *) With Windows Visual Studio builds, the .pdb files are installed
596 alongside the installed libraries and executables. For a static
597 library installation, ossl_static.pdb is the associate compiler
598 generated .pdb file to be used when linking programs.
601 *) Remove openssl.spec. Packaging files belong with the packagers.
604 *) Automatic Darwin/OSX configuration has had a refresh, it will now
605 recognise x86_64 architectures automatically. You can still decide
606 to build for a different bitness with the environment variable
607 KERNEL_BITS (can be 32 or 64), for example:
609 KERNEL_BITS=32 ./config
613 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
614 256 bit AES and HMAC with SHA256.
617 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
620 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
623 *) To enable users to have their own config files and build file templates,
624 Configure looks in the directory indicated by the environment variable
625 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
626 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
627 name and is used as is.
630 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
631 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
632 X509_CERT_FILE_CTX was removed.
635 *) "shared" builds are now the default. To create only static libraries use
636 the "no-shared" Configure option.
639 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
640 All of these option have not worked for some while and are fundamental
644 *) Make various cleanup routines no-ops and mark them as deprecated. Most
645 global cleanup functions are no longer required because they are handled
646 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
647 Explicitly de-initing can cause problems (e.g. where a library that uses
648 OpenSSL de-inits, but an application is still using it). The affected
649 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
650 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
651 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
655 *) --strict-warnings no longer enables runtime debugging options
656 such as REF_DEBUG. Instead, debug options are automatically
657 enabled with '--debug' builds.
658 [Andy Polyakov, Emilia Käsper]
660 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
661 have been moved out of the public header files. New functions for managing
662 these have been added.
665 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
666 objects have been moved out of the public header files. New
667 functions for managing these have been added.
670 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
671 have been moved out of the public header files. New functions for managing
672 these have been added.
675 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
676 moved out of the public header files. New functions for managing these
680 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
683 *) Removed the mk1mf build scripts.
686 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
687 it is always safe to #include a header now.
690 *) Removed the aged BC-32 config and all its supporting scripts
693 *) Removed support for Ultrix, Netware, and OS/2.
696 *) Add support for HKDF.
699 *) Add support for blake2b and blake2s
702 *) Added support for "pipelining". Ciphers that have the
703 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
704 encryptions/decryptions simultaneously. There are currently no built-in
705 ciphers with this property but the expectation is that engines will be able
706 to offer it to significantly improve throughput. Support has been extended
707 into libssl so that multiple records for a single connection can be
708 processed in one go (for >=TLS 1.1).
711 *) Added the AFALG engine. This is an async capable engine which is able to
712 offload work to the Linux kernel. In this initial version it only supports
713 AES128-CBC. The kernel must be version 4.1.0 or greater.
716 *) OpenSSL now uses a new threading API. It is no longer necessary to
717 set locking callbacks to use OpenSSL in a multi-threaded environment. There
718 are two supported threading models: pthreads and windows threads. It is
719 also possible to configure OpenSSL at compile time for "no-threads". The
720 old threading API should no longer be used. The functions have been
721 replaced with "no-op" compatibility macros.
722 [Alessandro Ghedini, Matt Caswell]
724 *) Modify behavior of ALPN to invoke callback after SNI/servername
725 callback, such that updates to the SSL_CTX affect ALPN.
728 *) Add SSL_CIPHER queries for authentication and key-exchange.
731 *) Changes to the DEFAULT cipherlist:
732 - Prefer (EC)DHE handshakes over plain RSA.
733 - Prefer AEAD ciphers over legacy ciphers.
734 - Prefer ECDSA over RSA when both certificates are available.
735 - Prefer TLSv1.2 ciphers/PRF.
736 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
740 *) Change the ECC default curve list to be this, in order: x25519,
741 secp256r1, secp521r1, secp384r1.
744 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
745 disabled by default. They can be re-enabled using the
746 enable-weak-ssl-ciphers option to Configure.
749 *) If the server has ALPN configured, but supports no protocols that the
750 client advertises, send a fatal "no_application_protocol" alert.
751 This behaviour is SHALL in RFC 7301, though it isn't universally
752 implemented by other servers.
755 *) Add X25519 support.
756 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
757 for public and private key encoding using the format documented in
758 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
759 key generation and key derivation.
761 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
765 *) Deprecate SRP_VBASE_get_by_user.
766 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
767 In order to fix an unavoidable memory leak (CVE-2016-0798),
768 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
769 seed, even if the seed is configured.
771 Users should use SRP_VBASE_get1_by_user instead. Note that in
772 SRP_VBASE_get1_by_user, caller must free the returned value. Note
773 also that even though configuring the SRP seed attempts to hide
774 invalid usernames by continuing the handshake with fake
775 credentials, this behaviour is not constant time and no strong
776 guarantees are made that the handshake is indistinguishable from
777 that of a valid user.
780 *) Configuration change; it's now possible to build dynamic engines
781 without having to build shared libraries and vice versa. This
782 only applies to the engines in engines/, those in crypto/engine/
783 will always be built into libcrypto (i.e. "static").
785 Building dynamic engines is enabled by default; to disable, use
786 the configuration option "disable-dynamic-engine".
788 The only requirements for building dynamic engines are the
789 presence of the DSO module and building with position independent
790 code, so they will also automatically be disabled if configuring
791 with "disable-dso" or "disable-pic".
793 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
794 are also taken away from openssl/opensslconf.h, as they are
798 *) Configuration change; if there is a known flag to compile
799 position independent code, it will always be applied on the
800 libcrypto and libssl object files, and never on the application
801 object files. This means other libraries that use routines from
802 libcrypto / libssl can be made into shared libraries regardless
803 of how OpenSSL was configured.
805 If this isn't desirable, the configuration options "disable-pic"
806 or "no-pic" can be used to disable the use of PIC. This will
807 also disable building shared libraries and dynamic engines.
810 *) Removed JPAKE code. It was experimental and has no wide use.
813 *) The INSTALL_PREFIX Makefile variable has been renamed to
814 DESTDIR. That makes for less confusion on what this variable
815 is for. Also, the configuration option --install_prefix is
819 *) Heartbeat for TLS has been removed and is disabled by default
820 for DTLS; configure with enable-heartbeats. Code that uses the
821 old #define's might need to be updated.
822 [Emilia Käsper, Rich Salz]
824 *) Rename REF_CHECK to REF_DEBUG.
827 *) New "unified" build system
829 The "unified" build system is aimed to be a common system for all
830 platforms we support. With it comes new support for VMS.
832 This system builds supports building in a different directory tree
833 than the source tree. It produces one Makefile (for unix family
834 or lookalikes), or one descrip.mms (for VMS).
836 The source of information to make the Makefile / descrip.mms is
837 small files called 'build.info', holding the necessary
838 information for each directory with source to compile, and a
839 template in Configurations, like unix-Makefile.tmpl or
842 With this change, the library names were also renamed on Windows
843 and on VMS. They now have names that are closer to the standard
844 on Unix, and include the major version number, and in certain
845 cases, the architecture they are built for. See "Notes on shared
846 libraries" in INSTALL.
848 We rely heavily on the perl module Text::Template.
851 *) Added support for auto-initialisation and de-initialisation of the library.
852 OpenSSL no longer requires explicit init or deinit routines to be called,
853 except in certain circumstances. See the OPENSSL_init_crypto() and
854 OPENSSL_init_ssl() man pages for further information.
857 *) The arguments to the DTLSv1_listen function have changed. Specifically the
858 "peer" argument is now expected to be a BIO_ADDR object.
860 *) Rewrite of BIO networking library. The BIO library lacked consistent
861 support of IPv6, and adding it required some more extensive
862 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
863 which hold all types of addresses and chains of address information.
864 It also introduces a new API, with functions like BIO_socket,
865 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
866 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
867 have been adapted accordingly.
870 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
874 *) CRIME protection: disable compression by default, even if OpenSSL is
875 compiled with zlib enabled. Applications can still enable compression
876 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
877 using the SSL_CONF library to configure compression.
880 *) The signature of the session callback configured with
881 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
882 was explicitly marked as 'const unsigned char*' instead of
886 *) Always DPURIFY. Remove the use of uninitialized memory in the
887 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
890 *) Removed many obsolete configuration items, including
891 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
892 MD2_CHAR, MD2_INT, MD2_LONG
894 IDEA_SHORT, IDEA_LONG
895 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
896 [Rich Salz, with advice from Andy Polyakov]
898 *) Many BN internals have been moved to an internal header file.
899 [Rich Salz with help from Andy Polyakov]
901 *) Configuration and writing out the results from it has changed.
902 Files such as Makefile include/openssl/opensslconf.h and are now
903 produced through general templates, such as Makefile.in and
904 crypto/opensslconf.h.in and some help from the perl module
907 Also, the center of configuration information is no longer
908 Makefile. Instead, Configure produces a perl module in
909 configdata.pm which holds most of the config data (in the hash
910 table %config), the target data that comes from the target
911 configuration in one of the Configurations/*.conf files (in
915 *) To clarify their intended purposes, the Configure options
916 --prefix and --openssldir change their semantics, and become more
917 straightforward and less interdependent.
919 --prefix shall be used exclusively to give the location INSTALLTOP
920 where programs, scripts, libraries, include files and manuals are
921 going to be installed. The default is now /usr/local.
923 --openssldir shall be used exclusively to give the default
924 location OPENSSLDIR where certificates, private keys, CRLs are
925 managed. This is also where the default openssl.cnf gets
927 If the directory given with this option is a relative path, the
928 values of both the --prefix value and the --openssldir value will
929 be combined to become OPENSSLDIR.
930 The default for --openssldir is INSTALLTOP/ssl.
932 Anyone who uses --openssldir to specify where OpenSSL is to be
933 installed MUST change to use --prefix instead.
936 *) The GOST engine was out of date and therefore it has been removed. An up
937 to date GOST engine is now being maintained in an external repository.
938 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
939 support for GOST ciphersuites (these are only activated if a GOST engine
943 *) EGD is no longer supported by default; use enable-egd when
945 [Ben Kaduk and Rich Salz]
947 *) The distribution now has Makefile.in files, which are used to
948 create Makefile's when Configure is run. *Configure must be run
949 before trying to build now.*
952 *) The return value for SSL_CIPHER_description() for error conditions
956 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
958 Obtaining and performing DNSSEC validation of TLSA records is
959 the application's responsibility. The application provides
960 the TLSA records of its choice to OpenSSL, and these are then
961 used to authenticate the peer.
963 The TLSA records need not even come from DNS. They can, for
964 example, be used to implement local end-entity certificate or
965 trust-anchor "pinning", where the "pin" data takes the form
966 of TLSA records, which can augment or replace verification
967 based on the usual WebPKI public certification authorities.
970 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
971 continues to support deprecated interfaces in default builds.
972 However, applications are strongly advised to compile their
973 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
974 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
975 or the 1.1.0 releases.
977 In environments in which all applications have been ported to
978 not use any deprecated interfaces OpenSSL's Configure script
979 should be used with the --api=1.1.0 option to entirely remove
980 support for the deprecated features from the library and
981 unconditionally disable them in the installed headers.
982 Essentially the same effect can be achieved with the "no-deprecated"
983 argument to Configure, except that this will always restrict
984 the build to just the latest API, rather than a fixed API
987 As applications are ported to future revisions of the API,
988 they should update their compile-time OPENSSL_API_COMPAT define
989 accordingly, but in most cases should be able to continue to
990 compile with later releases.
992 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
993 0x10000000L and 0x00908000L, respectively. However those
994 versions did not support the OPENSSL_API_COMPAT feature, and
995 so applications are not typically tested for explicit support
996 of just the undeprecated features of either release.
999 *) Add support for setting the minimum and maximum supported protocol.
1000 It can bet set via the SSL_set_min_proto_version() and
1001 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1002 MaxProtcol. It's recommended to use the new APIs to disable
1003 protocols instead of disabling individual protocols using
1004 SSL_set_options() or SSL_CONF's Protocol. This change also
1005 removes support for disabling TLS 1.2 in the OpenSSL TLS
1006 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1009 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1012 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1013 and integrates ECDSA and ECDH functionality into EC. Implementations can
1014 now redirect key generation and no longer need to convert to or from
1017 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1018 include the ec.h header file instead.
1021 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1022 ciphers who are no longer supported and drops support the ephemeral RSA key
1023 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1026 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1027 opaque. For HMAC_CTX, the following constructors and destructors
1030 HMAC_CTX *HMAC_CTX_new(void);
1031 void HMAC_CTX_free(HMAC_CTX *ctx);
1033 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1034 destroy such methods has been added. See EVP_MD_meth_new(3) and
1035 EVP_CIPHER_meth_new(3) for documentation.
1038 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1039 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1040 EVP_MD_CTX_reset() should be called instead to reinitialise
1041 an already created structure.
1042 2) For consistency with the majority of our object creators and
1043 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1044 EVP_MD_CTX_(new|free). The old names are retained as macros
1045 for deprecated builds.
1048 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1049 cryptographic operations to be performed asynchronously as long as an
1050 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1051 further details. Libssl has also had this capability integrated with the
1052 introduction of the new mode SSL_MODE_ASYNC and associated error
1053 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1054 pages. This work was developed in partnership with Intel Corp.
1057 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1058 always enabled now. If you want to disable the support you should
1059 exclude it using the list of supported ciphers. This also means that the
1060 "-no_ecdhe" option has been removed from s_server.
1063 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1064 SSL_{CTX_}set1_curves() which can set a list.
1067 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1068 curve you want to support using SSL_{CTX_}set1_curves().
1071 *) State machine rewrite. The state machine code has been significantly
1072 refactored in order to remove much duplication of code and solve issues
1073 with the old code (see ssl/statem/README for further details). This change
1074 does have some associated API changes. Notably the SSL_state() function
1075 has been removed and replaced by SSL_get_state which now returns an
1076 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1077 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1081 *) All instances of the string "ssleay" in the public API were replaced
1082 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1083 Some error codes related to internal RSA_eay API's were renamed.
1086 *) The demo files in crypto/threads were moved to demo/threads.
1089 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1091 [Matt Caswell, Rich Salz]
1093 *) New ASN.1 embed macro.
1095 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1096 structure is not allocated: it is part of the parent. That is instead of
1104 This reduces memory fragmentation and make it impossible to accidentally
1105 set a mandatory field to NULL.
1107 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1108 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1109 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1113 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1116 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1117 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1118 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1119 DES and RC4 ciphersuites.
1122 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1123 This changes the decoding behaviour for some invalid messages,
1124 though the change is mostly in the more lenient direction, and
1125 legacy behaviour is preserved as much as possible.
1128 *) Fix no-stdio build.
1129 [ David Woodhouse <David.Woodhouse@intel.com> and also
1130 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1132 *) New testing framework
1133 The testing framework has been largely rewritten and is now using
1134 perl and the perl modules Test::Harness and an extended variant of
1135 Test::More called OpenSSL::Test to do its work. All test scripts in
1136 test/ have been rewritten into test recipes, and all direct calls to
1137 executables in test/Makefile have become individual recipes using the
1138 simplified testing OpenSSL::Test::Simple.
1140 For documentation on our testing modules, do:
1142 perldoc test/testlib/OpenSSL/Test/Simple.pm
1143 perldoc test/testlib/OpenSSL/Test.pm
1147 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1148 are used; the latter aborts on memory leaks (usually checked on exit).
1149 Some undocumented "set malloc, etc., hooks" functions were removed
1150 and others were changed. All are now documented.
1153 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1155 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1157 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1158 from RFC4279, RFC4785, RFC5487, RFC5489.
1160 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1161 original RSA_PSK patch.
1164 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1165 era flag was never set throughout the codebase (only read). Also removed
1166 SSL3_FLAGS_POP_BUFFER which was only used if
1167 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1170 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1171 to be "oneline" instead of "compat".
1174 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1175 not aware of clients that still exhibit this bug, and the workaround
1176 hasn't been working properly for a while.
1179 *) The return type of BIO_number_read() and BIO_number_written() as well as
1180 the corresponding num_read and num_write members in the BIO structure has
1181 changed from unsigned long to uint64_t. On platforms where an unsigned
1182 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1186 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1187 OpenSSL without support for them. It also means that maintaining
1188 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1189 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1192 *) Removed support for the two export grade static DH ciphersuites
1193 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1194 were newly added (along with a number of other static DH ciphersuites) to
1195 1.0.2. However the two export ones have *never* worked since they were
1196 introduced. It seems strange in any case to be adding new export
1197 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1200 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1201 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1202 and turned into macros which simply call the new preferred function names
1203 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1204 should use the new names instead. Also as part of this change the ssl23.h
1205 header file has been removed.
1208 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1209 code and the associated standard is no longer considered fit-for-purpose.
1212 *) RT2547 was closed. When generating a private key, try to make the
1213 output file readable only by the owner. This behavior change might
1214 be noticeable when interacting with other software.
1216 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1220 *) Added HTTP GET support to the ocsp command.
1223 *) Changed default digest for the dgst and enc commands from MD5 to
1227 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1230 *) Added support for TLS extended master secret from
1231 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1232 initial patch which was a great help during development.
1235 *) All libssl internal structures have been removed from the public header
1236 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1237 now redundant). Users should not attempt to access internal structures
1238 directly. Instead they should use the provided API functions.
1241 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1242 Access to deprecated functions can be re-enabled by running config with
1243 "enable-deprecated". In addition applications wishing to use deprecated
1244 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1245 will, by default, disable some transitive includes that previously existed
1246 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1249 *) Added support for OCB mode. OpenSSL has been granted a patent license
1250 compatible with the OpenSSL license for use of OCB. Details are available
1251 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1252 for OCB can be removed by calling config with no-ocb.
1255 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1256 compatible client hello.
1259 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1260 done while fixing the error code for the key-too-small case.
1261 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1263 *) CA.sh has been removed; use CA.pl instead.
1266 *) Removed old DES API.
1269 *) Remove various unsupported platforms:
1275 Sinix/ReliantUNIX RM400
1280 16-bit platforms such as WIN16
1283 *) Clean up OPENSSL_NO_xxx #define's
1284 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1285 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1286 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1287 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1288 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1289 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1290 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1291 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1292 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1293 Remove MS_STATIC; it's a relic from platforms <32 bits.
1296 *) Cleaned up dead code
1297 Remove all but one '#ifdef undef' which is to be looked at.
1300 *) Clean up calling of xxx_free routines.
1301 Just like free(), fix most of the xxx_free routines to accept
1302 NULL. Remove the non-null checks from callers. Save much code.
1305 *) Add secure heap for storage of private keys (when possible).
1306 Add BIO_s_secmem(), CBIGNUM, etc.
1307 Contributed by Akamai Technologies under our Corporate CLA.
1310 *) Experimental support for a new, fast, unbiased prime candidate generator,
1311 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1312 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1314 *) New output format NSS in the sess_id command line tool. This allows
1315 exporting the session id and the master key in NSS keylog format.
1316 [Martin Kaiser <martin@kaiser.cx>]
1318 *) Harmonize version and its documentation. -f flag is used to display
1320 [mancha <mancha1@zoho.com>]
1322 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1323 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1324 [mancha <mancha1@zoho.com>]
1326 *) Fix some double frees. These are not thought to be exploitable.
1327 [mancha <mancha1@zoho.com>]
1329 *) A missing bounds check in the handling of the TLS heartbeat extension
1330 can be used to reveal up to 64k of memory to a connected client or
1333 Thanks for Neel Mehta of Google Security for discovering this bug and to
1334 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1335 preparing the fix (CVE-2014-0160)
1336 [Adam Langley, Bodo Moeller]
1338 *) Fix for the attack described in the paper "Recovering OpenSSL
1339 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1340 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1341 http://eprint.iacr.org/2014/140
1343 Thanks to Yuval Yarom and Naomi Benger for discovering this
1344 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1345 [Yuval Yarom and Naomi Benger]
1347 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1348 this fixes a limitation in previous versions of OpenSSL.
1351 *) Experimental encrypt-then-mac support.
1353 Experimental support for encrypt then mac from
1354 draft-gutmann-tls-encrypt-then-mac-02.txt
1356 To enable it set the appropriate extension number (0x42 for the test
1357 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1359 For non-compliant peers (i.e. just about everything) this should have no
1362 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1366 *) Add EVP support for key wrapping algorithms, to avoid problems with
1367 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1368 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1369 algorithms and include tests cases.
1372 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1376 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1377 MGF1 digest and OAEP label.
1380 *) Make openssl verify return errors.
1381 [Chris Palmer <palmer@google.com> and Ben Laurie]
1383 *) New function ASN1_TIME_diff to calculate the difference between two
1384 ASN1_TIME structures or one structure and the current time.
1387 *) Update fips_test_suite to support multiple command line options. New
1388 test to induce all self test errors in sequence and check expected
1392 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1393 sign or verify all in one operation.
1396 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1397 test programs and fips_test_suite. Includes functionality to parse
1398 the minimal script output of fipsalgest.pl directly.
1401 *) Add authorisation parameter to FIPS_module_mode_set().
1404 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1407 *) Use separate DRBG fields for internal and external flags. New function
1408 FIPS_drbg_health_check() to perform on demand health checking. Add
1409 generation tests to fips_test_suite with reduced health check interval to
1410 demonstrate periodic health checking. Add "nodh" option to
1411 fips_test_suite to skip very slow DH test.
1414 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1418 *) More extensive health check for DRBG checking many more failure modes.
1419 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1420 combination: call this in fips_test_suite.
1423 *) Add support for canonical generation of DSA parameter 'g'. See
1426 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1427 POST to handle HMAC cases.
1430 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1431 to return numerical and string versions of the FIPS module number.
1434 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1435 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1436 outside the validated module in the FIPS capable OpenSSL.
1439 *) Minor change to DRBG entropy callback semantics. In some cases
1440 there is no multiple of the block length between min_len and
1441 max_len. Allow the callback to return more than max_len bytes
1442 of entropy but discard any extra: it is the callback's responsibility
1443 to ensure that the extra data discarded does not impact the
1444 requested amount of entropy.
1447 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1448 information in FIPS186-3, SP800-57 and SP800-131A.
1451 *) CCM support via EVP. Interface is very similar to GCM case except we
1452 must supply all data in one chunk (i.e. no update, final) and the
1453 message length must be supplied if AAD is used. Add algorithm test
1457 *) Initial version of POST overhaul. Add POST callback to allow the status
1458 of POST to be monitored and/or failures induced. Modify fips_test_suite
1459 to use callback. Always run all selftests even if one fails.
1462 *) XTS support including algorithm test driver in the fips_gcmtest program.
1463 Note: this does increase the maximum key length from 32 to 64 bytes but
1464 there should be no binary compatibility issues as existing applications
1465 will never use XTS mode.
1468 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1469 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1470 performs algorithm blocking for unapproved PRNG types. Also do not
1471 set PRNG type in FIPS_mode_set(): leave this to the application.
1472 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1473 the standard OpenSSL PRNG: set additional data to a date time vector.
1476 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1477 This shouldn't present any incompatibility problems because applications
1478 shouldn't be using these directly and any that are will need to rethink
1479 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1482 *) Extensive self tests and health checking required by SP800-90 DRBG.
1483 Remove strength parameter from FIPS_drbg_instantiate and always
1484 instantiate at maximum supported strength.
1487 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1490 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1493 *) New function DH_compute_key_padded() to compute a DH key and pad with
1494 leading zeroes if needed: this complies with SP800-56A et al.
1497 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1498 anything, incomplete, subject to change and largely untested at present.
1501 *) Modify fipscanisteronly build option to only build the necessary object
1502 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1505 *) Add experimental option FIPSSYMS to give all symbols in
1506 fipscanister.o and FIPS or fips prefix. This will avoid
1507 conflicts with future versions of OpenSSL. Add perl script
1508 util/fipsas.pl to preprocess assembly language source files
1509 and rename any affected symbols.
1512 *) Add selftest checks and algorithm block of non-fips algorithms in
1513 FIPS mode. Remove DES2 from selftests.
1516 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1517 return internal method without any ENGINE dependencies. Add new
1518 tiny fips sign and verify functions.
1521 *) New build option no-ec2m to disable characteristic 2 code.
1524 *) New build option "fipscanisteronly". This only builds fipscanister.o
1525 and (currently) associated fips utilities. Uses the file Makefile.fips
1526 instead of Makefile.org as the prototype.
1529 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1530 Update fips_gcmtest to use IV generator.
1533 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1534 setting output buffer to NULL. The *Final function must be
1535 called although it will not retrieve any additional data. The tag
1536 can be set or retrieved with a ctrl. The IV length is by default 12
1537 bytes (96 bits) but can be set to an alternative value. If the IV
1538 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1542 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1543 underlying do_cipher function handles all cipher semantics itself
1544 including padding and finalisation. This is useful if (for example)
1545 an ENGINE cipher handles block padding itself. The behaviour of
1546 do_cipher is subtly changed if this flag is set: the return value
1547 is the number of characters written to the output buffer (zero is
1548 no longer an error code) or a negative error code. Also if the
1549 input buffer is NULL and length 0 finalisation should be performed.
1552 *) If a candidate issuer certificate is already part of the constructed
1553 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1556 *) Improve forward-security support: add functions
1558 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1559 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1561 for use by SSL/TLS servers; the callback function will be called whenever a
1562 new session is created, and gets to decide whether the session may be
1563 cached to make it resumable (return 0) or not (return 1). (As by the
1564 SSL/TLS protocol specifications, the session_id sent by the server will be
1565 empty to indicate that the session is not resumable; also, the server will
1566 not generate RFC 4507 (RFC 5077) session tickets.)
1568 A simple reasonable callback implementation is to return is_forward_secure.
1569 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1570 by the SSL/TLS server library, indicating whether it can provide forward
1572 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1574 *) New -verify_name option in command line utilities to set verification
1578 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1579 Add CMAC pkey methods.
1582 *) Experimental renegotiation in s_server -www mode. If the client
1583 browses /reneg connection is renegotiated. If /renegcert it is
1584 renegotiated requesting a certificate.
1587 *) Add an "external" session cache for debugging purposes to s_server. This
1588 should help trace issues which normally are only apparent in deployed
1589 multi-process servers.
1592 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1593 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1594 BIO_set_cipher() and some obscure PEM functions were changed so they
1595 can now return an error. The RAND changes required a change to the
1596 RAND_METHOD structure.
1599 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1600 a gcc attribute to warn if the result of a function is ignored. This
1601 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1602 whose return value is often ignored.
1605 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1606 These allow SCTs (signed certificate timestamps) to be requested and
1607 validated when establishing a connection.
1608 [Rob Percival <robpercival@google.com>]
1610 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1612 *) Prevent padding oracle in AES-NI CBC MAC check
1614 A MITM attacker can use a padding oracle attack to decrypt traffic
1615 when the connection uses an AES CBC cipher and the server support
1618 This issue was introduced as part of the fix for Lucky 13 padding
1619 attack (CVE-2013-0169). The padding check was rewritten to be in
1620 constant time by making sure that always the same bytes are read and
1621 compared against either the MAC or padding bytes. But it no longer
1622 checked that there was enough data to have both the MAC and padding
1625 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1629 *) Fix EVP_EncodeUpdate overflow
1631 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1632 Base64 encoding of binary data. If an attacker is able to supply very large
1633 amounts of input data then a length check can overflow resulting in a heap
1636 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1637 the PEM_write_bio* family of functions. These are mainly used within the
1638 OpenSSL command line applications, so any application which processes data
1639 from an untrusted source and outputs it as a PEM file should be considered
1640 vulnerable to this issue. User applications that call these APIs directly
1641 with large amounts of untrusted data may also be vulnerable.
1643 This issue was reported by Guido Vranken.
1647 *) Fix EVP_EncryptUpdate overflow
1649 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1650 is able to supply very large amounts of input data after a previous call to
1651 EVP_EncryptUpdate() with a partial block then a length check can overflow
1652 resulting in a heap corruption. Following an analysis of all OpenSSL
1653 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1654 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1655 the first called function after an EVP_EncryptInit(), and therefore that
1656 specific call must be safe. The second form is where the length passed to
1657 EVP_EncryptUpdate() can be seen from the code to be some small value and
1658 therefore there is no possibility of an overflow. Since all instances are
1659 one of these two forms, it is believed that there can be no overflows in
1660 internal code due to this problem. It should be noted that
1661 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1662 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1663 of these calls have also been analysed too and it is believed there are no
1664 instances in internal usage where an overflow could occur.
1666 This issue was reported by Guido Vranken.
1670 *) Prevent ASN.1 BIO excessive memory allocation
1672 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1673 a short invalid encoding can cause allocation of large amounts of memory
1674 potentially consuming excessive resources or exhausting memory.
1676 Any application parsing untrusted data through d2i BIO functions is
1677 affected. The memory based functions such as d2i_X509() are *not* affected.
1678 Since the memory based functions are used by the TLS library, TLS
1679 applications are not affected.
1681 This issue was reported by Brian Carpenter.
1687 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1688 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1689 in arbitrary stack data being returned in the buffer.
1691 This issue was reported by Guido Vranken.
1695 *) Modify behavior of ALPN to invoke callback after SNI/servername
1696 callback, such that updates to the SSL_CTX affect ALPN.
1699 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1703 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1704 methods are enabled and ssl2 is disabled the methods return NULL.
1707 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1709 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1710 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1711 provide any "EXPORT" or "LOW" strength ciphers.
1714 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1715 is by default disabled at build-time. Builds that are not configured with
1716 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1717 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1718 will need to explicitly call either of:
1720 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1722 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1724 as appropriate. Even if either of those is used, or the application
1725 explicitly uses the version-specific SSLv2_method() or its client and
1726 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1727 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1728 ciphers, and SSLv2 56-bit DES are no longer available.
1732 *) Fix a double-free in DSA code
1734 A double free bug was discovered when OpenSSL parses malformed DSA private
1735 keys and could lead to a DoS attack or memory corruption for applications
1736 that receive DSA private keys from untrusted sources. This scenario is
1739 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1744 *) Disable SRP fake user seed to address a server memory leak.
1746 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1748 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1749 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1750 was changed to ignore the "fake user" SRP seed, even if the seed
1753 Users should use SRP_VBASE_get1_by_user instead. Note that in
1754 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1755 also that even though configuring the SRP seed attempts to hide
1756 invalid usernames by continuing the handshake with fake
1757 credentials, this behaviour is not constant time and no strong
1758 guarantees are made that the handshake is indistinguishable from
1759 that of a valid user.
1763 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1765 In the BN_hex2bn function the number of hex digits is calculated using an
1766 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1767 large values of |i| this can result in |bn_expand| not allocating any
1768 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1769 field as NULL leading to a subsequent NULL ptr deref. For very large values
1770 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1771 In this case memory is allocated to the internal BIGNUM data field, but it
1772 is insufficiently sized leading to heap corruption. A similar issue exists
1773 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1774 is ever called by user applications with very large untrusted hex/dec data.
1775 This is anticipated to be a rare occurrence.
1777 All OpenSSL internal usage of these functions use data that is not expected
1778 to be untrusted, e.g. config file data or application command line
1779 arguments. If user developed applications generate config file data based
1780 on untrusted data then it is possible that this could also lead to security
1781 consequences. This is also anticipated to be rare.
1783 This issue was reported to OpenSSL by Guido Vranken.
1787 *) Fix memory issues in BIO_*printf functions
1789 The internal |fmtstr| function used in processing a "%s" format string in
1790 the BIO_*printf functions could overflow while calculating the length of a
1791 string and cause an OOB read when printing very long strings.
1793 Additionally the internal |doapr_outch| function can attempt to write to an
1794 OOB memory location (at an offset from the NULL pointer) in the event of a
1795 memory allocation failure. In 1.0.2 and below this could be caused where
1796 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1797 could be in processing a very long "%s" format string. Memory leaks can
1800 The first issue may mask the second issue dependent on compiler behaviour.
1801 These problems could enable attacks where large amounts of untrusted data
1802 is passed to the BIO_*printf functions. If applications use these functions
1803 in this way then they could be vulnerable. OpenSSL itself uses these
1804 functions when printing out human-readable dumps of ASN.1 data. Therefore
1805 applications that print this data could be vulnerable if the data is from
1806 untrusted sources. OpenSSL command line applications could also be
1807 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1808 as command line arguments.
1810 Libssl is not considered directly vulnerable. Additionally certificates etc
1811 received via remote connections via libssl are also unlikely to be able to
1812 trigger these issues because of message size limits enforced within libssl.
1814 This issue was reported to OpenSSL Guido Vranken.
1818 *) Side channel attack on modular exponentiation
1820 A side-channel attack was found which makes use of cache-bank conflicts on
1821 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1822 of RSA keys. The ability to exploit this issue is limited as it relies on
1823 an attacker who has control of code in a thread running on the same
1824 hyper-threaded core as the victim thread which is performing decryptions.
1826 This issue was reported to OpenSSL by Yuval Yarom, The University of
1827 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1828 Nadia Heninger, University of Pennsylvania with more information at
1829 http://cachebleed.info.
1833 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1834 if no keysize is specified with default_bits. This fixes an
1835 omission in an earlier change that changed all RSA/DSA key generation
1836 apps to use 2048 bits by default.
1839 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1840 *) DH small subgroups
1842 Historically OpenSSL only ever generated DH parameters based on "safe"
1843 primes. More recently (in version 1.0.2) support was provided for
1844 generating X9.42 style parameter files such as those required for RFC 5114
1845 support. The primes used in such files may not be "safe". Where an
1846 application is using DH configured with parameters based on primes that are
1847 not "safe" then an attacker could use this fact to find a peer's private
1848 DH exponent. This attack requires that the attacker complete multiple
1849 handshakes in which the peer uses the same private DH exponent. For example
1850 this could be used to discover a TLS server's private DH exponent if it's
1851 reusing the private DH exponent or it's using a static DH ciphersuite.
1853 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1854 TLS. It is not on by default. If the option is not set then the server
1855 reuses the same private DH exponent for the life of the server process and
1856 would be vulnerable to this attack. It is believed that many popular
1857 applications do set this option and would therefore not be at risk.
1859 The fix for this issue adds an additional check where a "q" parameter is
1860 available (as is the case in X9.42 based parameters). This detects the
1861 only known attack, and is the only possible defense for static DH
1862 ciphersuites. This could have some performance impact.
1864 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1865 default and cannot be disabled. This could have some performance impact.
1867 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1871 *) SSLv2 doesn't block disabled ciphers
1873 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1874 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1875 been disabled, provided that the SSLv2 protocol was not also disabled via
1878 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1879 and Sebastian Schinzel.
1883 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1885 *) BN_mod_exp may produce incorrect results on x86_64
1887 There is a carry propagating bug in the x86_64 Montgomery squaring
1888 procedure. No EC algorithms are affected. Analysis suggests that attacks
1889 against RSA and DSA as a result of this defect would be very difficult to
1890 perform and are not believed likely. Attacks against DH are considered just
1891 feasible (although very difficult) because most of the work necessary to
1892 deduce information about a private key may be performed offline. The amount
1893 of resources required for such an attack would be very significant and
1894 likely only accessible to a limited number of attackers. An attacker would
1895 additionally need online access to an unpatched system using the target
1896 private key in a scenario with persistent DH parameters and a private
1897 key that is shared between multiple clients. For example this can occur by
1898 default in OpenSSL DHE based SSL/TLS ciphersuites.
1900 This issue was reported to OpenSSL by Hanno Böck.
1904 *) Certificate verify crash with missing PSS parameter
1906 The signature verification routines will crash with a NULL pointer
1907 dereference if presented with an ASN.1 signature using the RSA PSS
1908 algorithm and absent mask generation function parameter. Since these
1909 routines are used to verify certificate signature algorithms this can be
1910 used to crash any certificate verification operation and exploited in a
1911 DoS attack. Any application which performs certificate verification is
1912 vulnerable including OpenSSL clients and servers which enable client
1915 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1919 *) X509_ATTRIBUTE memory leak
1921 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1922 memory. This structure is used by the PKCS#7 and CMS routines so any
1923 application which reads PKCS#7 or CMS data from untrusted sources is
1924 affected. SSL/TLS is not affected.
1926 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1931 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1932 This changes the decoding behaviour for some invalid messages,
1933 though the change is mostly in the more lenient direction, and
1934 legacy behaviour is preserved as much as possible.
1937 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1939 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1941 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1943 *) Alternate chains certificate forgery
1945 During certificate verification, OpenSSL will attempt to find an
1946 alternative certificate chain if the first attempt to build such a chain
1947 fails. An error in the implementation of this logic can mean that an
1948 attacker could cause certain checks on untrusted certificates to be
1949 bypassed, such as the CA flag, enabling them to use a valid leaf
1950 certificate to act as a CA and "issue" an invalid certificate.
1952 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1956 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1958 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1959 incompatibility in the handling of HMAC. The previous ABI has now been
1963 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1965 *) Malformed ECParameters causes infinite loop
1967 When processing an ECParameters structure OpenSSL enters an infinite loop
1968 if the curve specified is over a specially malformed binary polynomial
1971 This can be used to perform denial of service against any
1972 system which processes public keys, certificate requests or
1973 certificates. This includes TLS clients and TLS servers with
1974 client authentication enabled.
1976 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1980 *) Exploitable out-of-bounds read in X509_cmp_time
1982 X509_cmp_time does not properly check the length of the ASN1_TIME
1983 string and can read a few bytes out of bounds. In addition,
1984 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1987 An attacker can use this to craft malformed certificates and CRLs of
1988 various sizes and potentially cause a segmentation fault, resulting in
1989 a DoS on applications that verify certificates or CRLs. TLS clients
1990 that verify CRLs are affected. TLS clients and servers with client
1991 authentication enabled may be affected if they use custom verification
1994 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1995 independently by Hanno Böck.
1999 *) PKCS7 crash with missing EnvelopedContent
2001 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2002 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2003 with missing content and trigger a NULL pointer dereference on parsing.
2005 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2006 structures from untrusted sources are affected. OpenSSL clients and
2007 servers are not affected.
2009 This issue was reported to OpenSSL by Michal Zalewski (Google).
2013 *) CMS verify infinite loop with unknown hash function
2015 When verifying a signedData message the CMS code can enter an infinite loop
2016 if presented with an unknown hash function OID. This can be used to perform
2017 denial of service against any system which verifies signedData messages using
2019 This issue was reported to OpenSSL by Johannes Bauer.
2023 *) Race condition handling NewSessionTicket
2025 If a NewSessionTicket is received by a multi-threaded client when attempting to
2026 reuse a previous ticket then a race condition can occur potentially leading to
2027 a double free of the ticket data.
2031 *) Only support 256-bit or stronger elliptic curves with the
2032 'ecdh_auto' setting (server) or by default (client). Of supported
2033 curves, prefer P-256 (both).
2036 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2038 *) ClientHello sigalgs DoS fix
2040 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2041 invalid signature algorithms extension a NULL pointer dereference will
2042 occur. This can be exploited in a DoS attack against the server.
2044 This issue was was reported to OpenSSL by David Ramos of Stanford
2047 [Stephen Henson and Matt Caswell]
2049 *) Multiblock corrupted pointer fix
2051 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2052 feature only applies on 64 bit x86 architecture platforms that support AES
2053 NI instructions. A defect in the implementation of "multiblock" can cause
2054 OpenSSL's internal write buffer to become incorrectly set to NULL when
2055 using non-blocking IO. Typically, when the user application is using a
2056 socket BIO for writing, this will only result in a failed connection.
2057 However if some other BIO is used then it is likely that a segmentation
2058 fault will be triggered, thus enabling a potential DoS attack.
2060 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2064 *) Segmentation fault in DTLSv1_listen fix
2066 The DTLSv1_listen function is intended to be stateless and processes the
2067 initial ClientHello from many peers. It is common for user code to loop
2068 over the call to DTLSv1_listen until a valid ClientHello is received with
2069 an associated cookie. A defect in the implementation of DTLSv1_listen means
2070 that state is preserved in the SSL object from one invocation to the next
2071 that can lead to a segmentation fault. Errors processing the initial
2072 ClientHello can trigger this scenario. An example of such an error could be
2073 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2076 This issue was reported to OpenSSL by Per Allansson.
2080 *) Segmentation fault in ASN1_TYPE_cmp fix
2082 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2083 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2084 certificate signature algorithm consistency this can be used to crash any
2085 certificate verification operation and exploited in a DoS attack. Any
2086 application which performs certificate verification is vulnerable including
2087 OpenSSL clients and servers which enable client authentication.
2091 *) Segmentation fault for invalid PSS parameters fix
2093 The signature verification routines will crash with a NULL pointer
2094 dereference if presented with an ASN.1 signature using the RSA PSS
2095 algorithm and invalid parameters. Since these routines are used to verify
2096 certificate signature algorithms this can be used to crash any
2097 certificate verification operation and exploited in a DoS attack. Any
2098 application which performs certificate verification is vulnerable including
2099 OpenSSL clients and servers which enable client authentication.
2101 This issue was was reported to OpenSSL by Brian Carpenter.
2105 *) ASN.1 structure reuse memory corruption fix
2107 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2108 memory corruption via an invalid write. Such reuse is and has been
2109 strongly discouraged and is believed to be rare.
2111 Applications that parse structures containing CHOICE or ANY DEFINED BY
2112 components may be affected. Certificate parsing (d2i_X509 and related
2113 functions) are however not affected. OpenSSL clients and servers are
2118 *) PKCS7 NULL pointer dereferences fix
2120 The PKCS#7 parsing code does not handle missing outer ContentInfo
2121 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2122 missing content and trigger a NULL pointer dereference on parsing.
2124 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2125 otherwise parse PKCS#7 structures from untrusted sources are
2126 affected. OpenSSL clients and servers are not affected.
2128 This issue was reported to OpenSSL by Michal Zalewski (Google).
2132 *) DoS via reachable assert in SSLv2 servers fix
2134 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2135 servers that both support SSLv2 and enable export cipher suites by sending
2136 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2138 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2139 (OpenSSL development team).
2143 *) Empty CKE with client auth and DHE fix
2145 If client auth is used then a server can seg fault in the event of a DHE
2146 ciphersuite being selected and a zero length ClientKeyExchange message
2147 being sent by the client. This could be exploited in a DoS attack.
2151 *) Handshake with unseeded PRNG fix
2153 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2154 with an unseeded PRNG. The conditions are:
2155 - The client is on a platform where the PRNG has not been seeded
2156 automatically, and the user has not seeded manually
2157 - A protocol specific client method version has been used (i.e. not
2158 SSL_client_methodv23)
2159 - A ciphersuite is used that does not require additional random data from
2160 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2162 If the handshake succeeds then the client random that has been used will
2163 have been generated from a PRNG with insufficient entropy and therefore the
2164 output may be predictable.
2166 For example using the following command with an unseeded openssl will
2167 succeed on an unpatched platform:
2169 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2173 *) Use After Free following d2i_ECPrivatekey error fix
2175 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2176 could cause a use after free condition. This, in turn, could cause a double
2177 free in several private key parsing functions (such as d2i_PrivateKey
2178 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2179 for applications that receive EC private keys from untrusted
2180 sources. This scenario is considered rare.
2182 This issue was discovered by the BoringSSL project and fixed in their
2187 *) X509_to_X509_REQ NULL pointer deref fix
2189 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2190 the certificate key is invalid. This function is rarely used in practice.
2192 This issue was discovered by Brian Carpenter.
2196 *) Removed the export ciphers from the DEFAULT ciphers
2199 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2201 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2202 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2203 So far those who have to target multiple platforms would compromise
2204 and argue that binary targeting say ARMv5 would still execute on
2205 ARMv8. "Universal" build resolves this compromise by providing
2206 near-optimal performance even on newer platforms.
2209 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2210 (other platforms pending).
2211 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2213 *) Add support for the SignedCertificateTimestampList certificate and
2214 OCSP response extensions from RFC6962.
2217 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2218 for corner cases. (Certain input points at infinity could lead to
2219 bogus results, with non-infinity inputs mapped to infinity too.)
2222 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2223 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2224 common cases are optimized and there still is room for further
2225 improvements. Vector Permutation AES for Altivec is also added.
2228 *) Add support for little-endian ppc64 Linux target.
2229 [Marcelo Cerri (IBM)]
2231 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2232 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2233 are optimized and there still is room for further improvements.
2234 Both 32- and 64-bit modes are supported.
2235 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2237 *) Improved ARMv7 NEON support.
2240 *) Support for SPARC Architecture 2011 crypto extensions, first
2241 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2242 SHA256/512, MD5, GHASH and modular exponentiation.
2243 [Andy Polyakov, David Miller]
2245 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2247 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2249 *) Support for new and upcoming Intel processors, including AVX2,
2250 BMI and SHA ISA extensions. This includes additional "stitched"
2251 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2254 This work was sponsored by Intel Corp.
2257 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2258 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2259 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2262 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2263 this fixes a limitation in previous versions of OpenSSL.
2266 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2267 MGF1 digest and OAEP label.
2270 *) Add EVP support for key wrapping algorithms, to avoid problems with
2271 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2272 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2273 algorithms and include tests cases.
2276 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2278 [Douglas E. Engert, Steve Henson]
2280 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2281 difference in days and seconds between two tm or ASN1_TIME structures.
2284 *) Add -rev test option to s_server to just reverse order of characters
2285 received by client and send back to server. Also prints an abbreviated
2286 summary of the connection parameters.
2289 *) New option -brief for s_client and s_server to print out a brief summary
2290 of connection parameters.
2293 *) Add callbacks for arbitrary TLS extensions.
2294 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2296 *) New option -crl_download in several openssl utilities to download CRLs
2297 from CRLDP extension in certificates.
2300 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2303 *) New function X509_CRL_diff to generate a delta CRL from the difference
2304 of two full CRLs. Add support to "crl" utility.
2307 *) New functions to set lookup_crls function and to retrieve
2308 X509_STORE from X509_STORE_CTX.
2311 *) Print out deprecated issuer and subject unique ID fields in
2315 *) Extend OCSP I/O functions so they can be used for simple general purpose
2316 HTTP as well as OCSP. New wrapper function which can be used to download
2317 CRLs using the OCSP API.
2320 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2323 *) SSL_CONF* functions. These provide a common framework for application
2324 configuration using configuration files or command lines.
2327 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2328 message callback and prints the results. Needs compile time option
2329 "enable-ssl-trace". New options to s_client and s_server to enable
2333 *) New ctrl and macro to retrieve supported points extensions.
2334 Print out extension in s_server and s_client.
2337 *) New functions to retrieve certificate signature and signature
2341 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2345 *) New Suite B modes for TLS code. These use and enforce the requirements
2346 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2347 only use Suite B curves. The Suite B modes can be set by using the
2348 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2351 *) New chain verification flags for Suite B levels of security. Check
2352 algorithms are acceptable when flags are set in X509_verify_cert.
2355 *) Make tls1_check_chain return a set of flags indicating checks passed
2356 by a certificate chain. Add additional tests to handle client
2357 certificates: checks for matching certificate type and issuer name
2361 *) If an attempt is made to use a signature algorithm not in the peer
2362 preference list abort the handshake. If client has no suitable
2363 signature algorithms in response to a certificate request do not
2364 use the certificate.
2367 *) If server EC tmp key is not in client preference list abort handshake.
2370 *) Add support for certificate stores in CERT structure. This makes it
2371 possible to have different stores per SSL structure or one store in
2372 the parent SSL_CTX. Include distinct stores for certificate chain
2373 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2374 to build and store a certificate chain in CERT structure: returning
2375 an error if the chain cannot be built: this will allow applications
2376 to test if a chain is correctly configured.
2378 Note: if the CERT based stores are not set then the parent SSL_CTX
2379 store is used to retain compatibility with existing behaviour.
2383 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2384 mask based on the current session, check mask when sending client
2385 hello and checking the requested ciphersuite.
2388 *) New ctrls to retrieve and set certificate types in a certificate
2389 request message. Print out received values in s_client. If certificate
2390 types is not set with custom values set sensible values based on
2391 supported signature algorithms.
2394 *) Support for distinct client and server supported signature algorithms.
2397 *) Add certificate callback. If set this is called whenever a certificate
2398 is required by client or server. An application can decide which
2399 certificate chain to present based on arbitrary criteria: for example
2400 supported signature algorithms. Add very simple example to s_server.
2401 This fixes many of the problems and restrictions of the existing client
2402 certificate callback: for example you can now clear an existing
2403 certificate and specify the whole chain.
2406 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2407 the certificate can be used for (if anything). Set valid_flags field
2408 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2409 to have similar checks in it.
2411 Add new "cert_flags" field to CERT structure and include a "strict mode".
2412 This enforces some TLS certificate requirements (such as only permitting
2413 certificate signature algorithms contained in the supported algorithms
2414 extension) which some implementations ignore: this option should be used
2415 with caution as it could cause interoperability issues.
2418 *) Update and tidy signature algorithm extension processing. Work out
2419 shared signature algorithms based on preferences and peer algorithms
2420 and print them out in s_client and s_server. Abort handshake if no
2421 shared signature algorithms.
2424 *) Add new functions to allow customised supported signature algorithms
2425 for SSL and SSL_CTX structures. Add options to s_client and s_server
2429 *) New function SSL_certs_clear() to delete all references to certificates
2430 from an SSL structure. Before this once a certificate had been added
2431 it couldn't be removed.
2434 *) Integrate hostname, email address and IP address checking with certificate
2435 verification. New verify options supporting checking in openssl utility.
2438 *) Fixes and wildcard matching support to hostname and email checking
2439 functions. Add manual page.
2440 [Florian Weimer (Red Hat Product Security Team)]
2442 *) New functions to check a hostname email or IP address against a
2443 certificate. Add options x509 utility to print results of checks against
2447 *) Fix OCSP checking.
2448 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2450 *) Initial experimental support for explicitly trusted non-root CAs.
2451 OpenSSL still tries to build a complete chain to a root but if an
2452 intermediate CA has a trust setting included that is used. The first
2453 setting is used: whether to trust (e.g., -addtrust option to the x509
2457 *) Add -trusted_first option which attempts to find certificates in the
2458 trusted store even if an untrusted chain is also supplied.
2461 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2462 platform support for Linux and Android.
2465 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2468 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2469 When in FIPS mode the approved implementations are used as normal,
2470 when not in FIPS mode the internal unapproved versions are used instead.
2471 This means that the FIPS capable OpenSSL isn't forced to use the
2472 (often lower performance) FIPS implementations outside FIPS mode.
2475 *) Transparently support X9.42 DH parameters when calling
2476 PEM_read_bio_DHparameters. This means existing applications can handle
2477 the new parameter format automatically.
2480 *) Initial experimental support for X9.42 DH parameter format: mainly
2481 to support use of 'q' parameter for RFC5114 parameters.
2484 *) Add DH parameters from RFC5114 including test data to dhtest.
2487 *) Support for automatic EC temporary key parameter selection. If enabled
2488 the most preferred EC parameters are automatically used instead of
2489 hardcoded fixed parameters. Now a server just has to call:
2490 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2491 support ECDH and use the most appropriate parameters.
2494 *) Enhance and tidy EC curve and point format TLS extension code. Use
2495 static structures instead of allocation if default values are used.
2496 New ctrls to set curves we wish to support and to retrieve shared curves.
2497 Print out shared curves in s_server. New options to s_server and s_client
2498 to set list of supported curves.
2501 *) New ctrls to retrieve supported signature algorithms and
2502 supported curve values as an array of NIDs. Extend openssl utility
2503 to print out received values.
2506 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2507 between NIDs and the more common NIST names such as "P-256". Enhance
2508 ecparam utility and ECC method to recognise the NIST names for curves.
2511 *) Enhance SSL/TLS certificate chain handling to support different
2512 chains for each certificate instead of one chain in the parent SSL_CTX.
2515 *) Support for fixed DH ciphersuite client authentication: where both
2516 server and client use DH certificates with common parameters.
2519 *) Support for fixed DH ciphersuites: those requiring DH server
2523 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2525 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2526 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2527 X509_CINF_get_signature were reverted post internal team review.
2529 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2531 *) Build fixes for the Windows and OpenVMS platforms
2532 [Matt Caswell and Richard Levitte]
2534 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2536 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2537 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2538 dereference. This could lead to a Denial Of Service attack. Thanks to
2539 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2543 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2544 dtls1_buffer_record function under certain conditions. In particular this
2545 could occur if an attacker sent repeated DTLS records with the same
2546 sequence number but for the next epoch. The memory leak could be exploited
2547 by an attacker in a Denial of Service attack through memory exhaustion.
2548 Thanks to Chris Mueller for reporting this issue.
2552 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2553 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2554 method would be set to NULL which could later result in a NULL pointer
2555 dereference. Thanks to Frank Schmirler for reporting this issue.
2559 *) Abort handshake if server key exchange message is omitted for ephemeral
2562 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2563 reporting this issue.
2567 *) Remove non-export ephemeral RSA code on client and server. This code
2568 violated the TLS standard by allowing the use of temporary RSA keys in
2569 non-export ciphersuites and could be used by a server to effectively
2570 downgrade the RSA key length used to a value smaller than the server
2571 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2572 INRIA or reporting this issue.
2576 *) Fixed issue where DH client certificates are accepted without verification.
2577 An OpenSSL server will accept a DH certificate for client authentication
2578 without the certificate verify message. This effectively allows a client to
2579 authenticate without the use of a private key. This only affects servers
2580 which trust a client certificate authority which issues certificates
2581 containing DH keys: these are extremely rare and hardly ever encountered.
2582 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2587 *) Ensure that the session ID context of an SSL is updated when its
2588 SSL_CTX is updated via SSL_set_SSL_CTX.
2590 The session ID context is typically set from the parent SSL_CTX,
2591 and can vary with the CTX.
2594 *) Fix various certificate fingerprint issues.
2596 By using non-DER or invalid encodings outside the signed portion of a
2597 certificate the fingerprint can be changed without breaking the signature.
2598 Although no details of the signed portion of the certificate can be changed
2599 this can cause problems with some applications: e.g. those using the
2600 certificate fingerprint for blacklists.
2602 1. Reject signatures with non zero unused bits.
2604 If the BIT STRING containing the signature has non zero unused bits reject
2605 the signature. All current signature algorithms require zero unused bits.
2607 2. Check certificate algorithm consistency.
2609 Check the AlgorithmIdentifier inside TBS matches the one in the
2610 certificate signature. NB: this will result in signature failure
2611 errors for some broken certificates.
2613 Thanks to Konrad Kraszewski from Google for reporting this issue.
2615 3. Check DSA/ECDSA signatures use DER.
2617 Re-encode DSA/ECDSA signatures and compare with the original received
2618 signature. Return an error if there is a mismatch.
2620 This will reject various cases including garbage after signature
2621 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2622 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2623 (negative or with leading zeroes).
2625 Further analysis was conducted and fixes were developed by Stephen Henson
2626 of the OpenSSL core team.
2631 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2632 results on some platforms, including x86_64. This bug occurs at random
2633 with a very low probability, and is not known to be exploitable in any
2634 way, though its exact impact is difficult to determine. Thanks to Pieter
2635 Wuille (Blockstream) who reported this issue and also suggested an initial
2636 fix. Further analysis was conducted by the OpenSSL development team and
2637 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2638 the OpenSSL core team.
2642 *) Do not resume sessions on the server if the negotiated protocol
2643 version does not match the session's version. Resuming with a different
2644 version, while not strictly forbidden by the RFC, is of questionable
2645 sanity and breaks all known clients.
2646 [David Benjamin, Emilia Käsper]
2648 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2649 early CCS messages during renegotiation. (Note that because
2650 renegotiation is encrypted, this early CCS was not exploitable.)
2653 *) Tighten client-side session ticket handling during renegotiation:
2654 ensure that the client only accepts a session ticket if the server sends
2655 the extension anew in the ServerHello. Previously, a TLS client would
2656 reuse the old extension state and thus accept a session ticket if one was
2657 announced in the initial ServerHello.
2659 Similarly, ensure that the client requires a session ticket if one
2660 was advertised in the ServerHello. Previously, a TLS client would
2661 ignore a missing NewSessionTicket message.
2664 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2666 *) SRTP Memory Leak.
2668 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2669 sends a carefully crafted handshake message, to cause OpenSSL to fail
2670 to free up to 64k of memory causing a memory leak. This could be
2671 exploited in a Denial Of Service attack. This issue affects OpenSSL
2672 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2673 whether SRTP is used or configured. Implementations of OpenSSL that
2674 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2676 The fix was developed by the OpenSSL team.
2680 *) Session Ticket Memory Leak.
2682 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2683 integrity of that ticket is first verified. In the event of a session
2684 ticket integrity check failing, OpenSSL will fail to free memory
2685 causing a memory leak. By sending a large number of invalid session
2686 tickets an attacker could exploit this issue in a Denial Of Service
2691 *) Build option no-ssl3 is incomplete.
2693 When OpenSSL is configured with "no-ssl3" as a build option, servers
2694 could accept and complete a SSL 3.0 handshake, and clients could be
2695 configured to send them.
2697 [Akamai and the OpenSSL team]
2699 *) Add support for TLS_FALLBACK_SCSV.
2700 Client applications doing fallback retries should call
2701 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2703 [Adam Langley, Bodo Moeller]
2705 *) Add additional DigestInfo checks.
2707 Re-encode DigestInto in DER and check against the original when
2708 verifying RSA signature: this will reject any improperly encoded
2709 DigestInfo structures.
2711 Note: this is a precautionary measure and no attacks are currently known.
2715 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2717 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2718 SRP code can be overrun an internal buffer. Add sanity check that
2719 g, A, B < N to SRP code.
2721 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2722 Group for discovering this issue.
2726 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2727 TLS 1.0 instead of higher protocol versions when the ClientHello message
2728 is badly fragmented. This allows a man-in-the-middle attacker to force a
2729 downgrade to TLS 1.0 even if both the server and the client support a
2730 higher protocol version, by modifying the client's TLS records.
2732 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2733 researching this issue.
2737 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2738 to a denial of service attack. A malicious server can crash the client
2739 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2740 ciphersuite and sending carefully crafted handshake messages.
2742 Thanks to Felix Gröbert (Google) for discovering and researching this
2747 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2748 to leak memory. This can be exploited through a Denial of Service attack.
2749 Thanks to Adam Langley for discovering and researching this issue.
2753 *) An attacker can force openssl to consume large amounts of memory whilst
2754 processing DTLS handshake messages. This can be exploited through a
2755 Denial of Service attack.
2756 Thanks to Adam Langley for discovering and researching this issue.
2760 *) An attacker can force an error condition which causes openssl to crash
2761 whilst processing DTLS packets due to memory being freed twice. This
2762 can be exploited through a Denial of Service attack.
2763 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2768 *) If a multithreaded client connects to a malicious server using a resumed
2769 session and the server sends an ec point format extension it could write
2770 up to 255 bytes to freed memory.
2772 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2777 *) A malicious server can crash an OpenSSL client with a null pointer
2778 dereference (read) by specifying an SRP ciphersuite even though it was not
2779 properly negotiated with the client. This can be exploited through a
2780 Denial of Service attack.
2782 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2783 discovering and researching this issue.
2787 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2788 X509_name_oneline, X509_name_print_ex et al. to leak some information
2789 from the stack. Applications may be affected if they echo pretty printing
2790 output to the attacker.
2792 Thanks to Ivan Fratric (Google) for discovering this issue.
2794 [Emilia Käsper, and Steve Henson]
2796 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2797 for corner cases. (Certain input points at infinity could lead to
2798 bogus results, with non-infinity inputs mapped to infinity too.)
2801 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2803 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2804 handshake can force the use of weak keying material in OpenSSL
2805 SSL/TLS clients and servers.
2807 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2808 researching this issue. (CVE-2014-0224)
2809 [KIKUCHI Masashi, Steve Henson]
2811 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2812 OpenSSL DTLS client the code can be made to recurse eventually crashing
2815 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2817 [Imre Rad, Steve Henson]
2819 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2820 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2821 client or server. This is potentially exploitable to run arbitrary
2822 code on a vulnerable client or server.
2824 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2825 [Jüri Aedla, Steve Henson]
2827 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2828 are subject to a denial of service attack.
2830 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2831 this issue. (CVE-2014-3470)
2832 [Felix Gröbert, Ivan Fratric, Steve Henson]
2834 *) Harmonize version and its documentation. -f flag is used to display
2836 [mancha <mancha1@zoho.com>]
2838 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2839 in i2d_ECPrivateKey.
2840 [mancha <mancha1@zoho.com>]
2842 *) Fix some double frees. These are not thought to be exploitable.
2843 [mancha <mancha1@zoho.com>]
2845 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2847 *) A missing bounds check in the handling of the TLS heartbeat extension
2848 can be used to reveal up to 64k of memory to a connected client or
2851 Thanks for Neel Mehta of Google Security for discovering this bug and to
2852 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2853 preparing the fix (CVE-2014-0160)
2854 [Adam Langley, Bodo Moeller]
2856 *) Fix for the attack described in the paper "Recovering OpenSSL
2857 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2858 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2859 http://eprint.iacr.org/2014/140
2861 Thanks to Yuval Yarom and Naomi Benger for discovering this
2862 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2863 [Yuval Yarom and Naomi Benger]
2865 *) TLS pad extension: draft-agl-tls-padding-03
2867 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2868 TLS client Hello record length value would otherwise be > 255 and
2869 less that 512 pad with a dummy extension containing zeroes so it
2870 is at least 512 bytes long.
2872 [Adam Langley, Steve Henson]
2874 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2876 *) Fix for TLS record tampering bug. A carefully crafted invalid
2877 handshake could crash OpenSSL with a NULL pointer exception.
2878 Thanks to Anton Johansson for reporting this issues.
2881 *) Keep original DTLS digest and encryption contexts in retransmission
2882 structures so we can use the previous session parameters if they need
2883 to be resent. (CVE-2013-6450)
2886 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2887 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2888 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2889 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2890 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2891 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2892 [Rob Stradling, Adam Langley]
2894 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2896 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2897 supporting platforms or when small records were transferred.
2898 [Andy Polyakov, Steve Henson]
2900 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2902 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2904 This addresses the flaw in CBC record processing discovered by
2905 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2906 at: http://www.isg.rhul.ac.uk/tls/
2908 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2909 Security Group at Royal Holloway, University of London
2910 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2911 Emilia Käsper for the initial patch.
2913 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2915 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2916 ciphersuites which can be exploited in a denial of service attack.
2917 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2918 and detecting this bug and to Wolfgang Ettlinger
2919 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2923 *) Return an error when checking OCSP signatures when key is NULL.
2924 This fixes a DoS attack. (CVE-2013-0166)
2927 *) Make openssl verify return errors.
2928 [Chris Palmer <palmer@google.com> and Ben Laurie]
2930 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2931 the right response is stapled. Also change SSL_get_certificate()
2932 so it returns the certificate actually sent.
2933 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2934 [Rob Stradling <rob.stradling@comodo.com>]
2936 *) Fix possible deadlock when decoding public keys.
2939 *) Don't use TLS 1.0 record version number in initial client hello
2943 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2945 *) Sanity check record length before skipping explicit IV in TLS
2946 1.2, 1.1 and DTLS to fix DoS attack.
2948 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2949 fuzzing as a service testing platform.
2953 *) Initialise tkeylen properly when encrypting CMS messages.
2954 Thanks to Solar Designer of Openwall for reporting this issue.
2957 *) In FIPS mode don't try to use composite ciphers as they are not
2961 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2963 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2964 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2965 mean any application compiled against OpenSSL 1.0.0 headers setting
2966 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2967 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2968 0x10000000L Any application which was previously compiled against
2969 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2970 will need to be recompiled as a result. Letting be results in
2971 inability to disable specifically TLS 1.1 and in client context,
2972 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2975 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2976 disable just protocol X, but all protocols above X *if* there are
2977 protocols *below* X still enabled. In more practical terms it means
2978 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2979 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2980 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2984 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2986 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2987 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2988 in CRYPTO_realloc_clean.
2990 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2991 issue and to Adam Langley <agl@chromium.org> for fixing it.
2993 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2995 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2998 *) Workarounds for some broken servers that "hang" if a client hello
2999 record length exceeds 255 bytes.
3001 1. Do not use record version number > TLS 1.0 in initial client
3002 hello: some (but not all) hanging servers will now work.
3003 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3004 the number of ciphers sent in the client hello. This should be
3005 set to an even number, such as 50, for example by passing:
3006 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3007 Most broken servers should now work.
3008 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3009 TLS 1.2 client support entirely.
3012 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3015 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3017 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3018 STRING form instead of a DigestInfo.
3021 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3022 and the RSA_sign/RSA_verify functions. This was made more apparent when
3023 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3024 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3025 the correct format in RSA_verify so both forms transparently work.
3028 *) Some servers which support TLS 1.0 can choke if we initially indicate
3029 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3030 encrypted premaster secret. As a workaround use the maximum permitted
3031 client version in client hello, this should keep such servers happy
3032 and still work with previous versions of OpenSSL.
3035 *) Add support for TLS/DTLS heartbeats.
3036 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3038 *) Add support for SCTP.
3039 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3041 *) Improved PRNG seeding for VOS.
3042 [Paul Green <Paul.Green@stratus.com>]
3044 *) Extensive assembler packs updates, most notably:
3046 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3047 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3048 - x86_64: bit-sliced AES implementation;
3049 - ARM: NEON support, contemporary platforms optimizations;
3050 - s390x: z196 support;
3051 - *: GHASH and GF(2^m) multiplication implementations;
3055 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3056 (removal of unnecessary code)
3057 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3059 *) Add TLS key material exporter from RFC 5705.
3062 *) Add DTLS-SRTP negotiation from RFC 5764.
3065 *) Add Next Protocol Negotiation,
3066 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3067 disabled with a no-npn flag to config or Configure. Code donated
3069 [Adam Langley <agl@google.com> and Ben Laurie]
3071 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3072 NIST-P256, NIST-P521, with constant-time single point multiplication on
3073 typical inputs. Compiler support for the nonstandard type __uint128_t is
3074 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3075 Code made available under Apache License version 2.0.
3077 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3078 line to include this in your build of OpenSSL, and run "make depend" (or
3079 "make update"). This enables the following EC_METHODs:
3081 EC_GFp_nistp224_method()
3082 EC_GFp_nistp256_method()
3083 EC_GFp_nistp521_method()
3085 EC_GROUP_new_by_curve_name() will automatically use these (while
3086 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3088 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3090 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3091 all platforms. Move ssize_t definition from e_os.h to the public
3092 header file e_os2.h as it now appears in public header file cms.h
3095 *) New -sigopt option to the ca, req and x509 utilities. Additional
3096 signature parameters can be passed using this option and in
3100 *) Add RSA PSS signing function. This will generate and set the
3101 appropriate AlgorithmIdentifiers for PSS based on those in the
3102 corresponding EVP_MD_CTX structure. No application support yet.
3105 *) Support for companion algorithm specific ASN1 signing routines.
3106 New function ASN1_item_sign_ctx() signs a pre-initialised
3107 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3108 the appropriate parameters.
3111 *) Add new algorithm specific ASN1 verification initialisation function
3112 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3113 handling will be the same no matter what EVP_PKEY_METHOD is used.
3114 Add a PSS handler to support verification of PSS signatures: checked
3115 against a number of sample certificates.
3118 *) Add signature printing for PSS. Add PSS OIDs.
3119 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3121 *) Add algorithm specific signature printing. An individual ASN1 method
3122 can now print out signatures instead of the standard hex dump.
3124 More complex signatures (e.g. PSS) can print out more meaningful
3125 information. Include DSA version that prints out the signature
3129 *) Password based recipient info support for CMS library: implementing
3133 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3134 neatly separates the code into cipher and PBE sections and is required
3135 for some algorithms that split PBES2 into separate pieces (such as
3136 password based CMS).
3139 *) Session-handling fixes:
3140 - Fix handling of connections that are resuming with a session ID,
3141 but also support Session Tickets.
3142 - Fix a bug that suppressed issuing of a new ticket if the client
3143 presented a ticket with an expired session.
3144 - Try to set the ticket lifetime hint to something reasonable.
3145 - Make tickets shorter by excluding irrelevant information.
3146 - On the client side, don't ignore renewed tickets.
3147 [Adam Langley, Bodo Moeller (Google)]
3149 *) Fix PSK session representation.
3152 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3154 This work was sponsored by Intel.
3157 *) Add GCM support to TLS library. Some custom code is needed to split
3158 the IV between the fixed (from PRF) and explicit (from TLS record)
3159 portions. This adds all GCM ciphersuites supported by RFC5288 and
3160 RFC5289. Generalise some AES* cipherstrings to include GCM and
3161 add a special AESGCM string for GCM only.
3164 *) Expand range of ctrls for AES GCM. Permit setting invocation
3165 field on decrypt and retrieval of invocation field only on encrypt.
3168 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3169 As required by RFC5289 these ciphersuites cannot be used if for
3170 versions of TLS earlier than 1.2.
3173 *) For FIPS capable OpenSSL interpret a NULL default public key method
3174 as unset and return the appropriate default but do *not* set the default.
3175 This means we can return the appropriate method in applications that
3176 switch between FIPS and non-FIPS modes.
3179 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3180 ENGINE is used then we cannot handle that in the FIPS module so we
3181 keep original code iff non-FIPS operations are allowed.
3184 *) Add -attime option to openssl utilities.
3185 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3187 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3190 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3191 FIPS EC methods unconditionally for now.
3194 *) New build option no-ec2m to disable characteristic 2 code.
3197 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3198 all cases can be covered as some introduce binary incompatibilities.
3201 *) Redirect RSA operations to FIPS module including keygen,
3202 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3205 *) Add similar low level API blocking to ciphers.
3208 *) Low level digest APIs are not approved in FIPS mode: any attempt
3209 to use these will cause a fatal error. Applications that *really* want
3210 to use them can use the private_* version instead.
3213 *) Redirect cipher operations to FIPS module for FIPS builds.
3216 *) Redirect digest operations to FIPS module for FIPS builds.
3219 *) Update build system to add "fips" flag which will link in fipscanister.o
3220 for static and shared library builds embedding a signature if needed.
3223 *) Output TLS supported curves in preference order instead of numerical
3224 order. This is currently hardcoded for the highest order curves first.
3225 This should be configurable so applications can judge speed vs strength.
3228 *) Add TLS v1.2 server support for client authentication.
3231 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3235 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3236 FIPS modules versions.
3239 *) Add TLS v1.2 client side support for client authentication. Keep cache
3240 of handshake records longer as we don't know the hash algorithm to use
3241 until after the certificate request message is received.
3244 *) Initial TLS v1.2 client support. Add a default signature algorithms
3245 extension including all the algorithms we support. Parse new signature
3246 format in client key exchange. Relax some ECC signing restrictions for
3247 TLS v1.2 as indicated in RFC5246.
3250 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3251 to new signature format when needed using client digest preference.
3252 All server ciphersuites should now work correctly in TLS v1.2. No client
3253 support yet and no support for client certificates.
3256 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3257 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3258 ciphersuites. At present only RSA key exchange ciphersuites work with
3259 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3260 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3261 and version checking.
3264 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3265 with this defined it will not be affected by any changes to ssl internal
3266 structures. Add several utility functions to allow openssl application
3267 to work with OPENSSL_NO_SSL_INTERN defined.
3270 *) A long standing patch to add support for SRP from EdelWeb (Peter
3271 Sylvester and Christophe Renou) was integrated.
3272 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3273 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3276 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3279 *) Permit abbreviated handshakes when renegotiating using the function
3280 SSL_renegotiate_abbreviated().
3281 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3283 *) Add call to ENGINE_register_all_complete() to
3284 ENGINE_load_builtin_engines(), so some implementations get used
3285 automatically instead of needing explicit application support.
3288 *) Add support for TLS key exporter as described in RFC5705.
3289 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3291 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3292 a few changes are required:
3294 Add SSL_OP_NO_TLSv1_1 flag.
3295 Add TLSv1_1 methods.
3296 Update version checking logic to handle version 1.1.
3297 Add explicit IV handling (ported from DTLS code).
3298 Add command line options to s_client/s_server.
3301 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3303 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3304 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3305 content decryption and always return the same error. Note: this attack
3306 needs on average 2^20 messages so it only affects automated senders. The
3307 old behaviour can be re-enabled in the CMS code by setting the
3308 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3309 an MMA defence is not necessary.
3310 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3311 this issue. (CVE-2012-0884)
3314 *) Fix CVE-2011-4619: make sure we really are receiving a
3315 client hello before rejecting multiple SGC restarts. Thanks to
3316 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3319 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3321 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3322 Thanks to Antonio Martin, Enterprise Secure Access Research and
3323 Development, Cisco Systems, Inc. for discovering this bug and
3324 preparing a fix. (CVE-2012-0050)
3327 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3329 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3330 of the Vaudenay padding oracle attack on CBC mode encryption
3331 which enables an efficient plaintext recovery attack against
3332 the OpenSSL implementation of DTLS. Their attack exploits timing
3333 differences arising during decryption processing. A research
3334 paper describing this attack can be found at:
3335 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3336 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3337 Security Group at Royal Holloway, University of London
3338 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3339 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3340 for preparing the fix. (CVE-2011-4108)
3341 [Robin Seggelmann, Michael Tuexen]
3343 *) Clear bytes used for block padding of SSL 3.0 records.
3345 [Adam Langley (Google)]
3347 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3348 Kadianakis <desnacked@gmail.com> for discovering this issue and
3349 Adam Langley for preparing the fix. (CVE-2011-4619)
3350 [Adam Langley (Google)]
3352 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3353 [Andrey Kulikov <amdeich@gmail.com>]
3355 *) Prevent malformed RFC3779 data triggering an assertion failure.
3356 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3357 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3358 [Rob Austein <sra@hactrn.net>]
3360 *) Improved PRNG seeding for VOS.
3361 [Paul Green <Paul.Green@stratus.com>]
3363 *) Fix ssl_ciph.c set-up race.
3364 [Adam Langley (Google)]
3366 *) Fix spurious failures in ecdsatest.c.
3367 [Emilia Käsper (Google)]
3369 *) Fix the BIO_f_buffer() implementation (which was mixing different
3370 interpretations of the '..._len' fields).
3371 [Adam Langley (Google)]
3373 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3374 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3375 threads won't reuse the same blinding coefficients.
3377 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3378 lock to call BN_BLINDING_invert_ex, and avoids one use of
3379 BN_BLINDING_update for each BN_BLINDING structure (previously,
3380 the last update always remained unused).
3381 [Emilia Käsper (Google)]
3383 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3384 [Bob Buckholz (Google)]
3386 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3388 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3389 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3390 [Kaspar Brand <ossl@velox.ch>]
3392 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3393 for multi-threaded use of ECDH. (CVE-2011-3210)
3394 [Adam Langley (Google)]
3396 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3399 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3400 signature public key algorithm by using OID xref utilities instead.
3401 Before this you could only use some ECC ciphersuites with SHA1 only.
3404 *) Add protection against ECDSA timing attacks as mentioned in the paper
3405 by Billy Bob Brumley and Nicola Tuveri, see:
3407 http://eprint.iacr.org/2011/232.pdf
3409 [Billy Bob Brumley and Nicola Tuveri]
3411 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3413 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3414 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3416 *) Fix bug in string printing code: if *any* escaping is enabled we must
3417 escape the escape character (backslash) or the resulting string is
3421 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3423 *) Disable code workaround for ancient and obsolete Netscape browsers
3424 and servers: an attacker can use it in a ciphersuite downgrade attack.
3425 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3428 *) Fixed J-PAKE implementation error, originally discovered by
3429 Sebastien Martini, further info and confirmation from Stefan
3430 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3433 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3435 *) Fix extension code to avoid race conditions which can result in a buffer
3436 overrun vulnerability: resumed sessions must not be modified as they can
3437 be shared by multiple threads. CVE-2010-3864
3440 *) Fix WIN32 build system to correctly link an ENGINE directory into
3444 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3446 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3448 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3450 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3452 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3453 context. The operation can be customised via the ctrl mechanism in
3454 case ENGINEs want to include additional functionality.
3457 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3460 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3461 output hashes compatible with older versions of OpenSSL.
3462 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3464 *) Fix compression algorithm handling: if resuming a session use the
3465 compression algorithm of the resumed session instead of determining
3466 it from client hello again. Don't allow server to change algorithm.
3469 *) Add load_crls() function to apps tidying load_certs() too. Add option
3470 to verify utility to allow additional CRLs to be included.
3473 *) Update OCSP request code to permit adding custom headers to the request:
3474 some responders need this.
3477 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3479 [Julia Lawall <julia@diku.dk>]
3481 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3482 needlessly dereferenced structures, used obsolete functions and
3483 didn't handle all updated verify codes correctly.
3486 *) Disable MD2 in the default configuration.
3489 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3490 indicate the initial BIO being pushed or popped. This makes it possible
3491 to determine whether the BIO is the one explicitly called or as a result
3492 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3493 it handles reference counts correctly and doesn't zero out the I/O bio
3494 when it is not being explicitly popped. WARNING: applications which
3495 included workarounds for the old buggy behaviour will need to be modified
3496 or they could free up already freed BIOs.
3499 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3500 renaming to all platforms (within the 0.9.8 branch, this was
3501 done conditionally on Netware platforms to avoid a name clash).
3502 [Guenter <lists@gknw.net>]
3504 *) Add ECDHE and PSK support to DTLS.
3505 [Michael Tuexen <tuexen@fh-muenster.de>]
3507 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3511 *) Add "missing" function EVP_MD_flags() (without this the only way to
3512 retrieve a digest flags is by accessing the structure directly. Update
3513 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3514 or cipher is registered as in the "from" argument. Print out all
3515 registered digests in the dgst usage message instead of manually
3516 attempting to work them out.
3519 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3520 this allows the use of compression and extensions. Change default cipher
3521 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3522 by default unless an application cipher string requests it.
3525 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3526 key ids to find matching certificates and keys but some PKCS#12 files
3527 don't follow the (somewhat unwritten) rules and this strategy fails.
3528 Now just gather all certificates together and the first private key
3529 then look for the first certificate that matches the key.
3532 *) Support use of registered digest and cipher names for dgst and cipher
3533 commands instead of having to add each one as a special case. So now
3540 openssl dgst -sha256 foo
3542 and this works for ENGINE based algorithms too.
3546 *) Update Gost ENGINE to support parameter files.
3547 [Victor B. Wagner <vitus@cryptocom.ru>]
3549 *) Support GeneralizedTime in ca utility.
3550 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3552 *) Enhance the hash format used for certificate directory links. The new
3553 form uses the canonical encoding (meaning equivalent names will work
3554 even if they aren't identical) and uses SHA1 instead of MD5. This form
3555 is incompatible with the older format and as a result c_rehash should
3556 be used to rebuild symbolic links.
3559 *) Make PKCS#8 the default write format for private keys, replacing the
3560 traditional format. This form is standardised, more secure and doesn't
3561 include an implicit MD5 dependency.
3564 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3565 committed to OpenSSL should pass this lot as a minimum.
3568 *) Add session ticket override functionality for use by EAP-FAST.
3569 [Jouni Malinen <j@w1.fi>]
3571 *) Modify HMAC functions to return a value. Since these can be implemented
3572 in an ENGINE errors can occur.
3575 *) Type-checked OBJ_bsearch_ex.
3578 *) Type-checked OBJ_bsearch. Also some constification necessitated
3579 by type-checking. Still to come: TXT_DB, bsearch(?),
3580 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3584 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3585 seconds to a tm structure directly, instead of going through OS
3586 specific date routines. This avoids any issues with OS routines such
3587 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3588 and X509_time_adj_ex() to cover the extended range. The existing
3589 X509_time_adj() is still usable and will no longer have any date issues.
3592 *) Delta CRL support. New use deltas option which will attempt to locate
3593 and search any appropriate delta CRLs available.
3595 This work was sponsored by Google.
3598 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3599 code and add additional score elements. Validate alternate CRL paths
3600 as part of the CRL checking and indicate a new error "CRL path validation
3601 error" in this case. Applications wanting additional details can use
3602 the verify callback and check the new "parent" field. If this is not
3603 NULL CRL path validation is taking place. Existing applications won't
3604 see this because it requires extended CRL support which is off by
3607 This work was sponsored by Google.
3610 *) Support for freshest CRL extension.
3612 This work was sponsored by Google.
3615 *) Initial indirect CRL support. Currently only supported in the CRLs
3616 passed directly and not via lookup. Process certificate issuer
3617 CRL entry extension and lookup CRL entries by bother issuer name
3618 and serial number. Check and process CRL issuer entry in IDP extension.
3620 This work was sponsored by Google.
3623 *) Add support for distinct certificate and CRL paths. The CRL issuer
3624 certificate is validated separately in this case. Only enabled if
3625 an extended CRL support flag is set: this flag will enable additional
3626 CRL functionality in future.
3628 This work was sponsored by Google.
3631 *) Add support for policy mappings extension.
3633 This work was sponsored by Google.
3636 *) Fixes to pathlength constraint, self issued certificate handling,
3637 policy processing to align with RFC3280 and PKITS tests.
3639 This work was sponsored by Google.
3642 *) Support for name constraints certificate extension. DN, email, DNS
3643 and URI types are currently supported.
3645 This work was sponsored by Google.
3648 *) To cater for systems that provide a pointer-based thread ID rather
3649 than numeric, deprecate the current numeric thread ID mechanism and
3650 replace it with a structure and associated callback type. This
3651 mechanism allows a numeric "hash" to be extracted from a thread ID in
3652 either case, and on platforms where pointers are larger than 'long',
3653 mixing is done to help ensure the numeric 'hash' is usable even if it
3654 can't be guaranteed unique. The default mechanism is to use "&errno"
3655 as a pointer-based thread ID to distinguish between threads.
3657 Applications that want to provide their own thread IDs should now use
3658 CRYPTO_THREADID_set_callback() to register a callback that will call
3659 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3661 Note that ERR_remove_state() is now deprecated, because it is tied
3662 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3663 to free the current thread's error state should be replaced by
3664 ERR_remove_thread_state(NULL).
3666 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3667 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3668 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3669 application was previously providing a numeric thread callback that
3670 was inappropriate for distinguishing threads, then uniqueness might
3671 have been obtained with &errno that happened immediately in the
3672 intermediate development versions of OpenSSL; this is no longer the
3673 case, the numeric thread callback will now override the automatic use
3675 [Geoff Thorpe, with help from Bodo Moeller]
3677 *) Initial support for different CRL issuing certificates. This covers a
3678 simple case where the self issued certificates in the chain exist and
3679 the real CRL issuer is higher in the existing chain.
3681 This work was sponsored by Google.
3684 *) Removed effectively defunct crypto/store from the build.
3687 *) Revamp of STACK to provide stronger type-checking. Still to come:
3688 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3689 ASN1_STRING, CONF_VALUE.
3692 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3693 RAM on SSL connections. This option can save about 34k per idle SSL.
3696 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3697 STACK, TXT_DB, bsearch, qsort.
3700 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3701 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3702 support for data, signedData, compressedData, digestedData and
3703 encryptedData, envelopedData types included. Scripts to check against
3704 RFC4134 examples draft and interop and consistency checks of many
3705 content types and variants.
3708 *) Add options to enc utility to support use of zlib compression BIO.
3711 *) Extend mk1mf to support importing of options and assembly language
3712 files from Configure script, currently only included in VC-WIN32.
3713 The assembly language rules can now optionally generate the source
3714 files from the associated perl scripts.
3717 *) Implement remaining functionality needed to support GOST ciphersuites.
3718 Interop testing has been performed using CryptoPro implementations.
3719 [Victor B. Wagner <vitus@cryptocom.ru>]
3721 *) s390x assembler pack.
3724 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3728 *) Implement Opaque PRF Input TLS extension as specified in
3729 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3730 official specification yet and no extension type assignment by
3731 IANA exists, this extension (for now) will have to be explicitly
3732 enabled when building OpenSSL by providing the extension number
3733 to use. For example, specify an option
3735 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3737 to the "config" or "Configure" script to enable the extension,
3738 assuming extension number 0x9527 (which is a completely arbitrary
3739 and unofficial assignment based on the MD5 hash of the Internet
3740 Draft). Note that by doing so, you potentially lose
3741 interoperability with other TLS implementations since these might
3742 be using the same extension number for other purposes.
3744 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3745 opaque PRF input value to use in the handshake. This will create
3746 an interal copy of the length-'len' string at 'src', and will
3747 return non-zero for success.
3749 To get more control and flexibility, provide a callback function
3752 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3753 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3757 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3760 Callback function 'cb' will be called in handshakes, and is
3761 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3762 Argument 'arg' is for application purposes (the value as given to
3763 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3764 be provided to the callback function). The callback function
3765 has to return non-zero to report success: usually 1 to use opaque
3766 PRF input just if possible, or 2 to enforce use of the opaque PRF
3767 input. In the latter case, the library will abort the handshake
3768 if opaque PRF input is not successfully negotiated.
3770 Arguments 'peerinput' and 'len' given to the callback function
3771 will always be NULL and 0 in the case of a client. A server will
3772 see the client's opaque PRF input through these variables if
3773 available (NULL and 0 otherwise). Note that if the server
3774 provides an opaque PRF input, the length must be the same as the
3775 length of the client's opaque PRF input.
3777 Note that the callback function will only be called when creating
3778 a new session (session resumption can resume whatever was
3779 previously negotiated), and will not be called in SSL 2.0
3780 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3781 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3782 for applications that need to enforce opaque PRF input.
3786 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3789 [Victor B. Wagner <vitus@cryptocom.ru>]
3791 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3792 RFC4507bis. The encrypted ticket format is an encrypted encoded
3793 SSL_SESSION structure, that way new session features are automatically
3796 If a client application caches session in an SSL_SESSION structure
3797 support is transparent because tickets are now stored in the encoded
3800 The SSL_CTX structure automatically generates keys for ticket
3801 protection in servers so again support should be possible
3802 with no application modification.
3804 If a client or server wishes to disable RFC4507 support then the option
3805 SSL_OP_NO_TICKET can be set.
3807 Add a TLS extension debugging callback to allow the contents of any client
3808 or server extensions to be examined.
3810 This work was sponsored by Google.
3813 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3814 OpenSSL should now compile cleanly on gcc 4.2
3815 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3817 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3818 support including streaming MAC support: this is required for GOST
3819 ciphersuite support.
3820 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3822 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3823 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3824 to output in BER and PEM format.
3827 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3828 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3829 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3830 ENGINE support for HMAC keys which are unextractable. New -mac and
3831 -macopt options to dgst utility.
3834 *) New option -sigopt to dgst utility. Update dgst to use
3835 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3836 alternative signing parameters such as X9.31 or PSS in the dgst
3840 *) Change ssl_cipher_apply_rule(), the internal function that does
3841 the work each time a ciphersuite string requests enabling
3842 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3843 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3844 the order of disabled ciphersuites such that those ciphersuites
3845 that most recently went from enabled to disabled not only stay
3846 in order with respect to each other, but also have higher priority
3847 than other disabled ciphersuites the next time ciphersuites are
3850 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3851 the same ciphersuites as with "HIGH" alone, but in a specific
3852 order where the PSK ciphersuites come first (since they are the
3853 most recently disabled ciphersuites when "HIGH" is parsed).
3855 Also, change ssl_create_cipher_list() (using this new
3856 funcionality) such that between otherwise identical
3857 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3861 *) Change ssl_create_cipher_list() so that it automatically
3862 arranges the ciphersuites in reasonable order before starting
3863 to process the rule string. Thus, the definition for "DEFAULT"
3864 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3865 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3866 This makes it much easier to arrive at a reasonable default order
3867 in applications for which anonymous ciphers are OK (meaning
3868 that you can't actually use DEFAULT).
3869 [Bodo Moeller; suggested by Victor Duchovni]
3871 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3872 processing) into multiple integers instead of setting
3873 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3874 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3875 (These masks as well as the individual bit definitions are hidden
3876 away into the non-exported interface ssl/ssl_locl.h, so this
3877 change to the definition of the SSL_CIPHER structure shouldn't
3878 affect applications.) This give us more bits for each of these
3879 categories, so there is no longer a need to coagulate AES128 and
3880 AES256 into a single algorithm bit, and to coagulate Camellia128
3881 and Camellia256 into a single algorithm bit, which has led to all
3884 Thus, among other things, the kludge introduced in 0.9.7m and
3885 0.9.8e for masking out AES256 independently of AES128 or masking
3886 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3888 With the change, we also introduce new ciphersuite aliases that
3889 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3893 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3894 Use the leftmost N bytes of the signature input if the input is
3895 larger than the prime q (with N being the size in bytes of q).
3898 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3899 it yet and it is largely untested.
3902 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3905 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3906 some compilers (gcc 4.2 and later) reject their use. Safestack is
3907 reimplemented. Update ASN1 to avoid use of legacy functions.
3910 *) Win32/64 targets are linked with Winsock2.
3913 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3914 to external functions. This can be used to increase CRL handling
3915 efficiency especially when CRLs are very large by (for example) storing
3916 the CRL revoked certificates in a database.
3919 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3920 new CRLs added to a directory can be used. New command line option
3921 -verify_return_error to s_client and s_server. This causes real errors
3922 to be returned by the verify callback instead of carrying on no matter
3923 what. This reflects the way a "real world" verify callback would behave.
3926 *) GOST engine, supporting several GOST algorithms and public key formats.
3927 Kindly donated by Cryptocom.
3930 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3931 partitioned by DP are handled but no indirect CRL or reason partitioning
3932 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3933 selected via a scoring technique which handles IDP and AKID in CRLs.
3936 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3937 will ultimately be used for all verify operations: this will remove the
3938 X509_STORE dependency on certificate verification and allow alternative
3939 lookup methods. X509_STORE based implementations of these two callbacks.
3942 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3943 Modify get_crl() to find a valid (unexpired) CRL if possible.
3946 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3947 this would be called X509_CRL_cmp() but that name is already used by
3948 a function that just compares CRL issuer names. Cache several CRL
3949 extensions in X509_CRL structure and cache CRLDP in X509.
3952 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3953 this maps equivalent X509_NAME structures into a consistent structure.
3954 Name comparison can then be performed rapidly using memcmp().
3957 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3961 *) Allow digests to supply their own micalg string for S/MIME type using
3962 the ctrl EVP_MD_CTRL_MICALG.
3965 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3966 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3967 ctrl. It can then customise the structure before and/or after signing
3971 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3972 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3973 to free up any added signature OIDs.
3976 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3977 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3978 digest and cipher tables. New options added to openssl utility:
3979 list-message-digest-algorithms and list-cipher-algorithms.
3982 *) Change the array representation of binary polynomials: the list
3983 of degrees of non-zero coefficients is now terminated with -1.
3984 Previously it was terminated with 0, which was also part of the
3985 value; thus, the array representation was not applicable to
3986 polynomials where t^0 has coefficient zero. This change makes
3987 the array representation useful in a more general context.
3990 *) Various modifications and fixes to SSL/TLS cipher string
3991 handling. For ECC, the code now distinguishes between fixed ECDH
3992 with RSA certificates on the one hand and with ECDSA certificates
3993 on the other hand, since these are separate ciphersuites. The
3994 unused code for Fortezza ciphersuites has been removed.
3996 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3997 (not "ECDHE"). For consistency with the code for DH
3998 certificates, use of ECDH certificates is now considered ECDH
3999 authentication, not RSA or ECDSA authentication (the latter is
4000 merely the CA's signing algorithm and not actively used in the
4003 The temporary ciphersuite alias "ECCdraft" is no longer
4004 available, and ECC ciphersuites are no longer excluded from "ALL"
4005 and "DEFAULT". The following aliases now exist for RFC 4492
4006 ciphersuites, most of these by analogy with the DH case:
4008 kECDHr - ECDH cert, signed with RSA
4009 kECDHe - ECDH cert, signed with ECDSA
4010 kECDH - ECDH cert (signed with either RSA or ECDSA)
4011 kEECDH - ephemeral ECDH
4012 ECDH - ECDH cert or ephemeral ECDH
4018 AECDH - anonymous ECDH
4019 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4023 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4024 Use correct micalg parameters depending on digest(s) in signed message.
4027 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4028 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4031 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4032 an engine to register a method. Add ENGINE lookups for methods and
4033 functional reference processing.
4036 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
4037 EVP_{Sign,Verify}* which allow an application to customise the signature
4041 *) New -resign option to smime utility. This adds one or more signers
4042 to an existing PKCS#7 signedData structure. Also -md option to use an
4043 alternative message digest algorithm for signing.
4046 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4047 create PKCS7 structures containing multiple signers. Update smime
4048 application to support multiple signers.
4051 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4055 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4056 Reorganize PBE internals to lookup from a static table using NIDs,
4057 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4058 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4059 PRF which will be automatically used with PBES2.
4062 *) Replace the algorithm specific calls to generate keys in "req" with the
4066 *) Update PKCS#7 enveloped data routines to use new API. This is now
4067 supported by any public key method supporting the encrypt operation. A
4068 ctrl is added to allow the public key algorithm to examine or modify
4069 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4073 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4074 a default digest type to use. In most cases this will be SHA1 but some
4075 algorithms (such as GOST) need to specify an alternative digest. The
4076 return value indicates how strong the preference is 1 means optional and
4077 2 is mandatory (that is it is the only supported type). Modify
4078 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4079 use the default md. Update openssl utilities to use the default digest
4080 type for signing if it is not explicitly indicated.
4083 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4084 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4085 signing method from the key type. This effectively removes the link
4086 between digests and public key types.
4089 *) Add an OID cross reference table and utility functions. Its purpose is to
4090 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4091 rsaEncryption. This will allow some of the algorithm specific hackery
4092 needed to use the correct OID to be removed.
4095 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4096 structures for PKCS7_sign(). They are now set up by the relevant public
4100 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4103 *) Add support for key derivation (agreement) in the API, DH method and
4107 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4108 public and private key formats. As a side effect these add additional
4109 command line functionality not previously available: DSA signatures can be
4110 generated and verified using pkeyutl and DH key support and generation in
4115 [Oliver Tappe <zooey@hirschkaefer.de>]
4117 *) New make target "install_html_docs" installs HTML renditions of the
4119 [Oliver Tappe <zooey@hirschkaefer.de>]
4121 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4122 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4123 support key and parameter generation and add initial key generation
4124 functionality for RSA.
4127 *) Add functions for main EVP_PKEY_method operations. The undocumented
4128 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4129 EVP_PKEY_{encrypt,decrypt}_old.
4132 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4133 key API, doesn't do much yet.
4136 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4137 public key algorithms. New option to openssl utility:
4138 "list-public-key-algorithms" to print out info.
4141 *) Implement the Supported Elliptic Curves Extension for
4142 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4145 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4146 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4149 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4150 utilities such as rsa, dsa, dsaparam etc except they process any key
4154 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4155 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4156 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4160 *) Initial support for pluggable public key ASN1.
4161 De-spaghettify the public key ASN1 handling. Move public and private
4162 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4163 algorithm specific handling to a single module within the relevant
4164 algorithm directory. Add functions to allow (near) opaque processing
4165 of public and private key structures.
4168 *) Implement the Supported Point Formats Extension for
4169 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4172 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4173 for the psk identity [hint] and the psk callback functions to the
4174 SSL_SESSION, SSL and SSL_CTX structure.
4177 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4181 SSL_CTX_use_psk_identity_hint
4182 SSL_get_psk_identity_hint
4183 SSL_get_psk_identity
4184 SSL_use_psk_identity_hint
4186 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4188 *) Add RFC 3161 compliant time stamp request creation, response generation
4189 and response verification functionality.
4190 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4192 *) Add initial support for TLS extensions, specifically for the server_name
4193 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4194 have new members for a host name. The SSL data structure has an
4195 additional member SSL_CTX *initial_ctx so that new sessions can be
4196 stored in that context to allow for session resumption, even after the
4197 SSL has been switched to a new SSL_CTX in reaction to a client's
4198 server_name extension.
4200 New functions (subject to change):
4202 SSL_get_servername()
4203 SSL_get_servername_type()
4206 New CTRL codes and macros (subject to change):
4208 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4209 - SSL_CTX_set_tlsext_servername_callback()
4210 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4211 - SSL_CTX_set_tlsext_servername_arg()
4212 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4214 openssl s_client has a new '-servername ...' option.
4216 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4217 '-key2 ...', '-servername_fatal' (subject to change). This allows
4218 testing the HostName extension for a specific single host name ('-cert'
4219 and '-key' remain fallbacks for handshakes without HostName
4220 negotiation). If the unrecognized_name alert has to be sent, this by
4221 default is a warning; it becomes fatal with the '-servername_fatal'
4224 [Peter Sylvester, Remy Allais, Christophe Renou]
4226 *) Whirlpool hash implementation is added.
4229 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4230 bn(64,32). Because of instruction set limitations it doesn't have
4231 any negative impact on performance. This was done mostly in order
4232 to make it possible to share assembler modules, such as bn_mul_mont
4233 implementations, between 32- and 64-bit builds without hassle.
4236 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4237 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4241 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4242 dedicated Montgomery multiplication procedure, is introduced.
4243 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4244 "64-bit" performance on certain 32-bit targets.
4247 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4248 in SSL structures. New SSL ctrl to set maximum send fragment size.
4249 Save memory by seeting the I/O buffer sizes dynamically instead of
4250 using the maximum available value.
4253 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4254 in addition to the text details.
4257 *) Very, very preliminary EXPERIMENTAL support for printing of general
4258 ASN1 structures. This currently produces rather ugly output and doesn't
4259 handle several customised structures at all.
4262 *) Integrated support for PVK file format and some related formats such
4263 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4264 these in the 'rsa' and 'dsa' utilities.
4267 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4270 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4271 place for the (very old) "NETSCAPE" format certificates which are now
4272 handled using new ASN1 code equivalents.
4275 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4276 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4277 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4280 *) Modify CRL distribution points extension code to print out previously
4281 unsupported fields. Enhance extension setting code to allow setting of
4285 *) Add print and set support for Issuing Distribution Point CRL extension.
4288 *) Change 'Configure' script to enable Camellia by default.
4291 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4293 *) When rejecting SSL/TLS records due to an incorrect version number, never
4294 update s->server with a new major version number. As of
4295 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4296 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4297 the previous behavior could result in a read attempt at NULL when
4298 receiving specific incorrect SSL/TLS records once record payload
4299 protection is active. (CVE-2010-0740)
4300 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4302 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4303 could be crashed if the relevant tables were not present (e.g. chrooted).
4304 [Tomas Hoger <thoger@redhat.com>]
4306 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4308 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
4309 [Martin Olsson, Neel Mehta]
4311 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4312 accommodate for stack sorting, always a write lock!).
4315 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4316 excessive delays in the RAND_poll(): over a minute. As a workaround
4317 include a time check in the inner Heap32Next loop too.
4320 *) The code that handled flushing of data in SSL/TLS originally used the
4321 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4322 the problem outlined in PR#1949. The fix suggested there however can
4323 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4324 of Apache). So instead simplify the code to flush unconditionally.
4325 This should be fine since flushing with no data to flush is a no op.
4328 *) Handle TLS versions 2.0 and later properly and correctly use the
4329 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4330 off ancient servers have a habit of sticking around for a while...
4333 *) Modify compression code so it frees up structures without using the
4334 ex_data callbacks. This works around a problem where some applications
4335 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4336 restarting) then use compression (e.g. SSL with compression) later.
4337 This results in significant per-connection memory leaks and
4338 has caused some security issues including CVE-2008-1678 and
4342 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4343 change when encrypting or decrypting.
4346 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4347 connect and renegotiate with servers which do not support RI.
4348 Until RI is more widely deployed this option is enabled by default.
4351 *) Add "missing" ssl ctrls to clear options and mode.
4354 *) If client attempts to renegotiate and doesn't support RI respond with
4355 a no_renegotiation alert as required by RFC5746. Some renegotiating
4356 TLS clients will continue a connection gracefully when they receive
4357 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4358 waiting for a server hello which it will never receive. Now we treat a
4359 received no_renegotiation alert as a fatal error. This is because
4360 applications requesting a renegotiation might well expect it to succeed
4361 and would have no code in place to handle the server denying it so the
4362 only safe thing to do is to terminate the connection.
4365 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4366 peer supports secure renegotiation and 0 otherwise. Print out peer
4367 renegotiation support in s_client/s_server.
4370 *) Replace the highly broken and deprecated SPKAC certification method with
4371 the updated NID creation version. This should correctly handle UTF8.
4374 *) Implement RFC5746. Re-enable renegotiation but require the extension
4375 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4376 turns out to be a bad idea. It has been replaced by
4377 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4378 SSL_CTX_set_options(). This is really not recommended unless you
4379 know what you are doing.
4380 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4382 *) Fixes to stateless session resumption handling. Use initial_ctx when
4383 issuing and attempting to decrypt tickets in case it has changed during
4384 servername handling. Use a non-zero length session ID when attempting
4385 stateless session resumption: this makes it possible to determine if
4386 a resumption has occurred immediately after receiving server hello
4387 (several places in OpenSSL subtly assume this) instead of later in
4391 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4392 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4393 fixes for a few places where the return code is not checked
4395 [Julia Lawall <julia@diku.dk>]
4397 *) Add --strict-warnings option to Configure script to include devteam
4398 warnings in other configurations.
4401 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4402 makes it possible to install openssl libraries in locations which
4403 have names other than "lib", for example "/usr/lib64" which some
4405 [Steve Henson, based on patch from Jeremy Utley]
4407 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4408 X690 8.9.12 and can produce some misleading textual output of OIDs.
4409 [Steve Henson, reported by Dan Kaminsky]
4411 *) Delete MD2 from algorithm tables. This follows the recommendation in
4412 several standards that it is not used in new applications due to
4413 several cryptographic weaknesses. For binary compatibility reasons
4414 the MD2 API is still compiled in by default.
4417 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4421 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4422 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4424 [Guenter <lists@gknw.net>]
4426 *) Fix the server certificate chain building code to use X509_verify_cert(),
4427 it used to have an ad-hoc builder which was unable to cope with anything
4428 other than a simple chain.
4429 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4431 *) Don't check self signed certificate signatures in X509_verify_cert()
4432 by default (a flag can override this): it just wastes time without
4433 adding any security. As a useful side effect self signed root CAs
4434 with non-FIPS digests are now usable in FIPS mode.
4437 *) In dtls1_process_out_of_seq_message() the check if the current message
4438 is already buffered was missing. For every new message was memory
4439 allocated, allowing an attacker to perform an denial of service attack
4440 with sending out of seq handshake messages until there is no memory
4441 left. Additionally every future messege was buffered, even if the
4442 sequence number made no sense and would be part of another handshake.
4443 So only messages with sequence numbers less than 10 in advance will be
4444 buffered. (CVE-2009-1378)
4445 [Robin Seggelmann, discovered by Daniel Mentz]
4447 *) Records are buffered if they arrive with a future epoch to be
4448 processed after finishing the corresponding handshake. There is
4449 currently no limitation to this buffer allowing an attacker to perform
4450 a DOS attack with sending records with future epochs until there is no
4451 memory left. This patch adds the pqueue_size() function to determine
4452 the size of a buffer and limits the record buffer to 100 entries.
4454 [Robin Seggelmann, discovered by Daniel Mentz]
4456 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4457 parent structure is freed. (CVE-2009-1379)
4460 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4461 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4464 [Ilya O. <vrghost@gmail.com>]
4466 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4468 *) Disable renegotiation completely - this fixes a severe security
4469 problem (CVE-2009-3555) at the cost of breaking all
4470 renegotiation. Renegotiation can be re-enabled by setting
4471 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4472 run-time. This is really not recommended unless you know what
4476 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4478 *) Don't set val to NULL when freeing up structures, it is freed up by
4479 underlying code. If sizeof(void *) > sizeof(long) this can result in
4480 zeroing past the valid field. (CVE-2009-0789)
4481 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4483 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4484 checked correctly. This would allow some invalid signed attributes to
4485 appear to verify correctly. (CVE-2009-0591)
4486 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4488 *) Reject UniversalString and BMPString types with invalid lengths. This
4489 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4490 a legal length. (CVE-2009-0590)
4493 *) Set S/MIME signing as the default purpose rather than setting it
4494 unconditionally. This allows applications to override it at the store
4498 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4499 to handle some structures.
4502 *) Improve efficiency of mem_gets: don't search whole buffer each time
4504 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4506 *) New -hex option for openssl rand.
4509 *) Print out UTF8String and NumericString when parsing ASN1.
4512 *) Support NumericString type for name components.
4515 *) Allow CC in the environment to override the automatically chosen
4516 compiler. Note that nothing is done to ensure flags work with the
4520 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4522 *) Properly check EVP_VerifyFinal() and similar return values
4524 [Ben Laurie, Bodo Moeller, Google Security Team]
4526 *) Enable TLS extensions by default.
4529 *) Allow the CHIL engine to be loaded, whether the application is
4530 multithreaded or not. (This does not release the developer from the
4531 obligation to set up the dynamic locking callbacks.)
4532 [Sander Temme <sander@temme.net>]
4534 *) Use correct exit code if there is an error in dgst command.
4535 [Steve Henson; problem pointed out by Roland Dirlewanger]
4537 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4538 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4541 *) Add experimental JPAKE support, including demo authentication in
4542 s_client and s_server.
4545 *) Set the comparison function in v3_addr_canonize().
4546 [Rob Austein <sra@hactrn.net>]
4548 *) Add support for XMPP STARTTLS in s_client.
4549 [Philip Paeps <philip@freebsd.org>]
4551 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4552 to ensure that even with this option, only ciphersuites in the
4553 server's preference list will be accepted. (Note that the option
4554 applies only when resuming a session, so the earlier behavior was
4555 just about the algorithm choice for symmetric cryptography.)
4558 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4560 *) Fix NULL pointer dereference if a DTLS server received
4561 ChangeCipherSpec as first record (CVE-2009-1386).
4564 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4565 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4568 *) The fix in 0.9.8c that supposedly got rid of unsafe
4569 double-checked locking was incomplete for RSA blinding,
4570 addressing just one layer of what turns out to have been
4571 doubly unsafe triple-checked locking.
4573 So now fix this for real by retiring the MONT_HELPER macro
4574 in crypto/rsa/rsa_eay.c.
4576 [Bodo Moeller; problem pointed out by Marius Schilder]
4578 *) Various precautionary measures:
4580 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4582 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4583 (NB: This would require knowledge of the secret session ticket key
4584 to exploit, in which case you'd be SOL either way.)
4586 - Change bn_nist.c so that it will properly handle input BIGNUMs
4587 outside the expected range.
4589 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4592 [Neel Mehta, Bodo Moeller]
4594 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4595 the load fails. Useful for distros.
4596 [Ben Laurie and the FreeBSD team]
4598 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4601 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4604 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4606 This work was sponsored by Logica.
4609 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4610 keystores. Support for SSL/TLS client authentication too.
4611 Not compiled unless enable-capieng specified to Configure.
4613 This work was sponsored by Logica.
4616 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4617 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4618 attribute creation routines such as certificate requests and PKCS#12
4622 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4624 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4625 handshake which could lead to a cilent crash as found using the
4626 Codenomicon TLS test suite (CVE-2008-1672)
4627 [Steve Henson, Mark Cox]
4629 *) Fix double free in TLS server name extensions which could lead to
4630 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4633 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4635 Clear the error queue to ensure that error entries left from
4636 older function calls do not interfere with the correct operation.
4637 [Lutz Jaenicke, Erik de Castro Lopo]
4639 *) Remove root CA certificates of commercial CAs:
4641 The OpenSSL project does not recommend any specific CA and does not
4642 have any policy with respect to including or excluding any CA.
4643 Therefore it does not make any sense to ship an arbitrary selection
4644 of root CA certificates with the OpenSSL software.
4647 *) RSA OAEP patches to fix two separate invalid memory reads.
4648 The first one involves inputs when 'lzero' is greater than
4649 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4650 before the beginning of from). The second one involves inputs where
4651 the 'db' section contains nothing but zeroes (there is a one-byte
4652 invalid read after the end of 'db').
4653 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4655 *) Partial backport from 0.9.9-dev:
4657 Introduce bn_mul_mont (dedicated Montgomery multiplication
4658 procedure) as a candidate for BIGNUM assembler implementation.
4659 While 0.9.9-dev uses assembler for various architectures, only
4660 x86_64 is available by default here in the 0.9.8 branch, and
4661 32-bit x86 is available through a compile-time setting.
4663 To try the 32-bit x86 assembler implementation, use Configure
4664 option "enable-montasm" (which exists only for this backport).
4666 As "enable-montasm" for 32-bit x86 disclaims code stability
4667 anyway, in this constellation we activate additional code
4668 backported from 0.9.9-dev for further performance improvements,
4669 namely BN_from_montgomery_word. (To enable this otherwise,
4670 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4672 [Andy Polyakov (backport partially by Bodo Moeller)]
4674 *) Add TLS session ticket callback. This allows an application to set
4675 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4676 values. This is useful for key rollover for example where several key
4677 sets may exist with different names.
4680 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4681 This was broken until now in 0.9.8 releases, such that the only way
4682 a registered ENGINE could be used (assuming it initialises
4683 successfully on the host) was to explicitly set it as the default
4684 for the relevant algorithms. This is in contradiction with 0.9.7
4685 behaviour and the documentation. With this fix, when an ENGINE is
4686 registered into a given algorithm's table of implementations, the
4687 'uptodate' flag is reset so that auto-discovery will be used next
4688 time a new context for that algorithm attempts to select an
4690 [Ian Lister (tweaked by Geoff Thorpe)]
4692 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4693 implementation in the following ways:
4695 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4698 Lack of BER streaming support means one pass streaming processing is
4699 only supported if data is detached: setting the streaming flag is
4700 ignored for embedded content.
4702 CMS support is disabled by default and must be explicitly enabled
4703 with the enable-cms configuration option.
4706 *) Update the GMP engine glue to do direct copies between BIGNUM and
4707 mpz_t when openssl and GMP use the same limb size. Otherwise the
4708 existing "conversion via a text string export" trick is still used.
4709 [Paul Sheer <paulsheer@gmail.com>]
4711 *) Zlib compression BIO. This is a filter BIO which compressed and
4712 uncompresses any data passed through it.
4715 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4716 RFC3394 compatible AES key wrapping.
4719 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4720 sets string data without copying. X509_ALGOR_set0() and
4721 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4722 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4723 from an X509_ATTRIBUTE structure optionally checking it occurs only
4724 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4728 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4729 to get the expected BN_FLG_CONSTTIME behavior.
4730 [Bodo Moeller (Google)]
4734 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4735 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4736 - added some more tests to do_tests.pl
4737 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4738 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4739 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4740 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4741 - various changes to netware.pl to enable gcc-cross builds on Win32
4743 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4744 - various changes to fix missing prototype warnings
4745 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4746 - added AES, WHIRLPOOL and CPUID assembler code to build files
4747 - added missing AES assembler make rules to mk1mf.pl
4748 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4749 [Guenter Knauf <eflash@gmx.net>]
4751 *) Implement certificate status request TLS extension defined in RFC3546.
4752 A client can set the appropriate parameters and receive the encoded
4753 OCSP response via a callback. A server can query the supplied parameters
4754 and set the encoded OCSP response in the callback. Add simplified examples
4755 to s_client and s_server.
4758 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4760 *) Fix various bugs:
4761 + Binary incompatibility of ssl_ctx_st structure
4762 + DTLS interoperation with non-compliant servers
4763 + Don't call get_session_cb() without proposed session
4764 + Fix ia64 assembler code
4765 [Andy Polyakov, Steve Henson]
4767 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4769 *) DTLS Handshake overhaul. There were longstanding issues with
4770 OpenSSL DTLS implementation, which were making it impossible for
4771 RFC 4347 compliant client to communicate with OpenSSL server.
4772 Unfortunately just fixing these incompatibilities would "cut off"
4773 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4774 server keeps tolerating non RFC compliant syntax. The opposite is
4775 not true, 0.9.8f client can not communicate with earlier server.
4776 This update even addresses CVE-2007-4995.
4779 *) Changes to avoid need for function casts in OpenSSL: some compilers
4780 (gcc 4.2 and later) reject their use.
4781 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4784 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4785 RFC4507bis. The encrypted ticket format is an encrypted encoded
4786 SSL_SESSION structure, that way new session features are automatically
4789 If a client application caches session in an SSL_SESSION structure
4790 support is transparent because tickets are now stored in the encoded
4793 The SSL_CTX structure automatically generates keys for ticket
4794 protection in servers so again support should be possible
4795 with no application modification.
4797 If a client or server wishes to disable RFC4507 support then the option
4798 SSL_OP_NO_TICKET can be set.
4800 Add a TLS extension debugging callback to allow the contents of any client
4801 or server extensions to be examined.
4803 This work was sponsored by Google.
4806 *) Add initial support for TLS extensions, specifically for the server_name
4807 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4808 have new members for a host name. The SSL data structure has an
4809 additional member SSL_CTX *initial_ctx so that new sessions can be
4810 stored in that context to allow for session resumption, even after the
4811 SSL has been switched to a new SSL_CTX in reaction to a client's
4812 server_name extension.
4814 New functions (subject to change):
4816 SSL_get_servername()
4817 SSL_get_servername_type()
4820 New CTRL codes and macros (subject to change):
4822 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4823 - SSL_CTX_set_tlsext_servername_callback()
4824 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4825 - SSL_CTX_set_tlsext_servername_arg()
4826 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4828 openssl s_client has a new '-servername ...' option.
4830 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4831 '-key2 ...', '-servername_fatal' (subject to change). This allows
4832 testing the HostName extension for a specific single host name ('-cert'
4833 and '-key' remain fallbacks for handshakes without HostName
4834 negotiation). If the unrecognized_name alert has to be sent, this by
4835 default is a warning; it becomes fatal with the '-servername_fatal'
4838 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4840 *) Add AES and SSE2 assembly language support to VC++ build.
4843 *) Mitigate attack on final subtraction in Montgomery reduction.
4846 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4847 (which previously caused an internal error).
4850 *) Squeeze another 10% out of IGE mode when in != out.
4853 *) AES IGE mode speedup.
4854 [Dean Gaudet (Google)]
4856 *) Add the Korean symmetric 128-bit cipher SEED (see
4857 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4858 add SEED ciphersuites from RFC 4162:
4860 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4861 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4862 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4863 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4865 To minimize changes between patchlevels in the OpenSSL 0.9.8
4866 series, SEED remains excluded from compilation unless OpenSSL
4867 is configured with 'enable-seed'.
4868 [KISA, Bodo Moeller]
4870 *) Mitigate branch prediction attacks, which can be practical if a
4871 single processor is shared, allowing a spy process to extract
4872 information. For detailed background information, see
4873 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4874 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4875 and Necessary Software Countermeasures"). The core of the change
4876 are new versions BN_div_no_branch() and
4877 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4878 respectively, which are slower, but avoid the security-relevant
4879 conditional branches. These are automatically called by BN_div()
4880 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4881 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4882 remove a conditional branch.
4884 BN_FLG_CONSTTIME is the new name for the previous
4885 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4886 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4887 in the exponent causes BN_mod_exp_mont() to use the alternative
4888 implementation in BN_mod_exp_mont_consttime().) The old name
4889 remains as a deprecated alias.
4891 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4892 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4893 constant-time implementations for more than just exponentiation.
4894 Here too the old name is kept as a deprecated alias.
4896 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4897 the BN_BLINDING structure gets an independent copy of the
4898 modulus. This means that the previous "BIGNUM *m" argument to
4899 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4900 essentially becomes "const BIGNUM *m", although we can't actually
4901 change this in the header file before 0.9.9. It allows
4902 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4903 enable BN_FLG_CONSTTIME.
4905 [Matthew D Wood (Intel Corp)]
4907 *) In the SSL/TLS server implementation, be strict about session ID
4908 context matching (which matters if an application uses a single
4909 external cache for different purposes). Previously,
4910 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4911 set. This did ensure strict client verification, but meant that,
4912 with applications using a single external cache for quite
4913 different requirements, clients could circumvent ciphersuite
4914 restrictions for a given session ID context by starting a session
4915 in a different context.
4918 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4919 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4920 authentication-only ciphersuites.
4923 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4924 not complete and could lead to a possible single byte overflow
4925 (CVE-2007-5135) [Ben Laurie]
4927 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4929 *) Since AES128 and AES256 (and similarly Camellia128 and
4930 Camellia256) share a single mask bit in the logic of
4931 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4932 kludge to work properly if AES128 is available and AES256 isn't
4933 (or if Camellia128 is available and Camellia256 isn't).
4936 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4937 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4938 When a point or a seed is encoded in a BIT STRING, we need to
4939 prevent the removal of trailing zero bits to get the proper DER
4940 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4941 of a NamedBitList, for which trailing 0 bits need to be removed.)
4944 *) Have SSL/TLS server implementation tolerate "mismatched" record
4945 protocol version while receiving ClientHello even if the
4946 ClientHello is fragmented. (The server can't insist on the
4947 particular protocol version it has chosen before the ServerHello
4948 message has informed the client about his choice.)
4951 *) Add RFC 3779 support.
4952 [Rob Austein for ARIN, Ben Laurie]
4954 *) Load error codes if they are not already present instead of using a
4955 static variable. This allows them to be cleanly unloaded and reloaded.
4956 Improve header file function name parsing.
4959 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4960 or CAPABILITY handshake as required by RFCs.
4963 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4965 *) Introduce limits to prevent malicious keys being able to
4966 cause a denial of service. (CVE-2006-2940)
4967 [Steve Henson, Bodo Moeller]
4969 *) Fix ASN.1 parsing of certain invalid structures that can result
4970 in a denial of service. (CVE-2006-2937) [Steve Henson]
4972 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4973 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4975 *) Fix SSL client code which could crash if connecting to a
4976 malicious SSLv2 server. (CVE-2006-4343)
4977 [Tavis Ormandy and Will Drewry, Google Security Team]
4979 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4980 match only those. Before that, "AES256-SHA" would be interpreted
4981 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4982 the same strength classification in 0.9.7h) as we currently only
4983 have a single AES bit in the ciphersuite description bitmap.
4984 That change, however, also applied to ciphersuite strings such as
4985 "RC4-MD5" that intentionally matched multiple ciphersuites --
4986 namely, SSL 2.0 ciphersuites in addition to the more common ones
4987 from SSL 3.0/TLS 1.0.
4989 So we change the selection algorithm again: Naming an explicit
4990 ciphersuite selects this one ciphersuite, and any other similar
4991 ciphersuite (same bitmap) from *other* protocol versions.
4992 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4993 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4995 Since SSL 2.0 does not have any ciphersuites for which the
4996 128/256 bit distinction would be relevant, this works for now.
4997 The proper fix will be to use different bits for AES128 and
4998 AES256, which would have avoided the problems from the beginning;
4999 however, bits are scarce, so we can only do this in a new release
5000 (not just a patchlevel) when we can change the SSL_CIPHER
5001 definition to split the single 'unsigned long mask' bitmap into
5002 multiple values to extend the available space.
5006 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5008 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5009 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5011 *) Add AES IGE and biIGE modes.
5014 *) Change the Unix randomness entropy gathering to use poll() when
5015 possible instead of select(), since the latter has some
5016 undesirable limitations.
5017 [Darryl Miles via Richard Levitte and Bodo Moeller]
5019 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5020 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5021 cannot be implicitly activated as part of, e.g., the "AES" alias.
5022 However, please upgrade to OpenSSL 0.9.9[-dev] for
5023 non-experimental use of the ECC ciphersuites to get TLS extension
5024 support, which is required for curve and point format negotiation
5025 to avoid potential handshake problems.
5028 *) Disable rogue ciphersuites:
5030 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5031 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5032 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5034 The latter two were purportedly from
5035 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5038 Also deactivate the remaining ciphersuites from
5039 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5040 unofficial, and the ID has long expired.
5043 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5044 dual-core machines) and other potential thread-safety issues.
5047 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5048 versions), which is now available for royalty-free use
5049 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5050 Also, add Camellia TLS ciphersuites from RFC 4132.
5052 To minimize changes between patchlevels in the OpenSSL 0.9.8
5053 series, Camellia remains excluded from compilation unless OpenSSL
5054 is configured with 'enable-camellia'.
5057 *) Disable the padding bug check when compression is in use. The padding
5058 bug check assumes the first packet is of even length, this is not
5059 necessarily true if compresssion is enabled and can result in false
5060 positives causing handshake failure. The actual bug test is ancient
5061 code so it is hoped that implementations will either have fixed it by
5062 now or any which still have the bug do not support compression.
5065 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5067 *) When applying a cipher rule check to see if string match is an explicit
5068 cipher suite and only match that one cipher suite if it is.
5071 *) Link in manifests for VC++ if needed.
5072 [Austin Ziegler <halostatue@gmail.com>]
5074 *) Update support for ECC-based TLS ciphersuites according to
5075 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5076 TLS extensions, which are supported starting with the 0.9.9
5077 branch, not in the OpenSSL 0.9.8 branch).
5080 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5081 opaque EVP_CIPHER_CTX handling.
5084 *) Fixes and enhancements to zlib compression code. We now only use
5085 "zlib1.dll" and use the default __cdecl calling convention on Win32
5086 to conform with the standards mentioned here:
5087 http://www.zlib.net/DLL_FAQ.txt
5088 Static zlib linking now works on Windows and the new --with-zlib-include
5089 --with-zlib-lib options to Configure can be used to supply the location
5090 of the headers and library. Gracefully handle case where zlib library
5094 *) Several fixes and enhancements to the OID generation code. The old code
5095 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5096 handle numbers larger than ULONG_MAX, truncated printing and had a
5097 non standard OBJ_obj2txt() behaviour.
5100 *) Add support for building of engines under engine/ as shared libraries
5101 under VC++ build system.
5104 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5105 Hopefully, we will not see any false combination of paths any more.
5108 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5110 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5111 (part of SSL_OP_ALL). This option used to disable the
5112 countermeasure against man-in-the-middle protocol-version
5113 rollback in the SSL 2.0 server implementation, which is a bad
5114 idea. (CVE-2005-2969)
5116 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5117 for Information Security, National Institute of Advanced Industrial
5118 Science and Technology [AIST], Japan)]
5120 *) Add two function to clear and return the verify parameter flags.
5123 *) Keep cipherlists sorted in the source instead of sorting them at
5124 runtime, thus removing the need for a lock.
5127 *) Avoid some small subgroup attacks in Diffie-Hellman.
5128 [Nick Mathewson and Ben Laurie]
5130 *) Add functions for well-known primes.
5133 *) Extended Windows CE support.
5134 [Satoshi Nakamura and Andy Polyakov]
5136 *) Initialize SSL_METHOD structures at compile time instead of during
5137 runtime, thus removing the need for a lock.
5140 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5141 attempting to decrypt each encrypted key in turn. Add support to
5145 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5147 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5150 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5153 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5154 key into the same file any more.
5157 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5160 *) Add -utf8 command line and config file option to 'ca'.
5161 [Stefan <stf@udoma.org]
5163 *) Removed the macro des_crypt(), as it seems to conflict with some
5164 libraries. Use DES_crypt().
5167 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5168 involves renaming the source and generated shared-libs for
5169 both. The engines will accept the corrected or legacy ids
5170 ('ncipher' and '4758_cca' respectively) when binding. NB,
5171 this only applies when building 'shared'.
5172 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5174 *) Add attribute functions to EVP_PKEY structure. Modify
5175 PKCS12_create() to recognize a CSP name attribute and
5176 use it. Make -CSP option work again in pkcs12 utility.
5179 *) Add new functionality to the bn blinding code:
5180 - automatic re-creation of the BN_BLINDING parameters after
5181 a fixed number of uses (currently 32)
5182 - add new function for parameter creation
5183 - introduce flags to control the update behaviour of the
5184 BN_BLINDING parameters
5185 - hide BN_BLINDING structure
5186 Add a second BN_BLINDING slot to the RSA structure to improve
5187 performance when a single RSA object is shared among several
5191 *) Add support for DTLS.
5192 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5194 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5195 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5198 *) Remove buggy and incomplete DH cert support from
5199 ssl/ssl_rsa.c and ssl/s3_both.c
5202 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5203 the apps/openssl applications.
5206 *) Compile clean with "-Wall -Wmissing-prototypes
5207 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5208 DEBUG_SAFESTACK must also be set.
5211 *) Change ./Configure so that certain algorithms can be disabled by default.
5212 The new counterpiece to "no-xxx" is "enable-xxx".
5214 The patented RC5 and MDC2 algorithms will now be disabled unless
5215 "enable-rc5" and "enable-mdc2", respectively, are specified.
5217 (IDEA remains enabled despite being patented. This is because IDEA
5218 is frequently required for interoperability, and there is no license
5219 fee for non-commercial use. As before, "no-idea" can be used to
5220 avoid this algorithm.)
5224 *) Add processing of proxy certificates (see RFC 3820). This work was
5225 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5226 EGEE (Enabling Grids for E-science in Europe).
5229 *) RC4 performance overhaul on modern architectures/implementations, such
5230 as Intel P4, IA-64 and AMD64.
5233 *) New utility extract-section.pl. This can be used specify an alternative
5234 section number in a pod file instead of having to treat each file as
5235 a separate case in Makefile. This can be done by adding two lines to the
5238 =for comment openssl_section:XXX
5240 The blank line is mandatory.
5244 *) New arguments -certform, -keyform and -pass for s_client and s_server
5245 to allow alternative format key and certificate files and passphrase
5249 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5250 update associated structures and add various utility functions.
5252 Add new policy related verify parameters, include policy checking in
5253 standard verify code. Enhance 'smime' application with extra parameters
5254 to support policy checking and print out.
5257 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5258 Nehemiah processors. These extensions support AES encryption in hardware
5259 as well as RNG (though RNG support is currently disabled).
5260 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5262 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5265 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5266 [Andy Polyakov and a number of other people]
5268 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5269 implementation contributed by IBM.
5270 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5272 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5273 exponent rather than 'unsigned long'. There is a corresponding change to
5274 the new 'rsa_keygen' element of the RSA_METHOD structure.
5275 [Jelte Jansen, Geoff Thorpe]
5277 *) Functionality for creating the initial serial number file is now
5278 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5280 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5281 number file to 1, which is bound to cause problems. To avoid
5282 the problems while respecting compatibility between different 0.9.7
5283 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5284 CA.pl for serial number initialization. With the new release 0.9.8,
5285 we can fix the problem directly in the 'ca' utility.)
5288 *) Reduced header interdepencies by declaring more opaque objects in
5289 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5290 give fewer recursive includes, which could break lazy source code - so
5291 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5292 developers should define this symbol when building and using openssl to
5293 ensure they track the recommended behaviour, interfaces, [etc], but
5294 backwards-compatible behaviour prevails when this isn't defined.
5297 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5300 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5301 This will generate a random key of the appropriate length based on the
5302 cipher context. The EVP_CIPHER can provide its own random key generation
5303 routine to support keys of a specific form. This is used in the des and
5304 3des routines to generate a key of the correct parity. Update S/MIME
5305 code to use new functions and hence generate correct parity DES keys.
5306 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5307 valid (weak or incorrect parity).
5310 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5311 as looking them up. This is useful when the verified structure may contain
5312 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5313 present unless the new PKCS7_NO_CRL flag is asserted.
5316 *) Extend ASN1 oid configuration module. It now additionally accepts the
5319 shortName = some long name, 1.2.3.4
5322 *) Reimplemented the BN_CTX implementation. There is now no more static
5323 limitation on the number of variables it can handle nor the depth of the
5324 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5325 information can now expand as required, and rather than having a single
5326 static array of bignums, BN_CTX now uses a linked-list of such arrays
5327 allowing it to expand on demand whilst maintaining the usefulness of
5328 BN_CTX's "bundling".
5331 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5332 to allow all RSA operations to function using a single BN_CTX.
5335 *) Preliminary support for certificate policy evaluation and checking. This
5336 is initially intended to pass the tests outlined in "Conformance Testing
5337 of Relying Party Client Certificate Path Processing Logic" v1.07.
5340 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5341 remained unused and not that useful. A variety of other little bignum
5342 tweaks and fixes have also been made continuing on from the audit (see
5346 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5347 associated ASN1, EVP and SSL functions and old ASN1 macros.
5350 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5351 and this should never fail. So the return value from the use of
5352 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5353 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5356 *) BN_CTX_get() should return zero-valued bignums, providing the same
5357 initialised value as BN_new().
5358 [Geoff Thorpe, suggested by Ulf Möller]
5360 *) Support for inhibitAnyPolicy certificate extension.
5363 *) An audit of the BIGNUM code is underway, for which debugging code is
5364 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5365 is considered valid when processing BIGNUMs, and causes execution to
5366 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5367 further steps are taken to deliberately pollute unused data in BIGNUM
5368 structures to try and expose faulty code further on. For now, openssl will
5369 (in its default mode of operation) continue to tolerate the inconsistent
5370 forms that it has tolerated in the past, but authors and packagers should
5371 consider trying openssl and their own applications when compiled with
5372 these debugging symbols defined. It will help highlight potential bugs in
5373 their own code, and will improve the test coverage for OpenSSL itself. At
5374 some point, these tighter rules will become openssl's default to improve
5375 maintainability, though the assert()s and other overheads will remain only
5376 in debugging configurations. See bn.h for more details.
5377 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5379 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5380 that can only be obtained through BN_CTX_new() (which implicitly
5381 initialises it). The presence of this function only made it possible
5382 to overwrite an existing structure (and cause memory leaks).
5385 *) Because of the callback-based approach for implementing LHASH as a
5386 template type, lh_insert() adds opaque objects to hash-tables and
5387 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5388 to clean up those corresponding objects before destroying the hash table
5389 (and losing the object pointers). So some over-zealous constifications in
5390 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5391 objects as "const" and the lh_doall[_arg] callback wrappers are not
5392 prototyped to have "const" restrictions on the object pointers they are
5393 given (and so aren't required to cast them away any more).
5396 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5397 (speed) prefers to use its own implementation. The two implementations
5398 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5399 its object type properly exposed (MS_TM) instead of casting to/from "char
5400 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5401 aren't necessarily the greatest nomenclatures - but this is what was used
5402 internally to the implementation so I've used that for now.
5405 *) Ensure that deprecated functions do not get compiled when
5406 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5407 the self-tests were still using deprecated key-generation functions so
5408 these have been updated also.
5411 *) Reorganise PKCS#7 code to separate the digest location functionality
5412 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5413 New function PKCS7_set_digest() to set the digest type for PKCS#7
5414 digestedData type. Add additional code to correctly generate the
5415 digestedData type and add support for this type in PKCS7 initialization
5419 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5420 structure of type "other".
5423 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5424 sure the loop does correctly stop and breaking ("division by zero")
5425 modulus operations are not performed. The (pre-generated) prime
5426 table crypto/bn/bn_prime.h was already correct, but it could not be
5427 re-generated on some platforms because of the "division by zero"
5428 situation in the script.
5429 [Ralf S. Engelschall]
5431 *) Update support for ECC-based TLS ciphersuites according to
5432 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5433 SHA-1 now is only used for "small" curves (where the
5434 representation of a field element takes up to 24 bytes); for
5435 larger curves, the field element resulting from ECDH is directly
5436 used as premaster secret.
5437 [Douglas Stebila (Sun Microsystems Laboratories)]
5439 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5440 curve secp160r1 to the tests.
5441 [Douglas Stebila (Sun Microsystems Laboratories)]
5443 *) Add the possibility to load symbols globally with DSO.
5444 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5446 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5447 control of the error stack.
5450 *) Add support for STORE in ENGINE.
5453 *) Add the STORE type. The intention is to provide a common interface
5454 to certificate and key stores, be they simple file-based stores, or
5455 HSM-type store, or LDAP stores, or...
5456 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5459 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5460 pass a list of arguments to any function as well as provide a way
5461 for a function to pass data back to the caller.
5464 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5465 works like BUF_strdup() but can be used to duplicate a portion of
5466 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5470 *) Add the function sk_find_ex() which works like sk_find(), but will
5471 return an index to an element even if an exact match couldn't be
5472 found. The index is guaranteed to point at the element where the
5473 searched-for key would be inserted to preserve sorting order.
5476 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5477 takes an extra flags argument for optional functionality. Currently,
5478 the following flags are defined:
5480 OBJ_BSEARCH_VALUE_ON_NOMATCH
5481 This one gets OBJ_bsearch_ex() to return a pointer to the first
5482 element where the comparing function returns a negative or zero
5485 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5486 This one gets OBJ_bsearch_ex() to return a pointer to the first
5487 element where the comparing function returns zero. This is useful
5488 if there are more than one element where the comparing function
5492 *) Make it possible to create self-signed certificates with 'openssl ca'
5493 in such a way that the self-signed certificate becomes part of the
5494 CA database and uses the same mechanisms for serial number generation
5495 as all other certificate signing. The new flag '-selfsign' enables
5496 this functionality. Adapt CA.sh and CA.pl.in.
5499 *) Add functionality to check the public key of a certificate request
5500 against a given private. This is useful to check that a certificate
5501 request can be signed by that key (self-signing).
5504 *) Make it possible to have multiple active certificates with the same
5505 subject in the CA index file. This is done only if the keyword
5506 'unique_subject' is set to 'no' in the main CA section (default
5507 if 'CA_default') of the configuration file. The value is saved
5508 with the database itself in a separate index attribute file,
5509 named like the index file with '.attr' appended to the name.
5512 *) Generate muti valued AVAs using '+' notation in config files for
5516 *) Support for nameConstraints certificate extension.
5519 *) Support for policyConstraints certificate extension.
5522 *) Support for policyMappings certificate extension.
5525 *) Make sure the default DSA_METHOD implementation only uses its
5526 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5527 and change its own handlers to be NULL so as to remove unnecessary
5528 indirection. This lets alternative implementations fallback to the
5529 default implementation more easily.
5532 *) Support for directoryName in GeneralName related extensions
5536 *) Make it possible to link applications using Makefile.shared.
5537 Make that possible even when linking against static libraries!
5540 *) Support for single pass processing for S/MIME signing. This now
5541 means that S/MIME signing can be done from a pipe, in addition
5542 cleartext signing (multipart/signed type) is effectively streaming
5543 and the signed data does not need to be all held in memory.
5545 This is done with a new flag PKCS7_STREAM. When this flag is set
5546 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5547 is done after the data is output (and digests calculated) in
5548 SMIME_write_PKCS7().
5551 *) Add full support for -rpath/-R, both in shared libraries and
5552 applications, at least on the platforms where it's known how
5556 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5557 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5558 will now compute a table of multiples of the generator that
5559 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5560 faster (notably in the case of a single point multiplication,
5561 scalar * generator).
5562 [Nils Larsch, Bodo Moeller]
5564 *) IPv6 support for certificate extensions. The various extensions
5565 which use the IP:a.b.c.d can now take IPv6 addresses using the
5566 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5570 *) Added an ENGINE that implements RSA by performing private key
5571 exponentiations with the GMP library. The conversions to and from
5572 GMP's mpz_t format aren't optimised nor are any montgomery forms
5573 cached, and on x86 it appears OpenSSL's own performance has caught up.
5574 However there are likely to be other architectures where GMP could
5575 provide a boost. This ENGINE is not built in by default, but it can be
5576 specified at Configure time and should be accompanied by the necessary
5577 linker additions, eg;
5578 ./config -DOPENSSL_USE_GMP -lgmp
5581 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5582 testing availability of engines with "-t" - the old behaviour is
5583 produced by increasing the feature's verbosity with "-tt".
5586 *) ECDSA routines: under certain error conditions uninitialized BN objects
5587 could be freed. Solution: make sure initialization is performed early
5588 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5592 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5593 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5594 software implementations. For DSA and DH, parameter generation can
5595 also be overridden by providing the appropriate method callbacks.
5598 *) Change the "progress" mechanism used in key-generation and
5599 primality testing to functions that take a new BN_GENCB pointer in
5600 place of callback/argument pairs. The new API functions have "_ex"
5601 postfixes and the older functions are reimplemented as wrappers for
5602 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5603 declarations of the old functions to help (graceful) attempts to
5604 migrate to the new functions. Also, the new key-generation API
5605 functions operate on a caller-supplied key-structure and return
5606 success/failure rather than returning a key or NULL - this is to
5607 help make "keygen" another member function of RSA_METHOD etc.
5609 Example for using the new callback interface:
5611 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5615 BN_GENCB_set(&my_cb, my_callback, my_arg);
5617 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5618 /* For the meaning of a, b in calls to my_callback(), see the
5619 * documentation of the function that calls the callback.
5620 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5621 * my_callback should return 1 if it wants BN_is_prime_ex()
5622 * to continue, or 0 to stop.
5627 *) Change the ZLIB compression method to be stateful, and make it
5628 available to TLS with the number defined in
5629 draft-ietf-tls-compression-04.txt.
5632 *) Add the ASN.1 structures and functions for CertificatePair, which
5633 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5635 CertificatePair ::= SEQUENCE {
5636 forward [0] Certificate OPTIONAL,
5637 reverse [1] Certificate OPTIONAL,
5638 -- at least one of the pair shall be present -- }
5640 Also implement the PEM functions to read and write certificate
5641 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5643 This needed to be defined, mostly for the sake of the LDAP
5644 attribute crossCertificatePair, but may prove useful elsewhere as
5648 *) Make it possible to inhibit symlinking of shared libraries in
5649 Makefile.shared, for Cygwin's sake.
5652 *) Extend the BIGNUM API by creating a function
5653 void BN_set_negative(BIGNUM *a, int neg);
5654 and a macro that behave like
5655 int BN_is_negative(const BIGNUM *a);
5657 to avoid the need to access 'a->neg' directly in applications.
5660 *) Implement fast modular reduction for pseudo-Mersenne primes
5661 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5662 EC_GROUP_new_curve_GFp() will now automatically use this
5664 [Nils Larsch <nla@trustcenter.de>]
5666 *) Add new lock type (CRYPTO_LOCK_BN).
5669 *) Change the ENGINE framework to automatically load engines
5670 dynamically from specific directories unless they could be
5671 found to already be built in or loaded. Move all the
5672 current engines except for the cryptodev one to a new
5674 The engines in engines/ are built as shared libraries if
5675 the "shared" options was given to ./Configure or ./config.
5676 Otherwise, they are inserted in libcrypto.a.
5677 /usr/local/ssl/engines is the default directory for dynamic
5678 engines, but that can be overridden at configure time through
5679 the usual use of --prefix and/or --openssldir, and at run
5680 time with the environment variable OPENSSL_ENGINES.
5681 [Geoff Thorpe and Richard Levitte]
5683 *) Add Makefile.shared, a helper makefile to build shared
5684 libraries. Adapt Makefile.org.
5687 *) Add version info to Win32 DLLs.
5688 [Peter 'Luna' Runestig" <peter@runestig.com>]
5690 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5691 can be added using this API to created arbitrary PKCS#12
5692 files while avoiding the low level API.
5694 New options to PKCS12_create(), key or cert can be NULL and
5695 will then be omitted from the output file. The encryption
5696 algorithm NIDs can be set to -1 for no encryption, the mac
5697 iteration count can be set to 0 to omit the mac.
5699 Enhance pkcs12 utility by making the -nokeys and -nocerts
5700 options work when creating a PKCS#12 file. New option -nomac
5701 to omit the mac, NONE can be set for an encryption algorithm.
5702 New code is modified to use the enhanced PKCS12_create()
5703 instead of the low level API.
5706 *) Extend ASN1 encoder to support indefinite length constructed
5707 encoding. This can output sequences tags and octet strings in
5708 this form. Modify pk7_asn1.c to support indefinite length
5709 encoding. This is experimental and needs additional code to
5710 be useful, such as an ASN1 bio and some enhanced streaming
5713 Extend template encode functionality so that tagging is passed
5714 down to the template encoder.
5717 *) Let 'openssl req' fail if an argument to '-newkey' is not
5718 recognized instead of using RSA as a default.
5721 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5722 As these are not official, they are not included in "ALL";
5723 the "ECCdraft" ciphersuite group alias can be used to select them.
5724 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5726 *) Add ECDH engine support.
5727 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5729 *) Add ECDH in new directory crypto/ecdh/.
5730 [Douglas Stebila (Sun Microsystems Laboratories)]
5732 *) Let BN_rand_range() abort with an error after 100 iterations
5733 without success (which indicates a broken PRNG).
5736 *) Change BN_mod_sqrt() so that it verifies that the input value
5737 is really the square of the return value. (Previously,
5738 BN_mod_sqrt would show GIGO behaviour.)
5741 *) Add named elliptic curves over binary fields from X9.62, SECG,
5742 and WAP/WTLS; add OIDs that were still missing.
5744 [Sheueling Chang Shantz and Douglas Stebila
5745 (Sun Microsystems Laboratories)]
5747 *) Extend the EC library for elliptic curves over binary fields
5748 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5751 EC_GF2m_simple_method
5755 EC_GROUP_new_curve_GF2m
5756 EC_GROUP_set_curve_GF2m
5757 EC_GROUP_get_curve_GF2m
5758 EC_POINT_set_affine_coordinates_GF2m
5759 EC_POINT_get_affine_coordinates_GF2m
5760 EC_POINT_set_compressed_coordinates_GF2m
5762 Point compression for binary fields is disabled by default for
5763 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5766 As binary polynomials are represented as BIGNUMs, various members
5767 of the EC_GROUP and EC_POINT data structures can be shared
5768 between the implementations for prime fields and binary fields;
5769 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5770 are essentially identical to their ..._GFp counterparts.
5771 (For simplicity, the '..._GFp' prefix has been dropped from
5772 various internal method names.)
5774 An internal 'field_div' method (similar to 'field_mul' and
5775 'field_sqr') has been added; this is used only for binary fields.
5777 [Sheueling Chang Shantz and Douglas Stebila
5778 (Sun Microsystems Laboratories)]
5780 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5781 through methods ('mul', 'precompute_mult').
5783 The generic implementations (now internally called 'ec_wNAF_mul'
5784 and 'ec_wNAF_precomputed_mult') remain the default if these
5785 methods are undefined.
5787 [Sheueling Chang Shantz and Douglas Stebila
5788 (Sun Microsystems Laboratories)]
5790 *) New function EC_GROUP_get_degree, which is defined through
5791 EC_METHOD. For curves over prime fields, this returns the bit
5792 length of the modulus.
5794 [Sheueling Chang Shantz and Douglas Stebila
5795 (Sun Microsystems Laboratories)]
5797 *) New functions EC_GROUP_dup, EC_POINT_dup.
5798 (These simply call ..._new and ..._copy).
5800 [Sheueling Chang Shantz and Douglas Stebila
5801 (Sun Microsystems Laboratories)]
5803 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5804 Polynomials are represented as BIGNUMs (where the sign bit is not
5805 used) in the following functions [macros]:
5808 BN_GF2m_sub [= BN_GF2m_add]
5809 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5810 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5811 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5813 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5814 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5815 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5816 BN_GF2m_cmp [= BN_ucmp]
5818 (Note that only the 'mod' functions are actually for fields GF(2^m).
5819 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5821 For some functions, an the irreducible polynomial defining a
5822 field can be given as an 'unsigned int[]' with strictly
5823 decreasing elements giving the indices of those bits that are set;
5824 i.e., p[] represents the polynomial
5825 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5827 p[0] > p[1] > ... > p[k] = 0.
5828 This applies to the following functions:
5833 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5834 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5836 BN_GF2m_mod_sqrt_arr
5837 BN_GF2m_mod_solve_quad_arr
5841 Conversion can be performed by the following functions:
5846 bntest.c has additional tests for binary polynomial arithmetic.
5848 Two implementations for BN_GF2m_mod_div() are available.
5849 The default algorithm simply uses BN_GF2m_mod_inv() and
5850 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5851 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5852 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5854 [Sheueling Chang Shantz and Douglas Stebila
5855 (Sun Microsystems Laboratories)]
5857 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5858 functionality is disabled at compile-time.
5859 [Douglas Stebila <douglas.stebila@sun.com>]
5861 *) Change default behaviour of 'openssl asn1parse' so that more
5862 information is visible when viewing, e.g., a certificate:
5864 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5865 mode the content of non-printable OCTET STRINGs is output in a
5866 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5867 avoid the appearance of a printable string.
5868 [Nils Larsch <nla@trustcenter.de>]
5870 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5872 EC_GROUP_set_asn1_flag()
5873 EC_GROUP_get_asn1_flag()
5874 EC_GROUP_set_point_conversion_form()
5875 EC_GROUP_get_point_conversion_form()
5876 These control ASN1 encoding details:
5877 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5878 has been set to OPENSSL_EC_NAMED_CURVE.
5879 - Points are encoded in uncompressed form by default; options for
5880 asn1_for are as for point2oct, namely
5881 POINT_CONVERSION_COMPRESSED
5882 POINT_CONVERSION_UNCOMPRESSED
5883 POINT_CONVERSION_HYBRID
5885 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5888 EC_GROUP_get0_seed()
5889 EC_GROUP_get_seed_len()
5890 This is used only for ASN1 purposes (so far).
5891 [Nils Larsch <nla@trustcenter.de>]
5893 *) Add 'field_type' member to EC_METHOD, which holds the NID
5894 of the appropriate field type OID. The new function
5895 EC_METHOD_get_field_type() returns this value.
5896 [Nils Larsch <nla@trustcenter.de>]
5901 EC_POINT_point2hex()
5902 EC_POINT_hex2point()
5903 providing useful interfaces to EC_POINT_point2oct() and
5904 EC_POINT_oct2point().
5905 [Nils Larsch <nla@trustcenter.de>]
5907 *) Change internals of the EC library so that the functions
5908 EC_GROUP_set_generator()
5909 EC_GROUP_get_generator()
5910 EC_GROUP_get_order()
5911 EC_GROUP_get_cofactor()
5912 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5913 to methods, which would lead to unnecessary code duplication when
5914 adding different types of curves.
5915 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5917 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5918 arithmetic, and such that modified wNAFs are generated
5919 (which avoid length expansion in many cases).
5922 *) Add a function EC_GROUP_check_discriminant() (defined via
5923 EC_METHOD) that verifies that the curve discriminant is non-zero.
5925 Add a function EC_GROUP_check() that makes some sanity tests
5926 on a EC_GROUP, its generator and order. This includes
5927 EC_GROUP_check_discriminant().
5928 [Nils Larsch <nla@trustcenter.de>]
5930 *) Add ECDSA in new directory crypto/ecdsa/.
5932 Add applications 'openssl ecparam' and 'openssl ecdsa'
5933 (these are based on 'openssl dsaparam' and 'openssl dsa').
5935 ECDSA support is also included in various other files across the
5936 library. Most notably,
5937 - 'openssl req' now has a '-newkey ecdsa:file' option;
5938 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5939 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5940 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5941 them suitable for ECDSA where domain parameters must be
5942 extracted before the specific public key;
5943 - ECDSA engine support has been added.
5944 [Nils Larsch <nla@trustcenter.de>]
5946 *) Include some named elliptic curves, and add OIDs from X9.62,
5947 SECG, and WAP/WTLS. Each curve can be obtained from the new
5949 EC_GROUP_new_by_curve_name(),
5950 and the list of available named curves can be obtained with
5951 EC_get_builtin_curves().
5952 Also add a 'curve_name' member to EC_GROUP objects, which can be
5954 EC_GROUP_set_curve_name()
5955 EC_GROUP_get_curve_name()
5956 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5958 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5959 was actually never needed) and in BN_mul(). The removal in BN_mul()
5960 required a small change in bn_mul_part_recursive() and the addition
5961 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5962 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5963 bn_sub_words() and bn_add_words() except they take arrays with
5967 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5969 *) Cleanse PEM buffers before freeing them since they may contain
5971 [Benjamin Bennett <ben@psc.edu>]
5973 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5974 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5975 authentication-only ciphersuites.
5978 *) Since AES128 and AES256 share a single mask bit in the logic of
5979 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5980 kludge to work properly if AES128 is available and AES256 isn't.
5983 *) Expand security boundary to match 1.1.1 module.
5986 *) Remove redundant features: hash file source, editing of test vectors
5987 modify fipsld to use external fips_premain.c signature.
5990 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5991 run algorithm test programs.
5994 *) Make algorithm test programs more tolerant of whitespace.
5997 *) Have SSL/TLS server implementation tolerate "mismatched" record
5998 protocol version while receiving ClientHello even if the
5999 ClientHello is fragmented. (The server can't insist on the
6000 particular protocol version it has chosen before the ServerHello
6001 message has informed the client about his choice.)
6004 *) Load error codes if they are not already present instead of using a
6005 static variable. This allows them to be cleanly unloaded and reloaded.
6008 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6010 *) Introduce limits to prevent malicious keys being able to
6011 cause a denial of service. (CVE-2006-2940)
6012 [Steve Henson, Bodo Moeller]
6014 *) Fix ASN.1 parsing of certain invalid structures that can result
6015 in a denial of service. (CVE-2006-2937) [Steve Henson]
6017 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6018 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6020 *) Fix SSL client code which could crash if connecting to a
6021 malicious SSLv2 server. (CVE-2006-4343)
6022 [Tavis Ormandy and Will Drewry, Google Security Team]
6024 *) Change ciphersuite string processing so that an explicit
6025 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6026 will no longer include "AES128-SHA"), and any other similar
6027 ciphersuite (same bitmap) from *other* protocol versions (so that
6028 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6029 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6030 changes from 0.9.8b and 0.9.8d.
6033 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6035 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6036 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6038 *) Change the Unix randomness entropy gathering to use poll() when
6039 possible instead of select(), since the latter has some
6040 undesirable limitations.
6041 [Darryl Miles via Richard Levitte and Bodo Moeller]
6043 *) Disable rogue ciphersuites:
6045 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6046 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6047 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6049 The latter two were purportedly from
6050 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6053 Also deactive the remaining ciphersuites from
6054 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6055 unofficial, and the ID has long expired.
6058 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6059 dual-core machines) and other potential thread-safety issues.
6062 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6064 *) Adapt fipsld and the build system to link against the validated FIPS
6065 module in FIPS mode.
6068 *) Fixes for VC++ 2005 build under Windows.
6071 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6072 from a Windows bash shell such as MSYS. It is autodetected from the
6073 "config" script when run from a VC++ environment. Modify standard VC++
6074 build to use fipscanister.o from the GNU make build.
6077 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6079 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6080 The value now differs depending on if you build for FIPS or not.
6081 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6082 safely run with a non-FIPSed libcrypto, as it may crash because of
6083 the difference induced by this change.
6086 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6088 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6089 (part of SSL_OP_ALL). This option used to disable the
6090 countermeasure against man-in-the-middle protocol-version
6091 rollback in the SSL 2.0 server implementation, which is a bad
6092 idea. (CVE-2005-2969)
6094 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6095 for Information Security, National Institute of Advanced Industrial
6096 Science and Technology [AIST], Japan)]
6098 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6099 mainly for FIPS compliance and not fully integrated at this stage.
6102 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6103 the exponentiation using a fixed-length exponent. (Otherwise,
6104 the information leaked through timing could expose the secret key
6105 after many signatures; cf. Bleichenbacher's attack on DSA with
6109 *) Make a new fixed-window mod_exp implementation the default for
6110 RSA, DSA, and DH private-key operations so that the sequence of
6111 squares and multiplies and the memory access pattern are
6112 independent of the particular secret key. This will mitigate
6113 cache-timing and potential related attacks.
6115 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6116 and this is automatically used by BN_mod_exp_mont() if the new flag
6117 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6118 will use this BN flag for private exponents unless the flag
6119 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6120 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6122 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6124 *) Change the client implementation for SSLv23_method() and
6125 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6126 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6127 (Previously, the SSL 2.0 backwards compatible Client Hello
6128 message format would be used even with SSL_OP_NO_SSLv2.)
6131 *) Add support for smime-type MIME parameter in S/MIME messages which some
6135 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6136 a threadsafe manner. Modify rsa code to use new function and add calls
6137 to dsa and dh code (which had race conditions before).
6140 *) Include the fixed error library code in the C error file definitions
6141 instead of fixing them up at runtime. This keeps the error code
6142 structures constant.
6145 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6147 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6150 *) Fixes for newer kerberos headers. NB: the casts are needed because
6151 the 'length' field is signed on one version and unsigned on another
6152 with no (?) obvious way to tell the difference, without these VC++
6153 complains. Also the "definition" of FAR (blank) is no longer included
6154 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6155 some needed definitions.
6158 *) Undo Cygwin change.
6161 *) Added support for proxy certificates according to RFC 3820.
6162 Because they may be a security thread to unaware applications,
6163 they must be explicitly allowed in run-time. See
6164 docs/HOWTO/proxy_certificates.txt for further information.
6167 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6169 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6170 server and client random values. Previously
6171 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6172 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6174 This change has negligible security impact because:
6176 1. Server and client random values still have 24 bytes of pseudo random
6179 2. Server and client random values are sent in the clear in the initial
6182 3. The master secret is derived using the premaster secret (48 bytes in
6183 size for static RSA ciphersuites) as well as client server and random
6186 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6189 [Stephen Henson, reported by UK NISCC]
6191 *) Use Windows randomness collection on Cygwin.
6194 *) Fix hang in EGD/PRNGD query when communication socket is closed
6195 prematurely by EGD/PRNGD.
6196 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6198 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6201 *) Back-port of selected performance improvements from development
6202 branch, as well as improved support for PowerPC platforms.
6205 *) Add lots of checks for memory allocation failure, error codes to indicate
6206 failure and freeing up memory if a failure occurs.
6207 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6209 *) Add new -passin argument to dgst.
6212 *) Perform some character comparisons of different types in X509_NAME_cmp:
6213 this is needed for some certificates that re-encode DNs into UTF8Strings
6214 (in violation of RFC3280) and can't or won't issue name rollover
6218 *) Make an explicit check during certificate validation to see that
6219 the CA setting in each certificate on the chain is correct. As a
6220 side effect always do the following basic checks on extensions,
6221 not just when there's an associated purpose to the check:
6223 - if there is an unhandled critical extension (unless the user
6224 has chosen to ignore this fault)
6225 - if the path length has been exceeded (if one is set at all)
6226 - that certain extensions fit the associated purpose (if one has
6230 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6232 *) Avoid a race condition when CRLs are checked in a multi threaded
6233 environment. This would happen due to the reordering of the revoked
6234 entries during signature checking and serial number lookup. Now the
6235 encoding is cached and the serial number sort performed under a lock.
6236 Add new STACK function sk_is_sorted().
6239 *) Add Delta CRL to the extension code.
6242 *) Various fixes to s3_pkt.c so alerts are sent properly.
6243 [David Holmes <d.holmes@f5.com>]
6245 *) Reduce the chances of duplicate issuer name and serial numbers (in
6246 violation of RFC3280) using the OpenSSL certificate creation utilities.
6247 This is done by creating a random 64 bit value for the initial serial
6248 number when a serial number file is created or when a self signed
6249 certificate is created using 'openssl req -x509'. The initial serial
6250 number file is created using 'openssl x509 -next_serial' in CA.pl
6251 rather than being initialized to 1.
6254 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6256 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6257 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6258 [Joe Orton, Steve Henson]
6260 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6262 [Joe Orton, Steve Henson]
6264 *) Make it possible to have multiple active certificates with the same
6265 subject in the CA index file. This is done only if the keyword
6266 'unique_subject' is set to 'no' in the main CA section (default
6267 if 'CA_default') of the configuration file. The value is saved
6268 with the database itself in a separate index attribute file,
6269 named like the index file with '.attr' appended to the name.
6272 *) X509 verify fixes. Disable broken certificate workarounds when
6273 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6274 keyUsage extension present. Don't accept CRLs with unhandled critical
6275 extensions: since verify currently doesn't process CRL extensions this
6276 rejects a CRL with *any* critical extensions. Add new verify error codes
6280 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6281 A clarification of RFC2560 will require the use of OCTET STRINGs and
6282 some implementations cannot handle the current raw format. Since OpenSSL
6283 copies and compares OCSP nonces as opaque blobs without any attempt at
6284 parsing them this should not create any compatibility issues.
6287 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6288 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6289 this HMAC (and other) operations are several times slower than OpenSSL
6293 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6294 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6296 *) Use the correct content when signing type "other".
6299 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6301 *) Fix various bugs revealed by running the NISCC test suite:
6303 Stop out of bounds reads in the ASN1 code when presented with
6304 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6306 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6308 If verify callback ignores invalid public key errors don't try to check
6309 certificate signature with the NULL public key.
6313 *) New -ignore_err option in ocsp application to stop the server
6314 exiting on the first error in a request.
6317 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6318 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6322 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6323 extra data after the compression methods not only for TLS 1.0
6324 but also for SSL 3.0 (as required by the specification).
6325 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6327 *) Change X509_certificate_type() to mark the key as exported/exportable
6328 when it's 512 *bits* long, not 512 bytes.
6331 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6332 blocks during encryption.
6335 *) Various fixes to base64 BIO and non blocking I/O. On write
6336 flushes were not handled properly if the BIO retried. On read
6337 data was not being buffered properly and had various logic bugs.
6338 This also affects blocking I/O when the data being decoded is a
6342 *) Various S/MIME bugfixes and compatibility changes:
6343 output correct application/pkcs7 MIME type if
6344 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6345 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6346 of files as .eml work). Correctly handle very long lines in MIME
6350 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6352 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6353 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6354 a protocol version number mismatch like a decryption error
6355 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6358 *) Turn on RSA blinding by default in the default implementation
6359 to avoid a timing attack. Applications that don't want it can call
6360 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6361 They would be ill-advised to do so in most cases.
6362 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6364 *) Change RSA blinding code so that it works when the PRNG is not
6365 seeded (in this case, the secret RSA exponent is abused as
6366 an unpredictable seed -- if it is not unpredictable, there
6367 is no point in blinding anyway). Make RSA blinding thread-safe
6368 by remembering the creator's thread ID in rsa->blinding and
6369 having all other threads use local one-time blinding factors
6370 (this requires more computation than sharing rsa->blinding, but
6371 avoids excessive locking; and if an RSA object is not shared
6372 between threads, blinding will still be very fast).
6375 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6376 ENGINE as defaults for all supported algorithms irrespective of
6377 the 'flags' parameter. 'flags' is now honoured, so applications
6378 should make sure they are passing it correctly.
6381 *) Target "mingw" now allows native Windows code to be generated in
6382 the Cygwin environment as well as with the MinGW compiler.
6385 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6387 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6388 via timing by performing a MAC computation even if incorrect
6389 block cipher padding has been found. This is a countermeasure
6390 against active attacks where the attacker has to distinguish
6391 between bad padding and a MAC verification error. (CVE-2003-0078)
6393 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6394 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6395 Martin Vuagnoux (EPFL, Ilion)]
6397 *) Make the no-err option work as intended. The intention with no-err
6398 is not to have the whole error stack handling routines removed from
6399 libcrypto, it's only intended to remove all the function name and
6400 reason texts, thereby removing some of the footprint that may not
6401 be interesting if those errors aren't displayed anyway.
6403 NOTE: it's still possible for any application or module to have it's
6404 own set of error texts inserted. The routines are there, just not
6405 used by default when no-err is given.
6408 *) Add support for FreeBSD on IA64.
6409 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6411 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6412 Kerberos function mit_des_cbc_cksum(). Before this change,
6413 the value returned by DES_cbc_cksum() was like the one from
6414 mit_des_cbc_cksum(), except the bytes were swapped.
6415 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6417 *) Allow an application to disable the automatic SSL chain building.
6418 Before this a rather primitive chain build was always performed in
6419 ssl3_output_cert_chain(): an application had no way to send the
6420 correct chain if the automatic operation produced an incorrect result.
6422 Now the chain builder is disabled if either:
6424 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6426 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6428 The reasoning behind this is that an application would not want the
6429 auto chain building to take place if extra chain certificates are
6430 present and it might also want a means of sending no additional
6431 certificates (for example the chain has two certificates and the
6435 *) Add the possibility to build without the ENGINE framework.
6436 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6438 *) Under Win32 gmtime() can return NULL: check return value in
6439 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6442 *) DSA routines: under certain error conditions uninitialized BN objects
6443 could be freed. Solution: make sure initialization is performed early
6444 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6445 Nils Larsch <nla@trustcenter.de> via PR#459)
6448 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6449 checked on reconnect on the client side, therefore session resumption
6450 could still fail with a "ssl session id is different" error. This
6451 behaviour is masked when SSL_OP_ALL is used due to
6452 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6453 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6454 followup to PR #377.
6457 *) IA-32 assembler support enhancements: unified ELF targets, support
6458 for SCO/Caldera platforms, fix for Cygwin shared build.
6461 *) Add support for FreeBSD on sparc64. As a consequence, support for
6462 FreeBSD on non-x86 processors is separate from x86 processors on
6463 the config script, much like the NetBSD support.
6464 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6466 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6468 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6471 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6472 code (06) was taken as the first octet of the session ID and the last
6473 octet was ignored consequently. As a result SSLv2 client side session
6474 caching could not have worked due to the session ID mismatch between
6476 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6480 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6481 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6485 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6486 seems that in spite of existing for more than a year, many application
6487 author have done nothing to provide the necessary callbacks, which
6488 means that this particular engine will not work properly anywhere.
6489 This is a very unfortunate situation which forces us, in the name
6490 of usability, to give the hw_ncipher.c a static lock, which is part
6492 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6493 appear in 0.9.8 or later. We EXPECT application authors to have
6494 dealt properly with this when 0.9.8 is released (unless we actually
6495 make such changes in the libcrypto locking code that changes will
6496 have to be made anyway).
6499 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6500 octets have been read, EOF or an error occurs. Without this change
6501 some truncated ASN1 structures will not produce an error.
6504 *) Disable Heimdal support, since it hasn't been fully implemented.
6505 Still give the possibility to force the use of Heimdal, but with
6506 warnings and a request that patches get sent to openssl-dev.
6509 *) Add the VC-CE target, introduce the WINCE sysname, and add
6510 INSTALL.WCE and appropriate conditionals to make it build.
6511 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6513 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6514 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6515 edit numbers of the version.
6516 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6518 *) Introduce safe string copy and catenation functions
6519 (BUF_strlcpy() and BUF_strlcat()).
6520 [Ben Laurie (CHATS) and Richard Levitte]
6522 *) Avoid using fixed-size buffers for one-line DNs.
6523 [Ben Laurie (CHATS)]
6525 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6526 resizing buffers containing secrets, and use where appropriate.
6527 [Ben Laurie (CHATS)]
6529 *) Avoid using fixed size buffers for configuration file location.
6530 [Ben Laurie (CHATS)]
6532 *) Avoid filename truncation for various CA files.
6533 [Ben Laurie (CHATS)]
6535 *) Use sizeof in preference to magic numbers.
6536 [Ben Laurie (CHATS)]
6538 *) Avoid filename truncation in cert requests.
6539 [Ben Laurie (CHATS)]
6541 *) Add assertions to check for (supposedly impossible) buffer
6543 [Ben Laurie (CHATS)]
6545 *) Don't cache truncated DNS entries in the local cache (this could
6546 potentially lead to a spoofing attack).
6547 [Ben Laurie (CHATS)]
6549 *) Fix various buffers to be large enough for hex/decimal
6550 representations in a platform independent manner.
6551 [Ben Laurie (CHATS)]
6553 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6554 resizing buffers containing secrets, and use where appropriate.
6555 [Ben Laurie (CHATS)]
6557 *) Add BIO_indent() to avoid much slightly worrying code to do
6559 [Ben Laurie (CHATS)]
6561 *) Convert sprintf()/BIO_puts() to BIO_printf().
6562 [Ben Laurie (CHATS)]
6564 *) buffer_gets() could terminate with the buffer only half
6566 [Ben Laurie (CHATS)]
6568 *) Add assertions to prevent user-supplied crypto functions from
6569 overflowing internal buffers by having large block sizes, etc.
6570 [Ben Laurie (CHATS)]
6572 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6574 [Ben Laurie (CHATS)]
6576 *) Eliminate unused copy of key in RC4.
6577 [Ben Laurie (CHATS)]
6579 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6580 [Ben Laurie (CHATS)]
6582 *) Fix off-by-one error in EGD path.
6583 [Ben Laurie (CHATS)]
6585 *) If RANDFILE path is too long, ignore instead of truncating.
6586 [Ben Laurie (CHATS)]
6588 *) Eliminate unused and incorrectly sized X.509 structure
6590 [Ben Laurie (CHATS)]
6592 *) Eliminate unused and dangerous function knumber().
6593 [Ben Laurie (CHATS)]
6595 *) Eliminate unused and dangerous structure, KSSL_ERR.
6596 [Ben Laurie (CHATS)]
6598 *) Protect against overlong session ID context length in an encoded
6599 session object. Since these are local, this does not appear to be
6601 [Ben Laurie (CHATS)]
6603 *) Change from security patch (see 0.9.6e below) that did not affect
6604 the 0.9.6 release series:
6606 Remote buffer overflow in SSL3 protocol - an attacker could
6607 supply an oversized master key in Kerberos-enabled versions.
6609 [Ben Laurie (CHATS)]
6611 *) Change the SSL kerb5 codes to match RFC 2712.
6614 *) Make -nameopt work fully for req and add -reqopt switch.
6615 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6617 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6618 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6620 *) Make sure tests can be performed even if the corresponding algorithms
6621 have been removed entirely. This was also the last step to make
6622 OpenSSL compilable with DJGPP under all reasonable conditions.
6623 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6625 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6626 to allow version independent disabling of normally unselected ciphers,
6627 which may be activated as a side-effect of selecting a single cipher.
6629 (E.g., cipher list string "RSA" enables ciphersuites that are left
6630 out of "ALL" because they do not provide symmetric encryption.
6631 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6632 [Lutz Jaenicke, Bodo Moeller]
6634 *) Add appropriate support for separate platform-dependent build
6635 directories. The recommended way to make a platform-dependent
6636 build directory is the following (tested on Linux), maybe with
6639 # Place yourself outside of the OpenSSL source tree. In
6640 # this example, the environment variable OPENSSL_SOURCE
6641 # is assumed to contain the absolute OpenSSL source directory.
6642 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6643 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6644 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6645 mkdir -p `dirname $F`
6646 ln -s $OPENSSL_SOURCE/$F $F
6649 To be absolutely sure not to disturb the source tree, a "make clean"
6650 is a good thing. If it isn't successful, don't worry about it,
6651 it probably means the source directory is very clean.
6654 *) Make sure any ENGINE control commands make local copies of string
6655 pointers passed to them whenever necessary. Otherwise it is possible
6656 the caller may have overwritten (or deallocated) the original string
6657 data when a later ENGINE operation tries to use the stored values.
6658 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6660 *) Improve diagnostics in file reading and command-line digests.
6661 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6663 *) Add AES modes CFB and OFB to the object database. Correct an
6664 error in AES-CFB decryption.
6667 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6668 allows existing EVP_CIPHER_CTX structures to be reused after
6669 calling EVP_*Final(). This behaviour is used by encryption
6670 BIOs and some applications. This has the side effect that
6671 applications must explicitly clean up cipher contexts with
6672 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6675 *) Check the values of dna and dnb in bn_mul_recursive before calling
6676 bn_mul_comba (a non zero value means the a or b arrays do not contain
6677 n2 elements) and fallback to bn_mul_normal if either is not zero.
6680 *) Fix escaping of non-ASCII characters when using the -subj option
6681 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6684 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6685 form for "surname", serialNumber has no short form.
6686 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6687 therefore remove "mail" short name for "internet 7".
6688 The OID for unique identifiers in X509 certificates is
6689 x500UniqueIdentifier, not uniqueIdentifier.
6690 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6693 *) Add an "init" command to the ENGINE config module and auto initialize
6694 ENGINEs. Without any "init" command the ENGINE will be initialized
6695 after all ctrl commands have been executed on it. If init=1 the
6696 ENGINE is initailized at that point (ctrls before that point are run
6697 on the uninitialized ENGINE and after on the initialized one). If
6698 init=0 then the ENGINE will not be iniatialized at all.
6701 *) Fix the 'app_verify_callback' interface so that the user-defined
6702 argument is actually passed to the callback: In the
6703 SSL_CTX_set_cert_verify_callback() prototype, the callback
6704 declaration has been changed from
6707 int (*cb)(X509_STORE_CTX *,void *);
6708 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6709 i=s->ctx->app_verify_callback(&ctx)
6710 has been changed into
6711 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6713 To update applications using SSL_CTX_set_cert_verify_callback(),
6714 a dummy argument can be added to their callback functions.
6715 [D. K. Smetters <smetters@parc.xerox.com>]
6717 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6718 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6720 *) Add and OPENSSL_LOAD_CONF define which will cause
6721 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6722 This allows older applications to transparently support certain
6723 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6724 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6725 load the config file and OPENSSL_add_all_algorithms_conf() which will
6726 always load it have also been added.
6729 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6730 Adjust NIDs and EVP layer.
6731 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6733 *) Config modules support in openssl utility.
6735 Most commands now load modules from the config file,
6736 though in a few (such as version) this isn't done
6737 because it couldn't be used for anything.
6739 In the case of ca and req the config file used is
6740 the same as the utility itself: that is the -config
6741 command line option can be used to specify an
6745 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6746 use "openssl_conf" if filename is NULL use default openssl config file.
6749 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6750 config section name. Add a new flag to tolerate a missing config file
6751 and move code to CONF_modules_load_file().
6754 *) Support for crypto accelerator cards from Accelerated Encryption
6755 Processing, www.aep.ie. (Use engine 'aep')
6756 The support was copied from 0.9.6c [engine] and adapted/corrected
6757 to work with the new engine framework.
6758 [AEP Inc. and Richard Levitte]
6760 *) Support for SureWare crypto accelerator cards from Baltimore
6761 Technologies. (Use engine 'sureware')
6762 The support was copied from 0.9.6c [engine] and adapted
6763 to work with the new engine framework.
6766 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6767 make the newer ENGINE framework commands for the CHIL engine work.
6768 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6770 *) Make it possible to produce shared libraries on ReliantUNIX.
6771 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6773 *) Add the configuration target debug-linux-ppro.
6774 Make 'openssl rsa' use the general key loading routines
6775 implemented in apps.c, and make those routines able to
6776 handle the key format FORMAT_NETSCAPE and the variant
6778 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6780 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6781 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6783 *) Add -keyform to rsautl, and document -engine.
6784 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6786 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6787 BIO_R_NO_SUCH_FILE error code rather than the generic
6788 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6791 *) Add new functions
6793 ERR_peek_last_error_line
6794 ERR_peek_last_error_line_data.
6795 These are similar to
6798 ERR_peek_error_line_data,
6799 but report on the latest error recorded rather than the first one
6800 still in the error queue.
6801 [Ben Laurie, Bodo Moeller]
6803 *) default_algorithms option in ENGINE config module. This allows things
6805 default_algorithms = ALL
6806 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6809 *) Preliminary ENGINE config module.
6812 *) New experimental application configuration code.
6815 *) Change the AES code to follow the same name structure as all other
6816 symmetric ciphers, and behave the same way. Move everything to
6817 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6818 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6820 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6821 [Ben Laurie and Theo de Raadt]
6823 *) Add option to output public keys in req command.
6824 [Massimiliano Pala madwolf@openca.org]
6826 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6827 (up to about 10% better than before for P-192 and P-224).
6830 *) New functions/macros
6832 SSL_CTX_set_msg_callback(ctx, cb)
6833 SSL_CTX_set_msg_callback_arg(ctx, arg)
6834 SSL_set_msg_callback(ssl, cb)
6835 SSL_set_msg_callback_arg(ssl, arg)
6837 to request calling a callback function
6839 void cb(int write_p, int version, int content_type,
6840 const void *buf, size_t len, SSL *ssl, void *arg)
6842 whenever a protocol message has been completely received
6843 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6844 protocol version according to which the SSL library interprets
6845 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6846 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6847 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6848 specification (change_cipher_spec(20), alert(21), handshake(22)).
6849 'buf' and 'len' point to the actual message, 'ssl' to the
6850 SSL object, and 'arg' is the application-defined value set by
6851 SSL[_CTX]_set_msg_callback_arg().
6853 'openssl s_client' and 'openssl s_server' have new '-msg' options
6854 to enable a callback that displays all protocol messages.
6857 *) Change the shared library support so shared libraries are built as
6858 soon as the corresponding static library is finished, and thereby get
6859 openssl and the test programs linked against the shared library.
6860 This still only happens when the keyword "shard" has been given to
6861 the configuration scripts.
6863 NOTE: shared library support is still an experimental thing, and
6864 backward binary compatibility is still not guaranteed.
6865 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6867 *) Add support for Subject Information Access extension.
6868 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6870 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6871 additional bytes when new memory had to be allocated, not just
6872 when reusing an existing buffer.
6875 *) New command line and configuration option 'utf8' for the req command.
6876 This allows field values to be specified as UTF8 strings.
6879 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6880 runs for the former and machine-readable output for the latter.
6883 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6884 of the e-mail address in the DN (i.e., it will go into a certificate
6885 extension only). The new configuration file option 'email_in_dn = no'
6886 has the same effect.
6887 [Massimiliano Pala madwolf@openca.org]
6889 *) Change all functions with names starting with des_ to be starting
6890 with DES_ instead. Add wrappers that are compatible with libdes,
6891 but are named _ossl_old_des_*. Finally, add macros that map the
6892 des_* symbols to the corresponding _ossl_old_des_* if libdes
6893 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6894 desired, the des_* symbols will be mapped to DES_*, with one
6897 Since we provide two compatibility mappings, the user needs to
6898 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6899 compatibility is desired. The default (i.e., when that macro
6900 isn't defined) is OpenSSL 0.9.6c compatibility.
6902 There are also macros that enable and disable the support of old
6903 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6904 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6905 are defined, the default will apply: to support the old des routines.
6907 In either case, one must include openssl/des.h to get the correct
6908 definitions. Do not try to just include openssl/des_old.h, that
6911 NOTE: This is a major break of an old API into a new one. Software
6912 authors are encouraged to switch to the DES_ style functions. Some
6913 time in the future, des_old.h and the libdes compatibility functions
6914 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6915 default), and then completely removed.
6918 *) Test for certificates which contain unsupported critical extensions.
6919 If such a certificate is found during a verify operation it is
6920 rejected by default: this behaviour can be overridden by either
6921 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6922 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6923 X509_supported_extension() has also been added which returns 1 if a
6924 particular extension is supported.
6927 *) Modify the behaviour of EVP cipher functions in similar way to digests
6928 to retain compatibility with existing code.
6931 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6932 compatibility with existing code. In particular the 'ctx' parameter does
6933 not have to be to be initialized before the call to EVP_DigestInit() and
6934 it is tidied up after a call to EVP_DigestFinal(). New function
6935 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6936 EVP_MD_CTX_copy() changed to not require the destination to be
6937 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6938 requires the destination to be valid.
6940 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6941 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6944 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6945 so that complete 'Handshake' protocol structures are kept in memory
6946 instead of overwriting 'msg_type' and 'length' with 'body' data.
6949 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6950 [Massimo Santin via Richard Levitte]
6952 *) Major restructuring to the underlying ENGINE code. This includes
6953 reduction of linker bloat, separation of pure "ENGINE" manipulation
6954 (initialisation, etc) from functionality dealing with implementations
6955 of specific crypto iterfaces. This change also introduces integrated
6956 support for symmetric ciphers and digest implementations - so ENGINEs
6957 can now accelerate these by providing EVP_CIPHER and EVP_MD
6958 implementations of their own. This is detailed in crypto/engine/README
6959 as it couldn't be adequately described here. However, there are a few
6960 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6961 were changed in the original introduction of ENGINE code have now
6962 reverted back - the hooking from this code to ENGINE is now a good
6963 deal more passive and at run-time, operations deal directly with
6964 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6965 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6966 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6967 they were not being used by the framework as there is no concept of a
6968 BIGNUM_METHOD and they could not be generalised to the new
6969 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6970 ENGINE_cpy() has been removed as it cannot be consistently defined in
6974 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6977 *) Change mkdef.pl to sort symbols that get the same entry number,
6978 and make sure the automatically generated functions ERR_load_*
6979 become part of libeay.num as well.
6982 *) New function SSL_renegotiate_pending(). This returns true once
6983 renegotiation has been requested (either SSL_renegotiate() call
6984 or HelloRequest/ClientHello received from the peer) and becomes
6985 false once a handshake has been completed.
6986 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6987 sends a HelloRequest, but does not ensure that a handshake takes
6988 place. SSL_renegotiate_pending() is useful for checking if the
6989 client has followed the request.)
6992 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6993 By default, clients may request session resumption even during
6994 renegotiation (if session ID contexts permit); with this option,
6995 session resumption is possible only in the first handshake.
6997 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6998 more bits available for options that should not be part of
6999 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7002 *) Add some demos for certificate and certificate request creation.
7005 *) Make maximum certificate chain size accepted from the peer application
7006 settable (SSL*_get/set_max_cert_list()), as proposed by
7007 "Douglas E. Engert" <deengert@anl.gov>.
7010 *) Add support for shared libraries for Unixware-7
7011 (Boyd Lynn Gerber <gerberb@zenez.com>).
7014 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7015 be done prior to destruction. Use this to unload error strings from
7016 ENGINEs that load their own error strings. NB: This adds two new API
7017 functions to "get" and "set" this destroy handler in an ENGINE.
7020 *) Alter all existing ENGINE implementations (except "openssl" and
7021 "openbsd") to dynamically instantiate their own error strings. This
7022 makes them more flexible to be built both as statically-linked ENGINEs
7023 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7024 Also, add stub code to each that makes building them as self-contained
7025 shared-libraries easier (see README.ENGINE).
7028 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7029 implementations into applications that are completely implemented in
7030 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7031 commands that can be used to configure what shared-library to load and
7032 to control aspects of the way it is handled. Also, made an update to
7033 the README.ENGINE file that brings its information up-to-date and
7034 provides some information and instructions on the "dynamic" ENGINE
7035 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7038 *) Make it possible to unload ranges of ERR strings with a new
7039 "ERR_unload_strings" function.
7042 *) Add a copy() function to EVP_MD.
7045 *) Make EVP_MD routines take a context pointer instead of just the
7046 md_data void pointer.
7049 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7050 that the digest can only process a single chunk of data
7051 (typically because it is provided by a piece of
7052 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7053 is only going to provide a single chunk of data, and hence the
7054 framework needn't accumulate the data for oneshot drivers.
7057 *) As with "ERR", make it possible to replace the underlying "ex_data"
7058 functions. This change also alters the storage and management of global
7059 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7060 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7061 index counters. The API functions that use this state have been changed
7062 to take a "class_index" rather than pointers to the class's local STACK
7063 and counter, and there is now an API function to dynamically create new
7064 classes. This centralisation allows us to (a) plug a lot of the
7065 thread-safety problems that existed, and (b) makes it possible to clean
7066 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7067 such data would previously have always leaked in application code and
7068 workarounds were in place to make the memory debugging turn a blind eye
7069 to it. Application code that doesn't use this new function will still
7070 leak as before, but their memory debugging output will announce it now
7071 rather than letting it slide.
7073 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7074 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7075 has a return value to indicate success or failure.
7078 *) Make it possible to replace the underlying "ERR" functions such that the
7079 global state (2 LHASH tables and 2 locks) is only used by the "default"
7080 implementation. This change also adds two functions to "get" and "set"
7081 the implementation prior to it being automatically set the first time
7082 any other ERR function takes place. Ie. an application can call "get",
7083 pass the return value to a module it has just loaded, and that module
7084 can call its own "set" function using that value. This means the
7085 module's "ERR" operations will use (and modify) the error state in the
7086 application and not in its own statically linked copy of OpenSSL code.
7089 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7090 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7091 the operation, and provides a more encapsulated way for external code
7092 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7093 to use these functions rather than manually incrementing the counts.
7095 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7098 *) Add EVP test program.
7101 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7104 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7105 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7106 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7107 These allow a CRL to be built without having to access X509_CRL fields
7108 directly. Modify 'ca' application to use new functions.
7111 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7112 bug workarounds. Rollback attack detection is a security feature.
7113 The problem will only arise on OpenSSL servers when TLSv1 is not
7114 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7115 Software authors not wanting to support TLSv1 will have special reasons
7116 for their choice and can explicitly enable this option.
7117 [Bodo Moeller, Lutz Jaenicke]
7119 *) Rationalise EVP so it can be extended: don't include a union of
7120 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7121 (similar to those existing for EVP_CIPHER_CTX).
7126 EVP_MD_CTX_init(&md); /* new function call */
7127 EVP_DigestInit(&md, EVP_sha1());
7128 EVP_DigestUpdate(&md, in, len);
7129 EVP_DigestFinal(&md, out, NULL);
7130 EVP_MD_CTX_cleanup(&md); /* new function call */
7134 *) Make DES key schedule conform to the usual scheme, as well as
7135 correcting its structure. This means that calls to DES functions
7136 now have to pass a pointer to a des_key_schedule instead of a
7137 plain des_key_schedule (which was actually always a pointer
7140 des_key_schedule ks;
7142 des_set_key_checked(..., &ks);
7143 des_ncbc_encrypt(..., &ks, ...);
7145 (Note that a later change renames 'des_...' into 'DES_...'.)
7148 *) Initial reduction of linker bloat: the use of some functions, such as
7149 PEM causes large amounts of unused functions to be linked in due to
7150 poor organisation. For example pem_all.c contains every PEM function
7151 which has a knock on effect of linking in large amounts of (unused)
7152 ASN1 code. Grouping together similar functions and splitting unrelated
7153 functions prevents this.
7156 *) Cleanup of EVP macros.
7159 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7160 correct _ecb suffix.
7163 *) Add initial OCSP responder support to ocsp application. The
7164 revocation information is handled using the text based index
7165 use by the ca application. The responder can either handle
7166 requests generated internally, supplied in files (for example
7167 via a CGI script) or using an internal minimal server.
7170 *) Add configuration choices to get zlib compression for TLS.
7173 *) Changes to Kerberos SSL for RFC 2712 compliance:
7174 1. Implemented real KerberosWrapper, instead of just using
7175 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7176 2. Implemented optional authenticator field of KerberosWrapper.
7178 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7179 and authenticator structs; see crypto/krb5/.
7181 Generalized Kerberos calls to support multiple Kerberos libraries.
7182 [Vern Staats <staatsvr@asc.hpc.mil>,
7183 Jeffrey Altman <jaltman@columbia.edu>
7184 via Richard Levitte]
7186 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7187 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7188 values for each of the key sizes rather than having just
7189 parameters (and 'speed' generating keys each time).
7192 *) Speed up EVP routines.
7195 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7196 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7197 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7198 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7200 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7201 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7202 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7205 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7207 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7210 *) Added the OS2-EMX target.
7211 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7213 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7214 to support NCONF routines in extension code. New function CONF_set_nconf()
7215 to allow functions which take an NCONF to also handle the old LHASH
7216 structure: this means that the old CONF compatible routines can be
7217 retained (in particular wrt extensions) without having to duplicate the
7218 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7221 *) Enhance the general user interface with mechanisms for inner control
7222 and with possibilities to have yes/no kind of prompts.
7225 *) Change all calls to low level digest routines in the library and
7226 applications to use EVP. Add missing calls to HMAC_cleanup() and
7227 don't assume HMAC_CTX can be copied using memcpy().
7228 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7230 *) Add the possibility to control engines through control names but with
7231 arbitrary arguments instead of just a string.
7232 Change the key loaders to take a UI_METHOD instead of a callback
7233 function pointer. NOTE: this breaks binary compatibility with earlier
7234 versions of OpenSSL [engine].
7235 Adapt the nCipher code for these new conditions and add a card insertion
7239 *) Enhance the general user interface with mechanisms to better support
7240 dialog box interfaces, application-defined prompts, the possibility
7241 to use defaults (for example default passwords from somewhere else)
7242 and interrupts/cancellations.
7245 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7246 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7249 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7250 tidy up some unnecessarily weird code in 'sk_new()').
7251 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7253 *) Change the key loading routines for ENGINEs to use the same kind
7254 callback (pem_password_cb) as all other routines that need this
7258 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7259 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7260 than this minimum value is recommended.
7263 *) New random seeder for OpenVMS, using the system process statistics
7264 that are easily reachable.
7267 *) Windows apparently can't transparently handle global
7268 variables defined in DLLs. Initialisations such as:
7270 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7272 won't compile. This is used by the any applications that need to
7273 declare their own ASN1 modules. This was fixed by adding the option
7274 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7275 needed for static libraries under Win32.
7278 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7279 setting of purpose and trust fields. New X509_STORE trust and
7280 purpose functions and tidy up setting in other SSL functions.
7283 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7284 structure. These are inherited by X509_STORE_CTX when it is
7285 initialised. This allows various defaults to be set in the
7286 X509_STORE structure (such as flags for CRL checking and custom
7287 purpose or trust settings) for functions which only use X509_STORE_CTX
7288 internally such as S/MIME.
7290 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7291 trust settings if they are not set in X509_STORE. This allows X509_STORE
7292 purposes and trust (in S/MIME for example) to override any set by default.
7294 Add command line options for CRL checking to smime, s_client and s_server
7298 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7299 are set then the CRL is looked up in the X509_STORE structure and
7300 its validity and signature checked, then if the certificate is found
7301 in the CRL the verify fails with a revoked error.
7303 Various new CRL related callbacks added to X509_STORE_CTX structure.
7305 Command line options added to 'verify' application to support this.
7307 This needs some additional work, such as being able to handle multiple
7308 CRLs with different times, extension based lookup (rather than just
7309 by subject name) and ultimately more complete V2 CRL extension
7313 *) Add a general user interface API (crypto/ui/). This is designed
7314 to replace things like des_read_password and friends (backward
7315 compatibility functions using this new API are provided).
7316 The purpose is to remove prompting functions from the DES code
7317 section as well as provide for prompting through dialog boxes in
7318 a window system and the like.
7321 *) Add "ex_data" support to ENGINE so implementations can add state at a
7322 per-structure level rather than having to store it globally.
7325 *) Make it possible for ENGINE structures to be copied when retrieved by
7326 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7327 This causes the "original" ENGINE structure to act like a template,
7328 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7329 operational state can be localised to each ENGINE structure, despite the
7330 fact they all share the same "methods". New ENGINE structures returned in
7331 this case have no functional references and the return value is the single
7332 structural reference. This matches the single structural reference returned
7333 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7337 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7338 needs to match any other type at all we need to manually clear the
7342 *) Changes to the "openssl engine" utility to include;
7343 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7344 about an ENGINE's available control commands.
7345 - executing control commands from command line arguments using the
7346 '-pre' and '-post' switches. '-post' is only used if '-t' is
7347 specified and the ENGINE is successfully initialised. The syntax for
7348 the individual commands are colon-separated, for example;
7349 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7352 *) New dynamic control command support for ENGINEs. ENGINEs can now
7353 declare their own commands (numbers), names (strings), descriptions,
7354 and input types for run-time discovery by calling applications. A
7355 subset of these commands are implicitly classed as "executable"
7356 depending on their input type, and only these can be invoked through
7357 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7358 can be based on user input, config files, etc). The distinction is
7359 that "executable" commands cannot return anything other than a boolean
7360 result and can only support numeric or string input, whereas some
7361 discoverable commands may only be for direct use through
7362 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7363 pointers, or other custom uses. The "executable" commands are to
7364 support parameterisations of ENGINE behaviour that can be
7365 unambiguously defined by ENGINEs and used consistently across any
7366 OpenSSL-based application. Commands have been added to all the
7367 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7368 control over shared-library paths without source code alterations.
7371 *) Changed all ENGINE implementations to dynamically allocate their
7372 ENGINEs rather than declaring them statically. Apart from this being
7373 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7374 this also allows the implementations to compile without using the
7375 internal engine_int.h header.
7378 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7379 'const' value. Any code that should be able to modify a RAND_METHOD
7380 should already have non-const pointers to it (ie. they should only
7381 modify their own ones).
7384 *) Made a variety of little tweaks to the ENGINE code.
7385 - "atalla" and "ubsec" string definitions were moved from header files
7386 to C code. "nuron" string definitions were placed in variables
7387 rather than hard-coded - allowing parameterisation of these values
7388 later on via ctrl() commands.
7389 - Removed unused "#if 0"'d code.
7390 - Fixed engine list iteration code so it uses ENGINE_free() to release
7391 structural references.
7392 - Constified the RAND_METHOD element of ENGINE structures.
7393 - Constified various get/set functions as appropriate and added
7394 missing functions (including a catch-all ENGINE_cpy that duplicates
7395 all ENGINE values onto a new ENGINE except reference counts/state).
7396 - Removed NULL parameter checks in get/set functions. Setting a method
7397 or function to NULL is a way of cancelling out a previously set
7398 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7399 and doesn't justify the extra error symbols and code.
7400 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7401 flags from engine_int.h to engine.h.
7402 - Changed prototypes for ENGINE handler functions (init(), finish(),
7403 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7406 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7407 to the algorithm using long division. The binary algorithm can be
7408 used only if the modulus is odd. On 32-bit systems, it is faster
7409 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7410 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7411 up to 450 bits. In 64-bit environments, the binary algorithm
7412 appears to be advantageous for much longer moduli; here we use it
7413 for moduli up to 2048 bits.
7416 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7417 could not support the combine flag in choice fields.
7420 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7421 extensions from a certificate request to the certificate.
7424 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7425 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7426 file: this allows the display of the certificate about to be
7427 signed to be customised, to allow certain fields to be included
7428 or excluded and extension details. The old system didn't display
7429 multicharacter strings properly, omitted fields not in the policy
7430 and couldn't display additional details such as extensions.
7433 *) Function EC_POINTs_mul for multiple scalar multiplication
7434 of an arbitrary number of elliptic curve points
7435 \sum scalars[i]*points[i],
7436 optionally including the generator defined for the EC_GROUP:
7437 scalar*generator + \sum scalars[i]*points[i].
7439 EC_POINT_mul is a simple wrapper function for the typical case
7440 that the point list has just one item (besides the optional
7444 *) First EC_METHODs for curves over GF(p):
7446 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7447 operations and provides various method functions that can also
7448 operate with faster implementations of modular arithmetic.
7450 EC_GFp_mont_method() reuses most functions that are part of
7451 EC_GFp_simple_method, but uses Montgomery arithmetic.
7453 [Bodo Moeller; point addition and point doubling
7454 implementation directly derived from source code provided by
7455 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7457 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7458 crypto/ec/ec_lib.c):
7460 Curves are EC_GROUP objects (with an optional group generator)
7461 based on EC_METHODs that are built into the library.
7463 Points are EC_POINT objects based on EC_GROUP objects.
7465 Most of the framework would be able to handle curves over arbitrary
7466 finite fields, but as there are no obvious types for fields other
7467 than GF(p), some functions are limited to that for now.
7470 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7471 that the file contains a complete HTTP response.
7474 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7475 change the def and num file printf format specifier from "%-40sXXX"
7476 to "%-39s XXX". The latter will always guarantee a space after the
7477 field while the former will cause them to run together if the field
7478 is 40 of more characters long.
7481 *) Constify the cipher and digest 'method' functions and structures
7482 and modify related functions to take constant EVP_MD and EVP_CIPHER
7486 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7487 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7490 *) Modify EVP_Digest*() routines so they now return values. Although the
7491 internal software routines can never fail additional hardware versions
7495 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7497 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7498 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7501 ERR_R_NESTED_ASN1_ERROR
7503 ERR_R_MISSING_ASN1_EOS
7504 were 4 .. 9, conflicting with
7505 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7507 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7508 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7510 Add new error code 'ERR_R_INTERNAL_ERROR'.
7513 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7517 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7518 sets the subject name for a new request or supersedes the
7519 subject name in a given request. Formats that can be parsed are
7520 'CN=Some Name, OU=myOU, C=IT'
7522 'CN=Some Name/OU=myOU/C=IT'.
7524 Add options '-batch' and '-verbose' to 'openssl req'.
7525 [Massimiliano Pala <madwolf@hackmasters.net>]
7527 *) Introduce the possibility to access global variables through
7528 functions on platform were that's the best way to handle exporting
7529 global variables in shared libraries. To enable this functionality,
7530 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7531 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7532 is normally done by Configure or something similar).
7534 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7535 in the source file (foo.c) like this:
7537 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7538 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7540 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7541 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7543 OPENSSL_DECLARE_GLOBAL(int,foo);
7544 #define foo OPENSSL_GLOBAL_REF(foo)
7545 OPENSSL_DECLARE_GLOBAL(double,bar);
7546 #define bar OPENSSL_GLOBAL_REF(bar)
7548 The #defines are very important, and therefore so is including the
7549 header file everywhere where the defined globals are used.
7551 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7552 of ASN.1 items, but that structure is a bit different.
7554 The largest change is in util/mkdef.pl which has been enhanced with
7555 better and easier to understand logic to choose which symbols should
7556 go into the Windows .def files as well as a number of fixes and code
7557 cleanup (among others, algorithm keywords are now sorted
7558 lexicographically to avoid constant rewrites).
7561 *) In BN_div() keep a copy of the sign of 'num' before writing the
7562 result to 'rm' because if rm==num the value will be overwritten
7563 and produce the wrong result if 'num' is negative: this caused
7564 problems with BN_mod() and BN_nnmod().
7567 *) Function OCSP_request_verify(). This checks the signature on an
7568 OCSP request and verifies the signer certificate. The signer
7569 certificate is just checked for a generic purpose and OCSP request
7573 *) Add OCSP_check_validity() function to check the validity of OCSP
7574 responses. OCSP responses are prepared in real time and may only
7575 be a few seconds old. Simply checking that the current time lies
7576 between thisUpdate and nextUpdate max reject otherwise valid responses
7577 caused by either OCSP responder or client clock inaccuracy. Instead
7578 we allow thisUpdate and nextUpdate to fall within a certain period of
7579 the current time. The age of the response can also optionally be
7580 checked. Two new options -validity_period and -status_age added to
7584 *) If signature or public key algorithm is unrecognized print out its
7585 OID rather that just UNKNOWN.
7588 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7589 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7590 ID to be generated from the issuer certificate alone which can then be
7591 passed to OCSP_id_issuer_cmp().
7594 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7595 ASN1 modules to export functions returning ASN1_ITEM pointers
7596 instead of the ASN1_ITEM structures themselves. This adds several
7597 new macros which allow the underlying ASN1 function/structure to
7598 be accessed transparently. As a result code should not use ASN1_ITEM
7599 references directly (such as &X509_it) but instead use the relevant
7600 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7601 use of the new ASN1 code on platforms where exporting structures
7602 is problematical (for example in shared libraries) but exporting
7603 functions returning pointers to structures is not.
7606 *) Add support for overriding the generation of SSL/TLS session IDs.
7607 These callbacks can be registered either in an SSL_CTX or per SSL.
7608 The purpose of this is to allow applications to control, if they wish,
7609 the arbitrary values chosen for use as session IDs, particularly as it
7610 can be useful for session caching in multiple-server environments. A
7611 command-line switch for testing this (and any client code that wishes
7612 to use such a feature) has been added to "s_server".
7613 [Geoff Thorpe, Lutz Jaenicke]
7615 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7616 of the form '#if defined(...) || defined(...) || ...' and
7617 '#if !defined(...) && !defined(...) && ...'. This also avoids
7618 the growing number of special cases it was previously handling.
7621 *) Make all configuration macros available for application by making
7622 sure they are available in opensslconf.h, by giving them names starting
7623 with "OPENSSL_" to avoid conflicts with other packages and by making
7624 sure e_os2.h will cover all platform-specific cases together with
7626 Additionally, it is now possible to define configuration/platform-
7627 specific names (called "system identities"). In the C code, these
7628 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7629 macro with the name beginning with "OPENSSL_SYS_", which is determined
7630 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7634 *) New option -set_serial to 'req' and 'x509' this allows the serial
7635 number to use to be specified on the command line. Previously self
7636 signed certificates were hard coded with serial number 0 and the
7637 CA options of 'x509' had to use a serial number in a file which was
7641 *) New options to 'ca' utility to support V2 CRL entry extensions.
7642 Currently CRL reason, invalidity date and hold instruction are
7643 supported. Add new CRL extensions to V3 code and some new objects.
7646 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7647 disable standard block padding (aka PKCS#5 padding) in the EVP
7648 API, which was previously mandatory. This means that the data is
7649 not padded in any way and so the total length much be a multiple
7650 of the block size, otherwise an error occurs.
7653 *) Initial (incomplete) OCSP SSL support.
7656 *) New function OCSP_parse_url(). This splits up a URL into its host,
7657 port and path components: primarily to parse OCSP URLs. New -url
7658 option to ocsp utility.
7661 *) New nonce behavior. The return value of OCSP_check_nonce() now
7662 reflects the various checks performed. Applications can decide
7663 whether to tolerate certain situations such as an absent nonce
7664 in a response when one was present in a request: the ocsp application
7665 just prints out a warning. New function OCSP_add1_basic_nonce()
7666 this is to allow responders to include a nonce in a response even if
7667 the request is nonce-less.
7670 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7671 skipped when using openssl x509 multiple times on a single input file,
7672 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7675 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7676 set string type: to handle setting ASN1_TIME structures. Fix ca
7677 utility to correctly initialize revocation date of CRLs.
7680 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7681 the clients preferred ciphersuites and rather use its own preferences.
7682 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7683 Internet Explorer by ensuring unchanged hash method during stepup.
7684 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7687 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7688 to aes and add a new 'exist' option to print out symbols that don't
7692 *) Additional options to ocsp utility to allow flags to be set and
7693 additional certificates supplied.
7696 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7697 OCSP client a number of certificate to only verify the response
7701 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7702 handle the new API. Currently only ECB, CBC modes supported. Add new
7705 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7706 Encryption Standard (AES) Ciphersuites for Transport Layer
7707 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7708 not enabled by default and were not part of the "ALL" ciphersuite
7709 alias because they were not yet official; they could be
7710 explicitly requested by specifying the "AESdraft" ciphersuite
7711 group alias. In the final release of OpenSSL 0.9.7, the group
7712 alias is called "AES" and is part of "ALL".)
7713 [Ben Laurie, Steve Henson, Bodo Moeller]
7715 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7716 request to response.
7719 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7720 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7721 extract information from a certificate request. OCSP_response_create()
7722 creates a response and optionally adds a basic response structure.
7723 OCSP_basic_add1_status() adds a complete single response to a basic
7724 response and returns the OCSP_SINGLERESP structure just added (to allow
7725 extensions to be included for example). OCSP_basic_add1_cert() adds a
7726 certificate to a basic response and OCSP_basic_sign() signs a basic
7727 response with various flags. New helper functions ASN1_TIME_check()
7728 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7729 (converts ASN1_TIME to GeneralizedTime).
7732 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7733 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7734 structure from a certificate. X509_pubkey_digest() digests the public_key
7735 contents: this is used in various key identifiers.
7738 *) Make sk_sort() tolerate a NULL argument.
7739 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7741 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7742 passed by the function are trusted implicitly. If any of them signed the
7743 response then it is assumed to be valid and is not verified.
7746 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7747 to data. This was previously part of the PKCS7 ASN1 code. This
7748 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7749 [Steve Henson, reported by Kenneth R. Robinette
7750 <support@securenetterm.com>]
7752 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7753 routines: without these tracing memory leaks is very painful.
7754 Fix leaks in PKCS12 and PKCS7 routines.
7757 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7758 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7759 effectively meant GeneralizedTime would never be used. Now it
7760 is initialised to -1 but X509_time_adj() now has to check the value
7761 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7762 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7763 [Steve Henson, reported by Kenneth R. Robinette
7764 <support@securenetterm.com>]
7766 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7767 result in a zero length in the ASN1_INTEGER structure which was
7768 not consistent with the structure when d2i_ASN1_INTEGER() was used
7769 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7770 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7771 where it did not print out a minus for negative ASN1_INTEGER.
7774 *) Add summary printout to ocsp utility. The various functions which
7775 convert status values to strings have been renamed to:
7776 OCSP_response_status_str(), OCSP_cert_status_str() and
7777 OCSP_crl_reason_str() and are no longer static. New options
7778 to verify nonce values and to disable verification. OCSP response
7779 printout format cleaned up.
7782 *) Add additional OCSP certificate checks. These are those specified
7783 in RFC2560. This consists of two separate checks: the CA of the
7784 certificate being checked must either be the OCSP signer certificate
7785 or the issuer of the OCSP signer certificate. In the latter case the
7786 OCSP signer certificate must contain the OCSP signing extended key
7787 usage. This check is performed by attempting to match the OCSP
7788 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7789 in the OCSP_CERTID structures of the response.
7792 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7793 and related routines. This uses the standard OpenSSL certificate
7794 verify routines to perform initial checks (just CA validity) and
7795 to obtain the certificate chain. Then additional checks will be
7796 performed on the chain. Currently the root CA is checked to see
7797 if it is explicitly trusted for OCSP signing. This is used to set
7798 a root CA as a global signing root: that is any certificate that
7799 chains to that CA is an acceptable OCSP signing certificate.
7802 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7803 extensions from a separate configuration file.
7804 As when reading extensions from the main configuration file,
7805 the '-extensions ...' option may be used for specifying the
7807 [Massimiliano Pala <madwolf@comune.modena.it>]
7809 *) New OCSP utility. Allows OCSP requests to be generated or
7810 read. The request can be sent to a responder and the output
7811 parsed, outputed or printed in text form. Not complete yet:
7812 still needs to check the OCSP response validity.
7815 *) New subcommands for 'openssl ca':
7816 'openssl ca -status <serial>' prints the status of the cert with
7817 the given serial number (according to the index file).
7818 'openssl ca -updatedb' updates the expiry status of certificates
7820 [Massimiliano Pala <madwolf@comune.modena.it>]
7822 *) New '-newreq-nodes' command option to CA.pl. This is like
7823 '-newreq', but calls 'openssl req' with the '-nodes' option
7824 so that the resulting key is not encrypted.
7825 [Damien Miller <djm@mindrot.org>]
7827 *) New configuration for the GNU Hurd.
7828 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7830 *) Initial code to implement OCSP basic response verify. This
7831 is currently incomplete. Currently just finds the signer's
7832 certificate and verifies the signature on the response.
7835 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7836 value of OPENSSLDIR. This is available via the new '-d' option
7837 to 'openssl version', and is also included in 'openssl version -a'.
7840 *) Allowing defining memory allocation callbacks that will be given
7841 file name and line number information in additional arguments
7842 (a const char* and an int). The basic functionality remains, as
7843 well as the original possibility to just replace malloc(),
7844 realloc() and free() by functions that do not know about these
7845 additional arguments. To register and find out the current
7846 settings for extended allocation functions, the following
7847 functions are provided:
7849 CRYPTO_set_mem_ex_functions
7850 CRYPTO_set_locked_mem_ex_functions
7851 CRYPTO_get_mem_ex_functions
7852 CRYPTO_get_locked_mem_ex_functions
7854 These work the same way as CRYPTO_set_mem_functions and friends.
7855 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7856 extended allocation function is enabled.
7857 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7858 a conventional allocation function is enabled.
7859 [Richard Levitte, Bodo Moeller]
7861 *) Finish off removing the remaining LHASH function pointer casts.
7862 There should no longer be any prototype-casting required when using
7863 the LHASH abstraction, and any casts that remain are "bugs". See
7864 the callback types and macros at the head of lhash.h for details
7865 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7868 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7869 If /dev/[u]random devices are not available or do not return enough
7870 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7872 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7873 /etc/entropy will be queried once each in this sequence, querying stops
7874 when enough entropy was collected without querying more sockets.
7877 *) Change the Unix RAND_poll() variant to be able to poll several
7878 random devices, as specified by DEVRANDOM, until a sufficient amount
7879 of data has been collected. We spend at most 10 ms on each file
7880 (select timeout) and read in non-blocking mode. DEVRANDOM now
7881 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7882 (previously it was just the string "/dev/urandom"), so on typical
7883 platforms the 10 ms delay will never occur.
7884 Also separate out the Unix variant to its own file, rand_unix.c.
7885 For VMS, there's a currently-empty rand_vms.c.
7888 *) Move OCSP client related routines to ocsp_cl.c. These
7889 provide utility functions which an application needing
7890 to issue a request to an OCSP responder and analyse the
7891 response will typically need: as opposed to those which an
7892 OCSP responder itself would need which will be added later.
7894 OCSP_request_sign() signs an OCSP request with an API similar
7895 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7896 response. OCSP_response_get1_basic() extracts basic response
7897 from response. OCSP_resp_find_status(): finds and extracts status
7898 information from an OCSP_CERTID structure (which will be created
7899 when the request structure is built). These are built from lower
7900 level functions which work on OCSP_SINGLERESP structures but
7901 won't normally be used unless the application wishes to examine
7902 extensions in the OCSP response for example.
7904 Replace nonce routines with a pair of functions.
7905 OCSP_request_add1_nonce() adds a nonce value and optionally
7906 generates a random value. OCSP_check_nonce() checks the
7907 validity of the nonce in an OCSP response.
7910 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7911 This doesn't copy the supplied OCSP_CERTID and avoids the
7912 need to free up the newly created id. Change return type
7913 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7914 This can then be used to add extensions to the request.
7915 Deleted OCSP_request_new(), since most of its functionality
7916 is now in OCSP_REQUEST_new() (and the case insensitive name
7917 clash) apart from the ability to set the request name which
7918 will be added elsewhere.
7921 *) Update OCSP API. Remove obsolete extensions argument from
7922 various functions. Extensions are now handled using the new
7923 OCSP extension code. New simple OCSP HTTP function which
7924 can be used to send requests and parse the response.
7927 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7928 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7929 uses the special reorder version of SET OF to sort the attributes
7930 and reorder them to match the encoded order. This resolves a long
7931 standing problem: a verify on a PKCS7 structure just after signing
7932 it used to fail because the attribute order did not match the
7933 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7934 it uses the received order. This is necessary to tolerate some broken
7935 software that does not order SET OF. This is handled by encoding
7936 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7937 to produce the required SET OF.
7940 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7941 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7942 files to get correct declarations of the ASN.1 item variables.
7945 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7946 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7947 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7948 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7949 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7950 ASN1_ITEM and no wrapper functions.
7953 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7954 replace the old function pointer based I/O routines. Change most of
7955 the *_d2i_bio() and *_d2i_fp() functions to use these.
7958 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7959 lines, recognice more "algorithms" that can be deselected, and make
7960 it complain about algorithm deselection that isn't recognised.
7963 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7964 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7965 to use new functions. Add NO_ASN1_OLD which can be set to remove
7966 some old style ASN1 functions: this can be used to determine if old
7967 code will still work when these eventually go away.
7970 *) New extension functions for OCSP structures, these follow the
7971 same conventions as certificates and CRLs.
7974 *) New function X509V3_add1_i2d(). This automatically encodes and
7975 adds an extension. Its behaviour can be customised with various
7976 flags to append, replace or delete. Various wrappers added for
7977 certificates and CRLs.
7980 *) Fix to avoid calling the underlying ASN1 print routine when
7981 an extension cannot be parsed. Correct a typo in the
7982 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7985 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7986 entries for variables.
7989 *) Add functionality to apps/openssl.c for detecting locking
7990 problems: As the program is single-threaded, all we have
7991 to do is register a locking callback using an array for
7992 storing which locks are currently held by the program.
7995 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7996 SSL_get_ex_data_X509_STORE_idx(), which is used in
7997 ssl_verify_cert_chain() and thus can be called at any time
7998 during TLS/SSL handshakes so that thread-safety is essential.
7999 Unfortunately, the ex_data design is not at all suited
8000 for multi-threaded use, so it probably should be abolished.
8003 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8004 [Broadcom, tweaked and integrated by Geoff Thorpe]
8006 *) Move common extension printing code to new function
8007 X509V3_print_extensions(). Reorganise OCSP print routines and
8008 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8011 *) New function X509_signature_print() to remove duplication in some
8015 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8016 set (this was treated exactly the same as SET OF previously). This
8017 is used to reorder the STACK representing the structure to match the
8018 encoding. This will be used to get round a problem where a PKCS7
8019 structure which was signed could not be verified because the STACK
8020 order did not reflect the encoded order.
8023 *) Reimplement the OCSP ASN1 module using the new code.
8026 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8027 for its ASN1 operations. The old style function pointers still exist
8028 for now but they will eventually go away.
8031 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8032 completely replaces the old ASN1 functionality with a table driven
8033 encoder and decoder which interprets an ASN1_ITEM structure describing
8034 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8035 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8036 has also been converted to the new form.
8039 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8040 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8041 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8042 for negative moduli.
8045 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8046 of not touching the result's sign bit.
8049 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8053 *) Changed the LHASH code to use prototypes for callbacks, and created
8054 macros to declare and implement thin (optionally static) functions
8055 that provide type-safety and avoid function pointer casting for the
8056 type-specific callbacks.
8059 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8061 [Veers Staats <staatsvr@asc.hpc.mil>,
8062 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8064 *) Reformat the FAQ so the different questions and answers can be divided
8065 in sections depending on the subject.
8068 *) Have the zlib compression code load ZLIB.DLL dynamically under
8072 *) New function BN_mod_sqrt for computing square roots modulo a prime
8073 (using the probabilistic Tonelli-Shanks algorithm unless
8074 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8075 be handled deterministically).
8076 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8078 *) Make BN_mod_inverse faster by explicitly handling small quotients
8079 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8080 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8083 *) New function BN_kronecker.
8086 *) Fix BN_gcd so that it works on negative inputs; the result is
8087 positive unless both parameters are zero.
8088 Previously something reasonably close to an infinite loop was
8089 possible because numbers could be growing instead of shrinking
8090 in the implementation of Euclid's algorithm.
8093 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8094 sign of the number in question.
8096 Fix BN_is_word(a,w) to work correctly for w == 0.
8098 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8099 because its test if the absolute value of 'a' equals 'w'.
8100 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8101 it exists mostly for use in the implementations of BN_is_zero(),
8102 BN_is_one(), and BN_is_word().
8105 *) New function BN_swap.
8108 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8109 the exponentiation functions are more likely to produce reasonable
8110 results on negative inputs.
8113 *) Change BN_mod_mul so that the result is always non-negative.
8114 Previously, it could be negative if one of the factors was negative;
8115 I don't think anyone really wanted that behaviour.
8118 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8119 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8120 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8121 and add new functions:
8130 BN_mod_lshift1_quick
8134 These functions always generate non-negative results.
8136 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8137 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8139 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8140 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8141 be reduced modulo m.
8142 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8145 The following entry accidentally appeared in the CHANGES file
8146 distributed with OpenSSL 0.9.7. The modifications described in
8147 it do *not* apply to OpenSSL 0.9.7.
8149 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8150 was actually never needed) and in BN_mul(). The removal in BN_mul()
8151 required a small change in bn_mul_part_recursive() and the addition
8152 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8153 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8154 bn_sub_words() and bn_add_words() except they take arrays with
8159 *) In 'openssl passwd', verify passwords read from the terminal
8160 unless the '-salt' option is used (which usually means that
8161 verification would just waste user's time since the resulting
8162 hash is going to be compared with some given password hash)
8163 or the new '-noverify' option is used.
8165 This is an incompatible change, but it does not affect
8166 non-interactive use of 'openssl passwd' (passwords on the command
8167 line, '-stdin' option, '-in ...' option) and thus should not
8171 *) Remove all references to RSAref, since there's no more need for it.
8174 *) Make DSO load along a path given through an environment variable
8175 (SHLIB_PATH) with shl_load().
8178 *) Constify the ENGINE code as a result of BIGNUM constification.
8179 Also constify the RSA code and most things related to it. In a
8180 few places, most notable in the depth of the ASN.1 code, ugly
8181 casts back to non-const were required (to be solved at a later
8185 *) Make it so the openssl application has all engines loaded by default.
8188 *) Constify the BIGNUM routines a little more.
8191 *) Add the following functions:
8193 ENGINE_load_cswift()
8195 ENGINE_load_atalla()
8197 ENGINE_load_builtin_engines()
8199 That way, an application can itself choose if external engines that
8200 are built-in in OpenSSL shall ever be used or not. The benefit is
8201 that applications won't have to be linked with libdl or other dso
8202 libraries unless it's really needed.
8204 Changed 'openssl engine' to load all engines on demand.
8205 Changed the engine header files to avoid the duplication of some
8206 declarations (they differed!).
8209 *) 'openssl engine' can now list capabilities.
8212 *) Better error reporting in 'openssl engine'.
8215 *) Never call load_dh_param(NULL) in s_server.
8218 *) Add engine application. It can currently list engines by name and
8219 identity, and test if they are actually available.
8222 *) Improve RPM specification file by forcing symbolic linking and making
8223 sure the installed documentation is also owned by root.root.
8224 [Damien Miller <djm@mindrot.org>]
8226 *) Give the OpenSSL applications more possibilities to make use of
8227 keys (public as well as private) handled by engines.
8230 *) Add OCSP code that comes from CertCo.
8233 *) Add VMS support for the Rijndael code.
8236 *) Added untested support for Nuron crypto accelerator.
8239 *) Add support for external cryptographic devices. This code was
8240 previously distributed separately as the "engine" branch.
8241 [Geoff Thorpe, Richard Levitte]
8243 *) Rework the filename-translation in the DSO code. It is now possible to
8244 have far greater control over how a "name" is turned into a filename
8245 depending on the operating environment and any oddities about the
8246 different shared library filenames on each system.
8249 *) Support threads on FreeBSD-elf in Configure.
8252 *) Fix for SHA1 assembly problem with MASM: it produces
8253 warnings about corrupt line number information when assembling
8254 with debugging information. This is caused by the overlapping
8256 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8259 NCONF_get_number() has no error checking at all. As a replacement,
8260 NCONF_get_number_e() is defined (_e for "error checking") and is
8261 promoted strongly. The old NCONF_get_number is kept around for
8262 binary backward compatibility.
8263 Make it possible for methods to load from something other than a BIO,
8264 by providing a function pointer that is given a name instead of a BIO.
8265 For example, this could be used to load configuration data from an
8269 *) Fix for non blocking accept BIOs. Added new I/O special reason
8270 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8271 with non blocking I/O was not possible because no retry code was
8272 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8276 *) Added the beginnings of Rijndael support.
8279 *) Fix for bug in DirectoryString mask setting. Add support for
8280 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8281 to allow certificate printing to more controllable, additional
8282 'certopt' option to 'x509' to allow new printing options to be
8286 *) Clean old EAY MD5 hack from e_os.h.
8289 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8291 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8292 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8293 [Joe Orton, Steve Henson]
8295 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8297 *) Fix additional bug revealed by the NISCC test suite:
8299 Stop bug triggering large recursion when presented with
8300 certain ASN.1 tags (CVE-2003-0851)
8303 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8305 *) Fix various bugs revealed by running the NISCC test suite:
8307 Stop out of bounds reads in the ASN1 code when presented with
8308 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8310 If verify callback ignores invalid public key errors don't try to check
8311 certificate signature with the NULL public key.
8315 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8316 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8320 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8321 extra data after the compression methods not only for TLS 1.0
8322 but also for SSL 3.0 (as required by the specification).
8323 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8325 *) Change X509_certificate_type() to mark the key as exported/exportable
8326 when it's 512 *bits* long, not 512 bytes.
8329 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8331 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8332 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8333 a protocol version number mismatch like a decryption error
8334 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8337 *) Turn on RSA blinding by default in the default implementation
8338 to avoid a timing attack. Applications that don't want it can call
8339 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8340 They would be ill-advised to do so in most cases.
8341 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8343 *) Change RSA blinding code so that it works when the PRNG is not
8344 seeded (in this case, the secret RSA exponent is abused as
8345 an unpredictable seed -- if it is not unpredictable, there
8346 is no point in blinding anyway). Make RSA blinding thread-safe
8347 by remembering the creator's thread ID in rsa->blinding and
8348 having all other threads use local one-time blinding factors
8349 (this requires more computation than sharing rsa->blinding, but
8350 avoids excessive locking; and if an RSA object is not shared
8351 between threads, blinding will still be very fast).
8354 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8356 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8357 via timing by performing a MAC computation even if incorrrect
8358 block cipher padding has been found. This is a countermeasure
8359 against active attacks where the attacker has to distinguish
8360 between bad padding and a MAC verification error. (CVE-2003-0078)
8362 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8363 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8364 Martin Vuagnoux (EPFL, Ilion)]
8366 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8368 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8369 memory from it's contents. This is done with a counter that will
8370 place alternating values in each byte. This can be used to solve
8371 two issues: 1) the removal of calls to memset() by highly optimizing
8372 compilers, and 2) cleansing with other values than 0, since those can
8373 be read through on certain media, for example a swap space on disk.
8376 *) Bugfix: client side session caching did not work with external caching,
8377 because the session->cipher setting was not restored when reloading
8378 from the external cache. This problem was masked, when
8379 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8380 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8383 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8384 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8385 [Zeev Lieber <zeev-l@yahoo.com>]
8387 *) Undo an undocumented change introduced in 0.9.6e which caused
8388 repeated calls to OpenSSL_add_all_ciphers() and
8389 OpenSSL_add_all_digests() to be ignored, even after calling
8393 *) Change the default configuration reader to deal with last line not
8394 being properly terminated.
8397 *) Change X509_NAME_cmp() so it applies the special rules on handling
8398 DN values that are of type PrintableString, as well as RDNs of type
8399 emailAddress where the value has the type ia5String.
8400 [stefank@valicert.com via Richard Levitte]
8402 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8403 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8404 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8405 the bitwise-OR of the two for use by the majority of applications
8406 wanting this behaviour, and update the docs. The documented
8407 behaviour and actual behaviour were inconsistent and had been
8408 changing anyway, so this is more a bug-fix than a behavioural
8410 [Geoff Thorpe, diagnosed by Nadav Har'El]
8412 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8413 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8416 *) Fix initialization code race conditions in
8417 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8418 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8419 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8420 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8421 ssl2_get_cipher_by_char(),
8422 ssl3_get_cipher_by_char().
8423 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8425 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8426 the cached sessions are flushed, as the remove_cb() might use ex_data
8427 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8428 (see [openssl.org #212]).
8429 [Geoff Thorpe, Lutz Jaenicke]
8431 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8432 length, instead of the encoding length to d2i_ASN1_OBJECT.
8435 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8437 *) [In 0.9.6g-engine release:]
8438 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8439 [Lynn Gazis <lgazis@rainbow.com>]
8441 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8443 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8444 and get fix the header length calculation.
8445 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8446 Alon Kantor <alonk@checkpoint.com> (and others),
8449 *) Use proper error handling instead of 'assertions' in buffer
8450 overflow checks added in 0.9.6e. This prevents DoS (the
8451 assertions could call abort()).
8452 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8454 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8456 *) Add various sanity checks to asn1_get_length() to reject
8457 the ASN1 length bytes if they exceed sizeof(long), will appear
8458 negative or the content length exceeds the length of the
8460 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8462 *) Fix cipher selection routines: ciphers without encryption had no flags
8463 for the cipher strength set and where therefore not handled correctly
8464 by the selection routines (PR #130).
8467 *) Fix EVP_dsa_sha macro.
8471 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8472 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8473 that was added in OpenSSL 0.9.6d.
8475 As the countermeasure turned out to be incompatible with some
8476 broken SSL implementations, the new option is part of SSL_OP_ALL.
8477 SSL_OP_ALL is usually employed when compatibility with weird SSL
8478 implementations is desired (e.g. '-bugs' option to 's_client' and
8479 's_server'), so the new option is automatically set in many
8483 *) Changes in security patch:
8485 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8486 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8487 Air Force Materiel Command, USAF, under agreement number
8490 *) Add various sanity checks to asn1_get_length() to reject
8491 the ASN1 length bytes if they exceed sizeof(long), will appear
8492 negative or the content length exceeds the length of the
8493 supplied buffer. (CVE-2002-0659)
8494 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8496 *) Assertions for various potential buffer overflows, not known to
8498 [Ben Laurie (CHATS)]
8500 *) Various temporary buffers to hold ASCII versions of integers were
8501 too small for 64 bit platforms. (CVE-2002-0655)
8502 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8504 *) Remote buffer overflow in SSL3 protocol - an attacker could
8505 supply an oversized session ID to a client. (CVE-2002-0656)
8506 [Ben Laurie (CHATS)]
8508 *) Remote buffer overflow in SSL2 protocol - an attacker could
8509 supply an oversized client master key. (CVE-2002-0656)
8510 [Ben Laurie (CHATS)]
8512 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8514 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8515 encoded as NULL) with id-dsa-with-sha1.
8516 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8518 *) Check various X509_...() return values in apps/req.c.
8519 [Nils Larsch <nla@trustcenter.de>]
8521 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8522 an end-of-file condition would erroneously be flagged, when the CRLF
8523 was just at the end of a processed block. The bug was discovered when
8524 processing data through a buffering memory BIO handing the data to a
8525 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8526 <ptsekov@syntrex.com> and Nedelcho Stanev.
8529 *) Implement a countermeasure against a vulnerability recently found
8530 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8531 before application data chunks to avoid the use of known IVs
8532 with data potentially chosen by the attacker.
8535 *) Fix length checks in ssl3_get_client_hello().
8538 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8539 to prevent ssl3_read_internal() from incorrectly assuming that
8540 ssl3_read_bytes() found application data while handshake
8541 processing was enabled when in fact s->s3->in_read_app_data was
8542 merely automatically cleared during the initial handshake.
8543 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8545 *) Fix object definitions for Private and Enterprise: they were not
8546 recognized in their shortname (=lowercase) representation. Extend
8547 obj_dat.pl to issue an error when using undefined keywords instead
8548 of silently ignoring the problem (Svenning Sorensen
8549 <sss@sss.dnsalias.net>).
8552 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8553 generators, i.e. generators other than 2 and 5. (Previously, the
8554 code did not properly initialise the 'add' and 'rem' values to
8555 BN_generate_prime().)
8557 In the new general case, we do not insist that 'generator' is
8558 actually a primitive root: This requirement is rather pointless;
8559 a generator of the order-q subgroup is just as good, if not
8563 *) Map new X509 verification errors to alerts. Discovered and submitted by
8564 Tom Wu <tom@arcot.com>.
8567 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8568 returning non-zero before the data has been completely received
8569 when using non-blocking I/O.
8570 [Bodo Moeller; problem pointed out by John Hughes]
8572 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8573 [Ben Laurie, Lutz Jaenicke]
8575 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8576 Yoram Zahavi <YoramZ@gilian.com>).
8579 *) Add information about CygWin 1.3 and on, and preserve proper
8580 configuration for the versions before that.
8581 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8583 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8584 check whether we deal with a copy of a session and do not delete from
8585 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8586 <izhar@checkpoint.com>.
8589 *) Do not store session data into the internal session cache, if it
8590 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8591 flag is set). Proposed by Aslam <aslam@funk.com>.
8594 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8598 *) [In 0.9.6d-engine release:]
8599 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8600 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8602 *) Add the configuration target linux-s390x.
8603 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8605 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8606 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8607 variable as an indication that a ClientHello message has been
8608 received. As the flag value will be lost between multiple
8609 invocations of ssl3_accept when using non-blocking I/O, the
8610 function may not be aware that a handshake has actually taken
8611 place, thus preventing a new session from being added to the
8614 To avoid this problem, we now set s->new_session to 2 instead of
8615 using a local variable.
8616 [Lutz Jaenicke, Bodo Moeller]
8618 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8619 if the SSL_R_LENGTH_MISMATCH error is detected.
8620 [Geoff Thorpe, Bodo Moeller]
8622 *) New 'shared_ldflag' column in Configure platform table.
8625 *) Fix EVP_CIPHER_mode macro.
8626 ["Dan S. Camper" <dan@bti.net>]
8628 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8629 type, we must throw them away by setting rr->length to 0.
8630 [D P Chang <dpc@qualys.com>]
8632 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8634 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8635 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8636 worked incorrectly for those cases where range = 10..._2 and
8637 3*range is two bits longer than range.)
8640 *) Only add signing time to PKCS7 structures if it is not already
8644 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8645 OBJ_ld_ce should be OBJ_id_ce.
8646 Also some ip-pda OIDs in crypto/objects/objects.txt were
8647 incorrect (cf. RFC 3039).
8648 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8650 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8651 returns early because it has nothing to do.
8652 [Andy Schneider <andy.schneider@bjss.co.uk>]
8654 *) [In 0.9.6c-engine release:]
8655 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8656 [Andy Schneider <andy.schneider@bjss.co.uk>]
8658 *) [In 0.9.6c-engine release:]
8659 Add support for Cryptographic Appliance's keyserver technology.
8660 (Use engine 'keyclient')
8661 [Cryptographic Appliances and Geoff Thorpe]
8663 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8664 is called via tools/c89.sh because arguments have to be
8665 rearranged (all '-L' options must appear before the first object
8667 [Richard Shapiro <rshapiro@abinitio.com>]
8669 *) [In 0.9.6c-engine release:]
8670 Add support for Broadcom crypto accelerator cards, backported
8672 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8674 *) [In 0.9.6c-engine release:]
8675 Add support for SureWare crypto accelerator cards from
8676 Baltimore Technologies. (Use engine 'sureware')
8677 [Baltimore Technologies and Mark Cox]
8679 *) [In 0.9.6c-engine release:]
8680 Add support for crypto accelerator cards from Accelerated
8681 Encryption Processing, www.aep.ie. (Use engine 'aep')
8682 [AEP Inc. and Mark Cox]
8684 *) Add a configuration entry for gcc on UnixWare.
8685 [Gary Benson <gbenson@redhat.com>]
8687 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8688 messages are stored in a single piece (fixed-length part and
8689 variable-length part combined) and fix various bugs found on the way.
8692 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8693 instead. BIO_gethostbyname() does not know what timeouts are
8694 appropriate, so entries would stay in cache even when they have
8696 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8698 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8699 faced with a pathologically small ClientHello fragment that does
8700 not contain client_version: Instead of aborting with an error,
8701 simply choose the highest available protocol version (i.e.,
8702 TLS 1.0 unless it is disabled). In practice, ClientHello
8703 messages are never sent like this, but this change gives us
8704 strictly correct behaviour at least for TLS.
8707 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8708 never resets s->method to s->ctx->method when called from within
8709 one of the SSL handshake functions.
8710 [Bodo Moeller; problem pointed out by Niko Baric]
8712 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8713 (sent using the client's version number) if client_version is
8714 smaller than the protocol version in use. Also change
8715 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8716 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8717 the client will at least see that alert.
8720 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8724 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8725 client receives HelloRequest while in a handshake.
8726 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8728 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8729 should end in 'break', not 'goto end' which circumvents various
8730 cleanups done in state SSL_ST_OK. But session related stuff
8731 must be disabled for SSL_ST_OK in the case that we just sent a
8734 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8735 before just sending a HelloRequest.
8736 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8738 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8739 reveal whether illegal block cipher padding was found or a MAC
8740 verification error occurred. (Neither SSLerr() codes nor alerts
8741 are directly visible to potential attackers, but the information
8742 may leak via logfiles.)
8744 Similar changes are not required for the SSL 2.0 implementation
8745 because the number of padding bytes is sent in clear for SSL 2.0,
8746 and the extra bytes are just ignored. However ssl/s2_pkt.c
8747 failed to verify that the purported number of padding bytes is in
8751 *) Add OpenUNIX-8 support including shared libraries
8752 (Boyd Lynn Gerber <gerberb@zenez.com>).
8755 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8756 'wristwatch attack' using huge encoding parameters (cf.
8757 James H. Manger's CRYPTO 2001 paper). Note that the
8758 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8759 encoding parameters and hence was not vulnerable.
8762 *) BN_sqr() bug fix.
8763 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8765 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8766 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8767 followed by modular reduction.
8768 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8770 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8771 equivalent based on BN_pseudo_rand() instead of BN_rand().
8774 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8775 This function was broken, as the check for a new client hello message
8776 to handle SGC did not allow these large messages.
8777 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8780 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8783 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8784 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8787 *) Rework the configuration and shared library support for Tru64 Unix.
8788 The configuration part makes use of modern compiler features and
8789 still retains old compiler behavior for those that run older versions
8790 of the OS. The shared library support part includes a variant that
8791 uses the RPATH feature, and is available through the special
8792 configuration target "alpha-cc-rpath", which will never be selected
8794 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8796 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8797 with the same message size as in ssl3_get_certificate_request().
8798 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8799 messages might inadvertently be reject as too long.
8800 [Petr Lampa <lampa@fee.vutbr.cz>]
8802 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8805 *) Modified SSL library such that the verify_callback that has been set
8806 specificly for an SSL object with SSL_set_verify() is actually being
8807 used. Before the change, a verify_callback set with this function was
8808 ignored and the verify_callback() set in the SSL_CTX at the time of
8809 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8810 to allow the necessary settings.
8813 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8814 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8815 done automatically (in contradiction to the requirements of the C
8816 standard). This made problems when used from OpenSSH.
8819 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8820 dh->length and always used
8822 BN_rand_range(priv_key, dh->p).
8824 BN_rand_range() is not necessary for Diffie-Hellman, and this
8825 specific range makes Diffie-Hellman unnecessarily inefficient if
8826 dh->length (recommended exponent length) is much smaller than the
8827 length of dh->p. We could use BN_rand_range() if the order of
8828 the subgroup was stored in the DH structure, but we only have
8833 BN_rand(priv_key, l, ...)
8835 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8841 RSA_eay_public_encrypt
8842 RSA_eay_private_decrypt
8843 RSA_eay_private_encrypt (signing)
8844 RSA_eay_public_decrypt (signature verification)
8846 (default implementations for RSA_public_encrypt,
8847 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8848 always reject numbers >= n.
8851 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8852 to synchronize access to 'locking_thread'. This is necessary on
8853 systems where access to 'locking_thread' (an 'unsigned long'
8854 variable) is not atomic.
8857 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8858 *before* setting the 'crypto_lock_rand' flag. The previous code had
8859 a race condition if 0 is a valid thread ID.
8860 [Travis Vitek <vitek@roguewave.com>]
8862 *) Add support for shared libraries under Irix.
8863 [Albert Chin-A-Young <china@thewrittenword.com>]
8865 *) Add configuration option to build on Linux on both big-endian and
8867 [Ralf Baechle <ralf@uni-koblenz.de>]
8869 *) Add the possibility to create shared libraries on HP-UX.
8872 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8874 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8875 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8876 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8877 PRNG state recovery was possible based on the output of
8878 one PRNG request appropriately sized to gain knowledge on
8879 'md' followed by enough consecutive 1-byte PRNG requests
8880 to traverse all of 'state'.
8882 1. When updating 'md_local' (the current thread's copy of 'md')
8883 during PRNG output generation, hash all of the previous
8884 'md_local' value, not just the half used for PRNG output.
8886 2. Make the number of bytes from 'state' included into the hash
8887 independent from the number of PRNG bytes requested.
8889 The first measure alone would be sufficient to avoid
8890 Markku-Juhani's attack. (Actually it had never occurred
8891 to me that the half of 'md_local' used for chaining was the
8892 half from which PRNG output bytes were taken -- I had always
8893 assumed that the secret half would be used.) The second
8894 measure makes sure that additional data from 'state' is never
8895 mixed into 'md_local' in small portions; this heuristically
8896 further strengthens the PRNG.
8899 *) Fix crypto/bn/asm/mips3.s.
8902 *) When only the key is given to "enc", the IV is undefined. Print out
8903 an error message in this case.
8906 *) Handle special case when X509_NAME is empty in X509 printing routines.
8909 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8910 positive and less than q.
8913 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8914 used: it isn't thread safe and the add_lock_callback should handle
8916 [Paul Rose <Paul.Rose@bridge.com>]
8918 *) Verify that incoming data obeys the block size in
8919 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8923 [Ulf Möller, Bodo Möller]
8925 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8926 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8927 when fixing the server behaviour for backwards-compatible 'client
8928 hello' messages. (Note that the attack is impractical against
8929 SSL 3.0 and TLS 1.0 anyway because length and version checking
8930 means that the probability of guessing a valid ciphertext is
8931 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8934 Before 0.9.5, the countermeasure (hide the error by generating a
8935 random 'decryption result') did not work properly because
8936 ERR_clear_error() was missing, meaning that SSL_get_error() would
8937 detect the supposedly ignored error.
8939 Both problems are now fixed.
8942 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8943 (previously it was 1024).
8946 *) Fix for compatibility mode trust settings: ignore trust settings
8947 unless some valid trust or reject settings are present.
8950 *) Fix for blowfish EVP: its a variable length cipher.
8953 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8954 parameters in DSA public key structures and return an error in the
8955 DSA routines if parameters are absent.
8958 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8959 in the current directory if neither $RANDFILE nor $HOME was set.
8960 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8961 caused some confusion to Windows users who haven't defined $HOME.
8962 Thus RAND_file_name() is changed again: e_os.h can define a
8963 DEFAULT_HOME, which will be used if $HOME is not set.
8964 For Windows, we use "C:"; on other platforms, we still require
8965 environment variables.
8967 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8968 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8969 having multiple threads call RAND_poll() concurrently.
8972 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8973 combination of a flag and a thread ID variable.
8974 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8975 flag), *other* threads can enter ssleay_add_bytes without obeying
8976 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8977 that they do not hold after the first thread unsets add_do_not_lock).
8980 *) Change bctest again: '-x' expressions are not available in all
8984 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8986 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8987 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8989 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8990 the default extension for executables, if any. Also, make the perl
8991 scripts that use symlink() to test if it really exists and use "cp"
8992 if it doesn't. All this made OpenSSL compilable and installable in
8996 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8997 If SEQUENCE is length is indefinite just set c->slen to the total
8998 amount of data available.
8999 [Steve Henson, reported by shige@FreeBSD.org]
9000 [This change does not apply to 0.9.7.]
9002 *) Change bctest to avoid here-documents inside command substitution
9003 (workaround for FreeBSD /bin/sh bug).
9004 For compatibility with Ultrix, avoid shell functions (introduced
9005 in the bctest version that searches along $PATH).
9008 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9009 with des_encrypt() defined on some operating systems, like Solaris
9013 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9014 On the Importance of Eliminating Errors in Cryptographic
9015 Computations, J. Cryptology 14 (2001) 2, 101-119,
9016 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9019 *) MIPS assembler BIGNUM division bug fix.
9022 *) Disabled incorrect Alpha assembler code.
9025 *) Fix PKCS#7 decode routines so they correctly update the length
9026 after reading an EOC for the EXPLICIT tag.
9028 [This change does not apply to 0.9.7.]
9030 *) Fix bug in PKCS#12 key generation routines. This was triggered
9031 if a 3DES key was generated with a 0 initial byte. Include
9032 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9033 (but broken) behaviour.
9036 *) Enhance bctest to search for a working bc along $PATH and print
9038 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9040 *) Fix memory leaks in err.c: free err_data string if necessary;
9041 don't write to the wrong index in ERR_set_error_data.
9044 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9048 *) Replace rdtsc with _emit statements for VC++ version 5.
9049 [Jeremy Cooper <jeremy@baymoo.org>]
9051 *) Make it possible to reuse SSLv2 sessions.
9054 *) In copy_email() check for >= 0 as a return value for
9055 X509_NAME_get_index_by_NID() since 0 is a valid index.
9056 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9058 *) Avoid coredump with unsupported or invalid public keys by checking if
9059 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9060 PKCS7_verify() fails with non detached data.
9063 *) Don't use getenv in library functions when run as setuid/setgid.
9064 New function OPENSSL_issetugid().
9067 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9068 due to incorrect handling of multi-threading:
9070 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9072 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9074 3. Count how many times MemCheck_off() has been called so that
9075 nested use can be treated correctly. This also avoids
9076 inband-signalling in the previous code (which relied on the
9077 assumption that thread ID 0 is impossible).
9080 *) Add "-rand" option also to s_client and s_server.
9083 *) Fix CPU detection on Irix 6.x.
9084 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9085 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9087 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9090 [This change does not apply to 0.9.7.]
9092 *) Use the cached encoding of an X509_NAME structure rather than
9093 copying it. This is apparently the reason for the libsafe "errors"
9094 but the code is actually correct.
9097 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9098 Bleichenbacher's DSA attack.
9099 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9100 to be set and top=0 forces the highest bit to be set; top=-1 is new
9101 and leaves the highest bit random.
9102 [Ulf Moeller, Bodo Moeller]
9104 *) In the NCONF_...-based implementations for CONF_... queries
9105 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9106 a temporary CONF structure with the data component set to NULL
9107 (which gives segmentation faults in lh_retrieve).
9108 Instead, use NULL for the CONF pointer in CONF_get_string and
9109 CONF_get_number (which may use environment variables) and directly
9110 return NULL from CONF_get_section.
9113 *) Fix potential buffer overrun for EBCDIC.
9116 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9117 keyUsage if basicConstraints absent for a CA.
9120 *) Make SMIME_write_PKCS7() write mail header values with a format that
9121 is more generally accepted (no spaces before the semicolon), since
9122 some programs can't parse those values properly otherwise. Also make
9123 sure BIO's that break lines after each write do not create invalid
9127 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9128 macros previously used would not encode an empty SEQUENCE OF
9129 and break the signature.
9131 [This change does not apply to 0.9.7.]
9133 *) Zero the premaster secret after deriving the master secret in
9137 *) Add some EVP_add_digest_alias registrations (as found in
9138 OpenSSL_add_all_digests()) to SSL_library_init()
9139 aka OpenSSL_add_ssl_algorithms(). This provides improved
9140 compatibility with peers using X.509 certificates
9141 with unconventional AlgorithmIdentifier OIDs.
9144 *) Fix for Irix with NO_ASM.
9145 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9147 *) ./config script fixes.
9148 [Ulf Moeller, Richard Levitte]
9150 *) Fix 'openssl passwd -1'.
9153 *) Change PKCS12_key_gen_asc() so it can cope with non null
9154 terminated strings whose length is passed in the passlen
9155 parameter, for example from PEM callbacks. This was done
9156 by adding an extra length parameter to asc2uni().
9157 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9159 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9160 call failed, free the DSA structure.
9163 *) Fix to uni2asc() to cope with zero length Unicode strings.
9164 These are present in some PKCS#12 files.
9167 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9168 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9169 when writing a 32767 byte record.
9170 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9172 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9173 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9175 (RSA objects have a reference count access to which is protected
9176 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9177 so they are meant to be shared between threads.)
9178 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9179 "Reddie, Steven" <Steven.Reddie@ca.com>]
9181 *) Fix a deadlock in CRYPTO_mem_leaks().
9184 *) Use better test patterns in bntest.
9187 *) rand_win.c fix for Borland C.
9190 *) BN_rshift bugfix for n == 0.
9193 *) Add a 'bctest' script that checks for some known 'bc' bugs
9194 so that 'make test' does not abort just because 'bc' is broken.
9197 *) Store verify_result within SSL_SESSION also for client side to
9198 avoid potential security hole. (Re-used sessions on the client side
9199 always resulted in verify_result==X509_V_OK, not using the original
9200 result of the server certificate verification.)
9203 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9204 SSL3_RT_APPLICATION_DATA, return 0.
9205 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9209 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9210 releases, have been re-implemented by renaming the previous
9211 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9212 and ssl3_read_internal, respectively, and adding 'peek' parameters
9213 to them. The new ssl[23]_{read,peek} functions are calls to
9214 ssl[23]_read_internal with the 'peek' flag set appropriately.
9215 A 'peek' parameter has also been added to ssl3_read_bytes, which
9216 does the actual work for ssl3_read_internal.
9219 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9220 the method-specific "init()" handler. Also clean up ex_data after
9221 calling the method-specific "finish()" handler. Previously, this was
9222 happening the other way round.
9225 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9226 The previous value, 12, was not always sufficient for BN_mod_exp().
9229 *) Make sure that shared libraries get the internal name engine with
9230 the full version number and not just 0. This should mark the
9231 shared libraries as not backward compatible. Of course, this should
9232 be changed again when we can guarantee backward binary compatibility.
9235 *) Fix typo in get_cert_by_subject() in by_dir.c
9236 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9238 *) Rework the system to generate shared libraries:
9240 - Make note of the expected extension for the shared libraries and
9241 if there is a need for symbolic links from for example libcrypto.so.0
9242 to libcrypto.so.0.9.7. There is extended info in Configure for
9245 - Make as few rebuilds of the shared libraries as possible.
9247 - Still avoid linking the OpenSSL programs with the shared libraries.
9249 - When installing, install the shared libraries separately from the
9253 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9255 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9256 and not in SSL_clear because the latter is also used by the
9257 accept/connect functions; previously, the settings made by
9258 SSL_set_read_ahead would be lost during the handshake.
9259 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9261 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9262 Previously, it would create entries for disabled algorithms no
9266 *) Added several new manual pages for SSL_* function.
9269 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9271 *) In ssl23_get_client_hello, generate an error message when faced
9272 with an initial SSL 3.0/TLS record that is too small to contain the
9273 first two bytes of the ClientHello message, i.e. client_version.
9274 (Note that this is a pathologic case that probably has never happened
9275 in real life.) The previous approach was to use the version number
9276 from the record header as a substitute; but our protocol choice
9277 should not depend on that one because it is not authenticated
9278 by the Finished messages.
9281 *) More robust randomness gathering functions for Windows.
9282 [Jeffrey Altman <jaltman@columbia.edu>]
9284 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9285 not set then we don't setup the error code for issuer check errors
9286 to avoid possibly overwriting other errors which the callback does
9287 handle. If an application does set the flag then we assume it knows
9288 what it is doing and can handle the new informational codes
9292 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9293 a general "ANY" type, as such it should be able to decode anything
9294 including tagged types. However it didn't check the class so it would
9295 wrongly interpret tagged types in the same way as their universal
9296 counterpart and unknown types were just rejected. Changed so that the
9297 tagged and unknown types are handled in the same way as a SEQUENCE:
9298 that is the encoding is stored intact. There is also a new type
9299 "V_ASN1_OTHER" which is used when the class is not universal, in this
9300 case we have no idea what the actual type is so we just lump them all
9304 *) On VMS, stdout may very well lead to a file that is written to
9305 in a record-oriented fashion. That means that every write() will
9306 write a separate record, which will be read separately by the
9307 programs trying to read from it. This can be very confusing.
9309 The solution is to put a BIO filter in the way that will buffer
9310 text until a linefeed is reached, and then write everything a
9311 line at a time, so every record written will be an actual line,
9312 not chunks of lines and not (usually doesn't happen, but I've
9313 seen it once) several lines in one record. BIO_f_linebuffer() is
9316 Currently, it's a VMS-only method, because that's where it has
9317 been tested well enough.
9320 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9321 it can return incorrect results.
9322 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9323 but it was in 0.9.6-beta[12].)
9326 *) Disable the check for content being present when verifying detached
9327 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9328 include zero length content when signing messages.
9331 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9332 BIO_ctrl (for BIO pairs).
9335 *) Add DSO method for VMS.
9338 *) Bug fix: Montgomery multiplication could produce results with the
9342 *) Add RPM specification openssl.spec and modify it to build three
9343 packages. The default package contains applications, application
9344 documentation and run-time libraries. The devel package contains
9345 include files, static libraries and function documentation. The
9346 doc package contains the contents of the doc directory. The original
9347 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9350 *) Add a large number of documentation files for many SSL routines.
9351 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9353 *) Add a configuration entry for Sony News 4.
9354 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9356 *) Don't set the two most significant bits to one when generating a
9357 random number < q in the DSA library.
9360 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9361 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9362 the underlying transport is blocking) if a handshake took place.
9363 (The default behaviour is needed by applications such as s_client
9364 and s_server that use select() to determine when to use SSL_read;
9365 but for applications that know in advance when to expect data, it
9366 just makes things more complicated.)
9369 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9373 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9374 work better on such systems.
9375 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9377 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9378 Update PKCS12_parse() so it copies the friendlyName and the
9379 keyid to the certificates aux info.
9382 *) Fix bug in PKCS7_verify() which caused an infinite loop
9383 if there was more than one signature.
9384 [Sven Uszpelkat <su@celocom.de>]
9386 *) Major change in util/mkdef.pl to include extra information
9387 about each symbol, as well as presenting variables as well
9388 as functions. This change means that there's n more need
9389 to rebuild the .num files when some algorithms are excluded.
9392 *) Allow the verify time to be set by an application,
9393 rather than always using the current time.
9396 *) Phase 2 verify code reorganisation. The certificate
9397 verify code now looks up an issuer certificate by a
9398 number of criteria: subject name, authority key id
9399 and key usage. It also verifies self signed certificates
9400 by the same criteria. The main comparison function is
9401 X509_check_issued() which performs these checks.
9403 Lot of changes were necessary in order to support this
9404 without completely rewriting the lookup code.
9406 Authority and subject key identifier are now cached.
9408 The LHASH 'certs' is X509_STORE has now been replaced
9409 by a STACK_OF(X509_OBJECT). This is mainly because an
9410 LHASH can't store or retrieve multiple objects with
9411 the same hash value.
9413 As a result various functions (which were all internal
9414 use only) have changed to handle the new X509_STORE
9415 structure. This will break anything that messed round
9416 with X509_STORE internally.
9418 The functions X509_STORE_add_cert() now checks for an
9419 exact match, rather than just subject name.
9421 The X509_STORE API doesn't directly support the retrieval
9422 of multiple certificates matching a given criteria, however
9423 this can be worked round by performing a lookup first
9424 (which will fill the cache with candidate certificates)
9425 and then examining the cache for matches. This is probably
9426 the best we can do without throwing out X509_LOOKUP
9427 entirely (maybe later...).
9429 The X509_VERIFY_CTX structure has been enhanced considerably.
9431 All certificate lookup operations now go via a get_issuer()
9432 callback. Although this currently uses an X509_STORE it
9433 can be replaced by custom lookups. This is a simple way
9434 to bypass the X509_STORE hackery necessary to make this
9435 work and makes it possible to use more efficient techniques
9436 in future. A very simple version which uses a simple
9437 STACK for its trusted certificate store is also provided
9438 using X509_STORE_CTX_trusted_stack().
9440 The verify_cb() and verify() callbacks now have equivalents
9441 in the X509_STORE_CTX structure.
9443 X509_STORE_CTX also has a 'flags' field which can be used
9444 to customise the verify behaviour.
9447 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9448 excludes S/MIME capabilities.
9451 *) When a certificate request is read in keep a copy of the
9452 original encoding of the signed data and use it when outputting
9453 again. Signatures then use the original encoding rather than
9454 a decoded, encoded version which may cause problems if the
9455 request is improperly encoded.
9458 *) For consistency with other BIO_puts implementations, call
9459 buffer_write(b, ...) directly in buffer_puts instead of calling
9462 In BIO_puts, increment b->num_write as in BIO_write.
9463 [Peter.Sylvester@EdelWeb.fr]
9465 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9466 BN_zero, we may not return a BIGNUM with an array consisting of
9470 *) Avoid calling abort() from within the library when problems are
9471 detected, except if preprocessor symbols have been defined
9472 (such as REF_CHECK, BN_DEBUG etc.).
9475 *) New openssl application 'rsautl'. This utility can be
9476 used for low level RSA operations. DER public key
9477 BIO/fp routines also added.
9480 *) New Configure entry and patches for compiling on QNX 4.
9481 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9483 *) A demo state-machine implementation was sponsored by
9484 Nuron (http://www.nuron.com/) and is now available in
9485 demos/state_machine.
9488 *) New options added to the 'dgst' utility for signature
9489 generation and verification.
9492 *) Unrecognized PKCS#7 content types are now handled via a
9493 catch all ASN1_TYPE structure. This allows unsupported
9494 types to be stored as a "blob" and an application can
9495 encode and decode it manually.
9498 *) Fix various signed/unsigned issues to make a_strex.c
9500 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9502 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9503 length if passed a buffer. ASN1_INTEGER_to_BN failed
9504 if passed a NULL BN and its argument was negative.
9505 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9507 *) Modification to PKCS#7 encoding routines to output definite
9508 length encoding. Since currently the whole structures are in
9509 memory there's not real point in using indefinite length
9510 constructed encoding. However if OpenSSL is compiled with
9511 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9514 *) Added BIO_vprintf() and BIO_vsnprintf().
9517 *) Added more prefixes to parse for in the the strings written
9518 through a logging bio, to cover all the levels that are available
9519 through syslog. The prefixes are now:
9521 PANIC, EMERG, EMR => LOG_EMERG
9522 ALERT, ALR => LOG_ALERT
9523 CRIT, CRI => LOG_CRIT
9524 ERROR, ERR => LOG_ERR
9525 WARNING, WARN, WAR => LOG_WARNING
9526 NOTICE, NOTE, NOT => LOG_NOTICE
9527 INFO, INF => LOG_INFO
9528 DEBUG, DBG => LOG_DEBUG
9530 and as before, if none of those prefixes are present at the
9531 beginning of the string, LOG_ERR is chosen.
9533 On Win32, the LOG_* levels are mapped according to this:
9535 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9536 LOG_WARNING => EVENTLOG_WARNING_TYPE
9537 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9541 *) Made it possible to reconfigure with just the configuration
9542 argument "reconf" or "reconfigure". The command line arguments
9543 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9544 and are retrieved from there when reconfiguring.
9548 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9550 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9553 *) The obj_dat.pl script was messing up the sorting of object
9554 names. The reason was that it compared the quoted version
9555 of strings as a result "OCSP" > "OCSP Signing" because
9556 " > SPACE. Changed script to store unquoted versions of
9557 names and add quotes on output. It was also omitting some
9558 names from the lookup table if they were given a default
9559 value (that is if SN is missing it is given the same
9560 value as LN and vice versa), these are now added on the
9561 grounds that if an object has a name we should be able to
9562 look it up. Finally added warning output when duplicate
9563 short or long names are found.
9566 *) Changes needed for Tandem NSK.
9567 [Scott Uroff <scott@xypro.com>]
9569 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9570 RSA_padding_check_SSLv23(), special padding was never detected
9571 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9572 version rollback attacks was not effective.
9574 In s23_clnt.c, don't use special rollback-attack detection padding
9575 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9576 client; similarly, in s23_srvr.c, don't do the rollback check if
9577 SSL 2.0 is the only protocol enabled in the server.
9580 *) Make it possible to get hexdumps of unprintable data with 'openssl
9581 asn1parse'. By implication, the functions ASN1_parse_dump() and
9582 BIO_dump_indent() are added.
9585 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9586 these print out strings and name structures based on various
9587 flags including RFC2253 support and proper handling of
9588 multibyte characters. Added options to the 'x509' utility
9589 to allow the various flags to be set.
9592 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9593 Also change the functions X509_cmp_current_time() and
9594 X509_gmtime_adj() work with an ASN1_TIME structure,
9595 this will enable certificates using GeneralizedTime in validity
9596 dates to be checked.
9599 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9600 negative public key encodings) on by default,
9601 NO_NEG_PUBKEY_BUG can be set to disable it.
9604 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9605 content octets. An i2c_ASN1_OBJECT is unnecessary because
9606 the encoding can be trivially obtained from the structure.
9609 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9610 not read locks (CRYPTO_r_[un]lock).
9613 *) A first attempt at creating official support for shared
9614 libraries through configuration. I've kept it so the
9615 default is static libraries only, and the OpenSSL programs
9616 are always statically linked for now, but there are
9617 preparations for dynamic linking in place.
9618 This has been tested on Linux and Tru64.
9621 *) Randomness polling function for Win9x, as described in:
9622 Peter Gutmann, Software Generation of Practically Strong
9626 *) Fix so PRNG is seeded in req if using an already existing
9630 *) New options to smime application. -inform and -outform
9631 allow alternative formats for the S/MIME message including
9632 PEM and DER. The -content option allows the content to be
9633 specified separately. This should allow things like Netscape
9634 form signing output easier to verify.
9637 *) Fix the ASN1 encoding of tags using the 'long form'.
9640 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9641 STRING types. These convert content octets to and from the
9642 underlying type. The actual tag and length octets are
9643 already assumed to have been read in and checked. These
9644 are needed because all other string types have virtually
9645 identical handling apart from the tag. By having versions
9646 of the ASN1 functions that just operate on content octets
9647 IMPLICIT tagging can be handled properly. It also allows
9648 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9649 and ASN1_INTEGER are identical apart from the tag.
9652 *) Change the handling of OID objects as follows:
9654 - New object identifiers are inserted in objects.txt, following
9655 the syntax given in objects.README.
9656 - objects.pl is used to process obj_mac.num and create a new
9658 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9661 This is currently kind of a hack, and the perl code in objects.pl
9662 isn't very elegant, but it works as I intended. The simplest way
9663 to check that it worked correctly is to look in obj_dat.h and
9664 check the array nid_objs and make sure the objects haven't moved
9665 around (this is important!). Additions are OK, as well as
9666 consistent name changes.
9669 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9672 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9673 The given file adds to whatever has already been seeded into the
9674 random pool through the RANDFILE configuration file option or
9675 environment variable, or the default random state file.
9678 *) mkstack.pl now sorts each macro group into lexical order.
9679 Previously the output order depended on the order the files
9680 appeared in the directory, resulting in needless rewriting
9684 *) Patches to make OpenSSL compile under Win32 again. Mostly
9685 work arounds for the VC++ problem that it treats func() as
9686 func(void). Also stripped out the parts of mkdef.pl that
9687 added extra typesafe functions: these no longer exist.
9690 *) Reorganisation of the stack code. The macros are now all
9691 collected in safestack.h . Each macro is defined in terms of
9692 a "stack macro" of the form SKM_<name>(type, a, b). The
9693 DEBUG_SAFESTACK is now handled in terms of function casts,
9694 this has the advantage of retaining type safety without the
9695 use of additional functions. If DEBUG_SAFESTACK is not defined
9696 then the non typesafe macros are used instead. Also modified the
9697 mkstack.pl script to handle the new form. Needs testing to see
9698 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9699 the default if no major problems. Similar behaviour for ASN1_SET_OF
9700 and PKCS12_STACK_OF.
9703 *) When some versions of IIS use the 'NET' form of private key the
9704 key derivation algorithm is different. Normally MD5(password) is
9705 used as a 128 bit RC4 key. In the modified case
9706 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9707 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9708 as the old Netscape_RSA functions except they have an additional
9709 'sgckey' parameter which uses the modified algorithm. Also added
9710 an -sgckey command line option to the rsa utility. Thanks to
9711 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9712 algorithm to openssl-dev.
9715 *) The evp_local.h macros were using 'c.##kname' which resulted in
9716 invalid expansion on some systems (SCO 5.0.5 for example).
9717 Corrected to 'c.kname'.
9718 [Phillip Porch <root@theporch.com>]
9720 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9721 a STACK of email addresses from a certificate or request, these look
9722 in the subject name and the subject alternative name extensions and
9723 omit any duplicate addresses.
9726 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9727 This makes DSA verification about 2 % faster.
9730 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9731 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9732 plus overhead for 1024 bit moduli).
9733 This makes exponentiations about 0.5 % faster for 1024 bit
9734 exponents (as measured by "openssl speed rsa2048").
9737 *) Rename memory handling macros to avoid conflicts with other
9739 Malloc => OPENSSL_malloc
9740 Malloc_locked => OPENSSL_malloc_locked
9741 Realloc => OPENSSL_realloc
9742 Free => OPENSSL_free
9745 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9746 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9749 *) CygWin32 support.
9750 [John Jarvie <jjarvie@newsguy.com>]
9752 *) The type-safe stack code has been rejigged. It is now only compiled
9753 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9754 by default all type-specific stack functions are "#define"d back to
9755 standard stack functions. This results in more streamlined output
9756 but retains the type-safety checking possibilities of the original
9760 *) The STACK code has been cleaned up, and certain type declarations
9761 that didn't make a lot of sense have been brought in line. This has
9762 also involved a cleanup of sorts in safestack.h to more correctly
9763 map type-safe stack functions onto their plain stack counterparts.
9764 This work has also resulted in a variety of "const"ifications of
9765 lots of the code, especially "_cmp" operations which should normally
9766 be prototyped with "const" parameters anyway.
9769 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9770 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9771 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9772 where all of 'md' is used each time the PRNG is used, but 'state'
9773 is used only indexed by a cyclic counter. As entropy may not be
9774 well distributed from the beginning, 'md' is important as a
9775 chaining variable. However, the output function chains only half
9776 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9777 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9778 in all of 'state' being rewritten, with the new values depending
9779 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9782 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9783 the handshake is continued after ssl_verify_cert_chain();
9784 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9785 can lead to 'unexplainable' connection aborts later.
9786 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9788 *) Major EVP API cipher revision.
9789 Add hooks for extra EVP features. This allows various cipher
9790 parameters to be set in the EVP interface. Support added for variable
9791 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9792 setting of RC2 and RC5 parameters.
9794 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9797 Remove lots of duplicated code from the EVP library. For example *every*
9798 cipher init() function handles the 'iv' in the same way according to the
9799 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9800 for CFB and OFB modes they zero ctx->num.
9802 New functionality allows removal of S/MIME code RC2 hack.
9804 Most of the routines have the same form and so can be declared in terms
9807 By shifting this to the top level EVP_CipherInit() it can be removed from
9808 all individual ciphers. If the cipher wants to handle IVs or keys
9809 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9812 Change lots of functions like EVP_EncryptUpdate() to now return a
9813 value: although software versions of the algorithms cannot fail
9814 any installed hardware versions can.
9817 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9818 this option is set, tolerate broken clients that send the negotiated
9819 protocol version number instead of the requested protocol version
9823 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9824 i.e. non-zero for export ciphersuites, zero otherwise.
9825 Previous versions had this flag inverted, inconsistent with
9826 rsa_tmp_cb (..._TMP_RSA_CB).
9827 [Bodo Moeller; problem reported by Amit Chopra]
9829 *) Add missing DSA library text string. Work around for some IIS
9830 key files with invalid SEQUENCE encoding.
9833 *) Add a document (doc/standards.txt) that list all kinds of standards
9834 and so on that are implemented in OpenSSL.
9837 *) Enhance c_rehash script. Old version would mishandle certificates
9838 with the same subject name hash and wouldn't handle CRLs at all.
9839 Added -fingerprint option to crl utility, to support new c_rehash
9843 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9846 *) Fix for SSL server purpose checking. Server checking was
9847 rejecting certificates which had extended key usage present
9848 but no ssl client purpose.
9849 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9851 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9852 is a little unclear about how a blank password is handled.
9853 Since the password in encoded as a BMPString with terminating
9854 double NULL a zero length password would end up as just the
9855 double NULL. However no password at all is different and is
9856 handled differently in the PKCS#12 key generation code. NS
9857 treats a blank password as zero length. MSIE treats it as no
9858 password on export: but it will try both on import. We now do
9859 the same: PKCS12_parse() tries zero length and no password if
9860 the password is set to "" or NULL (NULL is now a valid password:
9861 it wasn't before) as does the pkcs12 application.
9864 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9865 perror when PEM_read_bio_X509_REQ fails, the error message must
9866 be obtained from the error queue.
9869 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9870 it in ERR_remove_state if appropriate, and change ERR_get_state
9871 accordingly to avoid race conditions (this is necessary because
9872 thread_hash is no longer constant once set).
9875 *) Bugfix for linux-elf makefile.one.
9878 *) RSA_get_default_method() will now cause a default
9879 RSA_METHOD to be chosen if one doesn't exist already.
9880 Previously this was only set during a call to RSA_new()
9881 or RSA_new_method(NULL) meaning it was possible for
9882 RSA_get_default_method() to return NULL.
9885 *) Added native name translation to the existing DSO code
9886 that will convert (if the flag to do so is set) filenames
9887 that are sufficiently small and have no path information
9888 into a canonical native form. Eg. "blah" converted to
9889 "libblah.so" or "blah.dll" etc.
9892 *) New function ERR_error_string_n(e, buf, len) which is like
9893 ERR_error_string(e, buf), but writes at most 'len' bytes
9894 including the 0 terminator. For ERR_error_string_n, 'buf'
9896 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9898 *) CONF library reworked to become more general. A new CONF
9899 configuration file reader "class" is implemented as well as a
9900 new functions (NCONF_*, for "New CONF") to handle it. The now
9901 old CONF_* functions are still there, but are reimplemented to
9902 work in terms of the new functions. Also, a set of functions
9903 to handle the internal storage of the configuration data is
9904 provided to make it easier to write new configuration file
9905 reader "classes" (I can definitely see something reading a
9906 configuration file in XML format, for example), called _CONF_*,
9907 or "the configuration storage API"...
9909 The new configuration file reading functions are:
9911 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9912 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9914 NCONF_default, NCONF_WIN32
9916 NCONF_dump_fp, NCONF_dump_bio
9918 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9919 NCONF_new creates a new CONF object. This works in the same way
9920 as other interfaces in OpenSSL, like the BIO interface.
9921 NCONF_dump_* dump the internal storage of the configuration file,
9922 which is useful for debugging. All other functions take the same
9923 arguments as the old CONF_* functions wth the exception of the
9924 first that must be a `CONF *' instead of a `LHASH *'.
9926 To make it easer to use the new classes with the old CONF_* functions,
9927 the function CONF_set_default_method is provided.
9930 *) Add '-tls1' option to 'openssl ciphers', which was already
9931 mentioned in the documentation but had not been implemented.
9932 (This option is not yet really useful because even the additional
9933 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9936 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9937 OpenSSL-based applications) load shared libraries and bind to
9938 them in a portable way.
9939 [Geoff Thorpe, with contributions from Richard Levitte]
9941 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9943 *) Make sure _lrotl and _lrotr are only used with MSVC.
9945 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9946 (the default implementation of RAND_status).
9948 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9949 to '-clrext' (= clear extensions), as intended and documented.
9950 [Bodo Moeller; inconsistency pointed out by Michael Attili
9953 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9954 was larger than the MD block size.
9955 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9957 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9958 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9959 using the passed key: if the passed key was a private key the result
9960 of X509_print(), for example, would be to print out all the private key
9964 *) des_quad_cksum() byte order bug fix.
9965 [Ulf Möller, using the problem description in krb4-0.9.7, where
9966 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9968 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9970 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9972 *) For easily testing in shell scripts whether some command
9973 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9974 returns with exit code 0 iff no command of the given name is available.
9975 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9976 the output goes to stdout and nothing is printed to stderr.
9977 Additional arguments are always ignored.
9979 Since for each cipher there is a command of the same name,
9980 the 'no-cipher' compilation switches can be tested this way.
9982 ('openssl no-XXX' is not able to detect pseudo-commands such
9983 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9986 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9989 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9990 is set; it will be thrown away anyway because each handshake creates
9992 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9993 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9994 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9995 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9998 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9999 'Q' and 'R' lose their special meanings (quit/renegotiate).
10000 This is part of what -quiet does; unlike -quiet, -ign_eof
10001 does not suppress any output.
10004 *) Add compatibility options to the purpose and trust code. The
10005 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10006 accepts a certificate or CA, this was the previous behaviour,
10007 with all the associated security issues.
10009 X509_TRUST_COMPAT is the old trust behaviour: only and
10010 automatically trust self signed roots in certificate store. A
10011 new trust setting X509_TRUST_DEFAULT is used to specify that
10012 a purpose has no associated trust setting and it should instead
10013 use the value in the default purpose.
10016 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10017 and fix a memory leak.
10020 *) In util/mkerr.pl (which implements 'make errors'), preserve
10021 reason strings from the previous version of the .c file, as
10022 the default to have only downcase letters (and digits) in
10023 automatically generated reasons codes is not always appropriate.
10026 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10027 using strerror. Previously, ERR_reason_error_string() returned
10028 library names as reason strings for SYSerr; but SYSerr is a special
10029 case where small numbers are errno values, not library numbers.
10032 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10033 converts DSA parameters into DH parameters. (When creating parameters,
10034 DSA_generate_parameters is used.)
10037 *) Include 'length' (recommended exponent length) in C code generated
10038 by 'openssl dhparam -C'.
10041 *) The second argument to set_label in perlasm was already being used
10042 so couldn't be used as a "file scope" flag. Moved to third argument
10046 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10047 instead of RAND_bytes for encryption IVs and salts.
10050 *) Include RAND_status() into RAND_METHOD instead of implementing
10051 it only for md_rand.c Otherwise replacing the PRNG by calling
10052 RAND_set_rand_method would be impossible.
10055 *) Don't let DSA_generate_key() enter an infinite loop if the random
10056 number generation fails.
10059 *) New 'rand' application for creating pseudo-random output.
10062 *) Added configuration support for Linux/IA64
10063 [Rolf Haberrecker <rolf@suse.de>]
10065 *) Assembler module support for Mingw32.
10068 *) Shared library support for HPUX (in shlib/).
10069 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10071 *) Shared library support for Solaris gcc.
10072 [Lutz Behnke <behnke@trustcenter.de>]
10074 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10076 *) PKCS7_encrypt() was adding text MIME headers twice because they
10077 were added manually and by SMIME_crlf_copy().
10080 *) In bntest.c don't call BN_rand with zero bits argument.
10081 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10083 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10084 case was implemented. This caused BN_div_recp() to fail occasionally.
10087 *) Add an optional second argument to the set_label() in the perl
10088 assembly language builder. If this argument exists and is set
10089 to 1 it signals that the assembler should use a symbol whose
10090 scope is the entire file, not just the current function. This
10091 is needed with MASM which uses the format label:: for this scope.
10092 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10094 *) Change the ASN1 types so they are typedefs by default. Before
10095 almost all types were #define'd to ASN1_STRING which was causing
10096 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10100 *) Change names of new functions to the new get1/get0 naming
10101 convention: After 'get1', the caller owns a reference count
10102 and has to call ..._free; 'get0' returns a pointer to some
10103 data structure without incrementing reference counters.
10104 (Some of the existing 'get' functions increment a reference
10105 counter, some don't.)
10106 Similarly, 'set1' and 'add1' functions increase reference
10107 counters or duplicate objects.
10110 *) Allow for the possibility of temp RSA key generation failure:
10111 the code used to assume it always worked and crashed on failure.
10114 *) Fix potential buffer overrun problem in BIO_printf().
10115 [Ulf Möller, using public domain code by Patrick Powell; problem
10116 pointed out by David Sacerdote <das33@cornell.edu>]
10118 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10119 RAND_egd() and RAND_status(). In the command line application,
10120 the EGD socket can be specified like a seed file using RANDFILE
10124 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10125 Some CAs (e.g. Verisign) distribute certificates in this form.
10128 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10129 list to exclude them. This means that no special compilation option
10130 is needed to use anonymous DH: it just needs to be included in the
10134 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10135 EVP_MD_type. The old functionality is available in a new macro called
10136 EVP_MD_md(). Change code that uses it and update docs.
10139 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10140 where the 'void *' argument is replaced by a function pointer argument.
10141 Previously 'void *' was abused to point to functions, which works on
10142 many platforms, but is not correct. As these functions are usually
10143 called by macros defined in OpenSSL header files, most source code
10144 should work without changes.
10147 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10148 sections with information on -D... compiler switches used for
10149 compiling the library so that applications can see them. To enable
10150 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10151 must be defined. E.g.,
10152 #define OPENSSL_ALGORITHM_DEFINES
10153 #include <openssl/opensslconf.h>
10154 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10155 [Richard Levitte, Ulf and Bodo Möller]
10157 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10161 *) Change the 'other' type in certificate aux info to a STACK_OF
10162 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10163 the required ASN1 format: arbitrary types determined by an OID.
10166 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10167 argument to 'req'. This is not because the function is newer or
10168 better than others it just uses the work 'NEW' in the certificate
10169 request header lines. Some software needs this.
10172 *) Reorganise password command line arguments: now passwords can be
10173 obtained from various sources. Delete the PEM_cb function and make
10174 it the default behaviour: i.e. if the callback is NULL and the
10175 usrdata argument is not NULL interpret it as a null terminated pass
10176 phrase. If usrdata and the callback are NULL then the pass phrase
10177 is prompted for as usual.
10180 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10181 the support is automatically enabled. The resulting binaries will
10182 autodetect the card and use it if present.
10183 [Ben Laurie and Compaq Inc.]
10185 *) Work around for Netscape hang bug. This sends certificate request
10186 and server done in one record. Since this is perfectly legal in the
10187 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10188 the bugs/SSLv3 entry for more info.
10191 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10194 *) Add -rand argument to smime and pkcs12 applications and read/write
10198 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10201 *) Add command line password options to the remaining applications.
10204 *) Bug fix for BN_div_recp() for numerators with an even number of
10208 *) More tests in bntest.c, and changed test_bn output.
10211 *) ./config recognizes MacOS X now.
10214 *) Bug fix for BN_div() when the first words of num and divsor are
10215 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10218 *) Add support for various broken PKCS#8 formats, and command line
10219 options to produce them.
10222 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10223 get temporary BIGNUMs from a BN_CTX.
10226 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10230 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10231 include a #define from the old name to the new. The original intent
10232 was that statically linked binaries could for example just call
10233 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10234 link with digests. This never worked because SSLeay_add_all_digests()
10235 and SSLeay_add_all_ciphers() were in the same source file so calling
10236 one would link with the other. They are now in separate source files.
10239 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10242 *) Use a less unusual form of the Miller-Rabin primality test (it used
10243 a binary algorithm for exponentiation integrated into the Miller-Rabin
10244 loop, our standard modexp algorithms are faster).
10247 *) Support for the EBCDIC character set completed.
10248 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10250 *) Source code cleanups: use const where appropriate, eliminate casts,
10251 use void * instead of char * in lhash.
10254 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10255 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10256 this the server could overwrite ephemeral keys that the client
10260 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10261 using 50 iterations of the Rabin-Miller test.
10263 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10264 iterations of the Rabin-Miller test as required by the appendix
10265 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10266 As BN_is_prime_fasttest includes trial division, DSA parameter
10267 generation becomes much faster.
10269 This implies a change for the callback functions in DSA_is_prime
10270 and DSA_generate_parameters: The callback function is called once
10271 for each positive witness in the Rabin-Miller test, not just
10272 occasionally in the inner loop; and the parameters to the
10273 callback function now provide an iteration count for the outer
10274 loop rather than for the current invocation of the inner loop.
10275 DSA_generate_parameters additionally can call the callback
10276 function with an 'iteration count' of -1, meaning that a
10277 candidate has passed the trial division test (when q is generated
10278 from an application-provided seed, trial division is skipped).
10281 *) New function BN_is_prime_fasttest that optionally does trial
10282 division before starting the Rabin-Miller test and has
10283 an additional BN_CTX * argument (whereas BN_is_prime always
10284 has to allocate at least one BN_CTX).
10285 'callback(1, -1, cb_arg)' is called when a number has passed the
10286 trial division stage.
10289 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10293 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10296 *) New function BN_pseudo_rand().
10299 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10300 bignum version of BN_from_montgomery() with the working code from
10301 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10305 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10306 made it impossible to use the same SSL_SESSION data structure in
10307 SSL2 clients in multiple threads.
10310 *) The return value of RAND_load_file() no longer counts bytes obtained
10311 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10312 to seed the PRNG (previously an explicit byte count was required).
10313 [Ulf Möller, Bodo Möller]
10315 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10316 used (char *) instead of (void *) and had casts all over the place.
10319 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10322 *) Retain source code compatibility for BN_prime_checks macro:
10323 BN_is_prime(..., BN_prime_checks, ...) now uses
10324 BN_prime_checks_for_size to determine the appropriate number of
10325 Rabin-Miller iterations.
10328 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10329 DH_CHECK_P_NOT_SAFE_PRIME.
10330 (Check if this is true? OpenPGP calls them "strong".)
10333 *) Merge the functionality of "dh" and "gendh" programs into a new program
10334 "dhparam". The old programs are retained for now but will handle DH keys
10335 (instead of parameters) in future.
10338 *) Make the ciphers, s_server and s_client programs check the return values
10339 when a new cipher list is set.
10342 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10343 ciphers. Before when the 56bit ciphers were enabled the sorting was
10346 The syntax for the cipher sorting has been extended to support sorting by
10347 cipher-strength (using the strength_bits hard coded in the tables).
10348 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10350 Fix a bug in the cipher-command parser: when supplying a cipher command
10351 string with an "undefined" symbol (neither command nor alphanumeric
10352 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10353 an error is flagged.
10355 Due to the strength-sorting extension, the code of the
10356 ssl_create_cipher_list() function was completely rearranged. I hope that
10357 the readability was also increased :-)
10358 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10360 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10361 for the first serial number and places 2 in the serial number file. This
10362 avoids problems when the root CA is created with serial number zero and
10363 the first user certificate has the same issuer name and serial number
10367 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10368 the new code. Add documentation for this stuff.
10371 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10372 X509_*() to X509at_*() on the grounds that they don't handle X509
10373 structures and behave in an analogous way to the X509v3 functions:
10374 they shouldn't be called directly but wrapper functions should be used
10377 So we also now have some wrapper functions that call the X509at functions
10378 when passed certificate requests. (TO DO: similar things can be done with
10379 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10380 things. Some of these need some d2i or i2d and print functionality
10381 because they handle more complex structures.)
10384 *) Add missing #ifndefs that caused missing symbols when building libssl
10385 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10386 NO_RSA in ssl/s2*.c.
10387 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10389 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10390 has a return value which indicates the quality of the random data
10391 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10392 error queue. New function RAND_pseudo_bytes() generates output that is
10393 guaranteed to be unique but not unpredictable. RAND_add is like
10394 RAND_seed, but takes an extra argument for an entropy estimate
10395 (RAND_seed always assumes full entropy).
10398 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10399 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10400 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10401 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10402 false-positive rate of at most 2^-80 for random input.
10405 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10408 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10409 in the 0.9.5 release), this returns the chain
10410 from an X509_CTX structure with a dup of the stack and all
10411 the X509 reference counts upped: so the stack will exist
10412 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10415 Also make SSL_SESSION_print() print out the verify return
10419 *) Add manpage for the pkcs12 command. Also change the default
10420 behaviour so MAC iteration counts are used unless the new
10421 -nomaciter option is used. This improves file security and
10422 only older versions of MSIE (4.0 for example) need it.
10425 *) Honor the no-xxx Configure options when creating .DEF files.
10428 *) Add PKCS#10 attributes to field table: challengePassword,
10429 unstructuredName and unstructuredAddress. These are taken from
10430 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10431 international characters are used.
10433 More changes to X509_ATTRIBUTE code: allow the setting of types
10434 based on strings. Remove the 'loc' parameter when adding
10435 attributes because these will be a SET OF encoding which is sorted
10439 *) Initial changes to the 'req' utility to allow request generation
10440 automation. This will allow an application to just generate a template
10441 file containing all the field values and have req construct the
10444 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10445 used all over the place including certificate requests and PKCS#7
10446 structures. They are currently handled manually where necessary with
10447 some primitive wrappers for PKCS#7. The new functions behave in a
10448 manner analogous to the X509 extension functions: they allow
10449 attributes to be looked up by NID and added.
10451 Later something similar to the X509V3 code would be desirable to
10452 automatically handle the encoding, decoding and printing of the
10453 more complex types. The string types like challengePassword can
10454 be handled by the string table functions.
10456 Also modified the multi byte string table handling. Now there is
10457 a 'global mask' which masks out certain types. The table itself
10458 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10459 is useful when for example there is only one permissible type
10460 (as in countryName) and using the mask might result in no valid
10464 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10465 SSL_get_peer_finished to allow applications to obtain the latest
10466 Finished messages sent to the peer or expected from the peer,
10467 respectively. (SSL_get_peer_finished is usually the Finished message
10468 actually received from the peer, otherwise the protocol will be aborted.)
10470 As the Finished message are message digests of the complete handshake
10471 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10472 be used for external authentication procedures when the authentication
10473 provided by SSL/TLS is not desired or is not enough.
10476 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10477 the host supports BWX extension and if Compaq C is present on the
10478 $PATH. Just exploiting of the BWX extension results in 20-30%
10479 performance kick for some algorithms, e.g. DES and RC4 to mention
10480 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10484 *) Add support for MS "fast SGC". This is arguably a violation of the
10485 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10486 weak crypto and after checking the certificate is SGC a second one
10487 with strong crypto. MS SGC stops the first handshake after receiving
10488 the server certificate message and sends a second client hello. Since
10489 a server will typically do all the time consuming operations before
10490 expecting any further messages from the client (server key exchange
10491 is the most expensive) there is little difference between the two.
10493 To get OpenSSL to support MS SGC we have to permit a second client
10494 hello message after we have sent server done. In addition we have to
10495 reset the MAC if we do get this second client hello.
10498 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10499 if a DER encoded private key is RSA or DSA traditional format. Changed
10500 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10501 format DER encoded private key. Newer code should use PKCS#8 format which
10502 has the key type encoded in the ASN1 structure. Added DER private key
10503 support to pkcs8 application.
10506 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10507 ciphersuites has been selected (as required by the SSL 3/TLS 1
10508 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10509 is set, we interpret this as a request to violate the specification
10510 (the worst that can happen is a handshake failure, and 'correct'
10511 behaviour would result in a handshake failure anyway).
10514 *) In SSL_CTX_add_session, take into account that there might be multiple
10515 SSL_SESSION structures with the same session ID (e.g. when two threads
10516 concurrently obtain them from an external cache).
10517 The internal cache can handle only one SSL_SESSION with a given ID,
10518 so if there's a conflict, we now throw out the old one to achieve
10522 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10523 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10524 some routines that use cipher OIDs: some ciphers do not have OIDs
10525 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10529 *) Simplify the trust setting structure and code. Now we just have
10530 two sequences of OIDs for trusted and rejected settings. These will
10531 typically have values the same as the extended key usage extension
10532 and any application specific purposes.
10534 The trust checking code now has a default behaviour: it will just
10535 check for an object with the same NID as the passed id. Functions can
10536 be provided to override either the default behaviour or the behaviour
10537 for a given id. SSL client, server and email already have functions
10538 in place for compatibility: they check the NID and also return "trusted"
10539 if the certificate is self signed.
10542 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10543 traditional format into an EVP_PKEY structure.
10546 *) Add a password callback function PEM_cb() which either prompts for
10547 a password if usr_data is NULL or otherwise assumes it is a null
10548 terminated password. Allow passwords to be passed on command line
10549 environment or config files in a few more utilities.
10552 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10553 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10554 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10555 Update documentation.
10558 *) Support for ASN1 "NULL" type. This could be handled before by using
10559 ASN1_TYPE but there wasn't any function that would try to read a NULL
10560 and produce an error if it couldn't. For compatibility we also have
10561 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10562 don't allocate anything because they don't need to.
10565 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10567 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10569 *) Rebuild of the memory allocation routines used by OpenSSL code and
10570 possibly others as well. The purpose is to make an interface that
10571 provide hooks so anyone can build a separate set of allocation and
10572 deallocation routines to be used by OpenSSL, for example memory
10573 pool implementations, or something else, which was previously hard
10574 since Malloc(), Realloc() and Free() were defined as macros having
10575 the values malloc, realloc and free, respectively (except for Win32
10576 compilations). The same is provided for memory debugging code.
10577 OpenSSL already comes with functionality to find memory leaks, but
10578 this gives people a chance to debug other memory problems.
10580 With these changes, a new set of functions and macros have appeared:
10582 CRYPTO_set_mem_debug_functions() [F]
10583 CRYPTO_get_mem_debug_functions() [F]
10584 CRYPTO_dbg_set_options() [F]
10585 CRYPTO_dbg_get_options() [F]
10586 CRYPTO_malloc_debug_init() [M]
10588 The memory debug functions are NULL by default, unless the library
10589 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10590 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10591 gives the standard debugging functions that come with OpenSSL) or
10592 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10593 provided by the library user) must be used. When the standard
10594 debugging functions are used, CRYPTO_dbg_set_options can be used to
10595 request additional information:
10596 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10597 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10599 Also, things like CRYPTO_set_mem_functions will always give the
10600 expected result (the new set of functions is used for allocation
10601 and deallocation) at all times, regardless of platform and compiler
10604 To finish it up, some functions that were never use in any other
10605 way than through macros have a new API and new semantic:
10607 CRYPTO_dbg_malloc()
10608 CRYPTO_dbg_realloc()
10611 All macros of value have retained their old syntax.
10612 [Richard Levitte and Bodo Moeller]
10614 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10615 ordering of SMIMECapabilities wasn't in "strength order" and there
10616 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10620 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10621 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10622 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10624 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10625 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10626 functionality to handle multipart/signed properly) and a utility
10627 called 'smime' to call all this stuff. This is based on code I
10628 originally wrote for Celo who have kindly allowed it to be
10629 included in OpenSSL.
10632 *) Add variants des_set_key_checked and des_set_key_unchecked of
10633 des_set_key (aka des_key_sched). Global variable des_check_key
10634 decides which of these is called by des_set_key; this way
10635 des_check_key behaves as it always did, but applications and
10636 the library itself, which was buggy for des_check_key == 1,
10637 have a cleaner way to pick the version they need.
10640 *) New function PKCS12_newpass() which changes the password of a
10644 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10645 dynamic mix. In both cases the ids can be used as an index into the
10646 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10647 functions so they accept a list of the field values and the
10648 application doesn't need to directly manipulate the X509_TRUST
10652 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10656 *) Modify the way the V3 extension code looks up extensions. This now
10657 works in a similar way to the object code: we have some "standard"
10658 extensions in a static table which is searched with OBJ_bsearch()
10659 and the application can add dynamic ones if needed. The file
10660 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10661 updated whenever a new extension is added to the core code and kept
10662 in ext_nid order. There is a simple program 'tabtest.c' which checks
10663 this. New extensions are not added too often so this file can readily
10664 be maintained manually.
10666 There are two big advantages in doing things this way. The extensions
10667 can be looked up immediately and no longer need to be "added" using
10668 X509V3_add_standard_extensions(): this function now does nothing.
10669 [Side note: I get *lots* of email saying the extension code doesn't
10670 work because people forget to call this function]
10671 Also no dynamic allocation is done unless new extensions are added:
10672 so if we don't add custom extensions there is no need to call
10673 X509V3_EXT_cleanup().
10676 *) Modify enc utility's salting as follows: make salting the default. Add a
10677 magic header, so unsalted files fail gracefully instead of just decrypting
10678 to garbage. This is because not salting is a big security hole, so people
10679 should be discouraged from doing it.
10682 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10683 digest to be passed on the command line but it only used this
10684 parameter when signing a certificate. Modified so all relevant
10685 operations are affected by the digest parameter including the
10686 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10687 DSA key was used because it didn't fix the digest.
10690 *) Initial certificate chain verify code. Currently tests the untrusted
10691 certificates for consistency with the verify purpose (which is set
10692 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10694 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10695 this is because it will reject chains with invalid extensions whereas
10696 every previous version of OpenSSL and SSLeay made no checks at all.
10698 Trust code: checks the root CA for the relevant trust settings. Trust
10699 settings have an initial value consistent with the verify purpose: e.g.
10700 if the verify purpose is for SSL client use it expects the CA to be
10701 trusted for SSL client use. However the default value can be changed to
10702 permit custom trust settings: one example of this would be to only trust
10703 certificates from a specific "secure" set of CAs.
10705 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10706 which should be used for version portability: especially since the
10707 verify structure is likely to change more often now.
10709 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10710 to set them. If not set then assume SSL clients will verify SSL servers
10713 Two new options to the verify program: -untrusted allows a set of
10714 untrusted certificates to be passed in and -purpose which sets the
10715 intended purpose of the certificate. If a purpose is set then the
10716 new chain verify code is used to check extension consistency.
10719 *) Support for the authority information access extension.
10722 *) Modify RSA and DSA PEM read routines to transparently handle
10723 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10724 public keys in a format compatible with certificate
10725 SubjectPublicKeyInfo structures. Unfortunately there were already
10726 functions called *_PublicKey_* which used various odd formats so
10727 these are retained for compatibility: however the DSA variants were
10728 never in a public release so they have been deleted. Changed dsa/rsa
10729 utilities to handle the new format: note no releases ever handled public
10730 keys so we should be OK.
10732 The primary motivation for this change is to avoid the same fiasco
10733 that dogs private keys: there are several incompatible private key
10734 formats some of which are standard and some OpenSSL specific and
10735 require various evil hacks to allow partial transparent handling and
10736 even then it doesn't work with DER formats. Given the option anything
10737 other than PKCS#8 should be dumped: but the other formats have to
10738 stay in the name of compatibility.
10740 With public keys and the benefit of hindsight one standard format
10741 is used which works with EVP_PKEY, RSA or DSA structures: though
10742 it clearly returns an error if you try to read the wrong kind of key.
10744 Added a -pubkey option to the 'x509' utility to output the public key.
10745 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10746 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10747 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10748 that do the same as the EVP_PKEY_assign_*() except they up the
10749 reference count of the added key (they don't "swallow" the
10753 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10754 CRLs would fail if the file contained no certificates or no CRLs:
10755 added a new function to read in both types and return the number
10756 read: this means that if none are read it will be an error. The
10757 DER versions of the certificate and CRL reader would always fail
10758 because it isn't possible to mix certificates and CRLs in DER format
10759 without choking one or the other routine. Changed this to just read
10760 a certificate: this is the best we can do. Also modified the code
10761 in apps/verify.c to take notice of return codes: it was previously
10762 attempting to read in certificates from NULL pointers and ignoring
10763 any errors: this is one reason why the cert and CRL reader seemed
10764 to work. It doesn't check return codes from the default certificate
10765 routines: these may well fail if the certificates aren't installed.
10768 *) Code to support otherName option in GeneralName.
10771 *) First update to verify code. Change the verify utility
10772 so it warns if it is passed a self signed certificate:
10773 for consistency with the normal behaviour. X509_verify
10774 has been modified to it will now verify a self signed
10775 certificate if *exactly* the same certificate appears
10776 in the store: it was previously impossible to trust a
10777 single self signed certificate. This means that:
10778 openssl verify ss.pem
10779 now gives a warning about a self signed certificate but
10780 openssl verify -CAfile ss.pem ss.pem
10784 *) For servers, store verify_result in SSL_SESSION data structure
10785 (and add it to external session representation).
10786 This is needed when client certificate verifications fails,
10787 but an application-provided verification callback (set by
10788 SSL_CTX_set_cert_verify_callback) allows accepting the session
10789 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10790 but returns 1): When the session is reused, we have to set
10791 ssl->verify_result to the appropriate error code to avoid
10793 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10795 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10796 case in PKCS7_dataInit() where the signed PKCS7 structure
10797 didn't contain any existing data because it was being created.
10798 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10800 *) Add a salt to the key derivation routines in enc.c. This
10801 forms the first 8 bytes of the encrypted file. Also add a
10802 -S option to allow a salt to be input on the command line.
10805 *) New function X509_cmp(). Oddly enough there wasn't a function
10806 to compare two certificates. We do this by working out the SHA1
10807 hash and comparing that. X509_cmp() will be needed by the trust
10811 *) SSL_get1_session() is like SSL_get_session(), but increments
10812 the reference count in the SSL_SESSION returned.
10813 [Geoff Thorpe <geoff@eu.c2.net>]
10815 *) Fix for 'req': it was adding a null to request attributes.
10816 Also change the X509_LOOKUP and X509_INFO code to handle
10817 certificate auxiliary information.
10820 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10824 *) Add the possibility to add extra information to the memory leak
10825 detecting output, to form tracebacks, showing from where each
10826 allocation was originated: CRYPTO_push_info("constant string") adds
10827 the string plus current file name and line number to a per-thread
10828 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10829 is like calling CYRPTO_pop_info() until the stack is empty.
10830 Also updated memory leak detection code to be multi-thread-safe.
10833 *) Add options -text and -noout to pkcs7 utility and delete the
10834 encryption options which never did anything. Update docs.
10837 *) Add options to some of the utilities to allow the pass phrase
10838 to be included on either the command line (not recommended on
10839 OSes like Unix) or read from the environment. Update the
10840 manpages and fix a few bugs.
10843 *) Add a few manpages for some of the openssl commands.
10846 *) Fix the -revoke option in ca. It was freeing up memory twice,
10847 leaking and not finding already revoked certificates.
10850 *) Extensive changes to support certificate auxiliary information.
10851 This involves the use of X509_CERT_AUX structure and X509_AUX
10852 functions. An X509_AUX function such as PEM_read_X509_AUX()
10853 can still read in a certificate file in the usual way but it
10854 will also read in any additional "auxiliary information". By
10855 doing things this way a fair degree of compatibility can be
10856 retained: existing certificates can have this information added
10857 using the new 'x509' options.
10859 Current auxiliary information includes an "alias" and some trust
10860 settings. The trust settings will ultimately be used in enhanced
10861 certificate chain verification routines: currently a certificate
10862 can only be trusted if it is self signed and then it is trusted
10866 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10867 The problem was that one of the replacement routines had not been working
10868 since SSLeay releases. For now the offending routine has been replaced
10869 with non-optimised assembler. Even so, this now gives around 95%
10870 performance improvement for 1024 bit RSA signs.
10873 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10874 handling. Most clients have the effective key size in bits equal to
10875 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10876 A few however don't do this and instead use the size of the decrypted key
10877 to determine the RC2 key length and the AlgorithmIdentifier to determine
10878 the effective key length. In this case the effective key length can still
10879 be 40 bits but the key length can be 168 bits for example. This is fixed
10880 by manually forcing an RC2 key into the EVP_PKEY structure because the
10881 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10882 the key length and effective key length are equal.
10885 *) Add a bunch of functions that should simplify the creation of
10886 X509_NAME structures. Now you should be able to do:
10887 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10888 and have it automatically work out the correct field type and fill in
10889 the structures. The more adventurous can try:
10890 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10891 and it will (hopefully) work out the correct multibyte encoding.
10894 *) Change the 'req' utility to use the new field handling and multibyte
10895 copy routines. Before the DN field creation was handled in an ad hoc
10896 way in req, ca, and x509 which was rather broken and didn't support
10897 BMPStrings or UTF8Strings. Since some software doesn't implement
10898 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10899 using the dirstring_type option. See the new comment in the default
10900 openssl.cnf for more info.
10903 *) Make crypto/rand/md_rand.c more robust:
10904 - Assure unique random numbers after fork().
10905 - Make sure that concurrent threads access the global counter and
10906 md serializably so that we never lose entropy in them
10907 or use exactly the same state in multiple threads.
10908 Access to the large state is not always serializable because
10909 the additional locking could be a performance killer, and
10910 md should be large enough anyway.
10913 *) New file apps/app_rand.c with commonly needed functionality
10914 for handling the random seed file.
10916 Use the random seed file in some applications that previously did not:
10918 dsaparam -genkey (which also ignored its '-rand' option),
10921 x509 (when signing).
10922 Except on systems with /dev/urandom, it is crucial to have a random
10923 seed file at least for key creation, DSA signing, and for DH exchanges;
10924 for RSA signatures we could do without one.
10926 gendh and gendsa (unlike genrsa) used to read only the first byte
10927 of each file listed in the '-rand' option. The function as previously
10928 found in genrsa is now in app_rand.c and is used by all programs
10929 that support '-rand'.
10932 *) In RAND_write_file, use mode 0600 for creating files;
10933 don't just chmod when it may be too late.
10936 *) Report an error from X509_STORE_load_locations
10937 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10940 *) New function ASN1_mbstring_copy() this copies a string in either
10941 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10942 into an ASN1_STRING type. A mask of permissible types is passed
10943 and it chooses the "minimal" type to use or an error if not type
10947 *) Add function equivalents to the various macros in asn1.h. The old
10948 macros are retained with an M_ prefix. Code inside the library can
10949 use the M_ macros. External code (including the openssl utility)
10950 should *NOT* in order to be "shared library friendly".
10953 *) Add various functions that can check a certificate's extensions
10954 to see if it usable for various purposes such as SSL client,
10955 server or S/MIME and CAs of these types. This is currently
10956 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10957 verification. Also added a -purpose flag to x509 utility to
10958 print out all the purposes.
10961 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10965 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10966 for, obtain and decode and extension and obtain its critical flag.
10967 This allows all the necessary extension code to be handled in a
10968 single function call.
10971 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10972 platforms. See crypto/rc4/rc4_enc.c for further details.
10975 *) New -noout option to asn1parse. This causes no output to be produced
10976 its main use is when combined with -strparse and -out to extract data
10977 from a file (which may not be in ASN.1 format).
10980 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10981 when producing the local key id.
10982 [Richard Levitte <levitte@stacken.kth.se>]
10984 *) New option -dhparam in s_server. This allows a DH parameter file to be
10985 stated explicitly. If it is not stated then it tries the first server
10986 certificate file. The previous behaviour hard coded the filename
10990 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10991 a public key to be input or output. For example:
10992 openssl rsa -in key.pem -pubout -out pubkey.pem
10993 Also added necessary DSA public key functions to handle this.
10996 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10997 in the message. This was handled by allowing
10998 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10999 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11001 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11002 to the end of the strings whereas this didn't. This would cause problems
11003 if strings read with d2i_ASN1_bytes() were later modified.
11004 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11006 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11007 data and it contains EOF it will end up returning an error. This is
11008 caused by input 46 bytes long. The cause is due to the way base64
11009 BIOs find the start of base64 encoded data. They do this by trying a
11010 trial decode on each line until they find one that works. When they
11011 do a flag is set and it starts again knowing it can pass all the
11012 data directly through the decoder. Unfortunately it doesn't reset
11013 the context it uses. This means that if EOF is reached an attempt
11014 is made to pass two EOFs through the context and this causes the
11015 resulting error. This can also cause other problems as well. As is
11016 usual with these problems it takes *ages* to find and the fix is
11017 trivial: move one line.
11018 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11020 *) Ugly workaround to get s_client and s_server working under Windows. The
11021 old code wouldn't work because it needed to select() on sockets and the
11022 tty (for keypresses and to see if data could be written). Win32 only
11023 supports select() on sockets so we select() with a 1s timeout on the
11024 sockets and then see if any characters are waiting to be read, if none
11025 are present then we retry, we also assume we can always write data to
11026 the tty. This isn't nice because the code then blocks until we've
11027 received a complete line of data and it is effectively polling the
11028 keyboard at 1s intervals: however it's quite a bit better than not
11029 working at all :-) A dedicated Windows application might handle this
11030 with an event loop for example.
11033 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11034 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11035 will be called when RSA_sign() and RSA_verify() are used. This is useful
11036 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11037 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11038 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11039 This necessitated the support of an extra signature type NID_md5_sha1
11040 for SSL signatures and modifications to the SSL library to use it instead
11041 of calling RSA_public_decrypt() and RSA_private_encrypt().
11044 *) Add new -verify -CAfile and -CApath options to the crl program, these
11045 will lookup a CRL issuers certificate and verify the signature in a
11046 similar way to the verify program. Tidy up the crl program so it
11047 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11048 less strict. It will now permit CRL extensions even if it is not
11049 a V2 CRL: this will allow it to tolerate some broken CRLs.
11052 *) Initialize all non-automatic variables each time one of the openssl
11053 sub-programs is started (this is necessary as they may be started
11054 multiple times from the "OpenSSL>" prompt).
11055 [Lennart Bang, Bodo Moeller]
11057 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11058 removing all other RSA functionality (this is what NO_RSA does). This
11059 is so (for example) those in the US can disable those operations covered
11060 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11064 *) Non-copying interface to BIO pairs.
11065 (still largely untested)
11068 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11069 ASCII string. This was handled independently in various places before.
11072 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11073 UTF8 strings a character at a time.
11076 *) Use client_version from client hello to select the protocol
11077 (s23_srvr.c) and for RSA client key exchange verification
11078 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11081 *) Add various utility functions to handle SPKACs, these were previously
11082 handled by poking round in the structure internals. Added new function
11083 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11084 print, verify and generate SPKACs. Based on an original idea from
11085 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11088 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11091 *) Allow the config file extension section to be overwritten on the
11092 command line. Based on an original idea from Massimiliano Pala
11093 <madwolf@comune.modena.it>. The new option is called -extensions
11094 and can be applied to ca, req and x509. Also -reqexts to override
11095 the request extensions in req and -crlexts to override the crl extensions
11099 *) Add new feature to the SPKAC handling in ca. Now you can include
11100 the same field multiple times by preceding it by "XXXX." for example:
11103 this is the same syntax as used in the req config file.
11106 *) Allow certificate extensions to be added to certificate requests. These
11107 are specified in a 'req_extensions' option of the req section of the
11108 config file. They can be printed out with the -text option to req but
11109 are otherwise ignored at present.
11112 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11113 data read consists of only the final block it would not decrypted because
11114 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11115 A misplaced 'break' also meant the decrypted final block might not be
11116 copied until the next read.
11119 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11120 a few extra parameters to the DH structure: these will be useful if
11121 for example we want the value of 'q' or implement X9.42 DH.
11124 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11125 provides hooks that allow the default DSA functions or functions on a
11126 "per key" basis to be replaced. This allows hardware acceleration and
11127 hardware key storage to be handled without major modification to the
11128 library. Also added low level modexp hooks and CRYPTO_EX structure and
11129 associated functions.
11132 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11133 as "read only": it can't be written to and the buffer it points to will
11134 not be freed. Reading from a read only BIO is much more efficient than
11135 a normal memory BIO. This was added because there are several times when
11136 an area of memory needs to be read from a BIO. The previous method was
11137 to create a memory BIO and write the data to it, this results in two
11138 copies of the data and an O(n^2) reading algorithm. There is a new
11139 function BIO_new_mem_buf() which creates a read only memory BIO from
11140 an area of memory. Also modified the PKCS#7 routines to use read only
11144 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11145 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11146 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11147 but a retry condition occurred while trying to read the rest.
11150 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11151 NID_pkcs7_encrypted by default: this was wrong since this should almost
11152 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11153 the encrypted data type: this is a more sensible place to put it and it
11154 allows the PKCS#12 code to be tidied up that duplicated this
11158 *) Changed obj_dat.pl script so it takes its input and output files on
11159 the command line. This should avoid shell escape redirection problems
11163 *) Initial support for certificate extension requests, these are included
11164 in things like Xenroll certificate requests. Included functions to allow
11165 extensions to be obtained and added.
11168 *) -crlf option to s_client and s_server for sending newlines as
11169 CRLF (as required by many protocols).
11172 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11174 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11175 [Ralf S. Engelschall]
11177 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11178 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11180 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11184 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11185 DH parameters/keys (q is lost during that conversion, but the resulting
11186 DH parameters contain its length).
11188 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11189 much faster than DH_generate_parameters (which creates parameters
11190 where p = 2*q + 1), and also the smaller q makes DH computations
11191 much more efficient (160-bit exponentiation instead of 1024-bit
11192 exponentiation); so this provides a convenient way to support DHE
11193 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11194 utter importance to use
11195 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11197 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11198 when such DH parameters are used, because otherwise small subgroup
11199 attacks may become possible!
11202 *) Avoid memory leak in i2d_DHparams.
11205 *) Allow the -k option to be used more than once in the enc program:
11206 this allows the same encrypted message to be read by multiple recipients.
11209 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11210 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11211 it will always use the numerical form of the OID, even if it has a short
11215 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11216 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11217 otherwise bn_mod_exp was called. In the case of hardware keys for example
11218 no private key components need be present and it might store extra data
11219 in the RSA structure, which cannot be accessed from bn_mod_exp.
11220 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11221 private key operations.
11224 *) Added support for SPARC Linux.
11227 *) pem_password_cb function type incompatibly changed from
11228 typedef int pem_password_cb(char *buf, int size, int rwflag);
11230 ....(char *buf, int size, int rwflag, void *userdata);
11231 so that applications can pass data to their callbacks:
11232 The PEM[_ASN1]_{read,write}... functions and macros now take an
11233 additional void * argument, which is just handed through whenever
11234 the password callback is called.
11235 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11237 New function SSL_CTX_set_default_passwd_cb_userdata.
11239 Compatibility note: As many C implementations push function arguments
11240 onto the stack in reverse order, the new library version is likely to
11241 interoperate with programs that have been compiled with the old
11242 pem_password_cb definition (PEM_whatever takes some data that
11243 happens to be on the stack as its last argument, and the callback
11244 just ignores this garbage); but there is no guarantee whatsoever that
11247 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11248 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11249 problems not only on Windows, but also on some Unix platforms.
11250 To avoid problematic command lines, these definitions are now in an
11251 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11252 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11255 *) MIPS III/IV assembler module is reimplemented.
11258 *) More DES library cleanups: remove references to srand/rand and
11259 delete an unused file.
11262 *) Add support for the the free Netwide assembler (NASM) under Win32,
11263 since not many people have MASM (ml) and it can be hard to obtain.
11264 This is currently experimental but it seems to work OK and pass all
11265 the tests. Check out INSTALL.W32 for info.
11268 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11269 without temporary keys kept an extra copy of the server key,
11270 and connections with temporary keys did not free everything in case
11274 *) New function RSA_check_key and new openssl rsa option -check
11275 for verifying the consistency of RSA keys.
11276 [Ulf Moeller, Bodo Moeller]
11278 *) Various changes to make Win32 compile work:
11279 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11280 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11281 comparison" warnings.
11282 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11285 *) Add a debugging option to PKCS#5 v2 key generation function: when
11286 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11287 derived keys are printed to stderr.
11290 *) Copy the flags in ASN1_STRING_dup().
11291 [Roman E. Pavlov <pre@mo.msk.ru>]
11293 *) The x509 application mishandled signing requests containing DSA
11294 keys when the signing key was also DSA and the parameters didn't match.
11296 It was supposed to omit the parameters when they matched the signing key:
11297 the verifying software was then supposed to automatically use the CA's
11298 parameters if they were absent from the end user certificate.
11300 Omitting parameters is no longer recommended. The test was also
11301 the wrong way round! This was probably due to unusual behaviour in
11302 EVP_cmp_parameters() which returns 1 if the parameters match.
11303 This meant that parameters were omitted when they *didn't* match and
11304 the certificate was useless. Certificates signed with 'ca' didn't have
11306 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11308 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11309 The interface is as follows:
11310 Applications can use
11311 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11312 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11313 "off" is now the default.
11314 The library internally uses
11315 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11316 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11317 to disable memory-checking temporarily.
11319 Some inconsistent states that previously were possible (and were
11320 even the default) are now avoided.
11322 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11323 with each memory chunk allocated; this is occasionally more helpful
11324 than just having a counter.
11326 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11328 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11332 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11333 which largely parallels "options", but is for changing API behaviour,
11334 whereas "options" are about protocol behaviour.
11335 Initial "mode" flags are:
11337 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11338 a single record has been written.
11339 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11340 retries use the same buffer location.
11341 (But all of the contents must be
11345 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11348 *) Fix problems with no-hmac etc.
11349 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11351 *) New functions RSA_get_default_method(), RSA_set_method() and
11352 RSA_get_method(). These allows replacement of RSA_METHODs without having
11353 to mess around with the internals of an RSA structure.
11356 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11357 Also really enable memory leak checks in openssl.c and in some
11359 [Chad C. Mulligan, Bodo Moeller]
11361 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11362 up the length of negative integers. This has now been simplified to just
11363 store the length when it is first determined and use it later, rather
11364 than trying to keep track of where data is copied and updating it to
11366 [Steve Henson, reported by Brien Wheeler
11367 <bwheeler@authentica-security.com>]
11369 *) Add a new function PKCS7_signatureVerify. This allows the verification
11370 of a PKCS#7 signature but with the signing certificate passed to the
11371 function itself. This contrasts with PKCS7_dataVerify which assumes the
11372 certificate is present in the PKCS#7 structure. This isn't always the
11373 case: certificates can be omitted from a PKCS#7 structure and be
11374 distributed by "out of band" means (such as a certificate database).
11377 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11378 function prototypes in pem.h, also change util/mkdef.pl to add the
11379 necessary function names.
11382 *) mk1mf.pl (used by Windows builds) did not properly read the
11383 options set by Configure in the top level Makefile, and Configure
11384 was not even able to write more than one option correctly.
11385 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11388 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11389 file to be loaded from a BIO or FILE pointer. The BIO version will
11390 for example allow memory BIOs to contain config info.
11393 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11394 Whoever hopes to achieve shared-library compatibility across versions
11395 must use this, not the compile-time macro.
11396 (Exercise 0.9.4: Which is the minimum library version required by
11398 Note: All this applies only to multi-threaded programs, others don't
11402 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11403 through a BIO pair triggered the default case, i.e.
11404 SSLerr(...,SSL_R_UNKNOWN_STATE).
11407 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11408 can use the SSL library even if none of the specific BIOs is
11412 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11413 for the encoded length.
11414 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11416 *) Add initial documentation of the X509V3 functions.
11419 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11420 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11421 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11422 secure PKCS#8 private key format with a high iteration count.
11425 *) Fix determination of Perl interpreter: A perl or perl5
11426 _directory_ in $PATH was also accepted as the interpreter.
11427 [Ralf S. Engelschall]
11429 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11430 wrong with it but it was very old and did things like calling
11431 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11432 unusual formatting.
11435 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11436 to use the new extension code.
11439 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11440 with macros. This should make it easier to change their form, add extra
11441 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11445 *) Add to configuration table a new entry that can specify an alternative
11446 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11447 according to Mark Crispin <MRC@Panda.COM>.
11451 *) DES CBC did not update the IV. Weird.
11454 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11455 Changing the behaviour of the former might break existing programs --
11456 where IV updating is needed, des_ncbc_encrypt can be used.
11459 *) When bntest is run from "make test" it drives bc to check its
11460 calculations, as well as internally checking them. If an internal check
11461 fails, it needs to cause bc to give a non-zero result or make test carries
11462 on without noticing the failure. Fixed.
11465 *) DES library cleanups.
11468 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11469 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11470 ciphers. NOTE: although the key derivation function has been verified
11471 against some published test vectors it has not been extensively tested
11472 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11476 *) Instead of "mkdir -p", which is not fully portable, use new
11477 Perl script "util/mkdir-p.pl".
11480 *) Rewrite the way password based encryption (PBE) is handled. It used to
11481 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11482 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11483 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11484 the 'parameter' field of the AlgorithmIdentifier is passed to the
11485 underlying key generation function so it must do its own ASN1 parsing.
11486 This has also changed the EVP_PBE_CipherInit() function which now has a
11487 'parameter' argument instead of literal salt and iteration count values
11488 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11491 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11492 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11493 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11494 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11495 value was just used as a "magic string" and not used directly its
11496 value doesn't matter.
11499 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11503 *) "linux-sparc64" configuration (ultrapenguin).
11504 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11505 "linux-sparc" configuration.
11506 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11508 *) config now generates no-xxx options for missing ciphers.
11511 *) Support the EBCDIC character set (work in progress).
11512 File ebcdic.c not yet included because it has a different license.
11513 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11515 *) Support BS2000/OSD-POSIX.
11516 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11518 *) Make callbacks for key generation use void * instead of char *.
11521 *) Make S/MIME samples compile (not yet tested).
11524 *) Additional typesafe stacks.
11527 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11531 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11533 *) New configuration variant "sco5-gcc".
11535 *) Updated some demos.
11536 [Sean O Riordain, Wade Scholine]
11538 *) Add missing BIO_free at exit of pkcs12 application.
11541 *) Fix memory leak in conf.c.
11544 *) Updates for Win32 to assembler version of MD5.
11547 *) Set #! path to perl in apps/der_chop to where we found it
11548 instead of using a fixed path.
11551 *) SHA library changes for irix64-mips4-cc.
11554 *) Improvements for VMS support.
11558 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11560 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11561 This also avoids the problems with SC4.2 and unpatched SC5.
11562 [Andy Polyakov <appro@fy.chalmers.se>]
11564 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11565 These are required because of the typesafe stack would otherwise break
11566 existing code. If old code used a structure member which used to be STACK
11567 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11568 sk_num or sk_value it would produce an error because the num, data members
11569 are not present in STACK_OF. Now it just produces a warning. sk_set
11570 replaces the old method of assigning a value to sk_value
11571 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11572 that does this will no longer work (and should use sk_set instead) but
11573 this could be regarded as a "questionable" behaviour anyway.
11576 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11577 correctly handle encrypted S/MIME data.
11580 *) Change type of various DES function arguments from des_cblock
11581 (which means, in function argument declarations, pointer to char)
11582 to des_cblock * (meaning pointer to array with 8 char elements),
11583 which allows the compiler to do more typechecking; it was like
11584 that back in SSLeay, but with lots of ugly casts.
11586 Introduce new type const_des_cblock.
11589 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11590 problems: find RecipientInfo structure that matches recipient certificate
11591 and initialise the ASN1 structures properly based on passed cipher.
11594 *) Belatedly make the BN tests actually check the results.
11597 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11598 to and from BNs: it was completely broken. New compilation option
11599 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11600 key elements as negative integers.
11603 *) Reorganize and speed up MD5.
11604 [Andy Polyakov <appro@fy.chalmers.se>]
11607 [Richard Levitte <richard@levitte.org>]
11609 *) New option -out to asn1parse to allow the parsed structure to be
11610 output to a file. This is most useful when combined with the -strparse
11611 option to examine the output of things like OCTET STRINGS.
11614 *) Make SSL library a little more fool-proof by not requiring any longer
11615 that SSL_set_{accept,connect}_state be called before
11616 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11617 in many applications because usually everything *appeared* to work as
11618 intended anyway -- now it really works as intended).
11621 *) Move openssl.cnf out of lib/.
11624 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11625 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11626 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11627 [Ralf S. Engelschall]
11629 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11630 handle PKCS#7 enveloped data properly.
11631 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11633 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11634 copying pointers. The cert_st handling is changed by this in
11635 various ways (and thus what used to be known as ctx->default_cert
11636 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11637 any longer when s->cert does not give us what we need).
11638 ssl_cert_instantiate becomes obsolete by this change.
11639 As soon as we've got the new code right (possibly it already is?),
11640 we have solved a couple of bugs of the earlier code where s->cert
11641 was used as if it could not have been shared with other SSL structures.
11643 Note that using the SSL API in certain dirty ways now will result
11644 in different behaviour than observed with earlier library versions:
11645 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11646 does not influence s as it used to.
11648 In order to clean up things more thoroughly, inside SSL_SESSION
11649 we don't use CERT any longer, but a new structure SESS_CERT
11650 that holds per-session data (if available); currently, this is
11651 the peer's certificate chain and, for clients, the server's certificate
11652 and temporary key. CERT holds only those values that can have
11653 meaningful defaults in an SSL_CTX.
11656 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11657 from the internal representation. Various PKCS#7 fixes: remove some
11658 evil casts and set the enc_dig_alg field properly based on the signing
11662 *) Allow PKCS#12 password to be set from the command line or the
11663 environment. Let 'ca' get its config file name from the environment
11664 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11668 *) Allow certificate policies extension to use an IA5STRING for the
11669 organization field. This is contrary to the PKIX definition but
11670 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11674 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11675 without disallowing inline assembler and the like for non-pedantic builds.
11678 *) Support Borland C++ builder.
11679 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11681 *) Support Mingw32.
11684 *) SHA-1 cleanups and performance enhancements.
11685 [Andy Polyakov <appro@fy.chalmers.se>]
11687 *) Sparc v8plus assembler for the bignum library.
11688 [Andy Polyakov <appro@fy.chalmers.se>]
11690 *) Accept any -xxx and +xxx compiler options in Configure.
11693 *) Update HPUX configuration.
11696 *) Add missing sk_<type>_unshift() function to safestack.h
11697 [Ralf S. Engelschall]
11699 *) New function SSL_CTX_use_certificate_chain_file that sets the
11700 "extra_cert"s in addition to the certificate. (This makes sense
11701 only for "PEM" format files, as chains as a whole are not
11705 *) Support verify_depth from the SSL API.
11706 x509_vfy.c had what can be considered an off-by-one-error:
11707 Its depth (which was not part of the external interface)
11708 was actually counting the number of certificates in a chain;
11709 now it really counts the depth.
11712 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11713 instead of X509err, which often resulted in confusing error
11714 messages since the error codes are not globally unique
11715 (e.g. an alleged error in ssl3_accept when a certificate
11716 didn't match the private key).
11718 *) New function SSL_CTX_set_session_id_context that allows to set a default
11719 value (so that you don't need SSL_set_session_id_context for each
11720 connection using the SSL_CTX).
11723 *) OAEP decoding bug fix.
11726 *) Support INSTALL_PREFIX for package builders, as proposed by
11730 *) New Configure options "threads" and "no-threads". For systems
11731 where the proper compiler options are known (currently Solaris
11732 and Linux), "threads" is the default.
11735 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11738 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11739 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11740 such as /usr/local/bin.
11743 *) "make linux-shared" to build shared libraries.
11744 [Niels Poppe <niels@netbox.org>]
11746 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11749 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11750 extension adding in x509 utility.
11753 *) Remove NOPROTO sections and error code comments.
11756 *) Partial rewrite of the DEF file generator to now parse the ANSI
11760 *) New Configure options --prefix=DIR and --openssldir=DIR.
11763 *) Complete rewrite of the error code script(s). It is all now handled
11764 by one script at the top level which handles error code gathering,
11765 header rewriting and C source file generation. It should be much better
11766 than the old method: it now uses a modified version of Ulf's parser to
11767 read the ANSI prototypes in all header files (thus the old K&R definitions
11768 aren't needed for error creation any more) and do a better job of
11769 translating function codes into names. The old 'ASN1 error code imbedded
11770 in a comment' is no longer necessary and it doesn't use .err files which
11771 have now been deleted. Also the error code call doesn't have to appear all
11772 on one line (which resulted in some large lines...).
11775 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11778 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11779 0 (which usually indicates a closed connection), but continue reading.
11782 *) Fix some race conditions.
11785 *) Add support for CRL distribution points extension. Add Certificate
11786 Policies and CRL distribution points documentation.
11789 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11792 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11793 8 of keying material. Merlin has also confirmed interop with this fix
11794 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11795 [Merlin Hughes <merlin@baltimore.ie>]
11797 *) Fix lots of warnings.
11798 [Richard Levitte <levitte@stacken.kth.se>]
11800 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11801 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11802 [Richard Levitte <levitte@stacken.kth.se>]
11804 *) Fix problems with sizeof(long) == 8.
11805 [Andy Polyakov <appro@fy.chalmers.se>]
11807 *) Change functions to ANSI C.
11810 *) Fix typos in error codes.
11811 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11813 *) Remove defunct assembler files from Configure.
11816 *) SPARC v8 assembler BIGNUM implementation.
11817 [Andy Polyakov <appro@fy.chalmers.se>]
11819 *) Support for Certificate Policies extension: both print and set.
11820 Various additions to support the r2i method this uses.
11823 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11824 return a const string when you are expecting an allocated buffer.
11827 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11828 types DirectoryString and DisplayText.
11831 *) Add code to allow r2i extensions to access the configuration database,
11832 add an LHASH database driver and add several ctx helper functions.
11835 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11836 fail when they extended the size of a BIGNUM.
11839 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11840 support typesafe stack.
11843 *) Fix typo in SSL_[gs]et_options().
11844 [Nils Frostberg <nils@medcom.se>]
11846 *) Delete various functions and files that belonged to the (now obsolete)
11847 old X509V3 handling code.
11850 *) New Configure option "rsaref".
11853 *) Don't auto-generate pem.h.
11856 *) Introduce type-safe ASN.1 SETs.
11859 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11860 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11862 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11863 that links with OpenSSL (well at least cause lots of warnings), but fear
11864 not: the conversion is trivial, and it eliminates loads of evil casts. A
11865 few STACKed things have been converted already. Feel free to convert more.
11866 In the fullness of time, I'll do away with the STACK type altogether.
11869 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11870 specified in <certfile> by updating the entry in the index.txt file.
11871 This way one no longer has to edit the index.txt file manually for
11872 revoking a certificate. The -revoke option does the gory details now.
11873 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11875 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11876 `-text' option at all and this way the `-noout -text' combination was
11877 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11878 [Ralf S. Engelschall]
11880 *) Make sure a corresponding plain text error message exists for the
11881 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11882 verify callback function determined that a certificate was revoked.
11883 [Ralf S. Engelschall]
11885 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11886 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11887 all available cipers including rc5, which was forgotten until now.
11888 In order to let the testing shell script know which algorithms
11889 are available, a new (up to now undocumented) command
11890 "openssl list-cipher-commands" is used.
11893 *) Bugfix: s_client occasionally would sleep in select() when
11894 it should have checked SSL_pending() first.
11897 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11898 the raw DSA values prior to ASN.1 encoding.
11901 *) Tweaks to Configure
11902 [Niels Poppe <niels@netbox.org>]
11904 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11908 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11911 *) New config option to avoid instructions that are illegal on the 80386.
11912 The default code is faster, but requires at least a 486.
11915 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11916 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11917 same as SSL2_VERSION anyway.
11920 *) New "-showcerts" option for s_client.
11923 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11924 application. Various cleanups and fixes.
11927 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11928 modify error routines to work internally. Add error codes and PBE init
11929 to library startup routines.
11932 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11933 packing functions to asn1 and evp. Changed function names and error
11934 codes along the way.
11937 *) PKCS12 integration: and so it begins... First of several patches to
11938 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11939 objects to objects.h
11942 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11943 and display support for Thawte strong extranet extension.
11946 *) Add LinuxPPC support.
11947 [Jeff Dubrule <igor@pobox.org>]
11949 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11950 bn_div_words in alpha.s.
11951 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11953 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11954 OAEP isn't supported when OpenSSL is built with RSAref.
11955 [Ulf Moeller <ulf@fitug.de>]
11957 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11958 so they no longer are missing under -DNOPROTO.
11959 [Soren S. Jorvang <soren@t.dk>]
11962 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11964 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11965 doesn't work when the session is reused. Coming soon!
11968 *) Fix a security hole, that allows sessions to be reused in the wrong
11969 context thus bypassing client cert protection! All software that uses
11970 client certs and session caches in multiple contexts NEEDS PATCHING to
11971 allow session reuse! A fuller solution is in the works.
11972 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11974 *) Some more source tree cleanups (removed obsolete files
11975 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11976 permission on "config" script to be executable) and a fix for the INSTALL
11978 [Ulf Moeller <ulf@fitug.de>]
11980 *) Remove some legacy and erroneous uses of malloc, free instead of
11982 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11984 *) Make rsa_oaep_test return non-zero on error.
11985 [Ulf Moeller <ulf@fitug.de>]
11987 *) Add support for native Solaris shared libraries. Configure
11988 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11989 if someone would make that last step automatic.
11990 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11992 *) ctx_size was not built with the right compiler during "make links". Fixed.
11995 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11996 except NULL ciphers". This means the default cipher list will no longer
11997 enable NULL ciphers. They need to be specifically enabled e.g. with
11998 the string "DEFAULT:eNULL".
12001 *) Fix to RSA private encryption routines: if p < q then it would
12002 occasionally produce an invalid result. This will only happen with
12003 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12006 *) Be less restrictive and allow also `perl util/perlpath.pl
12007 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12008 because this way one can also use an interpreter named `perl5' (which is
12009 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12010 installed as `perl').
12011 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12013 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12014 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12016 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12017 advapi32.lib to Win32 build and change the pem test comparison
12018 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12019 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12020 and crypto/des/ede_cbcm_enc.c.
12023 *) DES quad checksum was broken on big-endian architectures. Fixed.
12026 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12027 Win32 test batch file so it (might) work again. The Win32 test batch file
12028 is horrible: I feel ill....
12031 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12032 in e_os.h. Audit of header files to check ANSI and non ANSI
12033 sections: 10 functions were absent from non ANSI section and not exported
12034 from Windows DLLs. Fixed up libeay.num for new functions.
12037 *) Make `openssl version' output lines consistent.
12038 [Ralf S. Engelschall]
12040 *) Fix Win32 symbol export lists for BIO functions: Added
12041 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12042 to ms/libeay{16,32}.def.
12043 [Ralf S. Engelschall]
12045 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12046 fine under Unix and passes some trivial tests I've now added. But the
12047 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12048 added to make sure no one expects that this stuff really works in the
12049 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12050 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12052 [Ralf S. Engelschall]
12054 *) Fix the generation of two part addresses in perl.
12055 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12057 *) Add config entry for Linux on MIPS.
12058 [John Tobey <jtobey@channel1.com>]
12060 *) Make links whenever Configure is run, unless we are on Windoze.
12063 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12064 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12068 *) Add a useful kludge to allow package maintainers to specify compiler and
12069 other platforms details on the command line without having to patch the
12070 Configure script everytime: One now can use ``perl Configure
12071 <id>:<details>'', i.e. platform ids are allowed to have details appended
12072 to them (separated by colons). This is treated as there would be a static
12073 pre-configured entry in Configure's %table under key <id> with value
12074 <details> and ``perl Configure <id>'' is called. So, when you want to
12075 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12076 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12077 now, which overrides the FreeBSD-elf entry on-the-fly.
12078 [Ralf S. Engelschall]
12080 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12083 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12084 on the `perl Configure ...' command line. This way one can compile
12085 OpenSSL libraries with Position Independent Code (PIC) which is needed
12086 for linking it into DSOs.
12087 [Ralf S. Engelschall]
12089 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12093 *) Cleaned up the LICENSE document: The official contact for any license
12094 questions now is the OpenSSL core team under openssl-core@openssl.org.
12095 And add a paragraph about the dual-license situation to make sure people
12096 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12097 to the OpenSSL toolkit.
12098 [Ralf S. Engelschall]
12100 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12101 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12102 Additionally cleaned up the `make links' target: Remove unnecessary
12103 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12104 to speed processing and no longer clutter the display with confusing
12105 stuff. Instead only the actually done links are displayed.
12106 [Ralf S. Engelschall]
12108 *) Permit null encryption ciphersuites, used for authentication only. It used
12109 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12110 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12114 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12115 signed attributes when verifying signatures (this would break them),
12116 the detached data encoding was wrong and public keys obtained using
12117 X509_get_pubkey() weren't freed.
12120 *) Add text documentation for the BUFFER functions. Also added a work around
12121 to a Win95 console bug. This was triggered by the password read stuff: the
12122 last character typed gets carried over to the next fread(). If you were
12123 generating a new cert request using 'req' for example then the last
12124 character of the passphrase would be CR which would then enter the first
12128 *) Added the new `Includes OpenSSL Cryptography Software' button as
12129 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12130 button and can be used by applications based on OpenSSL to show the
12131 relationship to the OpenSSL project.
12132 [Ralf S. Engelschall]
12134 *) Remove confusing variables in function signatures in files
12135 ssl/ssl_lib.c and ssl/ssl.h.
12136 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12138 *) Don't install bss_file.c under PREFIX/include/
12139 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12141 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12142 functions that return function pointers and has support for NT specific
12143 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12144 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12145 unsigned to signed types: this was killing the Win32 compile.
12148 *) Add new certificate file to stack functions,
12149 SSL_add_dir_cert_subjects_to_stack() and
12150 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12151 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12152 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12153 This means that Apache-SSL and similar packages don't have to mess around
12154 to add as many CAs as they want to the preferred list.
12157 *) Experiment with doxygen documentation. Currently only partially applied to
12159 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12160 openssl.doxy as the configuration file.
12163 *) Get rid of remaining C++-style comments which strict C compilers hate.
12164 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12166 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12167 compiled in by default: it has problems with large keys.
12170 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12171 DH private keys and/or callback functions which directly correspond to
12172 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12173 is needed for applications which have to configure certificates on a
12174 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12176 For the RSA certificate situation is makes no difference, but
12177 for the DSA certificate situation this fixes the "no shared cipher"
12178 problem where the OpenSSL cipher selection procedure failed because the
12179 temporary keys were not overtaken from the context and the API provided
12180 no way to reconfigure them.
12181 The new functions now let applications reconfigure the stuff and they
12182 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12183 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12184 non-public-API function ssl_cert_instantiate() is used as a helper
12185 function and also to reduce code redundancy inside ssl_rsa.c.
12186 [Ralf S. Engelschall]
12188 *) Move s_server -dcert and -dkey options out of the undocumented feature
12189 area because they are useful for the DSA situation and should be
12190 recognized by the users.
12191 [Ralf S. Engelschall]
12193 *) Fix the cipher decision scheme for export ciphers: the export bits are
12194 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12195 SSL_EXP_MASK. So, the original variable has to be used instead of the
12196 already masked variable.
12197 [Richard Levitte <levitte@stacken.kth.se>]
12199 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12200 [Richard Levitte <levitte@stacken.kth.se>]
12202 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12203 from `int' to `unsigned int' because it's a length and initialized by
12204 EVP_DigestFinal() which expects an `unsigned int *'.
12205 [Richard Levitte <levitte@stacken.kth.se>]
12207 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12208 script. Instead use the usual Shell->Perl transition trick.
12209 [Ralf S. Engelschall]
12211 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12212 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12213 -noout -modulus' as it's already the case for `openssl rsa -noout
12214 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12215 currently the public key is printed (a decision which was already done by
12216 `openssl dsa -modulus' in the past) which serves a similar purpose.
12217 Additionally the NO_RSA no longer completely removes the whole -modulus
12218 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12220 [Ralf S. Engelschall]
12222 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12223 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12224 [Arne Ansper <arne@ats.cyber.ee>]
12226 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12227 to be added. Now both 'req' and 'ca' can use new objects defined in the
12231 *) Add cool BIO that does syslog (or event log on NT).
12232 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12234 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12235 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12236 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12237 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12240 *) Add preliminary config info for new extension code.
12243 *) Make RSA_NO_PADDING really use no padding.
12244 [Ulf Moeller <ulf@fitug.de>]
12246 *) Generate errors when private/public key check is done.
12249 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12250 for some CRL extensions and new objects added.
12253 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12254 key usage extension and fuller support for authority key id.
12257 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12258 padding method for RSA, which is recommended for new applications in PKCS
12259 #1 v2.0 (RFC 2437, October 1998).
12260 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12261 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12262 against Bleichbacher's attack on RSA.
12263 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12266 *) Updates to the new SSL compression code
12267 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12269 *) Fix so that the version number in the master secret, when passed
12270 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12271 (because the server will not accept higher), that the version number
12272 is 0x03,0x01, not 0x03,0x00
12273 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12275 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12276 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12277 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12280 *) Support for RAW extensions where an arbitrary extension can be
12281 created by including its DER encoding. See apps/openssl.cnf for
12285 *) Make sure latest Perl versions don't interpret some generated C array
12286 code as Perl array code in the crypto/err/err_genc.pl script.
12287 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12289 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12290 not many people have the assembler. Various Win32 compilation fixes and
12291 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12292 build instructions.
12295 *) Modify configure script 'Configure' to automatically create crypto/date.h
12296 file under Win32 and also build pem.h from pem.org. New script
12297 util/mkfiles.pl to create the MINFO file on environments that can't do a
12298 'make files': perl util/mkfiles.pl >MINFO should work.
12301 *) Major rework of DES function declarations, in the pursuit of correctness
12302 and purity. As a result, many evil casts evaporated, and some weirdness,
12303 too. You may find this causes warnings in your code. Zapping your evil
12304 casts will probably fix them. Mostly.
12307 *) Fix for a typo in asn1.h. Bug fix to object creation script
12308 obj_dat.pl. It considered a zero in an object definition to mean
12309 "end of object": none of the objects in objects.h have any zeros
12310 so it wasn't spotted.
12311 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12313 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12314 Masking (CBCM). In the absence of test vectors, the best I have been able
12315 to do is check that the decrypt undoes the encrypt, so far. Send me test
12316 vectors if you have them.
12319 *) Correct calculation of key length for export ciphers (too much space was
12320 allocated for null ciphers). This has not been tested!
12323 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12324 message is now correct (it understands "crypto" and "ssl" on its
12325 command line). There is also now an "update" option. This will update
12326 the util/ssleay.num and util/libeay.num files with any new functions.
12328 perl util/mkdef.pl crypto ssl update
12329 it will update them.
12332 *) Overhauled the Perl interface (perl/*):
12333 - ported BN stuff to OpenSSL's different BN library
12334 - made the perl/ source tree CVS-aware
12335 - renamed the package from SSLeay to OpenSSL (the files still contain
12336 their history because I've copied them in the repository)
12337 - removed obsolete files (the test scripts will be replaced
12338 by better Test::Harness variants in the future)
12339 [Ralf S. Engelschall]
12341 *) First cut for a very conservative source tree cleanup:
12342 1. merge various obsolete readme texts into doc/ssleay.txt
12343 where we collect the old documents and readme texts.
12344 2. remove the first part of files where I'm already sure that we no
12345 longer need them because of three reasons: either they are just temporary
12346 files which were left by Eric or they are preserved original files where
12347 I've verified that the diff is also available in the CVS via "cvs diff
12348 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12349 the crypto/md/ stuff).
12350 [Ralf S. Engelschall]
12352 *) More extension code. Incomplete support for subject and issuer alt
12353 name, issuer and authority key id. Change the i2v function parameters
12354 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12355 what that's for :-) Fix to ASN1 macro which messed up
12356 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12359 *) Preliminary support for ENUMERATED type. This is largely copied from the
12363 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12364 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12366 *) Make sure `make rehash' target really finds the `openssl' program.
12367 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12369 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12370 like to hear about it if this slows down other processors.
12373 *) Add CygWin32 platform information to Configure script.
12374 [Alan Batie <batie@aahz.jf.intel.com>]
12376 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12377 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12379 *) New program nseq to manipulate netscape certificate sequences
12382 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12386 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12387 but the BN code had some problems that would cause failures when
12388 doing certificate verification and some other functions.
12389 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12391 *) Add ASN1 and PEM code to support netscape certificate sequences.
12394 *) Add ASN1 and PEM code to support netscape certificate sequences.
12397 *) Add several PKIX and private extended key usage OIDs.
12400 *) Modify the 'ca' program to handle the new extension code. Modify
12401 openssl.cnf for new extension format, add comments.
12404 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12405 and add a sample to openssl.cnf so req -x509 now adds appropriate
12409 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12410 error code, add initial support to X509_print() and x509 application.
12413 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12414 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12415 stuff is currently isolated and isn't even compiled yet.
12418 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12419 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12420 Removed the versions check from X509 routines when loading extensions:
12421 this allows certain broken certificates that don't set the version
12422 properly to be processed.
12425 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12426 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12427 can still be regenerated with "make depend".
12430 *) Spelling mistake in C version of CAST-128.
12431 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12433 *) Changes to the error generation code. The perl script err-code.pl
12434 now reads in the old error codes and retains the old numbers, only
12435 adding new ones if necessary. It also only changes the .err files if new
12436 codes are added. The makefiles have been modified to only insert errors
12437 when needed (to avoid needlessly modifying header files). This is done
12438 by only inserting errors if the .err file is newer than the auto generated
12439 C file. To rebuild all the error codes from scratch (the old behaviour)
12440 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12441 or delete all the .err files.
12444 *) CAST-128 was incorrectly implemented for short keys. The C version has
12445 been fixed, but is untested. The assembler versions are also fixed, but
12446 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12447 to regenerate it if needed.
12448 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12449 Hagino <itojun@kame.net>]
12451 *) File was opened incorrectly in randfile.c.
12452 [Ulf Möller <ulf@fitug.de>]
12454 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12455 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12456 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12457 al: it's just almost always a UTCTime. Note this patch adds new error
12458 codes so do a "make errors" if there are problems.
12461 *) Correct Linux 1 recognition in config.
12462 [Ulf Möller <ulf@fitug.de>]
12464 *) Remove pointless MD5 hash when using DSA keys in ca.
12465 [Anonymous <nobody@replay.com>]
12467 *) Generate an error if given an empty string as a cert directory. Also
12468 generate an error if handed NULL (previously returned 0 to indicate an
12469 error, but didn't set one).
12470 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12472 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12475 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12476 parameters. This was causing a warning which killed off the Win32 compile.
12479 *) Remove C++ style comments from crypto/bn/bn_local.h.
12480 [Neil Costigan <neil.costigan@celocom.com>]
12482 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12483 based on a text string, looking up short and long names and finally
12484 "dot" format. The "dot" format stuff didn't work. Added new function
12485 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12486 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12487 OID is not part of the table.
12490 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12491 X509_LOOKUP_by_alias().
12494 *) Sort openssl functions by name.
12497 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12498 encryption from sample DSA keys (in case anyone is interested the password
12502 *) Make _all_ *_free functions accept a NULL pointer.
12503 [Frans Heymans <fheymans@isaserver.be>]
12505 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12507 [Anonymous <nobody@replay.com>]
12509 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12510 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12512 *) Don't blow it for numeric -newkey arguments to apps/req.
12513 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12515 *) Temp key "for export" tests were wrong in s3_srvr.c.
12516 [Anonymous <nobody@replay.com>]
12518 *) Add prototype for temp key callback functions
12519 SSL_CTX_set_tmp_{rsa,dh}_callback().
12522 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12523 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12526 *) X509_name_add_entry() freed the wrong thing after an error.
12527 [Arne Ansper <arne@ats.cyber.ee>]
12529 *) rsa_eay.c would attempt to free a NULL context.
12530 [Arne Ansper <arne@ats.cyber.ee>]
12532 *) BIO_s_socket() had a broken should_retry() on Windoze.
12533 [Arne Ansper <arne@ats.cyber.ee>]
12535 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12536 [Arne Ansper <arne@ats.cyber.ee>]
12538 *) Make sure the already existing X509_STORE->depth variable is initialized
12539 in X509_STORE_new(), but document the fact that this variable is still
12540 unused in the certificate verification process.
12541 [Ralf S. Engelschall]
12543 *) Fix the various library and apps files to free up pkeys obtained from
12544 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12547 *) Fix reference counting in X509_PUBKEY_get(). This makes
12548 demos/maurice/example2.c work, amongst others, probably.
12549 [Steve Henson and Ben Laurie]
12551 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12552 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12553 are no longer created. This way we have a single and consistent command
12554 line interface `openssl <command>', similar to `cvs <command>'.
12555 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12557 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12558 BIT STRING wrapper always have zero unused bits.
12561 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12564 *) Make the top-level INSTALL documentation easier to understand.
12567 *) Makefiles updated to exit if an error occurs in a sub-directory
12568 make (including if user presses ^C) [Paul Sutton]
12570 *) Make Montgomery context stuff explicit in RSA data structure.
12573 *) Fix build order of pem and err to allow for generated pem.h.
12576 *) Fix renumbering bug in X509_NAME_delete_entry().
12579 *) Enhanced the err-ins.pl script so it makes the error library number
12580 global and can add a library name. This is needed for external ASN1 and
12581 other error libraries.
12584 *) Fixed sk_insert which never worked properly.
12587 *) Fix ASN1 macros so they can handle indefinite length constructed
12588 EXPLICIT tags. Some non standard certificates use these: they can now
12592 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12593 into a single doc/ssleay.txt bundle. This way the information is still
12594 preserved but no longer messes up this directory. Now it's new room for
12595 the new set of documentation files.
12596 [Ralf S. Engelschall]
12598 *) SETs were incorrectly DER encoded. This was a major pain, because they
12599 shared code with SEQUENCEs, which aren't coded the same. This means that
12600 almost everything to do with SETs or SEQUENCEs has either changed name or
12601 number of arguments.
12602 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12604 *) Fix test data to work with the above.
12607 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12608 was already fixed by Eric for 0.9.1 it seems.
12609 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12611 *) Autodetect FreeBSD3.
12614 *) Fix various bugs in Configure. This affects the following platforms:
12618 unixware-2.0-pentium
12622 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12623 before they are needed.
12626 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12630 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12632 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12633 changed SSLeay to OpenSSL in version strings.
12634 [Ralf S. Engelschall]
12636 *) Some fixups to the top-level documents.
12639 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12640 because the symlink to include/ was missing.
12641 [Ralf S. Engelschall]
12643 *) Incorporated the popular no-RSA/DSA-only patches
12644 which allow to compile a RSA-free SSLeay.
12645 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12647 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12648 when "ssleay" is still not found.
12649 [Ralf S. Engelschall]
12651 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12652 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12654 *) Updated the README file.
12655 [Ralf S. Engelschall]
12657 *) Added various .cvsignore files in the CVS repository subdirs
12658 to make a "cvs update" really silent.
12659 [Ralf S. Engelschall]
12661 *) Recompiled the error-definition header files and added
12662 missing symbols to the Win32 linker tables.
12663 [Ralf S. Engelschall]
12665 *) Cleaned up the top-level documents;
12666 o new files: CHANGES and LICENSE
12667 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12668 o merged COPYRIGHT into LICENSE
12669 o removed obsolete TODO file
12670 o renamed MICROSOFT to INSTALL.W32
12671 [Ralf S. Engelschall]
12673 *) Removed dummy files from the 0.9.1b source tree:
12674 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12675 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12676 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12677 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12678 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12679 [Ralf S. Engelschall]
12681 *) Added various platform portability fixes.
12684 *) The Genesis of the OpenSSL rpject:
12685 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12686 Young and Tim J. Hudson created while they were working for C2Net until
12688 [The OpenSSL Project]
12691 Changes between 0.9.0b and 0.9.1b [not released]
12693 *) Updated a few CA certificates under certs/
12696 *) Changed some BIGNUM api stuff.
12699 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12700 DGUX x86, Linux Alpha, etc.
12703 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12704 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12708 *) Add -strparse option to asn1pars program which parses nested
12710 [Dr Stephen Henson <shenson@bigfoot.com>]
12712 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12715 *) DSA fix for "ca" program.
12718 *) Added "-genkey" option to "dsaparam" program.
12721 *) Added RIPE MD160 (rmd160) message digest.
12724 *) Added -a (all) option to "ssleay version" command.
12727 *) Added PLATFORM define which is the id given to Configure.
12730 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12733 *) Extended the ASN.1 parser routines.
12736 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12739 *) Added a BN_CTX to the BN library.
12742 *) Fixed the weak key values in DES library
12745 *) Changed API in EVP library for cipher aliases.
12748 *) Added support for RC2/64bit cipher.
12751 *) Converted the lhash library to the crypto/mem.c functions.
12754 *) Added more recognized ASN.1 object ids.
12757 *) Added more RSA padding checks for SSL/TLS.
12760 *) Added BIO proxy/filter functionality.
12763 *) Added extra_certs to SSL_CTX which can be used
12764 send extra CA certificates to the client in the CA cert chain sending
12765 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12768 *) Now Fortezza is denied in the authentication phase because
12769 this is key exchange mechanism is not supported by SSLeay at all.
12772 *) Additional PKCS1 checks.
12775 *) Support the string "TLSv1" for all TLS v1 ciphers.
12778 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12779 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12782 *) Fixed a few memory leaks.
12785 *) Fixed various code and comment typos.
12788 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12789 bytes sent in the client random.
12790 [Edward Bishop <ebishop@spyglass.com>]
projects / oweals / openssl.git / blob